United States Office
of Research and Development (ORD) EPA/600/B-18/233
Environmental
Protection August 2018
Agency ![]() (RELEASE 3.2) MODELING
ENVIRONMENTAL FATE AND ECOLOGICAL
EFFECTS IN AQUATIC
ECOSYSTEMS VOLUME 1: USER’S MANUAL
|
|
{This Page Left
Blank, Back of Cover}
MODELING
ENVIRONMENTAL FATE
AND
ECOLOGICAL EFFECTS IN
AQUATIC
ECOSYSTEMS
Jonathan
S. Clough
August 2018
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON DC 20460
This document
describes the interface of the aquatic ecosystem model AQUATOX, Release 3.2. Anticipated users of this document
include persons who are interested in using the model, including but not
limited to researchers and regulators.
The model described in this document is not required, and the document
does not change any legal requirements or impose legally binding requirements
on EPA, states, tribes or the regulated community. This document has been approved for
publication by the Office of Science and Technology, Office of Water, U.S.
Environmental Protection Agency. Mention
of trade names, commercial products or organizations does not imply endorsement
or recommendation for use.
This model has been developed and documented by Dr. Richard
A. Park of Eco Modeling and by Jonathan S. Clough of Warren Pinnacle
Consulting, Inc. under subcontract to Eco Modeling, Horsley Witten, and CSRA
LLC. Work has been funded with Federal
funds from the U.S. Environmental Protection Agency, Office of Science and
Technology. Work was performed under EPA
contract number 68-C-01-0037 with AQUA TERRA Consultants and EPA contract number
EP-C-12-006 with Horsley Witten Group.
Integration of Interspecies Correlation Estimation (Web-ICE) was made
possible due to the work of US. EPA Office of Research and Development Gulf
Breeze, the University of Missouri-Columbia, and the US Geological Survey. Release 3.2 was developed under contract
HHSN316201200013W, Task Order EP-G16H-01256 “Scientific Models, Applications,
Visualizations, Computational Science, and Statistical Support (SMAVCS3),”
under contract to with CSRA LLC, with Henry Helgen as TDD Lead.
The assistance, advice, and comments of the EPA work assignment manager, Marjorie Coombs Wellman of the Standards and Health Protection Division, Office of Science and Technology have been of great value in developing this model and preparing this report. Dr. Nigel Pickering of Horsley Witten Group provided extensive comments and writing support. Further technical and financial support from David A. Mauriello, Rufus Morison, and Donald Rodier of the Office of Pollution Prevention and Toxics is gratefully acknowledged. Marietta Echeverría, Office of Pesticide Program, contributed to the integrity of the model through her careful analysis and comparison with EXAMS. Release 3.2 was developed with the thoughtful assistance of Rajbir Parmar of the National Exposure Research Laboratory (ORD), and Brenda Rashleigh of the USEPA Environmental Effects Research Laboratory (ORD).
Release 2 of this model underwent independent peer review by Donald DeAngelis, Robert Pastorok, and Frieda Taub; and Release 3 underwent peer review by Marty Matlock, Damian Preziosi, and Frieda Taub. Their diligence is greatly appreciated.
AQUATOX (RELEASE 3.2) VOLUME 1: USER’S MANUAL
Libraries
and Parameters in a Simulation.
Modifying
the State Variable List
Initial
Conditions and Loadings
Importing
and Exporting Loadings
Chemical
Properties and Fate Data Screen.
Output
Setup Screen (results tracking)
Perturbed
and Control Screens (running the model)
Control
vs. Perturbed Simulations
Export
to Release 2.2/3.0 Databases
"Taking
Parameters" from Another Study
Wizard:
Step 1- Simulation Type
Wizard:
Step 2- Simulation Period
Wizard:
Select Type of Fish Compartment(s)
Wizard:
Step 8- Site Characteristics
Wizard:
Step 10- Water Temperature
Wizard:
Step 12- Light Loading
Wizard:
Step 14- Inorganic Solids
Wizard:
Step 16- Inflow Loadings
Wizard:
Step 17- Direct Precipitation
Wizard:
Step 18- Point-Source Loadings.
Wizard:
Step 19- Nonpoint-Source Loads.
Details
about Text-Based Study Files
Setting Up Simulation, Single Segment
Starting
with a Surrogate Simulation
Using
Sediment Bed Models and Data Requirements
Working with a Linked Simulation
"Local"
and "Global" Parameters
Exporting
All Graphs to Word (Linked)
Setting Up a Linked Simulation
Additional
Data Requirements for Linked Mode
Excel
Template Import Capability
Setting
up Linkages, Complex Systems
Stratification
in a Linked System
Tutorial--
Step 1: Deleting and Adding a Plant
Tutorial--
Step 2: Setting an Initial Condition
Tutorial--
Step 3: Viewing Parameters
Tutorial--
Step 4: Viewing Toxicant Loadings
Tutorial--
Step 5: Running the Simulation
Tutorial--
Step 6: Viewing Output
Stream
Tutorial-- Step 1: Selecting an Existing Study as a Template
Stream
Tutorial-- Step 2: Modifying the Study with the AQUATOX Wizard
Stream
Tutorial-- Step 3: Additional Study Modifications and Running
Stream
Tutorial-- Step 4: Examining Model Output and Importing External Data
AQUATOX is a simulation model for aquatic systems.
AQUATOX predicts the fate of various pollutants, such as nutrients and organic
chemicals, and their effects on the ecosystem, including fish, invertebrates,
and aquatic plants. AQUATOX is a valuable tool for ecologists, biologists,
water quality modelers, and anyone involved in performing ecological risk
assessments for aquatic ecosystems.
AQUATOX simulates the transfer of biomass, energy
and chemicals from one compartment of the ecosystem to another. It does this by
simultaneously computing each of the most important chemical or biological
processes for each day of the simulation period; therefore it is known as a
process-based or mechanistic model. AQUATOX can predict not only the
environmental fate of chemicals in aquatic ecosystems, but also their direct
and indirect effects on the resident organisms. Therefore it has the potential
to establish causal links between chemical water quality and biological
response and aquatic life uses.
AQUATOX is the only general ecological risk model
that represents the combined environmental fate and effects of conventional
pollutants, such as nutrients and sediments, and toxic chemicals in aquatic
ecosystems. It considers several trophic levels, including attached and planktonic
algae and submerged aquatic vegetation, invertebrates, and forage,
bottom-feeding, and game fish; it also represents associated organic toxicants.
AQUATOX has been implemented for streams, ponds, lakes, estuaries, reservoirs,
and experimental enclosures.
The fate portion of the model,
which is applicable especially to organic toxicants, includes: partitioning
among organisms, suspended and sedimented detritus, suspended and sedimented
inorganic sediments, and water; volatilization; hydrolysis; photolysis;
ionization; and microbial degradation. The effects portion of the model
includes: acute toxicity to the various organisms modeled; and indirect effects
such as release of grazing and predation pressure, increase in detritus and
recycling of nutrients from killed organisms, dissolved oxygen sag due to
increased decomposition, and loss of food base for animals.
AQUATOX is the latest in a long
series of models, starting with the aquatic ecosystem model CLEAN (Park et al.,
1974) and subsequently improved in consultation with numerous researchers at
various European hydrobiological laboratories, resulting in the CLEANER series
(Park et al., 1975, 1979, 1980; Park, 1978; Scavia and Park, 1976) and
LAKETRACE (Collins and Park, 1989). The MACROPHYTE model, developed for the
U.S. Army Corps of Engineers (Collins et al., 1985), provided additional
capability for representing submersed aquatic vegetation. Another series
started with the toxic fate model PEST, developed to complement CLEANER (Park
et al., 1980, 1982), and continued with the TOXTRACE model (Park, 1984) and the
spreadsheet equilibrium fugacity PART model. AQUATOX combined algorithms from
these models with an ecotoxicological construct borrowed from the FGETS model
(Suárez and Barber, 1992); and additional code was written as required for a
truly integrative fate and effects model (Park, 1990, 1993). In the late 1990s, AQUATOX was restructured
and linked to Microsoft Windows interfaces to provide even greater flexibility,
capacity for additional compartments, and user friendliness.
To install the AQUATOX Model, an "InstallAware" installer is utilized.
Through its install program, AQUATOX will produce the following folder structure:
Important Note: Be sure that you have read-write access to each of these directories wherever you are installing AQUATOX. This usually means that the “Program Files” directory is not the best choice.
If you install AQUATOX in the same location as a previous installation, you will no longer be able to access the previous version a–nd any modifications you made to the databases or default study files (*.APS) may be lost.
Some additional notes:
The
binary APS file is the basic unit in AQUATOX; it contains site data, loadings,
and parameter values used in a simulation; and it may contain results from a
prior simulation. (As of Release 3.2 these values can also be saved
within an editable a text file.) Click
on File in the menu bar to get the pull-down file menu, and click on Open.
You will then be given a choice of AQUATOX study files to load.
When a study is opened, a new
window is opened within the AQUATOX multi window interface. You may open
many studies at the same time and switch back and forth between them by
clicking or using the Window menu at the top of the screen.
A single segment AQUATOX
simulation is designated as a *.APS
file while a multi-segment AQUATOX simulation is designated as *.ALS. Text based AQUATOX simulations default to the *.TXT extension whether single or multi-segment.
The main window includes the name
of the study, the list of state variables used, and buttons from which to
choose various operations. At the top of the screen is an editable Toolbar.
Let the mouse hover over each button to get a "hint" that describes
the function of each toolbar button.
The Study Name can be
edited; it is separate from the name of the file, which you loaded and which is
displayed at the top of the screen. The study name is used as a title in
graphical output, so is best capitalized.
The Status window tells
when the perturbed and control runs were made, and warns if they are incomplete
or outdated.
The Initial Conditions
button brings up a screen with all the state variable values at the beginning
of a simulation. You can view the initial conditions, but not edit them
on this screen.
The Chemical button brings
up the loadings screen for the organic toxicant, if modeled. Double
clicking on the state variable named "Dissolved org.
toxicant" that will appear at the top of the list of state and driving
variables (if a chemical is included in the simulation) has the same
effect.
The Site button loads the
site characteristic screen.
Setup allows the user to set the dates of the simulation, and to
specify various options such as the control setup, uncertainty analysis, and saving biologic rates.
Notes provides a window for writing comments on the study.
Birds, Mink… links to a
model of chemical uptake for shorebirds or any other animals that feed
exclusively on aquatic prey.
Food Web brings
up an editable matrix of trophic interactions for the study.
The Sed Layers button
brings up parameters and initial conditions for the multi-layer sediment model
or the sediment diagenesis model if they are included in the simulation.
Perturbed starts the simulation with changed conditions, such as
with a toxicant.
Control starts a simulation without the stressor; the user can use
Control Setup to specify what is changed and what is held
constant, or parameters may be changed between running the control and
perturbed simulations.
Output presents the results as a series of charts and
graphs.
The output can be exported as
database, spreadsheet, or text files by clicking on Export Results
or Export Control
The Use Wizard button
allows you to edit the current simulation with the AQUATOX Wizard.
If you hit the Help button
from the Main Screen, you will jump to this topic in the help screen. The Help button on other screens links to
the appropriate subject in the help files.
The State and Driving Variables
In Study show the full list of state variables within the simulation.
Variables can be added to or deleted from this list using the buttons at the
bottom of the list. Animals, plants, and detritus within this list can
have up to twenty organic chemicals associated with them.
When the main screen is viewed in
Linked Mode it represents a single linked segment. In this case, two additional buttons are
available. Stratification affects the setup of vertically stratified segments
in linked mode. Morphometry allows the user to specify a time series of cross section
areas for each segment.
To save a file, click on File
then Save or Save As on the menu bar; you will also be given an
opportunity to save an altered file before exiting or loading another
file. Study files range in size from 25 KB to well over 2 MB. If
you wish to minimize the size of a study—for example, to transmit to someone
else—you can strip out the results by clicking on Study and choosing Clear
Results from the menu bar. The study files distributed with AQUATOX
have been minimized in this way.
Working with
Existing Simulations
An APS file is the basic unit in which AQUATOX simulations are loaded and saved. Each APS file contains the following items:
Not included in the APS file are uncertainty output, which is saved in a comma-separated-variable format (*.csv) and sensitivity output which is in Excel format (*.xls). Libraries of parameters are not included in the APS file either, other than those parameters that have been "loaded" into a simulation and therefore have relevance to that simulation.
New to AQUATOX Release 3.2, all of the data in an APS file can now be output into text format. These files contain all of the model parameters, driving data, and outputs so that the model can now be manipulated with third-party software.
See the section Details about Text-Based Study Files for more information about the structure of these files.
There is a basic dichotomy in
working with AQUATOX. You have a choice of editing database files in the general
"library" or of opening and editing a particular study. Studies
are self-contained files with all the information on a particular simulation,
including initial conditions, loadings, parameter values, first and last dates
for the simulation, and simulation results. Parameter values can be
edited, but changes apply only to that study. The intent is to be
able to archive a model application so that all assumptions and results are
saved for future reference. This is especially important for regulatory
applications that are subject to later review. (Of course, you also should
archive the version of AQUATOX that was used.)
Parameter and site records that
will be used repeatedly should be saved in the appropriate library. Each
library is a database in SQLite format with records for each organism,
chemical, or site. Generally, editing of parameters should be done in the
library mode to maintain consistency among studies. In contrast, if a
site record is only going to be used for a single study, it may be desirable to
create it within the study. Study records can be copied into the library;
so the choice of where to edit parameters is up to the user. It is the user’s
responsibility, though, to synchronize parameter values among studies. This
can be done by saving a record to a library and then loading that record to
each study.
To create or edit a record for
general use, click on Library in the menu bar. You can then click
on the specific library from the pull-down menu.
See Also: Types of Libraries
State variables are those
ecosystem components that are being simulated. These include organic
toxicants, nutrients and dissolved gasses, organism and detrital compartments
and their associated toxicants, and other variables traditionally considered
driving variables, such as water inflow, temperature, pH, light, and
wind.
AQUATOX is a powerful model
because you can add or delete state variables. It is even possible to
remove all biotic components in order to model a tank or other sterile
system. In general, the fewer state variables, the better. In
particular, unnecessary state variables slow down the simulation and create
additional requirements for verification. This is especially true for
streams, which tend to be more dynamic and therefore slower to simulate.
Nevertheless, often it is desirable to model a food web rather than a food
chain, for example to examine the possibility of less tolerant organisms being
replaced by more tolerant organisms as environmental perturbations occur.
The choice of which state variables to model depends to a large extent on the
purpose of the modeling application and the availability of data pertaining to
the state variables.
To Delete a state variable, select the variable you wish to delete
from the state variables list (on the opening screen after opening an APS file)
by clicking on it. To select multiple state variables, hold down the
control key while clicking more than one variable on the list. Then,
click on the delete button and confirm the deletion. There are
several state variables (such as nutrients) that are basic to an AQUATOX
simulation, and that therefore cannot be deleted.
To Add a state variable, click on the Add button and choose
the variable you wish to add from the dialog box that appears. Note that
the names of the taxonomic groups and ecologic guilds on the main study screen
are followed by the names of the specific groups in brackets. After
clicking the Add button, you will first be prompted to choose the
taxonomic group, ecologic guild, or chemical compartment you wish to add.
Then, you will be prompted to load the chemical or species specific parameters
from the appropriate library.
You may also modify the state
variable list using the AQUATOX Wizard.
After animal or plant state
variables have been added or removed one critical step is to view the trophic interactions for the simulation to ensure that the resulting food web
is reasonable. AQUATOX does
have default trophic interactions within each organism, but not knowing which
organisms are going to be included in each simulation, it will usually require
modification. The best way to access trophic interactions is by selecting
"Edit Trophic
Interactions" under the "Study" menu, or by clicking on
"Trophic
Matrix" which is a button found within Animal state
variable loadings screens.
Initial values and loadings are
needed for all the state variables or compartments simulated. These are
input on the loadings screen by double-clicking on the name in the state
variables list. If one or more toxicants are modeled, then initial
concentrations associated with the biota can also be specified. Constant
loadings for plants and invertebrates can be considered as “seed” values,
although care should be taken to use small values or the loadings can dominate
the simulation. Even periphyton and zoobenthos may be maintained through
drift from upstream, and a constant loading is appropriate. Likewise,
macrophytes might die back in winter and sprout from rhizomes; because rhizomes
are not explicitly modeled, a small loading is the mechanism for reestablishing
the population in the simulation when environmental conditions become more
favorable.
Of course, upstream loadings could
be significant inputs to a reach or lake. These might be represented by
constant or dynamic (time-varying) loadings. AQUATOX has a very flexible
interpolation routine to obtain daily values from irregular data points and
even time series occurring or extending outside the simulation period.
Dynamic loadings can be entered directly on the loadings screen, or they can be
entered or obtained offline and imported into the model. Imported data
can be in a variety of formats, which are evident when the "Change" button is
used. Loadings can be altered by means of a multiplier (the "Multiply loading by" button). This
procedure is especially useful for analyzing various loading scenarios.
It is also a way to correct or convert data series. However, ordinarily
the multipliers are set to 1 for the Control simulation, so their use for other
than perturbations is discouraged.
Loadings in "inflow
water" are closely related to the volume of inflow water specified
(or calculated as a result of choices) in the water volume screen.
In other words, loadings in unit per liter of water must have an associated inflow
of water in order to be relevant to the simulation.
On the other hand, nonpoint-source
(NPS), and point-source (PS) loadings are input in units of grams per
day and are not affected by the quantity of inflow water. AQUATOX ignores
the quantity of water that is associated with NPS and PS loadings as a model
simplification; however, this water can be directly modeled in the
multi-segment model with the use of Tributary Input Segments.
Loadings associated with direct
precipitation are a function of the site's surface area with units of grams per
meter squared per day. This input field includes both wet and dry
precipitation.
When working with a system with a
relatively low retention time, such as a segment of a river, loading of
floating biotic state variables such as phytoplankton can be important to properly
characterize. These loadings are
generally entered in units of milligrams per liter.
Any of the time-series loadings
may be imported, exported, or cleared using the "Change" button
found directly below their listing.
Also see the section titled "Important Note about Dynamic
Loadings" describing how AQUATOX
interpolates between data points..
For carbon dioxide loadings the Import CO2 Equil. button is also
available. This enables a user to import
a time series of equilibrium carbon dioxide values (if air to water exchange
were perfect). This interface is
designed for linkage with CO2SYS, CO2calc or other similar salt-water carbon
chemistry models. More information about
this interface may be found in section 5.6 of the Technical Documentation.
Loadings and other time series can be imported from Excel format, Tab delimited format, or Comma Separated Value (CSV) format.
To change the format, move to the List Files of Type menu and select a different format. The required input format will be specified in the box at the upper left. For example, an Excel data file should have the date in column A of the first worksheet and the data to be imported in Column B.
When you select a file, a preview will appear in the lower box. You can "update" this preview by pressing the button above it.
For Excel only, you may specify where the data you wish to import appears. This is helpful when working with very large spreadsheets with multiple sets of data on them. The worksheet number is specified first and then the start row, date column, and value columns may also be specified as shown below.
To import from Excel, Microsoft Excel 2000 or a later version of Microsoft Excel must be installed on the computer.
Some data, for which there are units conversions available, may be converted on the fly using the conversion check-box.
Loadings may also be exported from AQUATOX by selecting the "Export" tab at the top of the screen. Time series that are no longer relevant may be cleared from AQUATOX by selecting the "Clear Data" tab at the top of the screen.
Dynamic loadings are loadings that
are variable over the simulation's time-period. These loadings are
entered using a list of dates and associated loadings.
During a simulation, if the date
that is being simulated appears on the input list of dates, the loading is
taken directly from the list. If the current date in a simulation occurs between two dates,
interpolation is used to determine the correct loading value. Because of
this interpolation, if the intent is to represent a spike such as from storm
runoff on a particular day, the spike loading should be bracketed by zero
(“0") loadings.
If the current date in a
simulation occurs before
the first date or after the last date of the loading time series, AQUATOX assumes
that the loadings “wrap around” with an annual cycle. Specifically,
the AQUATOX algorithm will step towards the input data in one-year
increments until the derived date falls within the input time
series. The model will then interpolate the results, if necessary, and
assign the results to the date being modeled. In this manner,
if you had two years of loadings but ran the model for eight years, the model
would repeat the second year of loadings in an annual cycle for the
last seven years of the simulation. (Note: if a different type of
annual cycle or interpolation is desired this can be derived outside of the
AQUATOX interface and then imported into the model.)
Exercise caution when modeling
multiple years using loadings data from only one or a few years. Sporadic
loadings, which would be expected in that one particular year, may
inappropriately be repeated in successive years. If you do not wish
loadings to be repeated, enter values (“0" or otherwise) for the first and
last days of the simulation.
If there is only one dynamic
loading point present, this is interpreted in the same manner as a constant
load.
To access this screen,
double-click on “Dissolved Org. tox.” in the state variables list in the main
window. A toxicity variable needs to be
defined before it shows on this list, see “Adding a State Variable.”
This initial conditions and loadings screen contains a few items that are unique to dissolved organic toxicants.
The gas-phase concentration input allows the user to enter a constant concentration of toxicant in the air (g/m3) that affects the degree of volatilization (potentially in either direction).
The biotransformation button may be used to access the biotransformation screen which is only relevant when multiple chemicals are modeled, and only then when parent compounds and daughter products are both included in the mix.
The toxicity data button allows the user to specify direct effects from the toxicant to biotic elements in the simulation in the chemical toxicity data screen.
To access this screen double-click
on “Total Ammonia as N,” “Nitrate as
N,” or “Total Soluble P” in the state variables list in the
main window.
The nutrient initial conditions and loadings screens include the capability to model Total N and Total P along with all of the items contained on other loadings screens.
When initial conditions, inflows, or other loadings are entered as Total N, by selecting the appropriate check-box (e.g. "Init. Cond. is Total N"), model inputs are located in the "Nitrate" initial conditions and loadings screen and loadings on the ammonia screen are grayed out as they are not relevant. Total N is assumed to be 12% ammonia for inflow and nonpoint-source loadings and 15% ammonia for point-source loadings.
When Total N or Total P are used as model inputs, AQUATOX calculates the dissolved content by subtracting out loading inputs for suspended and dissolved detritus and suspended algae as discussed in section 5.4 of the technical documentation.
To access this screen double-click
on the “Susp. and dissolved detritus” in the state variables list in the main
window.
A complex loading screen is
necessary for organic matter inputs in the water column. AQUATOX
simulates Organic Matter (dry weight); however, the user can input data
as Organic Carbon or Carbonaceous Biochemical Oxygen Demand (CBOD)
and the model will make the necessary conversions. See the technical
documentation for more information about how the model converts organic carbon
and CBOD loadings into organic matter.
Organic matter initial
conditions and loadings are divided into four compartments:
Initial conditions and loadings are parsed by
specifying % Particulate and %
Refractory which can be entered as constant or time-varying
percentages (0-100). Loadings of organic matter can be constant or
dynamic (time series) for concentrations in inflowing water (mg/L), and for
mass from point sources and non-point sources (g/d). Toxicants associated
with detritus also can be specified (ug/kg).
Organic matter loadings in "inflow
water" are closely related to the volume of inflow water specified
(or calculated as a result of choices) in the water volume screen.
In other words, loadings in unit per liter of water must have an associated
inflow of water in order to be relevant to a simulation.
Separate state variable input
screens are provided for refractory and labile organic matter within the
sediment bed. The initial conditions are given as g/m2, and
the loadings are given as mg/L. Associated toxicants are given as ug/kg
(ppb).
See section 5.1 of the technical
documentation for more information. Also
see the section titled “Important
Note about Dynamic Loadings.”
To access this screen double-click
on “Temperature” in the state
variables list in the main window.
The annual mean and range in
temperature from the site underlying data screen can be used or a time series
can be entered—in which case make sure that the complete time period being
simulated is covered. If the system stratifies, then temperatures must be
given for both epilimnion and hypolimnion.
If a system can thermally stratify,
then hypolimnion temperature data could also be entered here. Under the
model’s default behavior, thermal stratification will be assumed
when vertical temperature differences exceed three degrees; at all other
times a well-mixed system is assumed and modeled. However, there is an additional button on
this screen labeled “Stratification
Options” that allows a user to modify this behavior by specifying dates of
stratification, thermocline depth, and flow routing options (see Stratification Options).
For more information on
temperature modeling, see section 3.5 of the technical documentation. If time series are entered you may wish to
see the note about dynamic
loadings.
To access this screen double-click
on Wind Loading in the state
variables list in the main window.
Within the wind loadings screen,
you can either use a default time series for loadings, a constant wind loading,
or enter dynamic loadings.
Wind is an important driving
variable because it determines the stability of blue-green algal blooms,
affects reaeration or oxygen exchange, and controls volatilization of some
organic chemicals. Wind also can affect
the degree of mixing in estuaries. Wind
is usually measured at meteorological stations at a height of 10 m from the
ground surface and is expressed as m/s. Wind is less important for
rivers and streams.
For the default time series, wind
is computed using a complex Fourier series of sines and cosines for a 365-day
repeating period with a user-supplied mean value. The edit box for
the mean value (in m/s) appears in the upper right hand of the wind
loadings screen, visible when “Use
Default Time Series” is chosen.
For more information on wind, see
section 3.7 of the technical documentation.
To access this screen double-click
on “Light” in the state variables list in the main window.
When entering light data, the user
has three options: Constant, time series, or annual mean and range may be
given for light in Langleys/day. If annual mean and range are used, these
parameters must be filled-in in the site underlying data screen.
Also, photoperiod can be
auto-calculated from latitude or entered manually. The latter can be
useful when modeling experimental facilities.
Conversions into Langleys per day
(Ly/d) are provided using the "convert" button; imported time
series may also be converted.
See section 3.6 of the technical
documentation. Also see the section titled
“Important Note about
Dynamic Loadings.”
To access this screen double-click
on “pH” in the state variables list in the main window.
pH is important in AQUATOX for
several reasons.
A user may input a time series of
pH values here or calculate pH values using a simple semi-empirical formulation
that requires a Mean Total Alkalinity input on this screen (see section 5.7 of
the technical documentation for more information.)
When a time series is utilized the
rules for dynamic loadings apply.
The pH state variable cannot be
deleted from an AQUATOX simulation.
To access this screen double-click
on “Water Volume” on the state variables list in the main window.
Considerable flexibility exists to
compute or specify water volume. Depending on the method chosen, inflow
or discharge values might be required. The Manning’s equation can be used
to compute changing volumes in a stream. The simplest procedure is to
hold volume constant at the initial condition. Volume can also be
computed dynamically using both inflow and discharge, which are input on
this screen. Evaporation can also affect
water volumes—this rate can be input as an annual evaporation rate on the site
underlying data screen or using a time-series import available on the site-type
screen. There is an “Evaporation” button available on the
Water Volume Data screen to directly link to the time-series evaporation output.
When available, a known time series can be entered or imported. Flow
data can be imported in several formats, including USGS tab-delimited; however,
recent changes in the USGS format, including variable header material,
makes this prone to problems. If the data do not appear in the preview
window then the flow data will have to be converted in a spreadsheet from cfs
to m3/d, and the date column and flow column then exported as a tab-delimited
or comma-separated file suitable for importing into AQUATOX.
Notes:
The initial condition water volume
is determined from the inputs on this screen and not from the site
underlying data. However, a "Get
Initial Cond. from Site Data" button exists to allow the initial
condition to be copied from that screen.
The Manning’s Equation Method (streams
only) requires discharge data. Inflow data and site volume are calculated
using Manning's Equation. Careful attention should be given to the
"Channel Slope" and "Manning's Coefficient" parameters
entered in the "Stream Data" screen (within the site underlying data
screen.)
The Keep Constant at Initial
Condition Level method
requires inflow data. Discharge is calculated based on inflow and
evaporation.
If you choose to Calculate Dynamically,
volume is calculated based on inflow, outflow and evaporation.
The Utilize Known Values Method
requires a time series of known volumes and inflow data. Outflow is
calculated taking evaporation into account.
The differential equation that calculates
the water volume of the system is:
δVolume / δ t = Inflow - Outflow -
Evaporation
The “Stratification Options” button allows a user to modify default
model behavior regarding stratification by specifying dates of stratification,
thermocline depth, or flow routing options.
For more information on
morphometry and volume, see section 3.1 of the technical documentation.
This screen can be accessed by
clicking on the “Initial Conds.”
button on the main screen.
Initial conditions can be
displayed and printed for all state variables in this summary screen and
any associated toxicants.
These initial conditions cannot be
edited on this screen because of the complexities of the suspended and
dissolved detritus screen.
Parameters provide values for
coefficients in the process equations. Although default values are given,
the user has great flexibility in specifying values to represent site-specific
species or groups.
There are five libraries of
parameter values that may be loaded into a simulation. These libraries can be reached by clicking
on the Animal (A), Chemical (C), Plant (P), Site (S), and Remineralization (R)
buttons on the program toolbar.
The Chemical Library
contains parameters for organic chemicals that could be associated with a
simulation. Within a simulation, chemical parameters may be found by
double-clicking a Dissolved org. toxicant on the state-variable list and
then choosing Edit underlying data. See also: Chemical Data Screen.
The Animal Library contains
parameters for fish and invertebrates that could be associated with a
simulation. Within a simulation, animal parameters may be found by
double-clicking an animal on the state-variable list and then choosing Edit
underlying data. See also: Animal Data
Screen.
The Plant Library contains
parameters for algae and macrophytes that could be added to a simulation.
Within a simulation, plant parameters may be found by double-clicking a plant
on the state-variable list and then choosing Edit underlying data. See
also Plant Data Screen.
The Site Library contains
parameters for representative sites that could be modeled by AQUATOX.
Within a simulation, the site parameters can be found by clicking the Site button
and then by clicking on the Edit Underlying Site Data button that
appears. See also Site Data Screen.
The Remineralization
Library contains parameters about the detritus and nutrients associated with a
site. To find these parameters within a simulation click on the Site button
and then by clicking on the Remineralization button that appears.
When editing libraries, a GridMode
button may be pressed that toggles editing between a matrix of all
parameters in the library or a close-up look at the parameters, units, and
comments for one specific entry in the library.
When editing a particular
simulation, this button allows the user to see a matrix of all of the animals,
plants, or chemicals within that particular study. All parameters within the grid may be edited
and the full set of parameters may be exported to Excel. All libraries can be
searched using improved “Find,” “Search Names,” or “Search Scientific Names” interfaces. Searches are not case sensitive and a partial
list of search results is shown for any entries that match the partial or
complete search string entered.
To add a new library entry you may
press the New button and a blank set
of parameters is created for data entry.
Alternatively, after modifying a parameter set in a simulation, those
parameter values may be saved back into the library (either overwriting the
original entry or the name may be changed to add additional information to the
library.)
When you are editing the
underlying data that applies to a simulation, you can either load from or save
those parameters to the library file by clicking on the Load from Library
or the Save to Library buttons at the top left of the screen.
This screen can be accessed by
clicking on the “Chemical” button on the main screen or double-click on “Dissolved
org. tox.” on the state variables list in the main window.
Required chemical "underlying
parameters" and units can be found on this screen. These parameters
govern chemical fate and partitioning behavior.
Note that the
organic-sediment/detritus and water partition coefficient can be calculated
dynamically or entered manually by the user. Sorption to inorganic
sediments is only relevant when the multi-layer sediment model is incorporated.
Parameters on this screen govern
the chemical fate processes of
A few of the parameters on this
screen are “greyed out” as they are not currently utilized by the model (e.g.
“Solubility” and “Vapor Pressure”).
These parameters have not been completely suppressed so that data in the
database are not lost and in case these parameters become useful in the future.
You may also edit the toxicity
data for the relevant chemical by selecting the Toxicity Data button at the top
right of the screen.
To maximize comprehension,
parameters on this user interface screen are described with several English
words rather than symbolically. Appendix
B of the AQUATOX
Technical documentation contains a full description of each of
the parameters shown here under "Chemical
Underlying Data" as well as their manner of referral in
the equations of the technical documentation (often a shorter variable
name). In this way, a user can use this appendix as a reference to search
the technical documentation and find all equations and related parameters.
Advanced users can also easily find the parameters within the AQUATOX source
code as the "internal" variable names are also listed within
Appendix B.
For more information on modeling
organic chemicals, see chapter 8 of the technical documentation.
This screen is where all of the
chemical toxicity parameters are located. To get to this screen go to Edit Underlying Data and select the “Toxicity Data”
button. Or use the button on the chemical's initial conditions and
loadings screen.
There are multiple options for
entering uptake rate constant (k1), the elimination rate constant (k2) and the
bioconcentration factor (BCF) or allowing the model to calculate these
parameters (BCF=k1/k2)
Additionally, elimination rates for
both plants and animals may be estimated using the octanol water partition
coefficient (Kow).
If the user only has toxicity data
for a few species, an extensive library of regressions (Interspecies
Correlation Estimation, or Web-ICE) is built into the model (Asfaw and Mayer,
2003). This library can be accessed by using the button at the bottom of
the screen ("Interspecies Toxicity
Correlation Models").
By entering both LC50 and EC50
values for a species the application factor can be computed. The user has
the option of applying that same ratio to the rest of the species in the animal
or plant toxicity screen using the buttons Estimate
animal LC50s... and Estimate plant
EC50s... at the bottom of this screen. Animal and Plant toxicity require
different parameters so they are given separate matrices on the screen.
The animal toxicity data appears above the plant toxicity data.
The animal toxicity parameters are
as follows:
Animal Name, must match the "toxicity data" record in the
relevant Animal Data screen.
LC50, (µg/L), external concentration of toxicant at which 50%
of population is killed
LC50 exp.
time, (h), exposure time in toxicity
determination
K2 Elim.
Rate Const., (1/d), elimination rate
constant
K1 Uptake Const. (L/kg-d) optional
uptake rate constant (dry weight)
BCF, (L/kg) optional
Bioconcentration Factor (dry weight)
Biotrnsfm.
rate const, (1/d), Daily rate of
biotransformation of this toxicant
EC50 growth, (µg/L), external concentration of toxicant at which there
is a 50% reduction in growth
Growth exp.
time, (h), exposure time in toxicity
determination
EC50 repro., (µg/L), external concentration of toxicant at which there
is a 50% reduction in reproduction
Repro exp.
time, (h), exposure time in toxicity
determination
Ave wet wt., (g), mean wet weight of organism
Lipid Frac, (g lipid/g organism), fraction of lipid in organism (wet
weight)
Drift Threshold, (µg/L), concentration at which invertebrate drift is initiated
LC50 Slope, (species-specific slope at LC50 multiplied by LC50), LC50 slope factor for the external toxicity model. If left blank or set to zero the value from the chemical’s underlying data is used.
Note: there are several comment
fields provided for documentation of parameter source citation or other notes.
The plant toxicity parameters are
as follows
Plant Name, must match the "toxicity data" record in the
relevant Plant Data screen.
EC50 photo, (µg/L), external concentration of toxicant at which there
is 50% reduction in photosynthesis
EC50 exp.
time, (h), exposure time in toxicity
determination
EC50
dislodge, (µg/L), external concentration
of toxicant at which 50% of algae is dislodged (not applicable for
phytoplankton)
K2 Elim.
Rate Const, (1/d), elimination rate
constant
K1 Uptake Const. (L/kg-d) optional
uptake rate constant (dry weight)
BCF, (L/kg) optional
Bioconcentration Factor (dry weight)
Biotrnsfm.
rate const., (1/d), Biotransformation Rate
LC50, (µg/L), external concentration of toxicant at which 50%
of population is killed
LC50 exp
time, (h), exposure time in toxicity
determination
Lipid frac., (g lipid/g organism), fraction of lipid in organism (wet weight)
LC50 Slope, (species-specific slope at LC50 multiplied by LC50), LC50 slope factor for the external toxicity model. If left blank or set to zero the value from the chemical’s underlying data is used.
Buttons at the bottom of the
screen:
Estimate Animal or Plant K2s using Kow. Elimination rate constants may be estimated using the
chemical's KOW data. Animal estimates utilize the "wet weight"
and "lipid frac." fields, whereas plant estimates utilize only the “fraction
lipid” field within the plant's toxicity record. For a
macrophyte-specific calculation to be performed, the plant name must include
the substring "macroph."
In Release 3.1 and after, an
alternative K2 estimation is also available within the code, based on Barber
(2003) that also requires the "wet weight" and "lipid
frac." fields to be populated with data.
For more information on this option, see the section on “Elimination” in
section 8.7 of the Technical Documentation. When the user selects to Estimate Animal
K2s using Kow, the option to use the classic formulation or the Barber
alternative is presented to the user.
When the chemical being modeled is
a PFA, animal K1 and K2 estimations are available as a function of the mean
weight of the organism and the “chain length” of the chemical (equations 398,
401, 405, and 406 of the Technical Documentation). In AQUATOX the K1 and K2 values
produced by this estimation procedure are now editable. However, all animals and plants must be
linked to a chemical toxicity record with a populated K2 field for the model to
work properly.
LC50 / EC50 estimation: Using the EC50 to LC50 ratio for one species, the EC50 or
LC50 can be estimated for other species within the simulation.
Interspecies Toxicity Correlation Models: Utilize the integrated ICE regression databases.
Extensive information about the
AQUATOX ecotoxicology model may be found in chapter 9 of the technical documentation.
This screen allows the user to utilize a database of regressions to populate all of the species in their simulation with toxicity data.
Often LC50 data will only be available for one or two of the many species that a user wishes to include in a simulation. To alleviate this problem, a substantial database of regressions is available (ICE) as developed by the US. EPA Office of Research and Development, the University of Missouri-Columbia, and the US Geological Survey (Raimondo 2010). The ICE database includes over 2000 models with over 100 aquatic species as “surrogates.” Regressions may be made on the basis of species, families, or genera. The database also includes goodness of fit information for regressions so their suitability for a given application may be ascertained (see statistic definitions below).
The steps that a user can take to use ICE within AQUATOX are as follows:
You may also visit the Web-ICE Website using the URL or the button displayed at
the top of the screen in case an updated database is available.
Graphs may be displayed on a log-scale or arithmetic basis and the user may change the confidence interval plotted.
Statistics provided for goodness of fit are defined below (Raimondo 2010)
When multiple chemicals are modeled,
the biotransformation screen indicates which chemicals can be converted to
other chemicals within the simulation.
For the chemical indicated in the title
of the screen, the columns of the matrix indicate the ecological compartments
in which that chemical can break down and the rows of the matrix represent
the other chemicals within the simulation into which the chemical could be
transformed. Each cell indicates the percentage of the chemical that is
biotransformed into the daughter product (chemical indicated in the row
title). Each column must sum to 100 percent; the "other" row
autocalculates to make this possible.
The rate of biotransformation in
each species is governed by the biotransformation parameter found in the
chemical toxicity record screen.
Categories in which
biotransformation may be specified are:
In some cases, modeling a specific
biotransformation pattern within a particular species will be desirable.
In this case, using the "Add
Species Specific Data" button will make this possible. You will
be able to select a modeled species in the simulation and then specify the
appropriate biotransformation rate within that species.
To access this screen double-click on a plant name in the state variables list in the main window and choose Edit Underlying Data or by accessing the plant library.
Several fields near the top
of the screen require explanation. If you click on the drop-down arrow to
the right of Plant type, you will be given a choice. The choice of Plant
type is important because different types have different physical or
biological processes that apply to them. For instance, phytoplankton are
subject to sinking, but not periphyton, which are attached to a surface.
Conversely periphyton are limited somewhat by very slow current velocity; but
not phytoplankton, which are adapted to still water. If “macrophytes” are chosen the species must
be distinguished as “benthic,” “rooted floating,” “or “free-floating.” Table 5 in the Technical Documentation helps
clarify some of the differences between plant types.
If a plant is “surface floating”
it is assumed reside in the top 0.1 m of the system unless wind distributes it
to the top 3 m based on Langmuir circulation.
The averaging depth for surface floating plants, when calculating plant
concentrations (mg/L), is assumed to be the top three meters to more closely
correspond to monitoring data.
The capability to model internal
nutrients in algae was added just prior to release of AQUATOX 3.1plus.
Consequently, there was limited time to calibrate the enhancement. The user must consciously choose to model
internal nutrients by checking that option in the Setup screen. However, to
ease the transition, six parameters that are unique to modeling internal
nutrients are populated with default values taken from the documentation of
WASP7 (Ambrose et al. 2006).
The effects of modeling internal
nutrients were investigated for all studies supplied as examples with the
AQUATOX installation. As might be
expected, the nutrient-poor sites exhibited the greatest response to luxury
uptake of nutrients, but default parameters cannot be used as-is and require
the most recalibration. Where there were
sufficient observed data to evaluate the results, the applicability of the
internal-nutrient submodel is noted. In
general, the following sites benefited from modeling internal nutrients: Cahaba
River AL, Evers Reservoir FL, Lake Onondaga NY, Tenkiller Reservoir OK, and Rum
River MN. The following sites showed a tendency to over-predict algal biomass
(and dissolved oxygen): Blue Earth River MN, Crow Wing River MN, DeGray
Reservoir AR, and Lake Jesup FL.
Although Plant type is
important for determining which processes do or do not apply to the state
variable, the Taxonomic Type field is included as an organizing tool and
does not affect model output.
The Toxicity Record field
within this screen links to the "plant name" within each chemical's
toxicity data. In this manner, several plants could link to the same
toxicity record if that is desired. You may select a record from the list
or type a new name if the desired plant name does not appear in the drop-down list.
Phytoplankton and periphyton
compartments may be linked together. When viewed in a simulation, a plant
will have an "Edit All Plant
Linkages" button available at the top of the screen. Periphyton
also have a "Periphyton Linkage"
button which allows you to edit the same information but only for the species
shown. It is considerably more powerful and user-friendly to use the
"Edit All" button.
A few notes regarding the
some of the most important parameters:
To maximize comprehension,
parameters on this user interface screen are described with several English
words rather than symbolically. Appendix
B of the AQUATOX
Technical documentation contains a full description of each of
the parameters shown here under "Plant
Underlying Data" as well as their manner of referral in
the equations of the technical documentation (often a shorter variable
name). In this way, a user can use this appendix as a reference to search
the technical documentation and find all equations in which each parameter is
utilized. Advanced users can also easily find the parameters within the
AQUATOX source code as the "internal" variable names are also
listed within Appendix B.
See section 4.1 and 4.2 of the
technical documentation for extensive discussion of modeling algae and
macrophytes.
Individual periphyton to phytoplankton linkages can be edited using the “Periphyton Linkage button.” Using the “Edit All” button allows you to edit these linkages for the entire simulation simultaneously. All periphyton in a simulation are listed here along with drop-down boxes that allow the user to link these to available phytoplankton compartments. For convenience sake, multiple periphyton compartments may be linked to the same phytoplankton compartment.
Periphyton may slough or be scoured, contributing to suspended algae; this may be reflected in the chlorophyll a observed in the water column. Using this screen, periphyton can be linked to a phytoplankton compartment so that chlorophyll a results reflect periphyton sloughing. One-third of periphyton is assumed to become the linked phytoplankton and two thirds are assumed to become suspended detritus in a sloughing event. When not linked, AQUATOX assumes that all sloughed periphyton becomes detritus.
Additionally, when phytoplankton undergoes sedimentation it will now be incorporated into the linked periphyton layer if such a linkage exists. If multiple periphyton species are linked to a single phytoplankton species, biomass is distributed to periphyton weighted by the mass of each periphyton compartment. If no linkage is present, settling phytoplankton are assumed to contribute to sedimented detritus. (At this time, it is not possible to divide sedimented phytoplankton between these two compartments.)
This screen displays all of the relevant parameters for the
animal that has been selected. To access
this screen double-click on an animal name in the state variables list in the
main window and choose Edit Underlying Data.
The
Toxicity Record field within this screen links to the "animal
name" within each chemical's toxicity data. In this manner, several
animals could link to the same toxicity record if that is desired. You
may select a record from the list or type a new name if the desired animal name
does not appear in the pull-down list. To edit toxicity linkages for all
plants and animals in a simulation simultaneously, click the "Edit All" button.
AQUATOX can model two size
classes for each fish species. Records for different size classes are
linked by clicking on Size-Class Links (under the Animal Name towards the top of the screen) and
choosing the correct record from the list given.
As was the case in the Plant Data
Screen, the choice of Animal type is important because different animal types
have different physical or biological processes that apply to them. For
instance, benthic insects are subject to emergence but other animal types
are not. Table 6 in the Technical
Documentation describes the differences between animal types.
For "Benthic
Invertebrates," a benthic designation box is available. This box does not affect model execution but
has effects on the calculation of Biological Metrics (see section 4.7 of the
technical documentation.)
Although Animal type is
important for determining which processes do or do not apply to the state
variable, the Taxonomic Type or Guild field is included as an organizing
tool and does not affect model output.
Sensitivity of animals to sediments comes into
this parameter screen in several places. The feeding effects are
found in the top of the screen with the feeding parameters. Lethal
effects and percent-embeddedness effects are found towards the bottom of the
screen with the other mortality parameters.
The fraction in water column field
is only relevant to models running the multi-layer sediment model and
determines how much of the organism is exposed to the water column as opposed
to pore waters in the top layer of sediment.
Bioaccumulation Data includes
the lifespan and fraction lipid. Uptake and maximum bioaccumulation of
organic chemicals are sensitive to these parameters.
Low Oxygen Effects parameters
are available for lethality and growth and reproduction effects. Ammonia toxicity
parameters follow below that followed by Salinity
effects. (These
parameters are not on the chemical toxicity screen since they are not
organic-toxicant specific.)
If an organism resides in a
stream, the habitat-type within the stream should be specified using the percent riffle, percent pool, and percent run parameters. This affects the water velocity the organism
is exposed to and also potentially excludes the organism from some water
segments on the basis of habitat availability.
Fish may spawn automatically using
a formulation based on the optimal
temperature parameter or specific spawning dates may be entered
under spawning
parameters.
For fish, maximum consumption and endogenous
(basal) respiration rates can be directly entered into the model (top of
screen) or allometric
models can be utilized (bottom of screen). Allometric
models within AQUATOX can calculate consumption and respiration rates
as functions of weight and temperature.
A few notes regarding the
some of the more important parameters:
To maximize comprehension, parameters on this user
interface screen are described with English words rather than
symbolically. Appendix
B of the AQUATOX
Technical documentation contains a full description of each of
the parameters shown here under "Animal
Underlying Data" as well as their manner of referral in
the equations of the technical documentation (often a shorter variable
name). In this way, a user can use this appendix as a reference to search
the technical documentation and find all equations in which each parameter is utilized.
Advanced users can also easily find the parameters within the AQUATOX source
code as the "internal" variable names are also listed within
Appendix B.
See section 4.3 of the technical documentation for
extensive discussion of modeling animals within AQUATOX.
Fish and predatory invertebrates may be modeled as a single "stand-alone" compartment, they may be linked by size-class, using the species data screen or an age-class fish may be modeled.
To get a good summary of the fish linkages currently utilized by the model, these can be viewed in Step-7 of the AQUATOX Wizard (example graphic below). This screen also provides a convenient interface to add size-class or age-class fish.
When you load an animal into an
AQUATOX simulation, you are also loading a trophic interaction matrix for that
particular animal. This trophic interaction table is important because it
defines food-web relationships and assimilation efficiencies. To get to
the single-species trophic interaction screen, within the Animal Library, click
on the Trophic Interactions button that appears at the top right of the
screen. To get to the trophic interaction matrix for an animal within a
simulation, click on that same button within the animal's underlying data.
Within AQUATOX, there are two ways
of viewing trophic interactions for a given simulation, on a
species-by-species basis or to view the trophic interactions matrix for the
whole simulation. The species-by-species manner of viewing trophic
interactions is generally less useful but is maintained so that users can edit
trophic interactions for libraries as well as for specific simulations.
This screen shows the preferences
and egestion factors for a single species. Within a simulation, you will
generally only be interested in viewing an animal's trophic interactions with
organisms within that simulation. To view preferences for all available
organism compartments, clicking the View All Data button at the bottom
of the screen.
The preference ratio indicates the
animal's preference to consume a particular compartment on the list. Each
preference is a fractional preference relative to the other food items on the
list, and it need not add to one. Each time-step, AQUATOX will normalize
these fractional preferences for the food sources that are available at that
time.
The egestion fraction represents
the portion of food that is not assimilated for a given animal and food
compartment combination.
The matrix enables the user to see all trophic
interactions for all animals in a particular species against all prey
items. Generally, this is a more useful interface for examining and
entering information about the food web. Preferences are normalized to
100% when the screen is first shown. After data are entered or modified,
these preferences can be renormalized by clicking the "renormalize"
button at the top of the screen.
By clicking on the radio buttons at the top of the
screen, the user may toggle between preferences, egestion coefficients, and
comments. The matrix can also be exported to Excel for printing or
sharing.
This screen, applicable to fish
only, is accessed by selecting "Size Class Links" from the Animal's
underlying data.
The species data screen allows the
user to represent a single species of fish with two state variables
representing two size classes of that fish, generally young of the year (YOY)
and adult.
Large gamefish can only be paired
with species loaded into a small gamefish compartment. Similarly, large
bottom fish can only be paired with small bottom fish and large forage fish can
only be paired with small forage fish compartments. Any species that can
be matched within the simulation will appear on the list within this screen.
To get a good summary of the
animal linkages currently utilized by the model, these can be viewed in Step-7
of the AQUATOX Wizard. This screen also provides a convenient interface
to add size-class or age-class fish.
A similar dialog box may also be
utilized to link periphyton to phytoplankton compartments, but the Plant Linkages screen (accessible by
using the "Edit All" button
in plant underlying data) is a preferable interface for this procedure as all
plant linkages may be viewed simultaneously.
AQUATOX can model one species of
fish as having multiple age classes with input from this series of specialized
screens. To access this screen one must first add a multi-age class fish
to the simulation (see below) and then double-click on "Multi. Age-Class Fish"
in the state variable list. When you add the Multi-Age fish, you must
select the name and age of the fish and select parameter sets for the
young-of-year and older age class fish.
Most fish parameters are not
assumed to change their values between age classes. However, the following parameters may be
specified as a function of fish age:
Initial Condition, Inflow Loadings, Toxicant Initial Condition, Toxicant
Loadings, Lipid Fraction, Mortality Coefficient, and Mean Weight. Tabs at the top of this screen enable
the user to switch between input screens for each of these parameters.
Click on each tab name to move from one tab to the next. For each
tab, the user can enter values for each age class or can choose one of several
distributions, characterized by user-supplied statistics. You can select a Uniform, Triangular, Normal,
or Log-normal distribution if these distributions could be used to describe the
pattern of fish parameters over its age class.
Uniform distributions (all age-classes having the same value) or
user-input data will probably be most useful.
The values can be graphed as well
to better display trends over the life of the fish.
Parameter screens and trophic
interaction screens can be accessed for young-of-the-year (YOY) and older fish.
When the "older fish" parameter screen and trophic interactions
screens are entered, the parameters are relevant for all age classes one year
old and older.
To add a multi-age class fish,
select to Add a
state variable from the main study window and then select "Multi.Age-Class Fish"
at the bottom of the available list of variables to add. A series of
dialogs follows allowing you to designate the number of age-classes modeled (Maximum Fish Age) and
to load parameters into the new fish. Alternatively, a multi-age fish can
be added in Step 7 of the AQUATOX Wizard.
When adding a fish-species, a dialog box will ask if you want to add a
single-compartment, size-class, or age-class fish. You will then be prompted for the number of
age classes and to load parameters into the fish as above.
When
size-class fish are included in a simulation, you can access the “Anadromous
Fish Setup” screen using the “Anadromous”
button on the fish loadings screen (accessible by double-clicking either of
the size-classes on the state variable list.)
This relatively simple model
allows the user to simulate the migration of fish into and out of the main
study area in order to approximate anadromous migration behavior. Anadromous fish live most of their adult life
in saltwater, but they return to freshwater to spawn, and juveniles grow for a
few months to a few years before going to saltwater. No additional exposure of the fish to organic
chemicals is assumed to occur to the fish while off-site, but growth dilution
and depuration of toxicants are assumed to occur.
When the radio-box to “Model Size-Class fish as Migrating Off-Site
and Returning as Adult” is checked then the model is activated and the user
must set certain variables describing the date of juvenile migration, the
fraction of biomass migrating, the date of adult return, the years spent off
site, and the fraction of mortality that occurs during the off-site years.
Based on these parameters and the
weight of the juvenile and adult organisms, the biomass returning to the
freshwater study area is calculated.
Additionally, the chemical concentration in these returning fish can
also be estimated, given the depuration rate for the chemical in the fish.
For more information see the
section entitled “Anadromous Migration Model” in Chapter 4 of the Technical
Documentation.
This screen displays all of the relevant parameters for the site. To access this screen click on the “Site” button on the main window.
This
screen allows a user to select which site type is being
simulated. Six site types are available: pond, lake, stream (creek
or river), reservoir, (experimental) enclosure, and estuary.
For standing water (ponds,
lakes, and reservoirs) site type is not currently a sensitive parameter
(these site-types can be used interchangeably).
Selecting to model a
"stream" has the following effects:
Selecting "Estuary" produces a salt-wedge model and permanent stratification.
For a linked version, a seventh site type is
available, that being a "Tributary
Input." State variables
are not solved within this type of segment; rather, loadings of nutrients,
organic matter, and biota to this type of segment are loaded from the segment directly
into the modeled system.
Additionally, a user may edit a site's
underlying data, remineralization
parameters, or stratification options from
this screen, or load default site or remineralization data from libraries.
Parameters that may be edited on
this screen in a time series or constant manner are accessed by the “Show Mean
Depth /Evaporation” and “Show Shading / Velocity” toggle button at the bottom
of the screen:
Each site can be characterized by
a relatively small number of site constants. These can be seen and edited
by clicking on Edit Underlying Data in the Site Data window, or
they can be loaded from the Library. There is some redundancy in
that Volume, Area, and Mean Depth all have to be
specified. Based on mean and maximum depth, the bathymetry of the site is
computed (see equation 8 in the Technical Documentation). Volume is a
state variable and can be computed in a variety of ways (accessible through the
volume loading screen
); however, one option is to set it to remain constant using the value
provided in the site screen.
The Max. Length is the
distance, usually the long axis, across which wave buildup can occur; it
determines the depth of mixing in stratified systems. This can also
affect phytoplankton retention time in flowing systems as well as the
calculation of cross-section area for velocity calculations.
The epilimnetic and hypolimnetic
temperature parameters are only used if the user has selected to
"use annual mean and range loadings" on the Temperature Screen.
If the user has selected to use this manner of calculating loadings,
then both epilimnetic and hypolimnetic temperature parameters have to be
specified on the site data screen, even for streams and ponds, where they can
be set equal.
Likewise, in the Light Screen,
if a user has selected to "use annual mean and range
loadings," that user must provide data about the Average Light
and the Average Light Range on
the Site Data screen, from which seasonal fluctuations are
computed. These are not computed from the latitude because of local and
regional differences in elevation, cloud cover, and maritime or continental
climatic conditions.
Latitude is used to compute the seasonal variation in day length, although
this can be overridden in the Light Screen. This override can be
useful for laboratory simulations.
Altitude is used in
the computation of oxygen saturation.
Enclosure wall area is
used for experimental enclosures only and affects a site’s morphometry.
(Specifically, it increases the fraction of the site area within the euphotic
zone.)
Baseline Percent Embeddedness is the initial condition percent embeddedness for a
site. This value is used to calculate
effects when organisms are sensitive to embeddedness, calculated as a function
of TSS. See the “Interstitial Sediments”
portion of section 4.3 of the Technical Documentation.
The Minimum Volume Frac. multiplied by the initial condition volume for
a site represents the minimum volume that a water body can attain within the
simulation. If the water volume drops
below this level, numerical modeling of state variables stops and the
simulation skips forward to the next period in time when the water volume is
calculated to be above the minimum level.
Toward the bottom of the site
screen are the parameters for the phytoplankton
retention function. Phytoplankton and zooplankton will
quickly wash out of a short reach, but might be able to grow over an extensive
reach of a river, including its tributaries. To solve this problem,
AQUATOX takes into account the “Total Length” of the river being simulated, as
opposed to the length of the river reach, or “Site Length” so that
phytoplankton and zooplankton production upstream can be estimated. The
assumption is that conditions upstream are similar to those in the reach being
modeled.
For estuaries, a series of tidal range parameters is
required (amplitudes and epochs). These are used to calculate
the daily average tidal range for the site, which affects the average depth of
the salt wedge on that day. These parameters are generally
available for download from NOAA databases.
If a stream is being simulated
(set in the Site Type panel on
Site screen) the Stream Data button in the upper right is
enabled. Clicking this button displays a series of important stream
parameters regarding site morphometry, habitats represented, and parameters for
the optional sand-silt-clay sediment model.
To maximize comprehension, parameters
on this user interface screen are described with several English words rather
than symbolically. Appendix
B of the AQUATOX
Technical documentation contains a full description of each of
the parameters shown here under "Site
Underlying Data" as well as their manner of referral in
the equations of the technical documentation (often a shorter variable
name). In this way, a user can use this appendix as a reference to search
the technical documentation and find all equations in which each parameter is
utilized. Advanced users can also easily find the parameters within the
AQUATOX source code as the "internal" variable names are also
listed within Appendix B.
The parameters on the remineralization screen primarily
govern the settling and degradation of organic matter and its nutrient
stoichiometry. These parameters are accessed from the “Remineralization” button at the lower
left of the screen.
Many of the parameters on the remineralization screen can be assumed to remain constant from one simulation to another. A few highlights may be worth examination because they can both be site-specific and can affect some types of model results:
Because of these differences, there are multiple Remineralization records included in the Release 3.2 model. :
If your site has very different characteristics governing
remineralization, you can edit one of the existing records or create your own.
Shade can be an important limitation to light, especially in riparian systems. A user input “Fraction of site that is Shaded (canopy covered)” parameter can be entered either as a constant or as a time-series within the “Site” input screen. AQUATOX assumes that 2% of incident radiation penetrates the canopy so a site that has 100% canopy cover still receives 2% light transmission. See section 3.6 of the Technical Documentation for more information on this assumption, the equations utilized, and modeling light in general.
This "fraction shaded" parameter can be left as zero to simulate no shading effects on light.
Velocity in cm/s may be calculated by AQUATOX or entered as a time series. Otherwise, velocity is calculated as a simple function of flow and cross-sectional area.
Velocity has wide reaching effects including:
Mean depth may be kept constant or entered as a time series. Mean depth has significant effects on the average light climate that plants are exposed to (especially periphyton, which reside on the bottom of the water column.)
As noted on the entry screen, if "Use Bathymetry" is not checked on the Site Underlying Data screen, the mean depth calculation is irrelevant. In this case, mean depth is volume divided by surface area.
To access this screen select “Stratification Options” from the Water Volume screen.
In the case of stratification, a site's thermocline depth can be calculated as a function of maximum depth and site-length. Alternatively, the depth of the thermocline may be entered as a constant or a time series. The model will interpolate between entered dates if necessary.
Inflow of water (and the loadings within that waterbody) can be routed to a specific vertical segment, as can the outflow of water.
AQUATOX calculates stratification based on the temperature difference between segments. When the temperature difference is less than 3 C, the system is assumed to be perfectly mixed, otherwise vertical stratification occurs. Alternatively, on this screen, users may enter the specific dates of stratification for their modeled systems. (A time-series may be entered. On each date where a “1” is entered, the system will stratify, if not already stratified. On each date where a “0” is entered, the system will undergo overturn and become completely mixed again.)
See section 3.4 of the technical documentation for more discussion of stratification and mixing.
The
“Simulation Setup” screen specifies important parameters about the simulation itself
(as opposed to any of the components of the simulation.) To access this
screen, click on the Setup button on
the main window.
At the top of the Simulation Setup
screen you can modify the first and last days of the simulation.
This defines the simulation period.
Default model behavior is to use a
variable step size Runke Kutta as described in section 2.1 of the technical
documentation. The Relative Error
is the acceptable error in the simulation; if it is not achieved in a
particular time step, the variable Runge-Kutta routine decreases the step size
and tries again. If the relative error is too large, the results may be
erroneous; if it is too small, the run time may be too long. Usually a value
between 0.005 and 0.0005 is appropriate, but you might wish to experiment for a
particular application.
A fixed step size may also be selected with time step varying from
one tenth to one hundredth of a day.
This may be useful when a user is precisely comparing “control” and
“perturbed” runs and so wishes to have their stepsizes be precisely
aligned. More information about this can
be found in section 2.1. of the technical documentation.
Daily and Hourly Simulations: A user may set the native model time step to one
day or one hour. If a daily simulation is utilized, average light
conditions are utilized throughout the day. If an hourly simulation
is selected, solar radiation is calculated as variable during the course of
each day. Rather than calculating daily average oxygen concentrations, if
an hourly simulation is selected, AQUATOX will simulate hourly average oxygen
concentrations within the water column. These concentrations will be
based on the hourly light climate (and optional hourly oxygen loadings).
These hourly predictions will then be used to calculate lethal and
non-lethal effects due to low oxygen. The user can output hourly values
of DO or set a larger Data
Storage Step and examine the minimum and maximum predictions
over that time interval. Hourly inflow
loadings can be input for all nutrients, carbon dioxide, oxygen, inorganic
suspended sediments, TSS, light, and organic matter.
Selecting the option to Model Nutrients Internally will
populate the model with additional state variables that track changes to
internal nutrient concentrations (N and P) in non-rooted plants. See Internal
Nutrients in section 4.1 of the AQUATOX Technical Documentation for more
information about this optional model.
Concentrations of internal nutrients will be output in units of ug/L or
the “N to org” and “P to org” ratios may be viewed as they change over
time. Some additional plant parameters
are required, see the section on the “Plant Data
Screen” in this document.
If the user chooses to “Run Model in Spin-up Mode", initial
conditions for biota will be set based on the model results on the last
time-step of the simulation. This can be useful in setting biotic initial
conditions when there are insufficient available data. As initial conditions will be automatically overwritten
it is best to first save a simulation before running in "spin-up
mode."
If the model is set to “Also spin up nutrients, suspended, and bed
sediments”, then all nutrients suspended and bed sediment initial
conditions are overwritten based on the spin up including initial conditions
for diagenesis bedded sediments. Organic
matter is also part of this enhanced spin up.
The four modeled suspended and dissolved organic matter compartments (see
section 5.1 in the Technical Documentation) are summed at the end of the simulation
period and copied into the AQUATOX interface as the initial condition for
“suspended and dissolved detritus.” (Note:
If inflow organic matter data are specified in CBOD or TOC units, the spin-up procedure
converts the AQUATOX final-step results from organic matter into the CBOD or
TOC units specified in the interface.)
With regards to nutrients, if TN
or TP initial conditions were initially utilized, these are replaced with TSP,
ammonia, and nitrate initial conditions at the end of the spin up. In this case, the "initial condition as
TN" and "initial condition as TP" checkboxes become unchecked.
By default, nitrogen fixation in
blue-green algae is assumed to occur when nitrogen concentrations are less than
half of the KN value (the
half-saturation constant for nitrogen). AQUATOX also provides an option
to trigger nitrogen fixation as a function of an editable inorganic N to inorganic
P ratio which may be selected and specified on this screen.
Ammonia is a driving variable may now be checked.
In this case, rather than solving AQUATOX differential equations,
ammonia time-series data may be used to drive model predictions. This is appropriate when the user is not
interested in nutrient dynamics but wants to specify ammonia conditions by
observed data rather than kinetic simulation.
If you wish to compute
steady-state BAFs, you might want to keep the freely dissolved toxicant
constant; this was done in an application concerning PCBs in Lake Ontario
(see Validation Studies document). Otherwise toxicant mass is balanced on
the basis of loading, partitioning, washout, and other toxicant loss terms.
The chemical concentration in
bottom-sediment organic matter can become a driving variable when “Tox. in Sed. Detr. is driving var.” is
checked. This allows the user to specify
a time-series of sediment chemical concentrations to drive the model and
examine the extent of bioaccumulation and effects on the ecosystem.
Toxic effects may be calculated
based on internal or
external concentrations. See Chapter 9 of the Technical
Documentation for more information about the consequences of this choice.
When calculating effects based on
external concentrations, internal concentrations may be less important.
For this reason, a user may wish to estimate
toxicant uptake based on BCF rather than using the default
kinetic model of toxicant uptake. In many cases this will speed up
simulation run time significantly.
If “T1 is an aggregate of all other toxicants in study” is checked,
then the organic chemical in compartment T1 is used to aggregate all of the
other chemicals being modeled as discussed in section 8.14 of the Release 3.2 Technical Documentation.
Older literature often does not
distinguish between freely dissolved contaminants and those complexed with
dissolved organic matter. You may choose to include the complexed
contaminant in computing BAFs so that the results are directly comparable
with the older literature values.
The Data Storage Step
represents the time period over which results are averaged. This
obviously can have a significant effect on the amount of output that is
produced by the model.
In plotting output for stratified
systems it is usually more pleasing to plot continuous values for the
hypolimnion, even when the system is not stratified. This is done by
duplicating epilimnion values for the hypolimnion when the system is well mixed
(by selecting the Write Hypolimnion Data When System not Stratified
option); however, this option takes additional storage, so you might choose not
to duplicate those data points, especially in systems that might not undergo
stratification.
If you click on Show
Integration Info. you will be able to see what time steps are used in
solving the differential equations and which rates and associated relative
errors are causing the integration to slow down while the model is
running.
The default method of averaging
results is by Trapezoidal
Integration, which calculates the average value that has
occurred since the last data storage step. A user may also bypass
integration altogether and “Output Instantaneous Concs",
that is to say, predictions that occur exactly at a given time-step.
For more information about the
other buttons on the screen, please see the Rate
Output Screen, Uncertainty Setup Screen, Control Setup
Screen, or
Output Setup Screen sections of
the help file.
In the Setup Screen, choose the Save
Rates radio button and click on Rate Specifications to designate
those state variables for which you want the additional output.
From this screen, you may
save state variable “rates” to be plotted in the output window. Each
element of a selected state variable's derivative will be integrated and saved
along with the results for the simulation. They can then be graphed or
viewed in tabular format.
Units for rates are
"percent" which is short for "percent of state variable
concentration per day." The only
exception is for limitations to photosynthesis (variables ending with “_LIM”)
which are expressed on a scale of 0.0 to 1.0 with 0.0 representing complete
limitation of photosynthesis and 1.0 representing no limitation to
photosynthesis.
For Release 3.2 two new light
limitation variables are available to signal the occurrence of photoinhibition (too
much light) as opposed to insufficient light.
When low-light limitation causes light conditions to be sub-optimal then
the LowLt_LIM is equal to the overall light limitation and the "high-light
limitation" is set to 1.0. When photoinhibition is occurring then the HighLtLIM
is equal to the overall light limitation and "low light limitation"
is set to 1.0.
AQUATOX can be run using point
estimates for all available parameters (deterministic
mode). Alternatively, a Latin Hypercube uncertainty analysis
can be utilized, or a nominal range sensitivity
analysis. In the “Uncertainty
Setup” screen, choose which mode to run the model in using the radio buttons
shown below.
Deterministic mode is the default mode for the AQUATOX model. In this case, a single simulation is used, and all parameters are represented by point-estimates. No further screen selections are necessary in this mode.
Uncertainty mode is an alternate mode for the AQUATOX model. This powerful feature can perform uncertainty or sensitivity analysis to provide probabilistic results. Latin hypercube sampling is performed, ensuring that all parts of the chosen distribution are sampled. Therefore, the number of iterations can be kept to a minimum, which is important because each iteration is a complete simulation. Twenty iterations is the default, meaning that the distribution is divided into 20 segments for purposes of sampling, and that twenty simulations will be run (besides the initial "deterministic" run). The number of iterations should be increased as the number of involved parameters increases.
The Uncertainty Setup screen is accessed from the Run in Uncertainty Mode radio button. This screen enables you to view all of the parameters and loadings that can be chosen, either singly or in combination.
Nearly all AQUATOX parameters and inputs can be modified in an uncertainty analysis. Click on the "+" or "-" symbols to expand or contract the various portions of the tree interface. Selecting the top branch of the tree interface allows you to view all possible distributions that could be utilized. These are the parameters that can be modeled as a distribution rather than a point estimate. Alternatively, distributions can be sorted by parameter type or by state variable. Double-click on a parameter to open the “Distribution Information” screen and select the Use Above Distribution radio button. The bottom branch of the tree interface shows the distributions that are currently included in the model's uncertainty run.
You may choose to utilize a non-random seed if you wish to be able to reproduce your simulation's results. A pseudo-random number generator requires a "seed" to start its sequence of "random" numbers. A random seed will be based on the exact clock time in which the simulation was started and this sequence will therefore not be reproduced in later simulations. A non-random seed will produce a repeatable sequence of random numbers.
When an uncertainty
simulation is run, a log file (text file) listing the full set of parameter
draws and the timing of model simulations is also produced. When Save
Each Iteration to CSV is selected the full set of AQUATOX outputs,
for each iteration, will be saved to a separate CSV (comma separated variable)
file that can be examined later using Excel or other software. This
allows a user to see the full set of time-series results associated with each
set of parameter draws. When viewing uncertainty output, statistics
regarding results at each time-step are plotted as opposed to individual runs.
A non-random set of
draws may be utilized so that parameters are sampled at even intervals over
their pre-defined range (also called a “statistical sensitivity analysis”).
This approach also can assist in multivariate sensitivity analyses in which
combined effects of parameter changes can be evaluated. If the Sample
Randomly within Intervals option is not selected, the uncertainty analysis
will sample the minimum probability for a distribution, the maximum
probability, and will also sample at even intervals in between. For example, a uniform distribution ranging
from 0.0 to 1.0 when sampled non-randomly with 11 iterations will produce the
following set of samples (though not sequentially ordered) [0.0, 0.1, 0.2,
0.3 ... 1.0]. For non-bounded distributions (normal
and log-normal) the minimum and maximum probabilities are estimated as (1/num_iterations)/2
and 1.0-(1/num_iterations)/2 respectively.
When uncertainty analysis is run on a multi-segment model,
choose either parameters that are relevant to all segments (animal or plant
parameters) or individual segments (initial conditions and loadings to those
segments) using the drop-down box that appears at the upper left of the
screen. A different set of parameters
will then be displayed depending on which segment has been chosen.
Correlations between modeled distributions may also be modeled, though this is not yet enabled for linked-mode uncertainty analyses.
When you have left this screen and choose to run an
uncertainty analysis, you will be prompted for a database file name in which to
store the results. Based on all of the iterations, the minimum, maximum,
mean, deterministic, and standard deviation result for each data-storage step
will be saved in these database files. (Sometimes uncertainty results
will not fit in one database file and will need to be added to one or
more additional successively numbered database files.) For more
information on viewing these files, see Viewing
Uncertainty Output.
To turn off the uncertainty portion of AQUATOX, select the "Deterministic Mode" radio button, below the list of distributions.
Also see section 2.5 of the technical documentation for more discussion of the AQUATOX uncertainty analysis.
This screen is accessed from the “Uncertainty Setup” screen, by double-clicking on a parameter to open the “Distribution Information” screen, and then selecting the Use Above Distribution radio button.
This screen shows whether a
parametric distribution has been chosen for a given variable or whether a
point estimate is used. The radio buttons at the bottom of the
screen are used to toggle back and forth.
The default is a normal
distribution with a mean of the point-estimate parameter value and a standard
deviation of 60% that value. In most cases this will need to be
overwritten with parameter-specific information.
A triangular, uniform, normal, or
lognormal distribution may be chosen and the nature of the parameters changes
based on the selected distribution. Graphs of the distribution will then
be produced and can be used to ensure that the parameters chosen are producing
the desired distribution. Please check these graphs (probability and
cumulative distribution) carefully.
Often environmental loadings are
well represented by lognormal distributions. If less is known about the
distribution, but minimum and maximum values and some central tendency can be defined,
then a triangular distribution may be appropriate. If only minimum and
maximum values are known, then a uniform distribution may sufficient, and the
number of iterations can be decreased.
Note for the lognormal
distribution that the parameters are "mean" and "standard
deviation" rather than geometric mean (GM) and geometric standard
deviation (GSD). To convert a GM or GSD to the required parameters, take
the natural logarithm of the GM or GSD.
Correlations
This screen is accessed from the “Uncertainty Setup” screen, by clicking on the Set Up Correlations button on the lower-right side of screen. You must have multiple distributions selected for uncertainty analysis before accessing the correlations screen.
This screen allows a user to specify correlations between two or more input distributions. All selected distributions for the uncertainty analysis are listed at the top of the screen. The user may then add a correlation using the Add a Correlation button. The user can also edit existing correlations using the Edit Selected Correlation button, or remove a correlation from the list of correlated distributions using the Remove Selected Correlation button.
Within AQUATOX, input distributions for uncertainty analyses are correlated using the method of Iman & Conover (1982). The Iman and Conover method induces a rank correlation between two randomly derived (stochastic) distributions such that the correlation of the two randomly derived distributions matches the correlation of the input data. Output distributions can be examined to verify that they closely match the correlation coefficient of the input distributions.
For the Iman and Conover method to work, correlation matrices must be "Positive Definite." In the authors' experiments with input matrices, we have found this to mean three things in practice:
To examine the input distribution correlation matrix and test for positive definite status, a user can use the Summarize Correlation Matrix button. Unless an error is indicated, the output will summarize the user-specified correlation matrix followed by the rank correlation of draws produced by the correlation method. To remove all correlations, the user can use the Remove All Correlations button.
For more information on this method, please see Iman and Conover (1982)
“Sensitivity” refers to the variation in output of a mathematical model with respect to changes in the values of the model inputs (Saltelli 2001). It provides a ranking of the model input assumptions with respect to their relative contribution to model output variability or uncertainty (U.S. Environmental Protection Agency 1997).
The Sensitivity Setup screen is accessed from the Run in Sensitivity Mode radio button. This screen enables you to view all of the parameters and loadings that can be chosen, either singly or in combination.
AQUATOX can automate a nominal range sensitivity analysis (also known as a "one-at-a-time" sensitivity analysis). This analysis assumes independence among tested parameters.
As was the case for the uncertainty analysis, nearly all AQUATOX parameters and inputs can be tested for sensitivity. Selecting the top branch of the tree interface (the "plus" next to All Distributions ") allows you to view all possible parameters that can be tested for sensitivity. Alternatively, parameters can be sorted by parameter type or by associated state variable. The bottom branch of the tree interface shows the distributions that are currently included in the model's sensitivity analysis. (Click on the "+" or "-" symbols to expand or contract the various portions of the tree interface.) If you double-click on any of these parameters, or press <Enter> while the parameter is highlighted the parameter will be selected (or deselected) for the sensitivity analysis.
To run the model in sensitivity mode, the user must
Choosing a large percent to vary could result in some parameters being set to an unreasonable value. For example, a 20% change in temperature would be highly unlikely except in the case of thermal discharge. Large percentage changes to log Kow for organic chemicals might also be unlikely.
When the model is then executed, the user will be prompted for an Excel file-name in which to save model results. The model will run a deterministic run and then will run the model twice for each parameter selected (testing sensitivity to changes in each parameter in the positive and negative directions in the magnitude of "percent to vary.") In this manner, if a user selects n parameters, the model will run (2n+1) iterations. It should be noted that the number of output variables tracked will not appreciably affect the simulation run time. However, the number of tested parameters obviously determines the number of iterations and therefore the total simulation time.
One important note about sensitivity analyses: Sensitivity analyses will calculate the percent difference between the results of the deterministic run and the modified simulation in the last averaged time-step of the simulation. In most cases the user will want to utilize a much longer averaging period within the model setup screen, for example, to calculate averaged sensitivities over the last year of the simulation or the entire simulation. If sensitivity at an intermediate date is desired, the simulation should be shortened so that the intermediate period is reflected in the last averaged time-step of the simulation.
Periphyton and phytoplankton may be linked within the AQUATOX model. If the "Link Periphyton/Phytoplankton" check-box is selected, whenever a linked periphyton or phytoplankton parameter is varied, the same parameter will be varied within the linked algal state-variable by the same percentage.
When sensitivity analysis is run on a multi-segment model, choose either parameters that are relevant to all segments (animal or plant parameters) or individual segments (initial conditions and loadings to those segments) using the drop-down box that appears at the upper left of the screen. A different set of parameters will then be displayed depending on which segment has been chosen. You may choose to vary any number of global or segment-specific parameters in a single sensitivity run. Excel files of output for each segment will be produced with the segment ID appended.
After the run is complete, the output file can be viewed in Excel. Additionally, AQUATOX can automatically produce tornado diagrams of the results. See Viewing Sensitivity Output for more information. Also see section 2.4 in the technical documentation.
There
are two separate banks of memory for model results, one which is labeled
"control" and one which is labeled
"perturbed." The default is for the control simulation to
have all organic toxicants zeroed out or omitted. However, there is
considerable flexibility in setting up the control run. For example,
toxicants can be kept and point-source nutrients can be omitted in the control
run. In fact, it is possible with a few judicial choices to set up a
factorial analysis to determine the effects of various combinations of
pollution control scenarios.
If none of the check-boxes are
selected, then the results of the control and perturbed simulations will be
identical. A user could, however, first run a "control"
simulation, then change parameters or loadings in the simulation as a
"perturbation," and finally run a "perturbed"
simulation. In this manner, the control and perturbed sets of results can
be used with considerable flexibility, without even utilizing the
"control-setup" screen.
Due to the multi-threaded nature
of AQUATOX, it is often useful to run the control and perturbed simulations
simultaneously, especially with multi-core computer processors.
The
“Output Setup” screen is accessed from the Output
Setup button on the “Setup” screen
This output screen allows the user to select which output from AQUATOX is
saved or not saved.
Because AQUATOX has the potential
for modeling so many state variables, a user has the option of not saving given
state variables, thus decreasing the memory requirements. This can be
especially important when running simulations on a decadal scale or when
outputting many data-points per day. However, given the RAM capacity of
modern computers, this screen rarely needs to be utilized.
Use the "<" and
">" buttons to move results into the "Results to Track"
or "Results NOT to Track" lists.
If "Save PPB Data" is
unchecked, the model will not save concentrations of chemicals in biota with
PPB units.
Note, when stratification is
likely to occur, assume that RAM memory utilization will roughly double.
To run an AQUATOX simulation, select the "Control" or "Perturbed" button in the main window.
The model will run for the amount of time specified in the setup screen. When the model has completed you can use the output button to view the model results.
You may work with the model while it is running and, for example, change parameters or open different models. If you change parameters though, AQUATOX will warn you when it is copying results back to the interface as your model results will no longer match the parameters or loadings associated with that study. In this case you will be able to optionally save the model results to a new file in which the model parameters will match the model output.
A powerful feature of AQUATOX is that it can run paired simulations for perturbed and control conditions. You can use the control setup screen to specify the differences between these two types of runs, or you can change parameters or loadings in between running the two different types of simulations.
The graphical output screen allows users to compare control and perturbed graphs and also to display difference graphs in which percentage differences are displayed.
The “Output” screen is accessed from
the Output button on the main screen.
AQUATOX has extensive capabilities
for graphical and tabular output. There are four tabs available at the
top of this window, each representing a different form of output. Click
on the tabs at the top of the output screen to move from one type of output to
the next.
The default window is the "Graph Library" window.
Graphical output is often the most useful manner to quickly view and
interpret model output. The AQUATOX graphing capabilities are quite
powerful as described in the following sections.
The first two tabs "Perturbed Simulation"
and "Control
Simulation" provide tables of output for the control and
perturbed results. The "Graph Library"
tab provides graphs of output for the control or perturbed results. The “Uncert. Sensitivity” tab on the right
leads to a window in which uncertainty and sensitivity results may be examined.
Within the output window you may
also save the current bank of results and the "graph
library" along with the current set of parameters using the "Save These Results" button.
This is no different from exiting the output window and saving the
existing "APS" file. Results may also be loaded from a
different study using the “Load Results
from File" button.
To determine which graphs have been produced for the given simulation, click
on the triangle at the top left of the graph. If there is more than one
graph included, a drop-down list of graphs will be shown. The user can select
a graph within this list or use the up and down arrows to the right of the list
to scroll through the graph library.
Each graph may be displayed in perturbed mode, control mode, or difference mode. By selecting from the drop-down menu next to the graph list, the user may toggle between these three modes.
The control graph will show a graph of the results saved as the "control" run and the perturbed graph will show the "perturbed" simulation's results.
One of the most useful forms of output is the Difference Graph, which plots the results of the perturbed simulation minus the control simulation as percent differences. The default minimum and maximum percentages are 400% and -100% (complete loss). However, these default extremes can be changed by using the "Edit" button.
If you click on any of the series shown on the graph, a pop-up box will appear that will tell you the value, the date, and the units for the point on which you clicked. For the difference graph it will describe the effect of the perturbation on that variable.
To add a new graph to the list of graphs in the graph library select the "New" button. To edit the existing graph select the "Edit" button. Many additional options for graphing are found in the Change Graph Variables screen which is activated by either of these two buttons.
When the "New" button is pressed, a user has a choice of creating a custom graph or creating from a list of pre-produced graphs, for example:
The custom graph will create a new blank graph that the user can modify using the Edit Graph interface.
By clicking on the "Edit" button above a graph, one can get to this Change
Graph Variables screen.
At the top of this window is the
capability to edit the graph's name. This affects how the graph appears
in the drop-down list.
Next, choose the output to place
on each axis. There are potentially hundreds of output variables
associated with each simulation. The filter tools help the user pick out
relevant variables from this considerable list (see the topic Selecting from a List of Output.)
State variables are organized in
order of trophic level, starting with organic matter and working upward through
plants, invertebrates, and fish.
Types of AQUATOX Output (in
order of output list)
•
Concentrations of
State Variables
–
toxicants in water
–
nutrients and gasses
–
organic matter,
plants, invertebrates, fish
•
Physical
Characteristic State Variables
–
water volume,
temperature, wind, light, pH
•
Mass of Toxicants
within State Variables (normalized to water volume)
–
T1-T20 in organic
matter, plants, invertebrates, and fish
•
Additional Model
Calculations
–
Secchi depth,
chlorophyll a, velocity, TN, TP, CBOD
•
Biological metrics
–
% EPT, Chironomids,
Amphipods, % Blue-Greens, Diatoms, Greens, Gross Primary Production, Turnover,
Trophic State Indices
•
Sediment diagenesis
state variables
•
Toxicant PPB
–
T1-T20 (PPB) in
organic matter, plants, invertebrates, and fish
•
Nitrogen and
Phosphorus Mass Tracking Variables
•
State Variable Rates
–
These include limitations
to photosynthesis
•
Bioaccumulation
Factors (BAFs) for all organisms
•
Uptake Rates (K1),
Depuration Rates (K2), and Bioconcentration Factors (BCFs), for each animal in
the simulation
•
Uptake, Depuration,
and Bioconcentration Factors
•
Observed data imported
by user
These output variables facilitate
detailed analyses of simulated responses. Mass loadings and losses and mass
balances are output for nutrients. K1, K2, and BCFs are output for toxicants.
The ability to output in tabular and graphical form all the state-variable
rates along with the limitations to photosynthesis is especially powerful.
The user can graph variables on
either one or two Y axes. Use the button under the results list to
toggle between one or two Y axes. Use the "<"
and ">" buttons to move results into and out of a
particular axes' results. The ">>" button moves
all variables into the selected axis. Note that all variables on the same axis must have the same
units.
The panel Graph Date Range in
the upper right sets the X-axis range and changed from the AQUATOX simulation date
range by typing dates in the edit boxes. Selecting the "reset" buton above the date range
will set the graph date range to the dates of available model results.
Scaling for each axis (Y1 and Y2) is
set in the “Y1 Axis Scale” panel and
may be entered manually or automatically calculated by AQUATOX. Under the
axis scale entry is A check-box option to use a Logarithmic Y Axis.
The check-box to Auto Scale to Control & Perturbed will scale the Y axes
to the minimum and maximum of both
the control and perturbed simulations (if they both have been run). This
enables the user to switch back and forth between control and perturbed graphs
while looking at axis-scales that remain constant between the two graphs,
enabling easier comparison.
By using the Graph Type drop-down
menu, you can choose between the four types of plots:
If the system has stratified
results, you can choose the vertical layer that you wish to display results for
by using the Epilimnion and
Hypolimnion radio
buttons. This may also be toggled in the graph menu.
If a duration or percent exceedance graphs is displayed, an check-box
option to “Repeat averaging period for all years in the simulation” is offered
in the change graph variables window. This
option will select data each year based on the starting and ending month and
day regardless of year. In the example
above, all data from June 1st to August 30th of each year
will be included in the percent exceedance or duration graphs.
In addition, on the graph itself, a
"Draw Threshold"
button is displayed on the lower right of the graph. When pressed, a
dialog box comes up that both displays duration of exceedance above a threshold
and draws a line on the graph.
There are potentially hundreds of
output variables associated with each simulation. To help the user pick
out relevant variables to be graphed, the "Filter By Substring"
radio button is available at the top of the screen. Typing a string in
the box provided limits the list of variables to those with names that include
the string. For example:
Type "peri" to find all periphyton variables
Type "mg/L" to find all variables with units of mg/L.
Type "ppb" to find all parts per billion output.
Type "detr" to find all results that pertain to detritus.
Type "T1" to find all results that pertain to the first organic
toxicant.
You can also exclude any results containing the substring
by checking the "Exclude Substring" box.
Use the "<"
and ">" buttons to move results into and out of the
"Available Results" and "Results to Display" columns.
Using "<<" and ">>" moves all
results into or out of the relevant column.
At the top right of the “Output” screen is a menu that may be accessed as a drop-down called “Menu” if it has not been used previously. After the menu has been used once, users may repeat their last action by clicking the "(repeat)" button that appears. Note that this menu can also be accessed by right-clicking on any AQUATOX graph.
The copy option allows a user to copy the current graph as a bitmap or a Windows metafile so it can be copied into presentations, documents, logs, or reports. Bitmaps are generally smaller files, but metafiles give better resolution.
The print option allows the user to print the current graph. A "preview" window will first appear allowing the user to select printer, orientation, margins, etc.
The graph-setup menu allows the user to change the colors, symbols, default text, and fonts used on a graph through a dialog box (see Graph Setup Screen topic).
Erase graph will permanently remove the current graph from the graph library (after prompting).
If there are results for both the epilimnion and hypolimnion segments, this menu option toggles between the two segments. Currently, in linked mode, to view another segment, the user must first go into that segment and then open the output window for that segment.
Allows the user to import observed data (see Importing Observed Data topic)
This option enables a user to write a set of Steinhaus similarity indices in Microsoft Excel format. The Steinhaus index (Legendre and Legendre 1998) measures the concordance in values between two samples for each species. A Steinhaus index of 1.0 indicates that all species have identical biomass in both simulations (i.e., the perturbed and control simulations); an index of 0.0 indicates a complete dissimilarity between the two simulations.
A time series of indices is written for each day of the simulation representing the overall species similarity on that date. Four separate indices are written out for all plants, all animals, invertebrates only, and fish only. Completed control and perturbed simulations (of the same length) must have been completed for this option to work.
AQUATOX will calculate and display trophic state indices (TSIs) for lakes using the methods identified in the EPA Nutrient Criteria Technical Guidance Manual (http://www.epa.gov/waterscience/criteria/nutrient/guidance/lakes/lakes.pdf). Separate TSIs are calculated based on average concentrations of Total N, Total P, and Chlorophyll a, as well as average Secchi Depths. The date range over which values are averaged may be specified in the TSI calculation screen as well as the AQUATOX scenario (control or perturbed) and the vertical segment (epilimnion or hypolimnion, if relevant).
All graphs will be exported into a new Microsoft Word document in Windows metafile" format. When this option is selected, Microsoft Word will start up and each of the graphs will be pasted into a new document. Microsoft Word 2000 or later must be installed on the user's system for this option to work properly. Note, the Windows metafile does not always work properly when computer monitors are in "portrait mode," due to what is apparently a minor glitch in Microsoft Windows (fonts become distorted).
When dealing with a large number of graphs, it can be very useful to arrange them in a particular order. This also affects the order of the graphs when they are exported to Microsoft Word. When the "Arrange Graphs" menu item is selected, a dialog box appears showing all available graphs for the simulation. The user may drag and drop the graph names in the list to reorganize them or use "up" and "down" arrows next to the list to reorder them. Graphs may also be deleted using a "delete this graph" button or by pressing the <delete> key when a graph is selected. All changes made in this window can be undone using the "Cancel" button.
The statistics dialog allows the user to examine and export a series of statistics based on the current graph. (See Output Statistics).
This screen is accessed by
choosing “Graph Setup” from the drop-down menu on the right of the “Output”
screen.
This screen allows a user to
change the appearance of the current graph. The graph can be shown in 3D,
and various grid-line options can be selected. The 3D option is somewhat
flashy but not particularly flexible and of limited utility.
Additionally, captions, text, and
font can be edited to personalize the graph. Uncheck the
relevant "Use
Defaults" check-box and edit the text under "Custom Captions."
Fonts may be modified using the “Font” buttons.
Finally, by selecting a
data-series from the "Series
Specific Characteristics" list, a user can select which
color, line-thickness, symbol size, and symbol represents the selected
series. The "SmallDot" symbol renders like a thin line, and is
the AQUATOX default. Sometimes it is useful to change the symbol or line-thickness
when it is desirable to emphasize one of several series on the graph.
All of the Graph Setup changes
will be saved along with the graph when the study (APS or ALS) file is saved.
This screen is accessed by
choosing “External Data” from the drop-down menu on the right of the
“Output” screen.
External "observed" data can be imported into AQUATOX for graphing against model results. After a data series has been imported, it will be displayed at the bottom of the list of data available for graphing in the edit graph dialog. You can also use the text filter within that dialog to find your observed data.
To add a series, use the Add button; the Delete button deletes the selected series.
When a new series has been added, you may edit the series name and the series units. The series units are important as you may only put items with precisely the same units on a single Y axis. So copy the text format of any units that you wish to plot against your imported data precisely. A reference or comment may be added below in which the source of the external data is identified.
Imported series may contain error bars, in which case vertical bars will be plotted against every point in which error bars are included. Imported series may also contain non-detects in which case red arrows will be plotted at the location of the value (which is assumed to represent the detection limit). You can also flag whether the series belongs within the Hypolimnion segment.
After setting these flags, click the Import Data button.
The interface to import external data is almost identical to the one used to import loadings for state variables from the “Main” screen.. The significant differences come because non-detects and error bars may be imported. To change the file type to import, click on the drop-down menu under "List Files of Type." Directions about the precise file-format required will be found in the rectangle at the bottom of the screen.
Generally, the date appears in the first column of the file, with the values to be imported in the second column; however when importing from Excel you can change the column designations. If you are importing non-detects, check the yellow "Import NDs" check box. Non-detects appear in the third column. If a value is not detected in the greater-than direction, use the ">" identifier, otherwise any non-blank value will be interpreted as a "lesser-than" non-detect which is much more common.
If you wish to import error bars, check the "Import Err Bars" check box. Minimum bar values will be imported from the fourth column with maximum bars imported from the fifth column.
After data have been imported, they will be displayed along with their dates in a table. This table can be edited by double-clicking on a cell and editing it. Additional rows may be added or rows may be deleted from the table using the directions that appear below it. (<Ctrl><Del> and <Ctrl><Ins>)
When you have finished importing and editing your observed data series, select the OK button and it will appear in the list of output available to graph. The external data, and any plots created with it, will be saved and loaded along with your AQUATOX simulation.
This screen is accessed by choosing “Statistics from Graph” from the drop-down menu on the right of the “Output” screen.
This dialog box enables the user to examine the following statistics that pertain to AQUATOX Output:
The date ranges, scenarios to export statistics for (i.e. "Control" vs. "Perturbed,") and the vertical segment (if relevant) may all be edited and the statistics will automatically update to reflect the changes from the various inputs and drop-down menus.
The “Repeat Averaging Period for all years in the simulation” checkbox allows you to export results on a seasonal basis. If this is selected, a user may put in a date range for a single year (e.g. “5-1-2011” to “9-1-2011”) and the model will extract results between those calendar dates regardless of the year selected. (The year chosen when inputting the seasonal date range is not important.)
AQUATOX displays these statistics in a fixed-width basis and they may be copied and pasted to any spreadsheet or database software. The width of numbers and the width of the output names can also be varied using the entry boxes at the bottom of the screen. Statistics can also be exported to Excel using a button at the bottom of the screen.
The derivation of statistics from external "observed" datasets is not enabled at this time.
Tables can be obtained for both
perturbed and control runs. Click on the “Perturbed Simulation” and
“Control Simulation” tabs at the top of the output screen to move from one type
of tabular output to the next.
Use the “Change Variables” button
to specify what variables are to be tabulated.
Use the “Save Table to Excel” button to
save the results to an Excel file. This button requires Microsoft Excel to
be installed on the user's machine to work properly. After the export is
completed, the user is prompted as to whether they wish to view the Excel file
immediately or just leave it saved to disk in the specified location.
This screen can be accessed from
the “Uncert. Sensitivity” tab on the
upper-right of the “Output” screen.
An additional form of
graphical output is the Uncertainty Graph. If an uncertainty
analysis has been performed, the results can be potted as a series of lines
representing the mean, minimum, maximum, mean - 1 standard deviation, mean + 1
standard deviation, and deterministic results as tracked through the simulation
time. Only one output variable can be viewed at a time, so click on View
a Different Variable to view another. The multi-purpose uncertainty
window may also used to view the results of a sensitivity analysis .
Uncertainty output is not saved as
part of an APS file but is instead saved in a comma-separated-variable output
(*.CSV) format. Immediately after running an uncertainty analysis, the
uncertainty window will bring up the primary data set file that is associated
with the uncertainty run. However, if the study is closed and opened
again, the pertinent database file will need to be re-opened using the View a Different Database button. This
will also be required if the output variable you wish to graph is in one of the
successively numbered database files.
When many iterations have been
run, the minimum and maximum output become less relevant (existing far out on
the tail of the output distribution, with very little probability of
occurrence.) For this reason, the minimum and maximum scenarios can be
toggled on or off again using the Show
Min & Max checkbox.
Of particular interest to risk
assessors is the Biomass
Risk Graph, which plots the probabilistic results as
"percent probabilities" against "percent declines" by the
end of the simulation. Any number of organisms can be plotted
simultaneously on the Risk Graph, so that the responses of both tolerant and
intolerant organisms can be analyzed. If an organism increases in biomass,
for example because of release of herbivory (for plants) or predation (for
animals), then the percent declines are shown as negative values. Data to
support the biomass risk graph are stored in CSV files with the suffix
"_decline" added to the end of the file-name. A different CSV
file may be loaded using the View
a different CSV file button, though the CSV file that is
relevant to the simulation will be initially displayed.
Uncertainty graphs may be copied
to the clipboard using the "Copy"
button, or printed using the "Print
Graph" button. The "Graph Setup" button
brings up the graph
setup window, allowing the appearance
of the graph to be edited.
The same screen that is used to view uncertainty output can automatically produce tornado diagrams, which are a visual manner of interpreting the results of a sensitivity analysis.
For a particular model output, a tornado diagram displays the parameters that cause the most sensitivity, sorted in decreasing order of sensitivity. Red bars plot results in which the parameter has been reduced by the given percent and blue bars plot results in which the parameter has been increased. The black vertical line in the middle of the blue and red bars represents the baseline model result (in which no sensitivity test was performed.) Clicking on any of the bars will give the user information about which parameter was tested and the result as compared to the baseline result.
The statistic shown in front of the parameter names is the "sensitivity" statistic which is a normalized average sensitivity for that parameter (see the equation in section 2.4 of the Technical Documentation). If a given parameter was varied by 20% in each direction and the output result also varied by 20% (on average) the "sensitivity" for that parameter/output pairing would be calculated as 100%.
Sensitivity analysis results are stored in Excel files. Immediately after running a sensitivity analysis, the proper Excel file will be loaded. Otherwise, the user will need to select "View a Different Database" change the file-type to "Sensitivity Results, (*.xls)" and load the file in which sensitivity results have been saved. If there are many results in the given Excel file loading the data from Excel could take several minutes.
All the model outputs that were tested for sensitivity may be scrolled through, using the drop-down menu at the top left of the tornado diagram. The user may also scroll through these tornado diagrams using the up and down arrows immediately to the right of the drop-down menu.
The number of rungs on the tornado diagram can be changed by changing the "Show ___ Most Sensitive Variables" edit box. For a full accounting of all parameters tested for this particular endpoint and their ordered sensitivity statistics press the "List Sensitivities" button
A user can also set up an alternative tornado diagram in which the effects of a single parameter change on all tracked outputs can be examined. To convert the diagrams, the user should click the "Show Effects" button. A diagram such as the one shown below will be produced. The drop-down menu can now be used to select from all of the tested parameters. Because output variables have many different units, this metric is tracked in terms of "percent change."
Clicking on an individual rung of the diagram will show precisely the percentage change predicted by a given perturbation. As was the case with the sensitivity tornado diagram, the number of effects may be edited and the individual effects can be listed in text format by clicking on the (renamed) "List Effects" button.
Control and perturbed results may be exported to a variety of formats by clicking the Export Results or Export Control buttons on the main study window.
Providing there are available results for export, a dialog box will appear allowing you to choose the output variables you wish to export. After output variables are chosen, the user is prompted for a file-name in which to save the results. Results may be saved in Excel format (*.xls), or Comma Separated Value format (*.csv). To change the format of the export, choose a different format from the "Save as type" pull-down menu.
An alternative means of exporting data is to save tabular output (viewed in the output window) to Excel.
As part of the BASINS linkage, when exporting results, if GenScn is installed on your machine, the "Export all Results to GenScn" button is available. This automatically starts the GenScn program with all available AQUATOX output attached. However, most of the GenScn functions are also now available within the AQUATOX interface.
AQUATOX
study files include results of any control or perturbed simulations that have
been run. When a long simulation has been executed, with a small data
storage step, these results can take up a lot of disk space. To
clear these results from memory, select Study and choose Clear
Results from the main menu bar. The results will be disposed of so that
your study file will be smaller but you will not be able to view output using
the output window until the simulation is run.
When a simulation is open, in the Study menu, there are two options to
"Export to Rel. 2.2
Databases" and "Export to Rel. 3.0 Databases."
This option will export all animal and plant parameters from the given study into
databases with the chosen file-format. The user will be prompted as to
which databases the data should be written to. Any available
database entries with the same plant or animal names will be overwritten as
part of this procedure, so it may be wise to back-up the original databases
before exporting.
When a simulation is open, in the Study menu there is an option to "Take Parameters from Another Study”." This menu option will copy all "non-site-specific" parameters from another study (which is required to have exactly the same state variable list). This procedure allows a user to test a single set of parameters at multiple sites.
Parameters that are copied include:
The user also will be prompted as to whether initial conditions should be copied into the given simulation.
Virtually
every function in AQUATOX can be accessed by clicking on the applicable icon on
the toolbar. For the experienced user this provides a quick way to bring
up a particular screen or to perform a function, such as saving a simulation,
without going through several layers of options in the menu bar or the “big
buttons.” (The big buttons can be suppressed entirely through the View
menu option.)
To determine the purpose of a
given button, hover the mouse cursor over the button for a moment and a
"hint" which describes the purpose of the button will become visible.
Toolbar icons can be added,
deleted, or moved by clicking on Edit Toolbar under the View menu
option. This is also a good way to learn the functions of the
icons. There are 32 icons that are listed and can be used, compared to
the 24 icons shown on the default toolbar. One can also use dividers to
visually group icons representing similar functions.
When
you have entered this screen, you may edit the toolbar by dragging and dropping
buttons onto the toolbar from the available list. Alternatively, you can
drop buttons from the toolbar into the large trash-can icon and they will no
longer appear on the toolbar.
Your custom toolbar will be saved
when you exit AQUATOX and reloaded when you start the program the next
time.
Dividers appear at the bottom of
the button list and they can be used to group tool-button functionality.
You cannot drag dividers from the toolbar into the trash, instead, the Remove
all Dividers button (at the bottom of the tool-button list) must be
selected and new dividers can then be dragged onto the toolbar.
In either linked mode or single-segment mode, a user may choose to "Export Parameters as Text" from the Study menu. This procedure will save all parameters and, optionally, all time-series loadings that make an AQUATOX simulation unique. This procedure can be useful for several reasons:
After this function has been selected, the user will be prompted as to whether to save time-series inputs to the file as well. This will increase the size of the text export, but will completely document the input variables that are driving simulation results. The filename in which to save results will then be solicited. In linked mode, segment-specific parameters and initial conditions for all segments will be included in the text output.
This menu option (in the Study menu) allows a user to pass data from HSPF “WDM files” into AQUATOX. This compliments the existing BASINS HSPF-to-AQUATOX linkage, which required that the user work with WinHSPF (i.e., the HSPF interface contained in BASINS) and then ask WinHSPF to produce special time series used for the linkage. In some recent HSPF-to-AQUATOX linkage applications, the HSPF simulation produced the needed boundary conditions in a form not compatible with the previous WinHSPF-AQUATOX linkage. This required time-consuming manual linkages of time-series, something that this automated linkage attempts to avoid.
This linkage has AQUATOX import, as inflow loadings, “in-stream concentrations” as derived from HSPF as opposed to the boundary condition calculations passed in the original linkage.
There was some concern that HSPF and AQUATOX would both be calculating the same in-stream processes so there could be double counting. However, when passing average-daily loadings into short reaches with low retention times, the HSPF in-stream concentration and the AQUATOX in-stream concentration will be dominated by the inflow loadings rather than in-stream processes. Our testing has indicated that linking HSPF in-stream concentrations as AQUATOX in-flow loadings for such short reaches introduces negligible error.
Additionally, one can design their study such that the AQUATOX boundary condition is represented by the end of, or outflow from, the HSPF reach being linked. This approach would eliminate any potential error from double-calculation of in-stream processes. The assumption would simply be that the well-mixed HSPF reach feeds directly into the AQUATOX reach being modeled below.
Specific Mechanics of the Linkage:
The steps taken by this linkage are summarized below:
1. The HSPF simulation must have been run and nutrients, organic matter and flow outputs must have been satisfactorily calibrated. The results must be in an accessible WDM file.
2. From AQUATOX, the user selects the WDM file, and then the relevant location, and scenario to be linked using the "study" menu (select "Import Data from HSPF WDM").
3. The user selects the date range for the linkage. (Note that this changes the AQUATOX first day and last day of simulation but this can be changed after the linkage is complete.)
4. The user is presented with options for importing phosphate. The following HSPF outputs may be input to maximize flexibility
a. TOTP -- Total Phosphorus in mg/L (Default is to link this to the AQUATOX Total P compartment, and AQUATOX will estimate Total Soluble P from Tot P)
b. Any combination of the below three compartments may be summed and imported as TP or TSP.
i. PO4-P – Ortho P concentration as P in mg/L
ii. PPO4 -- Adsorbed orthophosphate as P in mg/L
iii. TORP-- Total Organic Phosphorus in mg/L
5. The following items are loaded (Hourly data are converted to daily except for oxygen, CO2, organic matter, and nutrients.)
a. FLOW is read as AQUATOX “outflow water,” units are checked to be in cubic feet per second (cfs). If units are not specified the user is told that the assumption is that units are cfs. If different units are specified the user is prompted as to whether to continue or not. Units are converted to AQUATOX m3/d.
b. Average depth is read assuming the units are feet if unknown, converting, and prompting the user as above.
c. Ammonia, Phosphate, Oxygen, Nitrate,TOC/CBOD, and TSS are imported in units of mg/L.
d. CO2 is set to the default value of 0.7 mg/L. This can be changed (or not set) if that is desirable.
e. Temperature is imported and converted to Celsius units (C). if required.
f. For QA/QC, these imports may be examined as compared to the WDM file results.
6. The linkage also imports Chlorophyll a, Benthic Chlorophyll a, and CBOD as "external data" to plot against AQUATOX results (but not to drive the model).
With regards to the units of data being imported, if units are not specified in the WDM file the log-file alerts the user that it assumes they are using a certain type of unit (general HSPF defaults). If units are specified and the linkage finds unexpected units it raises an error.
This new capability can link both BASINS and HSPF, since both simulations read from, and write their data/results to WDM files, which is the primary medium of this linkage.
Following successful linkage a log file
describing all imported data and/or any errors encountered can be examined
through the AQUATOX interface.
The AQUATOX Setup Wizard is
intended to guide the user through the 19 steps necessary to set up a new
simulation. Even experienced users might find it to be a convenient
checklist, providing a measure of quality assurance not usually found in
models. Most steps consist of several parts, and one can move
systematically through these by making choices or entering values, then
clicking on Next.
Wizard
Components:
·
Summary Screen
·
Step 1: Simulation Type
·
Step 2: Simulation Period
·
Step 3: Nutrients
·
Step 4: Detritus
·
Step 5: Plants
·
Step 6: Invertebrates
·
Step 7: Fish
·
Step 8: Site Characteristics
·
Step 9: Water Volume
·
Step 10: Water Temperature
·
Step 11: Wind Loading
·
Step 12: Light Loading
·
Step 13: Water pH
·
Step 14: Inorganic Solids
·
Step 15: Chemicals
·
Step 16: Inflow Loadings
·
Step 17: Direct Precipitation
·
Step 18: Point-Source Loadings
·
Step 19: Non Point-Source Loads
The
Progress Screen lists all the Wizard's steps and shows which step is
currently active.
This sceen also provides a means
of skipping from one step to another by double-clicking on any step in order to
move there.
If a user decides to create a new
AQUATOX study from scratch, the wizard's progress screen also displays the
status of each step (whether the user has completed data entry for that step or
if data gaps still exist for the step).
The progress screen can be hidden
by selecting the Hide Progress button. It can be shown again at
any time by selecting the Show Progress button that is always available
on the main wizard window.
The Summary Screen provides
a list of the state and loading variables as they are changed while going
through the Wizard. This screen is for informational purposes only, to
give the user a sense of how their actions are changing the state variable list
they will see when the wizard is complete. The list is not editable through the
summary screen.
The summary screen can be hidden
by selecting the Hide Summary button. It can be shown again at any
time by selecting the Show Summary button always available on the main
wizard window.
In order to create a simulation from scratch, you must access the Wizard from the File menu (“New Simulation Wizard”). If you have an existing study open, selecting the Wizard “big button” means that you wish to edit that particular study with the wizard.
When creating a new simulation
with the AQUATOX setup wizard, you have the choice to start a simulation from
scratch, or work with a default study. When starting a study from
scratch, every parameter within the simulation will be specified by you or
loaded from the data libraries. Otherwise, a simulation will be based on
an existing simulation and you will have the opportunity to make modifications
to the default parameters.
The name for the simulation will
appear on the main screen and will be used as a heading in the output.
Six site types are available: pond, lake, stream (creek or river), reservoir,
experimental enclosure, and estuary.
This screen allows you to specify
the time period for the simulation.
The time period for the simulation
may be a few days, corresponding to an experiment, or a year, or even several
decades. The time period does not have to correspond to the loadings,
because the loadings can be interpolated automatically. However, it is advisable
to consider the correspondence between the start date and the initial
conditions; if the initial conditions are poorly known then a start date in the
middle of winter might allow the simulation to initialize before going into the
growing season. (This is different than "spin-up mode" from the
model-setup window which sets biotic initial conditions to the value at the end
of the simulation)
Initial conditions for dissolved
nutrients must be entered. Phosphate can be considered as soluble
reactive phosphate; by going into the phosphate loading screen (outside of the
Wizard) phosphate can be adjusted for availability. Because of
interchange with the atmosphere, the model is not very sensitive to the initial
conditions for carbon dioxide and dissolved oxygen.
This set of screens allows you to
specify initial conditions of detritus in the sediment bed and in the water
column.
Labile detritus is readily
decomposed and assimilated, while refractory detritus is resistant to
decomposition.
Initial conditions and loadings of
detritus in the water column can be input as Organic Matter (dry weight),
Organic Carbon, or Carbonaceous Biochemical Oxygen Demand (CBOD) and the model
will make the necessary conversions. Suspended and dissolved detritus
initial conditions and loadings are divided into four compartments: particulate
refractory and labile detritus and dissolved refractory and labile organic
matter. Initial conditions and loadings are parsed by specifying %
particulate and % refractory.
The user is provided with a list
of plants within each taxonomic group from which to choose. These
taxonomic groups can be scrolled through using the "next >>"
and "<< back" buttons. Drag the plant name from
the list provided into the simulation list to include it in the simulation.
After the plants have been
specified, an initial conditions entry screen is reached. Initial
conditions should be entered for each plant group; as with any biotic group, a
value of “0" will keep the group from being simulated. Note that the
units are sensitive to whether the plant is planktonic or benthic.
The list of available plants
corresponds to those species that have entries in the Plant Library. As
the library continues to expand with additional applications, the set of plants
available within the wizard also expands.
The user is presented with a list
of invertebrates for each ecological guild from which to choose. Some are
general taxonomic groups and some are genera and species. These
groups can be scrolled through using the "next >>" and "<<
back" buttons. Drag the invertebrate name from the list provided
into the simulation list to include it in the simulation.
After the invertebrates have been
specified, an initial-conditions entry screen is reached. The initial
conditions are either mg/L or g/m2, depending on the mode of life
(pelagic or benthic).
The list of available
invertebrates corresponds to those species that have entries in the Animal
Library. As the library continues to expand with additional
applications, the set of invertebrates available within the wizard also
expands.
Fish are classified as forage
fish, bottom fish, and game fish. At least two species can be modeled for
each general class. Furthermore, two size classes can be modeled for each
species, and one species can be modeled as multiple year classes.
For each trophic guild (forage
fish, bottom fish, and gamefish) the user can choose from a list of
appropriate species in the database.
After fish species are specified,
an initial conditions screen can be reached by selecting "next >>."
Initial conditions are given as g/m2 because it is easiest to
express biomass on an areal basis.
The list of available fish
corresponds to those species that have entries in the Animal Library. As
the library continues to expand with additional applications, the set of fish
available within the wizard will also expand.
More information on Fish Type and
Fish Class follow.
You can choose to model a fish
species either as a single state variable, as two state variables representing
size classes, or as up to fifteen state variables representing age classes.
Each fish species that is modeled
within AQUATOX must be classified as a forage fish, bottom fish, or game
fish. Fish are also designated as single compartment, size class, or age
class.
If you are modeling this species
as a single state variable you must also choose whether it is of large or small
size. When modeling a species as a single state variable this has little
effect on the model's results but food preferences for any predators of the
species must be set accordingly.
The most important morphometric
characteristic is mean depth because that controls light penetration,
volatilization, and attached plant distribution. Mean annual evaporation
is used for computing the water balance. If evaporation is variable over
the year, that time series can be input in the Site screen, outside of the
Wizard. Latitude is used to compute photoperiod for photosynthesis.
The data entered here is used to
parameterize the site underlying data.
The wall area is important for an
experimental enclosure because it represents additional area for attachment of
periphyton. This is only visible if your site type is an experimental
enclosure.
This screen presents the options
for modeling water flow in an AQUATOX simulation. Depending on the method
chosen, inflow or discharge values may be required. The Manning’s
equation can be used to compute changing volumes in a stream. The
simplest procedure is to hold volume constant at the initial condition.
Volume can be computed dynamically, given the inflow and outflow and factoring
in evaporation. Finally, time series of known values can be entered.
A constant temperature, annual
mean and range in temperature or time series can be entered. Depending on
your selection, you will be required to enter additional information about
temperature patterns in the next screens.
For more information about modeling temperature in AQUATOX see Temperature Data Screen.
Wind loadings can be constant, a
time series can be entered, or a default time series can be used. The
default is a 365-day record taken from the Buffalo NY airport. This
series is represented by a Fourier series, with a mean value that can be
specified by the user (the default is 3 m/s).
For more information about
modeling wind in AQUATOX see Wind Data.
Constant, time series, and annual
mean and range may be given for light in Langleys/day. If annual mean and
range are chosen, the user will be asked for photoperiod, which can either be
computed from the site latitude, or be constant.
When in a simulation, light
loadings may be altered by double clicking on "Light" within the
state variables list. This will bring you to the light loading screen.
The pH may be specified by the
user, either as a constant or as a time series. It has various
effects on simulations (see section 5.7 of the Technical Documentation).
When a time series is utilized the
rules for dynamic loadings apply.
Alternatively, through the
main pH entry screen, outside the Wizard, the user may specify that
AQUATOX should calculate pH values using a simple semi-empirical
formulation (this requires a Mean Total Alkalinity input, see section 5.7 of
the technical documentation for more information.)
There are three options for
simulating water-column inorganic solids within AQUATOX.
When the concentration of total
suspended solids (TSS) is known, TSS can be input to provide a measure of
inorganic solids. The model subtracts phytoplankton and detritus from the
TSS to estimate the inorganic solids; therefore, care should be taken to use
contemporaneous TSS and nutrient time series.
The second option, the
sand-silt-clay model, is only relevant for streams. This option requires
considerably more parameters and simulates the scour, deposition and transport
of sediments and calculates the concentration of sediments in the water column
and sediment bed within a river reach. More information about the
sand-silt-clay model can be found in the section titled Sand-Silt-Clay.
The third option is to utilize the
multi-layer sediment model, though this cannot be accessed through the AQUATOX
wizard.
AQUATOX can simulate as many as 20
different organic chemicals simultaneously. The assumption is that the
toxic effects are additive.
After a chemical has been selected
for a simulation, you may specify initial conditions for that chemical in the
water column and in all biotic and detrital compartments.
The Wizard compiles a list of all
variables that may be loaded as concentrations in inflowing water. The
units are sensitive to each given variable.
Click on one of the loadings from
the "Inflow Loadings in Simulation" box and you will be able to
select whether this is a constant or dynamic loading and edit the details about
the loading. If the “Ignore ALL Loadings for this State Variable”
check box is selected then inflow, direct precipitation, point-source, and
nonpoint-source loadings are all set to zero.
The Wizard compiles a list of all
variables that may be loaded as atmospheric deposition (“direct precipitation”
and dry fall). The units are on an areal basis because deposition is on
the surface of the water.
Click on one of the loadings from
the "Direct Precipitation Loadings" box and you will be able to
select whether this is a constant or dynamic loading and edit the details about
the loading.
Point-source loadings are entered
as mass per day (g/d) to the water body.
Click on one of the loadings from
the "Point-Source Loadings" box and you will be able to select
whether this is a constant or dynamic loading and edit the details about the
loading.
Nonpoint-source loadings are also
entered as mass per day (g/d) to the water body.
Click on one of the loadings from
the "Nonpoint-Source Loadings" box and you will be able to select
whether this is a constant or dynamic loading and edit the details about the
loading.
This screen appears if the user
originally selected to create a study from scratch. In order to do this,
the user must populate all of the parameters within the model.
At this point you have selected to
leave the wizard before providing all of the required parameters. You
must go back and enter values for these parameters or accept default parameters
provided by the program. The wizard
progress screen can help you to determine in which areas you have not entered
data. Arrows appear by the steps in
which you have not entered information.
This screen is accessed by
clicking the “Birds, Mink…” button on the “Main” screen.
This simple submodel, pertains to shorebirds and other terrestrial animals that feed exclusively from the water column. The model will estimate chemical concentrations in tissues based on dietary uptake from food-sources in the water column. Only one “aquatic-dependent vertebrate” may be included in an AQUATOX simulation.
In this screen, the user can specify a biomagnification factor (BMF) and clearance rate for each chemical, and the preferences of the animal for various food sources so that dietary exposures can be computed. Dietary preferences will be normalized to 100% by the model if they do not sum to 100% within this screen.
The predicted concentration of each chemical in the animal is based on the chemical concentration in prey at a given time-step. Uptake of toxicant is assumed to be instantaneous, but depuration of the chemical is governed by the user-input clearance rate.
After a run, when the aquatic-dependent vertebrates model is used, the list of model output will include one or more entries entitled: "[Toxicant in] Birds etc. (ug/kg wet)."
There
are applications where it is desirable to run a series of studies
automatically. This kind of run can be done by creating a Batch
subdirectory under the Study directory and placing studies with appropriately
chosen options in it. The subdirectory should also contain a text file
labeled “batch.txt” that lists the names of the studies to be run, one to a
line. On the menu bar you should click on Run Batch on the pull-down File
window. That will open a window that allows you to Run in Batch Mode.
You also can save the bioaccumulation factors (BAFs) and organic-matter
partition coefficients (KOMs) to a comma-separated text file batchout.csv.
(See chapter 8 in the Technical Documentation for discussion of bioaccumulation
of organic toxicants, BAFs and KOMs.)
Batch mode automatically runs each
simulation in “perturbed” mode. If the “Run
Control Simulations Too” checkbox is selected, then control and perturbed
simulations are both run (with differences reflecting the options selected in
the Control Setup Screen.)
IMPORTANT SPECIFICATIONS FOR
BATCH MODE: |
AQUATOX has a number of command line options that can be used either to start the GUI or to allow model execution without instantiating the GUI at all. Commands and filenames can be passed to AQUATOX to get a desired program behavior. Relative or absolute filenames can be passed to the software.
Note: when running from a Windows CMD prompt, AQUATOX doesn't tell the user that it is running and another command-line prompt instead appears. When the run has completed executing, a message is sent to the command-line window that tells the user of the successful run. A user should check the task manager to see if AQUATOX.EXE is running if they are unsure. When running the model in DOS batch-file mode users can use the START command if they want to run on multiple CPUs, or omit the START command if they want to run sequentially on a single CPU.
The full list of parameters that can be passed to AQUATOX and the resulting program behavior follows:
Command Line Options
·AQUATOX.EXE {FileName} -- open specified file
within the GUI
·AQUATOX.EXE EPERT {FileName} -- open specified file
within the GUI and execute perturbed
·AQUATOX.EXE ECONT {FileName} -- open specified file
within the GUI and execute control
·AQUATOX.EXE EPSAVE
{FileName} {SaveFileName} -- open specified file without the GUI, execute
perturbed, and save to the {SaveFileName} or over the existing version if
{SaveFileName} is not specified
·AQUATOX.EXE ECSAVE {FileName}
{SaveFileName}-- open specified file without the GUI, execute control, and
save to the save file name or over the existing version if {SaveFileName} is
not specified
·AQUATOX.EXE EPEXP {FileName}
{ExportResultsFileName} -- open specified file without the GUI, execute
perturbed, and export results to the specified file name (*.csv or *.xls) or
to a csv file with the same base name as {FileName} if
{ExportResultsFileName} is not specified.
Note, the xls export option is dramatically slower and requires
Excel to be installed as it uses Excel OLE automation.
·AQUATOX.EXE ECEXP -- Same as EPEXP (above) for
the control scenario
BASINS specific options below – called directly by BASINS
linkage
·
AQUATOX.EXE HSPF {GISBaseFile} {WDMFile} {Scenario}
{Location} {AvgString} -- open HSPF linkage file -- designed for and tested with
BASINS 3.2 linkage
o {GISBaseFile} = BASINS
base GIS filename (without the ‘.rch’,’.pdf’, or ‘.psr’ extensions)
o {WDMFile} = path to the
HSPF WDM file
o {Scenario} = HSPF scenario
string
o {Location} = HSPF location
string
o {AvgString} = distinguish
between summed daily data or hourly averages
·
AQUATOX.EXE UNKNOWN {GISBaseFile} -- open GIS linkage file
from BASINS 3.2 linkage -- designed for and tested with BASINS 3.2
Some valid example of command-line inputs follow:
·
aquatox.exe ECSave "..\studies\hcb tank.aps"
·
aquatox.exe ECSave "../studies/hcb tank.aps"
·
aquatox.exe EPSAVE c:\newtemp\tenktest.als
c:\newtemp\tenktest2.als
·
aquatox.exe EPSAVE c:\newtemp\tenktest.als
c:\newtemp\tenktest3.txt
·
aquatox.exe ECSave "../studies/blue earth r.mn.aps"
bec.txt
·
AQUATOX.EXE EPSave
c:\newtemp\hcrun.txt hptestrun.txt
·
AQUATOX.EXE ECSave "c:\newtemp\htest.txt" c:\newtemp\hcrun.txt
·
AQUATOX.EXE EPExp
c:\newtemp\hcrun.txt
·
AQUATOX.EXE ECExp
c:\newtemp\hcrun.txt hcruncexp.csv
·
AQUATOX.EXE EPExp
c:\newtemp\hcrun.txt c:/newtemp/hpruncexp.csv
·
aquatox.exe EPExp c:\newtemp\tenktest.als
c:\newtemp\tenktest2.csv
·
aquatox.exe ECExp c:\newtemp\tenktest.als c:\newtemp\tenktest2.xls
Text based study files contain all of the data in a binary file with the exception of user-specified colors and fonts on saved graphs. The output format is based on JSON (JavaScript Object Notation) format. Each object contains brackets and indents to help define its domain. Each variable is explicitly named in quotation marks followed by the relevant value.
The general order of AQUATOX
data stored within a text-based study file is as follows:
o TAQUATOXSegment
§ TStates (set of state variables within a
segment -- there can be two under stratified conditions)
o TCollection -- holds all of
the state variables in the simulation.
These are iterated through in the same order they are shown in the state
variable list
o 1. Organic
Toxicants
o 2. Nutrients
o 3. Gasses
o 3. Detritus
o 4. Plants
o 5. Invertebrates
o 6. Fish
o 7. Water Volume
o 8. Temperature
o 9. Wind
o 10. Light
o 11. pH
o 12. Organic
Toxicants sorbed to Detritus, Plants, Invertebrates, then Fish
§ These state
variables have SVType > 1 and are listed as “undisplayed” (PName^) as they
do not show on the state variable GUI list.
·
Discharge,
Shade, Thermocline, Velocity data
·
Results
o TCollection --
holds all the headers (THeader Objects) for the results objects including name
and type, etc.
o Collection of “ObjID
1010”-- holds results for a given day in order of the “PointIndex” of the
THeader object
·
Distributions -- for
uncertainty or sensitivity analyses
·
Control
Options Record
·
Graph
Records (defines saved graphs)
·
GullPref - data for
optional Aquatic Dependent Organism model
·
Setup
Record -- data from “setup” screen -- first day, etc.
·
PDiagenesis -- diagenesis
parameters if relevant
·
Uncertainty
Setup Record -- data for uncertainty run if relevant
·
Site
Record
·
Remin
Record
·
Chemical
Records for up to 20 organic toxicants
o Followed by Animal
and Plant Toxicity Records
o
Followed by biotransformation parameters if
relevant
For each state variable, the
following data are saved
o Initial
Condition
o State
(concentration)
o Name (PName^)
o Loading Record –
time series loadings
o Underlying Data
Records (plant, animal parameters)
o Animals – trophic
interaction array
It is worth noting that some
intermediate-calculation variables are saved in the text file that have no
impact on file execution. Some examples
of these are:
·
Differentiation
variables such as “yhold,” “yorig,” “stepres,” “yerror,” “yscale”
·
Normalized
feeding preferences saved in the TPreference object-- the “TrophInt” object is
normalized at the start of each time step as a function of available food.
These intermediate calculations are included in text output because the binary-file saving and text-file saving have the same codebase. (The binary file saving is used to duplicate state variables when systems stratify dynamically so some of these intermediate variables are saved to ensure that there is no problem moving forward.) These intermediate variables can eventually be removed from the text input/output with some additional logic that has not yet been put into place.
Setting up a single segment simulation generally
has the following types of steps
Data requirements in AQUATOX depend considerably on the site-type that is being modeled and the goal of the modeling study. A modeling study to assess the effects of changing nutrient levels would benefit from an excellent accounting of nutrients in the water column and boundary-condition loadings of nutrients. On the other hand, a study that is primarily about chemical effects might not require such detailed nutrient information. In this case, a user might wish to produce the required nutrient environment for a stable food-web and then detailed information about chemical loadings can be applied to the simulation to assess the effects.
Precise accounting of time-series loadings is especially important in rivers and streams. Due to low retention time, inflow loadings can be the dominant factor affecting water column conditions. For standing water, initial conditions might be more important than a perfect representation of daily loadings depending on both simulation time and retention time.
Here is a list of data requirements put together for a single-segment
simulation project. This list is not necessarily exhaustive but
contains what are generally the most important time-series inputs.
LIST OF IMPORTANT AQUATOX TIME-SERIES INPUTS
See Additional Data Requirements for Linked Mode for a more detailed list. Not all loadings need to be precise. However, it is useful to have at least some data for each of the various categories (nutrients, oxygen, organics, inorganic sediments or TSS, etc.)
A user will need to define the biotic state
variables in each segment as well. Default parameterizations may be used
for various organisms’ "Underlying Data" unless there's a pressing
need to use your own site-specific parameters.
If a chemical is modeled, requirements include initial concentrations in biota,
loadings from various sources, and toxicity data if you wish to model effects.
The site type may be set in the site data screen.
The following site-types are included.
When an estuarine model is selected, the modeled segment is assumed to be composed of two well-mixed layers; stratification is governed by salinity. Water balance between the two layers is computed using a salt-balance approach. For this reason, the salinity in each layer at the mouth of the estuary must be specified as a constant value or as a time series. A time series of salinity is especially important for the upper layer as this will permit a reasonable and variable set of water flows to be derived.
Water-column salinity can have effects on animal mortality, gamete loss, and algal photosynthesis, ingestion, respiration, and sinking. Estuary reaeration is also affected by salinity, especially through calculation of the level of saturation at which oxygen saturation occurs. Salinity can also affect the rate of volatilization of organic toxicants.
Various additional Tidal Range Model parameters are required at the bottom of the site "Underlying Data" screen to properly model the timing and extent of tidal amplitude, which especially affects salt-water intrusion. These parameters are generally available for download from online NOAA databases.
Water volume is modeled on a daily basis, rather than an hourly basis; hourly tidal fluxes are not modeled. For this reason, the overall daily-average water volume of the estuary is assumed to remain constant. The simplifying assumption is that the water volume of the estuary is not sensitive to the fresh-water inflow. The water volume input screen is therefore different for the estuary model, with inflow water representing fresh-water inflow. Salt-water inflow is derived from the tidal range model as discussed above. The size and depth of the fresh and salt wedges then vary as a function of the tidal range and fresh-water flows.
Although Release 3 of AQUATOX can run on an hourly basis, the estuary version of the model has not been modified to model tidal inputs and withdrawals on an hourly basis.
Nutrient concentrations in seawater flowing into the lower (saline) layer are set to temporally constant levels. The nutrients in seawater may be edited selecting the “Estuary Salt Water Nutrients” button available at the bottom of the initial conditions and loadings screen for each relevant variable (Ammonia, Nitrate, Phosphate, Oxygen, and Carbon Dioxide).
Chapter 10 of the technical documentation covers the AQUATOX estuary model in detail.
When applying AQUATOX to a new site it is usually most efficient to find a surrogate site that best matches the characteristics of the site to be modeled. The user can then modify that site's characteristics so that it matches the modeled site with respect to site morphometry, nutrients, organic matter, suspended sediments, biota, and organic chemicals (if relevant).
To assist in this process the file "Study descriptions Release 3.1.pdf" has been added to the STUDIES directory that describes the characteristics of each of the example model applications included with the model.
After the surrogate simulation has been loaded, the user should save the APS file to a new name and then examine the following components for compatibility with the site being modeled
Stepping through each step of the AQUATOX Wizard can be a useful method of examining the key parameters in a given simulation.
To start a simulation from scratch (i.e. without using a surrogate simulation) within AQUATOX, use the "New Simulation Wizard" in the file menu and then select "Create Simulation from Scratch" in Step 1 of the wizard.
Some users will want to enter each and every parameter and loading from scratch. In some senses this might be the "safest" approach as the user is required to assess the appropriateness of each and every parameter in the model for their individual site. However, there are many parameters that can safely be assumed to be "global" within the model and the user will lose the benefit of having those parameters automatically brought in to their simulation. Because of this, the model creators suggest that users who are concerned about bringing in external parameters examine each and every parameter of the surrogate simulation closely.
As noted in the Water Volume Data screen, there are many options as to how to compute or specify water volume; each requires a different set of input data. Often, the selected volume-modeling option is a function of the available data for the site being modeled.
Time series of stream volumes are quite rare whereas discharge data are more often available. For this reason, the Manning’s Equation, which derives river volumes based on discharge, slope, and width data, is often used for streams.
Standing water volumes are sometimes assumed to be kept constant; alternatively time series of known volumes may be used. See the Water Volume Data section for more information about modeling water volume.
Chemicals and biotic state variables can be added to the model, either through the AQUATOX wizard or through the main interface screen (by clicking the "Add" button under the list of state variables, which brings up the “Insert State Variable” screen and its list of available state variable compartments). When adding fish, the wizard is especially useful as size-class and age-class linkages will be automatically set up through the user interface.
If you are modeling an animal, chemical, or plant state variable for which there are no database parameters available, you will first need to create a new set of parameters for your state variable in the relevant AQUATOX library. Often it is best to use an existing database record most similar to your new addition and modify as necessary. See the section on parameters for more instructions on adding a new database entry.
To add a chemical in AQUATOX select one of the 20 available toxicant compartments (usually "Dissolved org. tox 1") on the “Insert State Variable” screen. You will then be prompted as to which chemical to load from the chemical library. Loading this chemical brings in underlying data for the chemical and also animal and plant toxicity databases.
To properly model toxicity for the chemical, each and every animal and plant in the simulation must be properly linked to one of the rows in these toxicity databases. If any plant or animal lacks the link to toxicity data an error message with be returned. In the animal and plant underlying data records, the toxicity linkages may be entered individually by choosing the appropriate group in the dropdown in the “Toxicity Record” field or "Edit All" may be selected, allowing you to link these organisms all at once.
To add a plant in AQUATOX, select one of the available plant compartments (Diatoms, Greens, Blue-greens, Other Algae, or Macrophytes) on the “Insert State Variable” screen and then choose the plant data you wish to load from the plant library. Initial conditions for the plant and any external loadings must then be specified. A "seed" loading may be specified (generally as a constant load) to prevent permanent extinction from occurring. Seed loadings are generally very small loadings (e.g. 1e-5 mg/L) that allow for reintroduction of an organism after environmental conditions have improved to allow that organism to be viable in the system again. The Wizard interface may also be used to add plant variables.
To add an animal in AQUATOX, select one of the available animal compartments (Shredders, Sediment Feeders, Suspended Feeders, Clams, Grazers, Snails, Predatory Invertebrates, Large and Small Forage Fish, Large and Small Bottom Fish, and Large and Small Game Fish) on the “Insert State Variable” screen and then choose the animal data you wish to load from the animal library. Initial conditions for the animal, any external loadings, and trophic interactions within the food web must also be specified. Size-class fish may be specified using the Animal Linkages interface. A "seed" loading may be specified as a constant load to prevent permanent extinction from occurring. The Wizard interface is especially useful for adding fish state variables.
If a four-compartment oyster simulation is desired (see section 4.4 of the Release 3.2 Technical Documentation.), the four oyster compartments (veliger, spat, seed, and sack) need to be added to the model. The seed and sack categories must be modeled using one of the four clam compartments in the model. Within the “Size-Class Links” available in these clam’s “Underlying Data,” the sack and the seed state variables must be linked together (this only needs to be done once.) These state variables should also have their taxonomic type set to “Seed oyster” or “Sack oyster” as appropriate. At that point, the model will model the oyster in four life stages as discussed in more detail in section 4.4 of the Release 3.2 Technical Documentation.
To add a multi-age fish, select "Multi. Age-Class Fish" from the bottom of the list of state variables. The user will be prompted for the number of age-classes to add and sources of parameters for the young of year (YOY) fish and all older fish. You will then be able to access the Multi-Age Fish interface by double-clicking on the "Age-Class Fish" item at the bottom of the state variable list on the Main screen. See the section titled “Multi Age Fish Screen.”
Inorganic sediments can have significant effects on light climate and inorganic sediment effects on biota can also be explicitly modeled. Release 3.2 of AQUATOX contains four levels of inorganic sediment submodels:
1.
TSS Model:
The TSS option is a simple model based on a regression relationship
between sediment deposition and total suspended sediments. This option also ensures that the effects of
inorganic sediments on the light climate of the water column are properly represented. This approach should be used when the only
inorganic sediment data available are TSS.
This option also assumes that TSS does not undergo any degradation or
other reactions while within the AQUATOX reach.
Add the "TSS" state variable to use this option.
2.
Sand-Silt-Clay Model: This option is an inorganic sediments
submodel that can be used to estimate the scour and deposition of inorganic
sediments at a site as a function of water flows. This model is only applicable to streams and
rivers. The sand-silt-clay model requires additional data about the types of
inorganic sediments (i.e., sand, silt, or clay) and their average rate of scour
and deposition under different water-flow regimes. This model may be selected under the sediment
menu by choosing "Add Sand Silt Clay Model."
3.
Multi-Layer Sediment Model: This option is a complex multiple-layer
sediment submodel described in Section 6.2 of the Technical Documentation. This model can be used to estimate the
sequestration of organic toxicants within the deeper layers of the sediments
and the potential for scour of such toxicants from these deep layers. The
multi-layer sediment model is a complicated model with considerable additional
data requirements. This model may be selected under the sediment menu by
choosing "Add Multi-Layer Sediment." Additional layers may also be
added or removed using the options listed under the sediment menu. This action
adds the “Sed Layer(s)” button to the main screen, which can be use to access
the multi-layer sediment parameters.
4. Sediment Diagenesis Model: The sediment diagenesis model implements the sediment diagenesis model as presented in Sediment Flux Modeling (Di Toro,2001). This model provides a more sophisticated accounting of the decay of organic matter and remineralization in an anaerobic sediment bed and the effects on sediment oxygen demand and nutrient flux into the water column. The diagenesis model assumes a depositional environment; scour of sediments is not incorporated. This model is best suited to represent predominantly anaerobic sediments in eutrophic lakes. This model may be selected under the sediment menu by choosing "Add Sediment Diagenesis." This action adds the “Sed Layer(s)” button to the main screen, which can be use to access the sediment diagenesis parameters.
The simplest manner is to use the AQUATOX "default" sediment bed model in which organic matter is modeled in the sediment bed as a food source and toxicant-sorption site but inorganic matter in the bed is ignored. The model will utilize this option if none of the four models above have been added to a simulation. The model will always track the remineralization of organic material within the sediment bed and the water column.
The default sediment bed model includes the following components
These detrital compartments are assumed to be located in the active layer and available for consumption and sorption-desorption of chemical from-to the water column. Through long-term deposition, these compartments may become deeply buried and join the buried detritus state variables. Buried detritus is considered to be non-reactive. Nutrients within the sediment detrital compartments are subject to release to the water column when the organic matter decomposes. For more information about this model, please see section 5.1 of the AQUATOX Technical Documentation.
Data requirements for this model are rather minimal, primarily being the initial condition of detritus in the sediment bed. Setting the initial condition for these state variables can be facilitated by using an Excel spreadsheet (AQUATOX_Sed_Bed_Inputs.xls) that is included in the Studies directory. This spreadsheet will convert data about the fraction organic carbon, depth, sediment density, and fraction labile into initial conditions for these two state variable compartments.
AQUATOX allows a user to input time-varying concentrations of TSS so that the light climate of the system being modeled can be accurately represented. TSS may represent either total suspended solids or total suspended sediments. A user can choose which of these options is being specified using the radio buttons on the upper right portion of this screen (accessed by double-clicking TSS on the state variable list):
If total suspended solids are chosen, AQUATOX will subtract phytoplankton and detritus loadings from the TSS loadings to estimate the inorganic solids loadings, therefore, care should be taken to use synoptic TSS and nutrient time series.
When TSS is included in the model, it is a driving variable that is not in any way linked to the sediment bed unlike the inorganic sediments in the Sand-Silt-Clay model.
The sediment transport component of AQUATOX simulates
scour, deposition and transport of sediments and calculates the concentration
of sediments in the water column and sediment bed within a river reach.
For running waters, the inorganic sediment model within AQUATOX is based
primarily on the algorithms in the Hydrologic Simulation Program in Fortran
(HSPF, US EPA 1991). Within river or stream simulations, sediment is
divided into sand, silt, and clay. Wash load (primarily clay and silt) is
deposited or eroded within the channel reach depending on the daily flow
regime. Sand transport is also computed within the channel reach.
Inorganic sediments in standing water are computed based on total suspended
solids loadings and not by the sand-silt-clay model.
Within AQUATOX, inorganic sediment
concentrations affect the extinction coefficient for water, and therefore
change the light climate for algae. Chemicals are assumed not to sorb to
inorganic sediments within the AQUATOX sand-silt-clay model, but chemicals
sorbed to organic matter are assumed to scour and deposit with the same characteristics
as the inorganic silt within the model. The fraction of detritus that is
being scoured or deposited within a river reach is assumed to equal the
fraction of silt that is being scoured or deposited.
There are additional data
requirements for this model. AQUATOX requires loadings information for
each of the inorganic sediment categories as well as an initial
condition. The initial fraction in bed sediments must also be
specified. Additional parameters for the inorganic sediment model are found
in the site underlying data (use the “Stream Data” button at the top of the
screen) or can be found in Step 14 of the setup Wizard. These parameters
include
To add the Sand-Silt-Clay model to
a simulation, select to "Add"
a state variable from under the state variable list and select Sand, Silt, or Clay
from the list of variables (directly under the chemicals) or use Step 14 of the
AQUATOX Wizard. All three variables are added; the model will not let you add
only one or two.
AQUATOX also includes a complex multiple-layer sediment model. This model can simulate up to ten bottom layers of sediment. Within each sediment layer, the state variables consist of inorganic solids in several size classes (cohesives and two sizes of non-cohesives), pore waters, labile and refractory dissolved organic matter in pore waters, and sedimented detritus. Each of these state variables can also have up to twenty organic toxicant concentrations associated with it.
Data requirements for this model are significant
This model was designed primarily to capture the long-term sequestration or release of chemicals from the sediment bed (complex nutrient processes, i.e. diagenesis, are not included in this model). As such, linkage to a complex model of sediment scour and deposition might be warranted. For more information on this model, please see section 6.2 of the AQUATOX technical documentation.
Di Toro’s (2001) model assumes a small aerobic
layer (L1) above a larger anaerobic layer (L2). For this reason, it is
best to apply this optional submodel in sites where anoxic sediments are
prevalent, otherwise the assumption of a very thin aerobic layer might
not be appropriate.
The additional 22 state variables added when the
sediment diagenesis model is enabled are as follows:
The sediment diagenesis model is only relevant for
depositional regimes. No direct scour of sediment from the sediment bed
is modeled.
AQUATOX includes the option to “Model Diagenesis Layer 1 as Steady State.” When this option is selected, state variables
in the upper-layer (between Layer 2 and the water column) are all modeled using
the assumption of steady-state. This is
the “native mode” of Di Toro’s (2001) sediment diagenesis model. As noted in the interface, this process
significantly speeds model execution (by up-to a factor of 10), but might have
an effect on the mass balance of nutrients.
However, effects on overall model results are generally quite minor. More information about this option is
available in Chapter 7 of the Technical Documentation.
Data requirements for this model are
moderate. Initial conditions for all variables specified above are
required but rapidly become a function of organic matter deposition and can be
derived by "spinning up" the model (running the model with arbitrary
initial conditions and then setting the initial conditions based on model
results which are a function of organic matter deposition). Initial conditions for the variables, with
the exception of COD, are entered on the Diagenesis Model Setup screen.
Selecting the Edit Diagenesis Parameters button on
the Diagenesis Model Setup screen brings you to a large set of "Diagenesis Parameters"
that also govern the behavior of the model. Most of these may be kept
constant from site-to-site but Di Toro (2001) suggests a few parameters are
site-specific. Partitioning to phosphorus, in particular, is subject to
considerable site specificity (kdPO42
and dKDPO41f).
For reporting and archival purposes, parameters can be saved to Excel using a
button on this screen.
Setting the initial condition for the organic
matter sediment diagenesis state variables can be facilitated by using an
Excel spreadsheet (AQUATOX_Sed_Bed_Inputs.xls)
that is included in the Studies directory. This spreadsheet will convert
data about the fraction organic carbon, depth, sediment density, fraction
labile and nonreactive sediment, as well as stoichiometry
assumptions, into initial conditions for these three state variable
compartments (POC, PON, and POP).
For more information about the equations and
concepts within the Sediment Diagenesis model, please see chapter 7 of the
Technical Documentation.
Model calibration is the process of modifying model parameters within their range of uncertainty (reasonable ranges as defined by literature if possible) such that model results match observed data or observed conditions.
The Import Observed Data function is quite useful for model calibration as all external data relevant to model output can be imported into the model and compared to model results. The graph gallery can be filled with plots of model results compared to observed data, and then parameters can be modified to best match all available data sets.
Parameters will have a varying degree of uncertainty depending on how measurable a parameter is, variability in field measurements, and uncertainty in lab equipment. Parameters will also have a varying degree of sensitivity depending on how much model outputs change as a function of the change in parameter value. Within AQUATOX, parameter sensitivity may be ascertained through an automated sensitivity analysis. "Calibration Parameters" might be defined as those parameters that simultaneously have some range of uncertainty and to which important output endpoints are sensitive.
Calibration can be like trying to solve a puzzle. It requires creative thought and a real understanding of why the model is behaving like it is.
The first step in AQUATOX model calibration is often to ensure that the food-web is stable and matches current conditions. This might be quantitative if biomass data exists for a site or it might be qualitative. For example, if the user knows that largemouth bass exist at the modeled site but this category does not persist in model results then some type of respecification of the trophic interactions (or the lower portions of the food-web) will be required. Once the food-web is stable and reasonable, then calibration to organic toxicant data can be undertaken.
The user should start by trying to specify all of the important components of the food-web simultaneously. Calibrating the primary producers alone might be misleading as this model would not include grazing pressures from higher in the food web. Next, produce a plot of all the plants and animals and ensure that there are no inappropriate crashes or expansions of biomass beyond what would be supported by observed data or, lacking that, common sense.
Some notes about calibrating plants:
Some notes about calibrating animals:
Additional
information about model calibration may be found in the following document:
EPA, 2009. AQUATOX Technical Note 1, A Calibrated Parameter Set for Simulation of Algae in Shallow Rivers, EPA-823-R-09-003 February 2009
Also
see Section 2.6 on "Calibration and Validation" in the AQUATOX
Technical Documentation and Section 2.4 on Sensitivity Analysis.
Model validation is a numerical cross-examination of the model against new data sets. By definition, a calibrated model is effective given the conditions it was calibrated against, but there is uncertainty as to whether it will correctly model changing conditions. If the model can be tested against a different data set than was used for calibration, this increases confidence that the model has the capability to effectively capture alternative scenarios.
Model validation can consist of extrapolating the model
Model validation exercises are often defined by, or limited by, the data available for validation.
If a model validation is not successful then the user can go back to the model calibration and seek an alternate calibration that will adapt to changing conditions more successfully. The model calibration and validation data sets could be combined for a more robust calibration, but at the expense of the loss of the validation data sets; this combined approach should be avoided if possible.
Also
see Section 2.6 on "Calibration and Validation" in the AQUATOX
Technical Documentation.
A linked AQUATOX simulation is essentially a
series of single segment simulations that are linked together with
flow-fields. All linked segments must have an identical set of state
variables. State variables that do not occur in an upstream segment may
be set to zero there. State variables can pass from segment to segment
through active migration, passive drift, diffusion, or bedload transport. Mass
balance of all state variables is maintained throughout a multi-segment
simulation.
Linked AQUATOX simulations have a file extension
of *.ALS. When a linked simulation is
loaded, the top left panel of the primary interface is modified to show the
segments or links.
Clicking the Show Segment Data radio
button on the top left of main screen displays all segments in a list.
Double clicking any of the segments will bring the user into the normal “Main”
screen for viewing that particular segment. Segment-specific parameters
and output can be viewed through that window.
Clicking on Show
Link Data will provide the user with a list of water-flow
linkages between modeled segments. Double clicking on any of the linkages
will bring a user to the Edit Linkage
Between Segments window.
The Linked Output button on the Main screen brings a user to a screen that will provide graphical or
tabular output for all segments simultaneously.
For reference, an optional map or image (in
bitmap format) may be attached to the simulation, though this map will have no
practical functionality within the interface.
Another option, within Linked mode, is to run several
single-segment simulations together, but without any linkages to one
another. This approach is an efficient way of calibrating several
single-segment simulations with the same set of parameters, since it can
be used to calibrate across sites or experiments (such as artificial
streams). In a linked simulation, changing a biotic parameter in one
segment changes that parameter for all segments. Therefore, if it is
desirable to test parameter changes over several simulations, loading these
simulations into a "linked simulation" but specifying no water
linkage between the segments is possible. However, the linked mode also
requires the state variable list to be identical in every segment.
To set up several single-segment simulations together for
cross-site calibration purposes, first select a “New Linked Simulation” from
the File menu. Then, select one of the
single-segment APS files you wish to link (as an example, you may select “Rum R
MN.aps”). Do not select “Use Excel Template to Populate Linked Segments.” You now have a “linked” segment
implementation with only one segment.
Next select “Add” another segment using the button under the segment
listing, and “Load segment from *.aps file”.
As an example, you might select “Blue Earth R.MN.aps”, which has the
same state variables as Rum River. Now
you have two segments in the same simulation that may be run concurrently. Any change to any of the plant or animal
parameters (underlying data) in one study will have the same effect on the
other study.
For more information on linked simulations, see section
3.8 in the Technical Documentation.
There are two types of linkages that may be specified between individual segments: “cascade links” and “feedback links”.
A “cascade” link is unidirectional; there is no potential for water or state variable flow back upstream. Segments that are linked together by cascade linkages are solved separately from one another moving from upstream to downstream. This is particularly efficient computationally when modeling faster flowing rivers and streams. "Passive" state variables such as nutrients and organic chemicals dissolved in water and plankton, move between segments as a function of water flow and diffusion. Benthic state variables and fish are assumed to stay in the reach in which they were initiated with the exception of specified fish migration and specified bed-load between segments (for benthic sediments).
A “feedback” link allows for water or state variables to flow in both directions. For bookkeeping purposes, water flows are required to be unidirectional (i.e. entered water flows over a feedback link must not be negative). However, two feedback links may be specified simultaneously (in opposite directions) to allow for bidirectional water flows. Segments that are linked together by feedback links are solved simultaneously. There may only be one contiguous set of segments linked together by feedback linkages within a simulation but there is no limit on the number of segments that may be linked together in this manner.
Each segment's water volume inflow is a function of any boundary condition loadings that are specified in the “Water Volume” screen (accessed by double-clicking the Water Volume state variable), and water from one or more linked segments. Each segment's water volume outflow is a function of any boundary condition outflows, as specified in the water volume screen, and outflow of water to one or more linked segments.
Feedback linkages also allow for passive diffusive exchange of state variables between segments. To model this type of state variable movement, the dispersion coefficient for each interface must be specified along with the cross section area and the characteristic length.
Finally, bed loads for inorganic sediments may be specified as a time series (g/d) over a given linkage if the multi-layer sediment model is included.
Linkages between segments should be assumed to be horizontal unless two segments have been specified as a stratified pair in which case vertical linkage is possible.
The linked output window displays results from all linked segments simultaneously.
As in the single-segment output window, a library of graphs can be specified and observed data can be plotted against them. The only difference between the single-segment and linked output screens is that every segment is plotted simultaneously, differentiated by segment ID.
To observe results from a single segment only, go into a specific segment by double-clicking on the list and then go into the output window from there.
Currently there is no way to specify a sub-set of linked segments, and the graph-setup screen is disabled in linked-output mode.
The tabular data in this screen will also display all segments' results simultaneously.
When a linked AQUATOX simulation (*.ALS) is loaded, and the Show Segment Data radio button is checked, all segments are displayed in a list. Double clicking any of the segments will bring the user into the “Main” window for viewing that particular segment. Segment-specific parameters and output may be viewed through that window.
Double-clicking on any of the state variables in the state variable list will display segment-specific initial conditions and segment-specific inflow loadings. Inflow loadings are associated with inflow boundary conditions as specified in the water volume screen. Inflow loadings from another modeled segment are a function of the internal processes in that segment and the flow of water over the linkage between segments.
Moving into underlying data for any of the state variables shows parameters that are global to all the animals, plants, or chemicals in a simulation. Changing any of these parameters will change the parameters for all segments in the system. Similarly, if a state variable is added or removed when the model is run in Linked-Mode, that change to the state variable list will occur within every modeled segment.
On the other hand, going into the ”Site” screen (accessed from the “Site” button on the “Main” screen) within a linked segment will shows parameters that are unique to the given segment. Site underlying parameters, remineralization parameters, water velocity, shade, and water depth information are all unique to the segment being modeled.
When the main linked interface is shown (map and list of segments) a user may export all of the graphs associated with the simulation into Microsoft Word at the same time, by going to the “Study” item on the main menu and selecting "Export All Graphs to MSWord". AQUATOX will step through each and every graph in the graph library of every linked segment and paste those graphs into MSWord in Windows metafile format.
This option can be a powerful tool if summary graphs and graphs of observed data vs. model results are available in each segment. The selection of a single menu item enables a graphical summary of the entire system.
For the linked version of AQUATOX "Tributary Input" is available on the “Site” screen, accessed from the “Site” button on the “Main” screen. State variables are not solved within this type of segment, rather, loadings of nutrients, organic matter, and biota to this type of segment are loaded directly into the modeled system.
In the single-segment formulation, point sources and non-point source loadings have no water associated with them, and loadings are given in grams per day. The tributary input segment allows a user to enter a set of nutrient, organics, toxicants, and/or biotic loadings as concentrations in water that are associated with a water flow. The loadings are entered as loadings to the tributary input segment itself. The water flows that those loadings are associated with are entered as linkages from the tributary input segments into the modeled segment.
Tributary inputs may be used to model tributaries, point sources, non-point sources, and even groundwater inputs into the modeled system.
Because tributary input segments are "dummy" segments, output from these segments are not included in the linked output screen or when linked results are exported.
Tributary input segments provide the user with considerable flexibility for evaluating "what-if" scenarios regarding a given source of loadings. For example, if a management practice could conceivably reduce both the amount of water released and the concentrations of pollutants within that water from a point-source, the effects of this change in the tributary drainage area can be directly modeled.
The primary data requirements difference between linked mode and single-segment mode is that a water volume and water movement model for the entire linked system must be specified and input as a combination of boundary condition loadings, boundary condition withdrawals, and movement of water from segment to segment. This model may be derived from an external model or through simple assumptions and Manning's equation for linked streams. Additionally, initial conditions and boundary conditions are required for the entire modeled system rather than a single modeled segment.
The overall data requirements for a linked project are represented by the following list. For each segment modeled, the following time series may be required:
Biotic state variables might need initial conditions set in each segment or might be “spun up” if the extensive biomass data are not available.
Note that this list is fairly exhaustive and includes many optional items so the user should not be put off by it-- there are many ways to fill these parameters using observed water quality data. In some cases, point-source and nonpoint-source loadings might not be required.
To create a new linked simulation, a user must first select an existing single-segment AQUATOX simulation to serve as a template for the rest of the simulation. This will be the first segment in the linked system. The state variable list within this segment must be matched by any additional segments that are added. However, state variables may be added and subtracted in linked-mode; they will simply be added or removed from all linked segments simultaneously.
Using the AQUATOX interface to set-up or modify a large linked-mode AQUATOX
simulation can be very time-consuming. This process is also not
“auditable,” that is, it is very difficult to ascertain whether a mistake was
made in the time-consuming click-intensive process. The wrong loadings
could be imported to the wrong segment and the only way of knowing this would
be by clicking into the loadings for that segment or carefully examining the “text
output” of AQUATOX parameters.
To facilitate the process of setting up large simulations and understanding the
parameters that go into them, a function to import loadings and parameters from
an “Excel template” was added to the linked mode of AQUATOX. This enables
a user to perform the following tasks:
The Excel template function is offered by AQUATOX interface when the user selects “File, New Linked Simulation”. In this case, the user will need to specify an existing single-segment simulation to reproduce as multiple segments of the new simulation. Alternatively, a template can be used to modify an existing linked simulation (“Add Time series from Excel” in the “Study” item of the “Main” menu). The user will need to provide the file name and location of the Excel template in either case. There is an example template included in the AQUATOX Studies directory as “import_template.xls.”
AQUATOX will then automatically read from the Excel template on a column by
column basis. Depending on the type of input being read, and whether
non-detects are included, each model input can be from one to three columns in
width. The first column (column “A”) of each worksheet is not read but
contains template input text (for reference only). This first column is
included in the example
“import_template.xls.” In cell “B1” of the spreadsheet is the first
“command” to AQUATOX and it explains what type of model input is located in its
columns and potentially in the next one or two columns to the right.
After those columns have been processed, AQUATOX reads the top row (Row “1”) of
the next column to obtain its next input or command.
The type of model inputs, their parameters, and their effects on a linked
simulation are documented below.
Loadings or valuations are added to the linked simulation or modified if they already exist. Loadings in water are relevant to boundary conditions only, as nutrients moving from modeled segment to modeled segment are internally calculated. As noted above, loadings applied to unmodeled “tributary input” segments will be passed directly through to the modeled segment that the tributary is linked to.
If no valid assignment is found in this cell, the software finds the phytoplankton compartment highest on the list of state variables. If there are no phytoplankton compartments in the simulation, the linkage raises an error.
To add a new segment to an AQUATOX linked simulation, the "Add" button may be used on the main linked study screen when "show segment data" is selected.
A dialog box will appear that allows the user to specify a segment ID (which is used organizationally, and to label output), a segment name (which describes each segment), and three options as to how to add a segment. The Segment ID must be four characters or less.
To link two segments together, the two segments must first be included in a simulation. Then the "Show Link Data" radio button (top left of “Main” screen) must be selected in the “Main” linked interface window. When "Add" is then selected, the “Edit Linkage Between Segments” screen will be visible which will enable the user to specify a "cascade" or "feedback" linkage and also to enter water flows and other relevant parameters and time series. For more information on editing linkages see "Viewing Linkages."
To add a tributary input (which can also be used to model point and non-point sources and groundwater inputs) a new segment must first be added. This is done by selecting “Add” when the “show segment data” radio button is active. Adding a "blank" segment is recommended. Then, the site type must be designated as a "Tributary Input" within the Site Type panel on the “Site” screen. Within this tributary-input segment, initial conditions, site underlying data, and boundary-condition water volume inflows and outflows should all be ignored as these are not relevant and will not be used by the simulation. Instead, the user should focus exclusively on the nutrient, oxygen, biota, and chemical loadings into the new tributary segment. These loadings will be passed right through this tributary segment and into the modeled segment that it is linked to. For this reason, a linkage of the tributary input into the modeled segment must also be added. Add a new linkage from the tributary input segment to the modeled segment and specify an associated water flow. This linkage must be a "cascade" linkage, as no feedback to the tributary input is possible.
On a given date, the water flow of the linkage multiplied by the concentration of the loadings specified into the tributary-input segment will provide the mass of the loading into the modeled segment. The volume of water specified will also be added into the modeled segment.
Tributary inputs can also be added through the Excel Template Import, which can simplify the process. For more information about tributary inputs see "Tributary Input Segments."
Setting up linkages in truly complex systems can be a time-consuming process. For this reason, the Excel Template Import Capability is especially powerful for these systems. This allows complex linkages to be set up in an auditable fashion and also allows these linkages to be modified and re-imported. Alternatively, boundary-condition loadings may be modified in the Excel template and then the template may again be re-imported to test alternative scenarios.
It is important to remember that inflow water for each segment is governed by two sources of model input:
Similarly, the outflow water for each segment has two sets of loadings:
New users of the AQUATOX linked-mode often have difficulties with water volumes going to zero, or water volumes going to unreasonably high levels.
Other than evaporation (which is often set to zero in linked-systems for simplicity), AQUATOX linked segments are modeled as "closed systems" where no water enters or leaves the system unless it is specifically accounted for. The outflow for each segment is defined as "Inflow - d Volume + inputs - withdrawals." The inflow of the next segment may be defined as the outflow from the previous segment.
For systems that take a long time to run, any changes to the water balance model can be tested by first saving the simulation, then saving the simulation to a different name, then removing all state variables and ensuring that the water volumes are behaving as required before returning to the full version with all of the biotic state variables.
An external "closed system" accounting of water flow is often quite useful for the use of the linked-mode model, whether it be a complex hydrological model or a simple spreadsheet model.
“Dynamic stratification” (i.e. stratification triggered by temperature and temperature differences) does not apply to the multi-segment model. Instead, a user must specify two unique linked segments as a vertically stratified pair. In this case, the segments must be linked together with a feedback linkage. A stratification screen within each segment’s main interface allows a user to specify whether a segment is part of a vertically stratified pair and, if so, whether it is the epilimnion or the hypolimnion segment. (The model needs to know if one segment is located "on top" of another segment so that sinking organic matter and plants are routed to the lower segment rather than moving directly to the sediment bed. This also affects light climate in the lower segment.)
Unlike the single-segment model, water flows must be specified between the two segments. Overturn may be specified by a high degree of mixing between the two segments whereas periods of stratification will have considerably lower flow. These flow-fields are how all mixing (or lack thereof) is indicated in linked-segment stratified segments.
When two segments are set up as stratified together, the thermocline area is defined by the user-entered cross section between them. If this data field is left blank then standard AQUATOX bathymetry equations are used (see equation (9) in the Tech Doc.) As noted above, annual cycles of stratification and overturn may be specified using the time-varying water flows and dispersion coefficients. As was the case in the dynamic stratification model, fish automatically migrate to the epilimnion in the case of hypoxia in the lower segment. Sinking phytoplankton and suspended detritus in the epilimnion segment pass into the designated hypolimnion segment. The light climate of the bottom segment is limited to that light which penetrates the segment defined as the epilimnion.
This basic tutorial introduces basic concepts and gets started using the AQUATOX interface.
Open the file Farm Pond MO Esfenval.
(see Loading a Study)
This study represents a generic
pond built to USDA specifications. Esfenvalerate
loadings are the worst-case scenario using runoff from an adjacent corn field
predicted by the PRZM model.
First remove the macrophyte
compartment by highlighting it in the list, clicking on the Delete
button, and confirming the deletion. The variable will disappear from the
state variable list.
Likewise, state variables can be
added by clicking on the Add button and choosing from the list.
Let’s add macrophytes back to the list of state variables.
Click on the Add Button and scroll
down the list until you see the Macrophyte1 to Macrophyte6
compartments. AQUATOX has the capacity to model up to six unique
macrophyte species. Select Macrophyte1 and click on OK.
You will then see another dialog
box open asking you to "Select Plant Entry to Load Into Macrophyte1."
Note that the names of the taxonomic groups and ecologic guilds on the main
study screen are followed by the names of the specific groups in
brackets. We therefore have to specify the type of macrophyte we are
adding. Select Myriophyllum on the list at the left and click on OK.
If you do not select a species when you run the simulation you will receive an
error message indicating that there is no data associated with the state
variable Macrophyte.
Next, Setting an Initial Condition
To continue with our macrophyte
example, we should enter a value for the biomass of macrophytes present at the
beginning of the simulation; if the value is left as 0 and there is no loading,
then macrophytes would not be simulated. The initial condition will
depend on when the simulation starts (which is specified in Setup).
In this example we will enter a value of 0.1 g/m2, which is
appropriate for Myriophyllum in a temperate pond at the beginning of the
growing season.
There are three things we are
looking at in this screen. First, the
initial condition for the state variable is shown at the top of the left-most
panel. Ongoing plant loadings from
up-river are shown below that. The right
panel pertains to chemical concentrations within the macrophyte and any
macrophytes being loaded into the system.
Ignore the chemical exposure panel to the right for now.
Double click on the state variable
list on Macrophyte1: Myriophyllum. Enter 0.1 as shown below.
As you can see from the picture
above, AQUATOX also automatically entered a loading of 1e-5 (0.00001) for
macrophytes in this simulation. Ordinarily we think of loadings as
pertaining only to chemicals and freely moving organisms such as plankton and fish.
However, it is usually desirable to enter a small constant loading to
serve as a “seed” if the population or group is killed off entirely by a
toxicant or adverse environmental conditions.
The “seed” will allow the biota to recover or recolonize, if and when
conditions become favorable. This is especially important for macrophytes
that suffer winter die-back
For all biota, AQUATOX uses a
value of 1E-5 g/m2 as a constant loading. This is small
enough that it will not affect the results during the growing season, but it is
large enough to prevent extinction.
Click the OK button to return to
the Main screen.
See Also: Initial Conditions and Loadings
next, Viewing Parameters
In the following examples we will
examine a record from each of the other libraries. A record can be
downloaded into a study from a library by choosing Load data on
the Edit State Variable Data screen.
We will examine first the
parameter screen for plants. From the main screen, choose Plants
from the Library menu at the top of the screen (or click on the big
button “P” on the toolbar), view the default library file, then move to Cyclotella
nana by clicking on the arrows at the top left of the screen.
We are now moving through a database of parameters that is provided with
AQUATOX. These parameters are not necessarily associated with the study
file that has been loaded, but can be loaded into the simulation. Note, to
see the parameters associated with the given study, double-click on any of the
animals or plants in the state variable list and then click the "Edit Underlying Data" button.
Two fields near the top of the
screen require explanation. If you click on the drop down menu to the
right of Plant type, you will be given a choice. The choice of Plant type is important
because different types have different physical or biological processes that
apply to them. For instance, phytoplankton are subject to sinking, but
not periphyton, which are attached to a surface. Conversely periphyton
are limited somewhat by very slow current velocity; but not phytoplankton,
which are adapted to still water.
Less obvious is the Toxicity
Record; again, clicking on the arrow to the right of the field will give
you several choices. The intent is to associate the organism record with
one of the limited number of organisms that have some toxicity data or
procedures for estimating toxicity. In this instance, if you choose Diatoms
the model will utilize the toxicity data (e.g., EC50) for Esfenvalerate to
Diatoms, as listed on the Toxicity Record portion of the Chemical
Properties screen.
The given parameter values are
provided to get you started; if you have more appropriate values, you should
use them. Those parameters that do not pertain to phytoplankton are inactive
and grayed out. If you try to enter a value for Reduction in still
water, for example, you will find that the field cannot be edited.
Next we will locate the record for
Chironomid from the Animal Library. Hit Cancel to leave the plant
screen and again go through the Library menu to get to the animal
library. (Alternatively you can select the "A" for animal on
the toolbar.) A pop-up will appear asking if you would like to discard
all edits. If you have made any changes and would like to save them, click
“Save.” If you have not made any changes, click “OK.” Once you have left the
plant library, scroll using the arrow buttons to find the Chironomid record or
you can click the find button and enter "Chironomid." (Alternatively you can use the “Search Names”
button.)
Click on Animal Type to see
the pull-down menu. Chironomids have aquatic larvae, so Benthic
insects is chosen; this is important because emergence is simulated by
AQUATOX for insects as a loss term, but does not apply to other animals.
Click the Trophic Interactions button to see the associated trophic
interaction table. The trophic interaction table is important because it
defines the food-web relationships and assimilation efficiencies. Here we
see that Chironomids primarily feed on labile detritus, and that they
assimilate 70% of it. See also Trophic Interactions.
After exiting the trophic
interactions screen, click on the scroll bar to the right to see the rest of
the Animal screen. The Bioaccumulation Data section contains
parameters relevant to bioaccumulation of organic toxicants, only one of which
(Fraction that is Lipid) is sensitive in insects. The model is not
sensitive to the longevity of the insects because emergence is a function of
growth rate, which depends on local, seasonally varying conditions.
Likewise, mean weight (higher up on the screen) can only be approximate across
all instars (developmental stages) because it will vary greatly during the
growing season.
Finally, we will examine the
remineralization screen for this study. To do this, return to the main
screen, select the Site
Button and then select the Remineralization button from within
that dialog box. . Note that the Site Button can be found under
“Data Operations” on the main window. The “S” button on the tool bar will take
you to the Site Library, which will not be necessary for this step. Many of
the parameters on this screen are global and there is little need to change
them for a site, unless the organic material is quite different or there is
some reason that the microflora might have adapted to abnormal conditions, such
as a thermal spring or acid mine drainage. For more information about the
three different types of Remineralization records included with AQUATOX see the
section entitled Remineralization.
Next, Viewing Toxicant Loadings
Following
the tutorial to this point, we should still have FarmPond MO Esfenval.aps
loaded from back in Step 1. Return to the main screen.
In the “State and Driving
Variables” list, double-click on Dissolved org. tox 1 or click on the Chemical
button on the main screen to bring up the Edit Chemical Data screen,
and to examine the options for loadings to the system. Pollutant loadings
can be entered as constant or dynamic loadings in several different forms. The
pollutant can be entered as a concentration in the dissolved phase or loosely
bound to suspended sediment; the water inflow and the site volume are then used
by the model to compute the loading per unit volume. The gas-phase concentration
is used to compute atmospheric exchange; ordinarily concentration in the
atmosphere can be considered to be 0, although some pollutants such as PCBs might
have significant atmospheric concentrations.
Point-source loadings are mass per
day (g/d) for the entire site; they are divided by the site volume to obtain
the loading per unit volume. In this example, dynamic loadings from a
point discharge as calculated by the PRZM model are entered. Note that
the dynamic loadings are interpolated, so if the intent is to represent a spike
such as from storm runoff on a particular day, then the loadings should be
bracketed by “0" loadings. The model assumes that the loadings “wrap
around” with an annual cycle and that the last loading can be interpolated to
the first loading as if it were in the succeeding year. Exercise
caution when modeling multiple years with only loadings data from one or a few
years. Sporadic loadings, which could only be expected in one particular
year might inappropriately be repeated in successive years. If you do not
wish loadings to be repeated, enter values (“0" or otherwise) for the
first and last days of the simulation. The dynamic loadings in this
example were entered by hand; an excellent alternative is to download or
prepare a file external to the model and import it into the study using the Import
button. See also Important
Note about Dynamic Loadings.
Another potential pollution
loading source is from direct precipitation. These loads are given as
g/m2 day because AQUATOX does not explicitly consider precipitation.
Click on N.P.S. to toggle to the non-point source screen, which is in
g/d.
Next, Running the Simulation
To run the simulation, select the Perturbed
or Control buttons from the main screen.
Perturbed starts the simulation with changed conditions, such as
with a toxicant.
Control starts a simulation without the stressor; the user can use
Control Setup to specify what is changed and what is held
constant, or parameters may be changed between running the control and
perturbed simulations.
Depending on the control setup
selected, a control run will omit toxicants or nutrients from the run.
Next, Viewing Output
When a study has been executed,
results for various state variables are stored in memory. To view this
output, you must either export the output to an appropriate file-type using the
Export buttons on the main screen or select the Output Button
from the main screen. The Output Button will spawn an additional
output window that will allow you to view output in graphical or tabular
format. For more information about exporting results, see the “Exporting Results”
section.
Also See: Viewing Graphical Output
This tutorial uses the AQUATOX Wizard to make a number of changes to a “template” study. This is the usual approach taken when applying the model.
An existing study that is as close as possible to the site you are trying to model should be chosen. In this case, open the file Rum R MN.aps (see Loading a Study), save as LBR Middleton ID.aps.
It is best to immediately save the template study with a new
name so that you do not overwrite the original file with your changes.
Click on large Use Wizard button on main screen. Use the Next>> Buttton to move through the multiple screens. The Show Progress Buttton activates a progress window that is useful for keeping track of progress when using the wizard.
Wizard Step 1: Change the study name to “L. Boise Middleton” This will be the heading for output. You will also be given the option to select the type of system to be simulated. Keep “Stream” selected and click “next.”
Wizard Step 2: Change the simulation period to 1/1/2000 to 12/31/2001.
Wizard Step 3: Keep the nutrient initial conditions; stream simulations are not sensitive to nutrient initial conditions but rather to loadings
Wizard Step 4: Also keep the sediment detrital initial conditions for lack of better data and click “Next” to keep the water-column CBOD (you can change this after importing loadings)
Wizard Step 5: Keep the following algal state variables and initial conditions; generally these will be applicable to any stream; however, remove the macrophyte Fontinalis, which is a moss by clicking on the macrophyte on the right and clicking “Remove From Simulation”. To do this click “Next >>” until you reach the “Macrophytes” screen.
Wizard Step 6: Keep the invertebrates without change until “Pred Inverts” when you should drag Stonefly to the simulate column on the right. Keep the initial conditions, but use 0.01 for Stonefly (last input box).
(Note, any changes that are made in the wizard may be undone using the
“cancel” button. However, any changes
made in any single wizard session must be undone at the same time. Therefore you may wish to periodically click
on “finish” to save your partial changes and then reenter the wizard. You may jump to whatever screen is
appropriate by double clicking on that step on the Wizard Progress window.)
Wizard Step 7: Although plant and
invertebrate groups do not vary appreciably over large geographic areas, fish
are often restricted to specific drainage basins; therefore, you will need to
make numerous changes. First, select and
remove: Bluegill, Smallmouth Bass (YOY and Lg) and Walleye.
Next add:
Single compartment, small forage fish Chiselmouth
and Dace;
Single compartment, large forage fish Pikeminnow;
and
Size-class gamefish Largemouth Bass,
Mtn Whitefish, and Rainbow Trout (Use YOY data record for small size class).
There are different parameter sets for
YOY (young of year) and adult fish in many cases. When adding a size-class fish, you are first
prompted to load parameters for the small or young-of-year fish. Then you are asked if you wish to use the
same parameter record for the large fish or not. In this tutorial, load a different parameter
set for the large size-class fish.
Change all fish initial conditions to 0.1; we will let the model adjust these
during iterative simulations.
Wizard Step 8: Change the site name to “Lower Boise”; enter
5 km for length
(arbitrary);
surface area = 1.8e5 m2 based on 36 m width
determined from Google Earth;
mean depth = 0.42 m;
max depth = 0.63 (1.5 X mean depth);
evaporation = 0 (is not important for a short stream reach);
latitude = 43.57.
Keep channel slope = 0.002
Keep Manning’s coefficient estimated based on natural stream;
Riffle = 47% and pool = 0% based on site observations and pebble count.
Wizard Step 9: Continue to “Use Manning’s
Equation” for simulating water volume and select “Next.”
For initial volume use 7.56E4 m3 based on mean depth of
0.42 and surface area given earlier; discharge data should be entered by
clicking on the Change button, then the
Import tab, then choosing Midd
discharge.xls (based on USGS data from the Middleton gage converted from
cfs to cu m/d) by clicking on the “Tutorial” folder and navigating to Midd Discharge.xls. Once this has been
selected, click the Import Button.
Note:
All Excel files to be imported in this tutorial may be found in the
“Tutorial” directory off of the AQUATOX “Studies” directory. Whenever one time-series is imported over
another, the original data is deleted prior to the import.
Wizard Step 10: Select to use time-varying temperature and then use time series imported from Midd temp.xls; these observed data are almost a year in length and are from a later period, but they represent the variable nature of temperature and are interpolated by the model for the simulation period. The initial temperature for January 1 may be set to 5 degrees based on the imported data.
Wizard Step 11: Wind is not important for a stream, so keep it constant at 1 m/s
Wizard Step 12: Select “Use Annual Mean
and Range” for light loading then use observed mean = 380 Ly/d and range = 516
Ly/d. Links to internet sources for these data change frequently so one
research approach would be to search for the nearest airport and “solar” to try
to find approximate values for light loadings (which may require units
conversion).
Select to compute photoperiod from latitude.
Wizard Step 13: Select “Use Time-Varying pH” and import observed pH from Midd pH.xls; keep initial pH at 8, which is approximate for January 1 based on the imported data.
Wizard Step 14: Sporadic observed TSS values are available from the USGS NWIS for the Middleton gage. A regression was developed for TSS as a function of discharge (see Midd discharge TSS.xls), but there was no significant relationship, so the raw observed values in Midd TSS.xls should be imported.
Wizard Step 15: Skip this step as no toxic chemicals will be simulated.
Wizard Step 16: Ignore loadings for ammonia; import nitrate loadings from Midd TN.xls and P from Midd TP.xls (you will have an opportunity to verify that these are specified as TN and TP later); keep inflow CO2 loadings constant at 0.7 mg/L; for inflow oxygen import Midd DO.xls; import suspended and dissolved detrital loadings from Midd BOD.xls (later you will be able check to make sure it is specified as CBOD); all the biotic loadings (below the suspended and dissolved detrital loadings) are seeds to prevent a group from becoming extinct and should be 1E-5.
Wizard Step 17: Direct precipitation is unimportant to a stream such as this, so all values should be either 0 or ignored.
Wizard Step 18: Point Source Loadings are unimportant for this reach and should be 0; these would be set for effluent such as from a wastewater treatment plant.
Wizard Step 19: Nonpoint-Source Loadings are also unimportant and should be 0 for this reach.
Wizard Step 20: Click the Finish Button and use File | Save to save the changes. You have now finished changing state variables, site constants, and loadings with the Wizard.
Revert to the Main Menu and now check to make sure that N was properly input as “TN,” P as “TP,” and Suspended and Dissolved Detritus as “CBOD” by double-clicking on each in the state variable list in the main screen (The variables to be clicked on are “Nitrate as N,” “Total Soluble P,” and “Susp. and dissolved detritus.”) Look for a check box in each screen.
While in the “Susp. and dissolved detritus.” screen, the % refractory and % particulate detrital allocations (accessed by the appropriate button) are 60% and 10%, which do not need to be adjusted; however, they could be quite different if a factor such as leaf fall were important. If seasonal differences or site-specific data are available, the user may enter time-varying refractory and particulate percentages. More information about these parameters and how to populate them may be found in section 5.1 of the Technical Documentation
Click on the Site button and then the Edit Underlying Site Data button to review and change other site constants. Altitude = 760 m; Total Length for Enhanced Phytoplankton/Zooplankton Retention = 63 km (distance to diversion dam). This last parameter is located toward the very bottom of the screen.
From the Main Screen, click on the Setup Button and choose Run model in Spin-up Mode under Biota Modeling Options so that the biotic end conditions will be used as the initial conditions in the next simulation. Do not choose to spin up nutrients and suspended and bed sediments. (Turn this off when you have a fully calibrated model.) Also, unless storage space is a problem, choose all the state variables for saving rates (click on the Save Rates radio button and Rate specifications button, then use the >> button to move all variables into the Track Rates column).
Be sure to save the study! Then click on Control to run the model. You will get a reminder that it is in spin-up mode, which you should ignore; you will also get a warning that there is zero habitat for adult largemouth bass—cancel, double-click on bass in the state variable list, and set “Percent in Pool” = 50% in Edit Underlying Data, about 2/3 down the screen. The once again click on Control.
When run is complete (which will take at least several minutes), click on Output. The observed data from Rum River will still display, so choose External Data from the pull-down menu in the upper right corner of the screen. Delete “Obs. chlorophyll a” and select “Obs. periphyton chl a”. Click on Import Data and choose “Mid peri chl norm.xls.” You may wish to open the file first in Excel to make sure where the data are located; in this case the default sheet 1, row 1, columns a and b are satisfied, so choose Import and overwrite data. Keep the series name as “Obs. periphyton chl a” and the units as “mg/sq.m.” The “Periphyton chl a” graph will now automatically update with these observed data. When the data are plotted the simulation results will span only a short period compared to the observed data. Click on Edit and Graph Date Range button called Reset (in the upper right corner), then click on the OK button. By choosing Statistics from Graph in the pull-down menu (on upper right), you can get the statistics for the modeled variables, but not the observed data. Also, from the pull-down menu (or right-clicking on the graph itself) you can Copy, Print, and alter the graph (using Graph Setup).
Click on New, choose Custom Graph and click OK, then label the Graph Name as “Depth” and select "Zmean (Dynamic)”, and use the “>” to add Zmean to the plot depth on the Y1 axis. Click the OK button and you will see that a constant, initial depth was used in the simulation. This is not realistic for a stream that is hydraulically regulated for purposes of seasonal irrigation. Therefore we will change the way that mean depth is modeled.
Exit Output and click on the large Site button and Show Mean Depth / Evap button; choose Change in the Depth panel, then Import “Midd Depth.xls” and re-run simulation (perhaps saving results as “perturbed” to help see the effects of this change). “Midd Depth.xls” was created using Manning’s equation with observed discharge in an Excel spreadsheet originally calibrated for the Glenwood Bridge site on the Lower Boise River (Glenwood Depth Discharge calibration.xls). The new model result fits the sparse periphyton data equally well, but is more defensible.
You may wish to run the simulation over again to obtain better initial conditions for the animals due to spin-up mode. Remember to turn off spin-up mode when you have obtained a calibrated model.
Also, click on the “Save-As” icon (4th from left on toolbar) often to save the study. Saving iterative versions of the model under different names along with a log file that describes changes between versions is recommended.
Ambrose, J., Robert B. , J. L.
Martin, and T. A. Wool. 2006. WASP7 Benthic Algae - Model Theory and User's
Guide, Supplement to Water Quality Analysis Simulation Program (WASP) User
Documentation Page 32. U.S. Environmental
Protection Agency, Washington, DC
Barber, M. C. 2003. A Review and Comparison of Models for Predicting Dynamic Chemical Bioconcentration in Fish. Environmental Toxicology and Chemistry, 22(9):1963-1992.
Di Toro, D. M. 2001. Sediment Flux Modeling. Wiley-Interscience, New York.
Iman, R. L., and W. J. Conover. 1982. A distribution-free approach to inducing rank correlation among input variables. Communications in Statistics B11: 311-334.
Park, R.A., R.V. O'Neill, J.A. Bloomfield, H.H. Shugart, Jr., R.S. Booth, J.F. Koonce, M.S. Adams, L.S. Clesceri, E.M. Colon, E.H. Dettman, R.A. Goldstein, J.A. Hoopes, D.D. Huff, S. Katz, J.F. Kitchell, R.C. Kohberger, E.J. LaRow, D.C, McNaught, J.L. Peterson, D. Scavia, J.E. Titus, P.R. Weiler, J.W. Wilkinson, and C.S. Zahorcak. 1974. A Generalized Model for Simulating Lake Ecosystems. Simulation, 23(2):30-50. Reprinted in Benchmark Papers in Ecology.
Park, R.A., D. Scavia, and N.L. Clesceri. 1975. CLEANER, The Lake George Model. In Ecological Modeling in a Management Context. Resources for the Future, Inc., Washington, D.C.
Park, R.A. 1978. A Model for Simulating Lake Ecosystems. Center for Ecological Modeling Report No. 3, Rensselaer Polytechnic Institute, Troy, New York, 19 pp.
Park, R.A., T.W. Groden, and C.J. Desormeau. 1979. Modifications to the Model CLEANER Requiring Further Research. In Perspectives on Lake Ecosystem Modeling, edited by D. Scavia and A. Robertson. Ann Arbor Science Publishers, Inc., 22 pp.
Park, R.A., C.D. Collins, C.I. Connolly, J.R. Albanese, and B.B. MacLeod. 1980. Documentation of the Aquatic Ecosystem Model MS.CLEANER, A Final Report for Grant No. R80504701, U.S. Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia. 112 pp.
Park, R.A., C.I. Connolly, J.R. Albanese, L.S. Clesceri, G.W. Heitzman, H.H. Herbrandson, B.H. Indyke, J.R. Loehe, S. Ross, D.D. Sharma, and W.W. Shuster. 1980. Modeling Transport and Behavior of Pesticides and Other Toxic Organic Materials in Aquatic Environments. Center for Ecological Modeling Report No. 7. Rensselaer Polytechnic Institute, Troy, New York. 163 pp.
Park, R.A., C.I. Connolly, J.R. Albanese, L.S. Clesceri, G.W. Heitzman, H.H. Herbrandson, B.H. Indyke, J.R. Loehe, S. Ross, D.D. Sharma, and W.W. Shuster. 1982. Modeling the Fate of Toxic Organic Materials in Aquatic Environments. U.S. Environmental Protection Agency Rept. EPA-6OO/S3-82-028, Athens, Georgia.
Park, R.A. 1984. TOXTRACE: A Model to Simulate the Fate and Transport of Toxic Chemicals in Terrestrial and Aquatic Environments. Acqua e Aria, No. 6, p. 599-607 (in Italian).
Park, R.A. 1990. AQUATOX, a Modular Toxic Effects Model for Aquatic Ecosystems. Final Report, EPA-026-87; U.S. Environmental Protection Agency, Corvallis, Oregon.
Raimondo, S., D.N.
Vivian, and M.G. Barron. 2010. Web-based
Interspecies Correlation Estimation (Web-ICE) for Acute Toxicity: User Manual.
Version 3.1. EPA/600/R-10/004. Office of Research and Development, U. S.
Environmental Protection Agency. Gulf Breeze, FL.
Saltelli,
A. 2001. Sensitivity Analysis for Importance Assessment. Paper read at
Sensitivity Analysis Methods, June 11-12, 2001, at North Carolina State
University, Raleigh NC
Scavia, D., and R.A. Park. 1976. Documentation of Selected Constructs and Parameter Values in the Aquatic Model CLEANER. Ecological Modelling 2(1):33-58.
Suárez, L.A., and M.C. Barber. 1992. PIRANHA Version 2.0, FGETS Version 3.0-11 User's Manual, In PIRANHA Pesticide and Industrial Chemical Risk Analysis and Hazard Assessment. Athens, Georgia: U.S. Environmental Protection Agency.
U.S. Environmental Protection Agency. 1997. Guiding Principles for Monte Carlo Analysis. Risk Assessment Forum. Washington, DC: U.S. Environmental Protection Agency
U.S. Environmental Protection Agency. 2009. AQUATOX Technical Note 1, A Calibrated Parameter Set for Simulation of Algae in Shallow Rivers, EPA-823-R-09-003 February 2009