Skip to main content
U.S. flag

An official website of the United States government

Here’s how you know

Dot gov

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

HTTPS

Secure .gov websites use HTTPS
A lock (LockA locked padlock) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

    • Environmental Topics
    • Air
    • Bed Bugs
    • Cancer
    • Chemicals, Toxics, and Pesticide
    • Emergency Response
    • Environmental Information by Location
    • Health
    • Land, Waste, and Cleanup
    • Lead
    • Mold
    • Radon
    • Research
    • Science Topics
    • Water Topics
    • A-Z Topic Index
    • Laws & Regulations
    • By Business Sector
    • By Topic
    • Compliance
    • Enforcement
    • Laws and Executive Orders
    • Regulations
    • Report a Violation
    • Environmental Violations
    • Fraud, Waste or Abuse
    • About EPA
    • Our Mission and What We Do
    • Headquarters Offices
    • Regional Offices
    • Labs and Research Centers
    • Planning, Budget, and Results
    • Organization Chart
    • EPA History

Breadcrumb

  1. Home
  2. Causal Analysis/Diagnosis Decision Information System (CADDIS)

Estimating Taxon-Environment Relationships: Non-Parametric Regressions

  • Overview
  • Using Taxon-Environment Relationships
  • Estimating Taxon-Environment Relationships
  • Computing Inferences
  • R Scripts

Non-Parametric Regressions

Helpful Links
Topics in Estimating Taxon-Environment Relationships
  • Overview
  • Data Requirements
  • Central Tendencies
  • Environmental Limits
  • Parametric Regressions
  • Non-Parametric Regressions
  • Assessing Model Fit
  • Measurement Error
  • Survey Weighting
  • Placing Taxa into Tolerance Categories
Other Pages and Websites
  • Using R for Non-Parametric Regression
  • Maximum Likelihood Inferences

PECBO Appendix Site Map

Many researchers have noted that unimodal relationships are unlikely for all taxa across all gradients (Austin and Meyers 1996, Oksanen and Minchin 2002). To address this issue, a modeling approach that requires only that the modeled function vary smoothly and slowly over the modeled range is often used. Here, the distribution of a given taxon is modeled in Equation 8:

EQUATION 8
EQUATION 8
where p is the probability of occurrence of the taxon of interest, s0 is a constant, and s represents a nonparametric smooth curve that is fit through the data.

The locations of the mean responses for each point along a nonparametric curve, s, are determined through an iterative procedure that uses data in a local neighborhood around each point. The local nature of the fit differs fundamentally from that of a parametric model, which computes a best fit for the entire set of data. Thus, nonparametric responses can capture smaller scale variations in response (Figure 7). Near an edge of the domain, though, less data are available on one side of the point of interest, and increasing amounts of data must be drawn from within the sampled range. Therefore, the width of the neighborhood broadens at the extremes of the environmental gradient, and the fit is less local than in the center of the domain (Hastie and Tibshirani 1999).

Probability of occurrence and stream temperature for Heterlimnius and Malenka.
Figure 7. Relationship between probability of occurrence and stream temperature (degrees C) for Heterlimnius and Malenka using a nonparametric model.The solid line shows the mean relationship between probability of occurrence and temperature. Dotted lines indicate the estimated 90% confidence limits about the location of the regression curve. Each open circle represents the average occurrence probability in approximately 20 samples surrounding the indicated temperature.

A commonly used method for fitting nonparametric curves is known as the generalized additive model (GAM) (Hastie and Tibshirani 1999). This method allows for more than one explanatory variable, each associated with its own nonparametric smooth curve.

The flexibility of nonparametric regressions also complicates this method's use because there are no parameters with which the modeled relationship can be represented. Instead, a numerical representation of the entire curve must be stored for further analysis (e.g., to compute inferences).

As with parametric regressions, several variables can be modeled simultaneously with non-parametric regressions. Each additional variable can be treated as independent, additive smooth functions, or combinations of variables can be smoothed simultaneously, which allows for interactions between variables (Figure 8).

Probability of occurrence and temperature (°C) and percent sand/fines (SED) for Heterlimnius and Malenka.
Figure 8. Non-parametric relationship between probability of occurrence and temperature (°C) and percent sand/fines (SED) for Heterlimnius and Malenka. Contours represent mean probabilities of occurrence.
Statistical scripts for computing single variable nonparametric regressions are available under the R Scripts tab of this section.
 
Once taxon-environment relationships have been estimated using parametric regression, the most appropriate method to use for computing inferences is a maximum likelihood approach.
 

References

  • Austin MP, Meyers JA (1996) Current approaches to modeling the environmental niche of eucalyptus: implication for management of forest biodiversity. Forest Ecology and Management 85:95-106.
  • Hastie TJ, Tibshirani RJ (1999) Generalized Additive Models. Chapman & Hall/CRC, Washington DC.
  • Oksanen J, Minchin PR (2002) Continuum theory revisited: what shape are species responses along ecological gradients. Ecological Modelling 157:119-129.

Causal Analysis/Diagnosis Decision Information System (CADDIS)

  • CADDIS Home
    • About CADDIS
    • Frequent Questions
    • Publications
    • Recent Additions
    • Related Links
    • CADDIS Glossary
  • Volume 1: Stressor Identification
    • About Causal Assessment
    • Getting Started
    • Step 1. Define the Case
    • Step 2. List Candidate Causes
    • Step 3. Evaluate Data from the Case
    • Step 4. Evaluate Data from Elsewhere
    • Step 5. Identify Probable Causes
  • Volume 2: Sources, Stressors and Responses
    • About Sources
      • Urbanization
    • About Stressors
  • Volume 3: Examples and Applications
    • Analytical Examples
    • Worksheet Examples
    • State Examples
    • Case Studies
    • Galleries
  • Volume 4: Data Analysis
    • Selecting an Analysis Approach
    • Getting Started
    • Basic Principles & Issues
    • Exploratory Data Analysis
    • Basic Analyses
    • Advanced Analyses
    • PECBO Appendix
    • Download Software
    • Data Analysis Topics (A -Z)
  • Volume 5: Causal Databases
    • Learn about CADLink
Contact Us about CADDIS
Contact Us about CADDIS to ask a question, provide feedback, or report a problem.
Last updated on February 13, 2025
  • Assistance
  • Spanish
  • Arabic
  • Chinese (simplified)
  • Chinese (traditional)
  • French
  • Haitian Creole
  • Korean
  • Portuguese
  • Russian
  • Tagalog
  • Vietnamese
United States Environmental Protection Agency

Discover.

  • Accessibility Statement
  • Budget & Performance
  • Contracting
  • EPA www Web Snapshot
  • Grants
  • No FEAR Act Data
  • Plain Writing
  • Privacy
  • Privacy and Security Notice

Connect.

  • Data
  • Inspector General
  • Jobs
  • Newsroom
  • Regulations.gov
  • Subscribe
  • USA.gov
  • White House

Ask.

  • Contact EPA
  • EPA Disclaimers
  • Hotlines
  • FOIA Requests
  • Frequent Questions
  • Site Feedback

Follow.