Web Access to RSIG Data

Researchers with sufficient programming experience can use a web service to obtain data from RSIG and Estuary Data Mapper (EDM) without having to use the RSIG or EDM applications.

End users who are interested in bulk data transfer may opt to use the web service directly, or incorporate it into an automated script.

You may also want to read more about web service queries and essential performance tips.

• About the web service
• Anatomy of a web service call
• Popular RSIG data sources
• Output data formats
• Command-line examples using cURL
• Code used to perform CMAQ column integration

About the RSIG API web service

The RSIG web service (rsigserver) is OGC-WCS/WMS-compliant.

The following links exit the site

• Open Geospatial Consortium's Web Coverage Service (WCS) standards 
• Open Geospatial Consortium's Web Map Service (WMS) 

Essential information about the web service is as follows:

• The purpose of rsigserver is to stream (over a network) aggregated (across multiple source data files) subsets (by time-range, variables, longitude-latitude-layer domain) of data into client applications (without the need for first writing files to disk, as in ftp).
• EMVL extends rsigserver each time a new data source is requested for RSIG.
• All data are GMT-hourly.
• Longitude, latitude, elevation (meters above MSL when available), and GMT timestamps are always included implicitly to facilitate geospatial-temporal data alignment.

Anatomy of a web service call

To create a webservice call you must first construct a URL of the proper form, which is outlined below. To get started you may want to obtain data with the RSIG3D desktop application, which prints the URL for every data request it makes. You can copy/paste the URL to use as a basis for constructing other customized URLs. Examples of full URLs are given in the Command Line Example section below.

The URL for typical RSIG API calls is composed of three main parts: a basename, mandatory key/value pairs, and optional key/value pairs. Some of the optional key/value pairs are specific to certain data types. The key/value pairs are always separated by ampersands (&) in the URL string. Some people may find the length of the URL string to be long and daunting, but if you break it down into its constituent components it is easy to understand.

Basename: The basename below specifies the RSIG server and web service. The question mark signifies the end of the basename and the beginning of a series of key/value pairs to follow that specify details about the data that is being requested.

• https://ofmpub.epa.gov/rsig/rsigserver?

Mandatory key/value pairs: The mandatory key/value pairs are required in part to make the RSIG API OGC-WCS compliant. They also describe basic details about the data request including the data source, the desired format,  and the spatial and temporal extents of the data being requested as shown in the table below.

Optional key/value pairs: Other key/value pairs can be specified to tailor your request.

Data specific key/value pairs: Some data sources have key/value pairs that affect data filtering. If omitted, the filter values will default to the values recommended by the respective science teams for general use.

Putting it all together: To construct a URL, start with the basename and then append key value pairs separated by ampersands. Make sure that there are no spaces or line breaks anywhere in the URL. The order of the key/value pairs does not matter, but standard practice is to list the mandatory ones first, followed by the optional ones. Here is an example that will request AirNow PM_2.5 data for August 1, 2025 in ASCII (plain text) format.  More examples can be found in the Command Line Examples below.

https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&COVERAGE=airnow.pm25&FORMAT=ascii&TIME=2025-08-01T00:00:00Z/2025-08-01T23:59:59Z&BBOX=-126,24,-66,50

Popular RSIG data sources

The table below lists some of the popular data sources available via RSIG, but is by no means exhaustive. RSIG can serve over 5,000 data sources and variables. See the full list of RSIG variables for more information. Note that a given source (e.g. airnow) may have many variables that can be requested (e.g. pm25, ozone, no2). 

WCS Output data formats

Command-line examples using cURL

To help craft WCS queries, you can use cURL, a command available on multiple platforms for transferring data. As summarized on the cURL home page :

curl is an open source command line tool and library for transferring data with URL syntax that compiles and runs under a wide variety of operating systems....

The easiest way to learn these WCS commands is to run RSIG3D and inspect the status window in the lower left hand corner after a data request is made. The exact WCS call used by the application is echoed to the status window, and can be used as a starting point for constructing custom WCS calls.

Note that in the examples below, the curl commands should all be entered on a single line.

Note for Windows users: The command-line examples below should be run in a Powershell terminal. You may need to replace 'curl' in the examples below with 'curl.exe'.

List available datasets from the RSIG WCS server

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCapabilities'

Listing just the names of available variables from a script

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCapabilities' |grep name | grep -v WCS | sed 's/<name>//g; s/<\/name>//g' | tr -d '[:blank:]'

Using DescribeCoverage to get more information about AirNow PM_2.5

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=DescribeCoverage&COVERAGE=airnow.pm25'

Getting metadata about TROPOMI NO₂

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetMetadata&COVERAGE=tropomi.offl.no2.nitrogendioxide_total_column&TIME=2026-02-01T17:00:00Z/2026-02-01T17:59:59Z&BBOX=-78,37,-76,38&FORMAT=ascii'

Getting ASCII format data from TROPOMI NO₂

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&COVERAGE=tropomi.offl.no2.nitrogendioxide_total_column&TIME=2026-02-01T17:00:00Z/2026-02-01T17:59:59Z&BBOX=-78,37,-76,38&FORMAT=ascii'

Getting XDR-format MODIS AOD data from NASA Goddard

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&FORMAT=xdr&TIME=2025-07-15T14:00:00Z/2025-07-15T14:59:59Z&BBOX=-126.0,24.0,-66.0,50.0&COVERAGE=modis.mod4.Optical_Depth_Land_And_Ocean' > modis_sample1.xdr

Compression/decompression is optional but recommended when streaming large amounts of data over long-distance networks. To use compression, add the &COMPRESS=1 option to return data that has been gzipped. Windows users may need to get a decompression utility such as 7zip to uncompress .gz files.

Same as above but including MODIS cell corner points

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&FORMAT=xdr&TIME=2025-07-15T14:00:00Z/2025-07-15T14:59:59Z&BBOX=-126.0,24.0,-66.0,50.0&COVERAGE=modis.mod4.Optical_Depth_Land_And_Ocean&CORNERS=1' > modis_sample2.xdr

The CORNERS=1 option will result in 8 additional variables: Longitude_SW, Longitude_SE, Longitude_NW, Longitude_NE, Latitude_SW, Latitude_SE, Latitude_NW, Latitude_NE which are the linearly interpolated/extrapolated corner points surrounding each MODIS ground point center to provide quadrillateral cells for each MODIS data point.

Getting CALIPSO satellite LIDAR data in ascii format

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&FORMAT=ascii&TIME=2023-06-15T00:00:00Z/2023-06-15T23:59:59Z&BBOX=-95.0,34.0,-89.0,38.0,0,40000&COVERAGE=calipso.l1.Total_Attenuated_Backscatter_532&MINIMUM_CAD=20&MAXIMUM_UNCERTAINTY=99'

Getting CMAQ wind data for the lowest grid layer (in IOAPI format)

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&COVERAGE=cmaq.equates.conus.metdot3d.wind&TIME=2005-08-28T20:00:00Z/2005-08-29T01:59:59Z&BBOX=-90,30,-88,32,1,1&FORMAT=netcdf-ioapi' > cmaq_sample.nc

To obtain data for all layers, omit the last two values of BBOX:

BBOX=-90,30,-88,32

To obtain the first three layers:

BBOX=-90,30,-88,32,1,3

Optional compression/decompression

...&COMPRESS=1' | gzip -d > sample.nc

Regridding one day of TEMPO NO2 data to layer one of the CMAQ EQUATES 12km grid in Lambert space (saved as NetCDF IOAPI format)

curl --silent --retry 0 -L --tcp-nodelay --max-time 0 'https://ofmpub.epa.gov/rsig/rsigserver?SERVICE=wcs&VERSION=1.0.0&REQUEST=GetCoverage&FORMAT=netcdf-ioapi&TIME=2026-03-01T17:00:00Z/2026-03-01T18:59:59Z&BBOX=-125,25,-65,50&COVERAGE=tempo.l2.no2.vertical_column_troposphere&CORNERS=1&MINIMUM_QUALITY=normal&MAXIMUM_CLOUD_FRACTION=0.15&MAXIMUM_SOLAR_ZENITH_ANGLE=70.0&REGRID=weighted&LAMBERT=33.0,45.0,-97.0,40.0&ELLIPSOID=6370000,6370000&GRID=459,299,-2556000,-1728000,12000,12000' > cmaq_sample_regridded.nc

Example Windows and UNIX scripts for automated downloads

Contact the RSIG team for a sample script that will automate downloads using cURL.

Code used to perform CMAQ column integration

The C code below is used to numerically integrate data layers (ppmV) and convert to molecules/cm².

/******************************************************************************
PURPOSE: integrateLayers - Integrate data (ppmV) over layers to molecules/cm2.
INPUTS:  const size_t timesteps  Number of timesteps.
        const size_t layers     Number of layers.
        const size_t rows       Number of rows.
        const size_t columns    Number of columns.
        const float* const zf   Layer thicknesses in meters:
                                zf[ timesteps ][ layers ][ rows ][ columns ]
        const float* const dens  Density of dry air in kg/m3:
                                 dens[ timesteps ][ layers ][ rows ][columns]
        float* const data        Concentration data (ppmV) to integrate:
                                 data[ timesteps ][ layers ][ rows ][columns]
OUTPUTS: float* const data        Layer-integrated data in molecules/cm2:
                                 data[ timesteps ][ 1 ][ rows ][ columns ]
NOTES:   Column integrated concentration won't handle BADVAL3 values.

The CMAS M3 package ignores the impact of water vapor variations on the molar
mass (and density) of air, so I will do the same.
This makes the calculation very easy.
Final units should be molecules per cm^2.
Trace gas column =
 Vertical integral of Full Layer height * air mass density / molar mass of air
 * gas mixing ratio * Avogadro’s number

CONC and METCRO Variables:
Layer Height = ZF(i) – ZF (i-1) where  ZF(0) =0;
Air mass density = DENS;
Gas mixing ratio = CONC;

Constants:
REAL, PARAMETER :: AVO   = 6.0221367e23 ! Avogadro's Constant [ number/mol ]
REAL, PARAMETER :: MWAIR = 28.9628 ! mean molecular weight for dry air [g/mol]
                            ! FSB: 78.06% N2, 21% O2, and 0.943% A on a mole
                            ! fraction basis ( Source : Hobbs, 1995) pp. 69-70

REAL, PARAMETER :: DENS_CONV = ( 1.0E3 * AVO / MWAIR ) * 1.0E-6
! convert from kg/m**3 to #/cc
REAL, PARAMETER :: PPM_MCM3  = 1.0E-06
! convert from ppm to molecules / cc mol_Spec/mol_Air = ppm * 1E-06
REAL, PARAMETER :: M2CM      = 1.0E2   ! meters to centimeters
REAL, PARAMETER :: M2CM1     = 1.0E-6  ! 1/ m**3 to 1/ cm**3

******************************************************************************/

static void integrateLayers( const size_t timesteps,
                            const size_t layers,
                            const size_t rows,
                            const size_t columns,
                            const float* const zf,
                            const float* const dens,
                            float* const data ) {

 PRE07( timesteps > 0, layers > 0, rows > 0, columns > 0, zf, dens, data );

 const double cm_per_m = 100.0; /* Centimenters per meter. */

 /* Cubic meters per cubic centimeter: */

 const double m3_per_cm3 = 1.0 / ( cm_per_m * cm_per_m * cm_per_m );

 const double g_per_kg = 1000.0; /* 1000 g air / kg air. */

 const double Avogadro_molecules_per_mol = 6.022140857e23; /* number/mol. */

 /* grams of air / mole of air: */

 const double mean_molecular_weight_of_dry_air_g_per_mol = 28.9628;

 const double kg_per_m3_to_moles_per_cm3 =
   m3_per_cm3 * g_per_kg / mean_molecular_weight_of_dry_air_g_per_mol;

 const double ppm_to_molecules_per_mole = 1e-6 * Avogadro_molecules_per_mol;

 const size_t layer_cells = rows * columns;
 const size_t cells = layers * layer_cells;
 size_t timestep = 0;

 for ( timestep = 0; timestep < timesteps; ++timestep ) {
   const size_t timestep_offset = timestep * cells;
   const size_t output_offset   = timestep * layer_cells;
   const float* timestep_height        = zf   + timestep_offset;
   const float* timestep_density       = dens + timestep_offset;
   const float* timestep_concentration = data + timestep_offset;
   float* timestep_integration         = data + output_offset;
   size_t layer_cell = 0;

   for ( layer_cell = 0; layer_cell < layer_cells; ++layer_cell,
         ++timestep_height, ++timestep_density, ++timestep_concentration,
         ++timestep_integration ) {
     size_t layer = 0;
     double previous_height_m = 0.0;
     double sum = 0.0;

     for ( layer = 0; layer < layers; ++layer ) {
       const size_t layer_offset = layer * layer_cells;

       const double height_m = timestep_height[ layer_offset ];
       const double cell_layer_thickness_m = height_m - previous_height_m;
       const double cell_layer_thickness_cm =
         cell_layer_thickness_m * cm_per_m;

       const double density_kg_per_m3 = timestep_density[ layer_offset ];
       const double density_moles_per_cm3 =
         density_kg_per_m3 * kg_per_m3_to_moles_per_cm3;

       const double concentration_ppm = timestep_concentration[ layer_offset];
       const double concentration_molecules_per_mole =
         concentration_ppm * ppm_to_molecules_per_mole;

       const double term_molecules_per_cm2 =
         cell_layer_thickness_cm *
         density_moles_per_cm3 *
         concentration_molecules_per_mole;
       sum += term_molecules_per_cm2;
       previous_height_m = height_m;
     }

     *timestep_integration = sum; /*Write layer-integrated result to layer 1*/
   }
 }
}