WASHINGTON, Oct. 28, 2021 – USDA’s National Agricultural Statistics Service and Agricultural Research Service have announced enhancements to the CropScape web app, allowing users to more easily conduct area and statistical analysis of planted U.S. commodities. Now known as CroplandCROS (
), the geospatial data product hosts the Cropland Data Layer (CDL). The app allows users to geolocate farms and map areas of interest. To aid users, the app features a user guide and instructional videos.
- What data is hosted on the CropScape website?
The geospatial data product called the Cropland Data Layer (CDL) is hosted on CropScape (https://nassgeodata.gmu.edu/CropScape/). The CDL is a raster, geo-referenced, crop-specific land cover data layer created annually for the continental United States using moderate resolution satellite imagery and extensive agricultural ground truth. All historical CDL products are available for use and free for download through CropScape. For more information about the CDL Program please refer to the metadata for the particular state and year you are interested at the following web page: (/Research_and_Science/Cropland/metadata/meta.php).
- Where can I find instructions on how to use the tools provided through the CropScape website?
There are four buttons in the upper right-hand corner of the CropScape website that offer tutorials and basic instructions. These options include a "Demo Video", "Help", "Developer Guide", and "FAQ".
- What future updates do you expect for CropScape?
Further CropScape enhancements will depend on user feedback and resources available.
- Who created the Cropland Data Layer and who developed the CropScape web service?
The Cropland Data Layer (CDL) was created by the USDA, National Agricultural Statistics Service, Research and Development Division, Geospatial Information Branch, Spatial Analysis Research Section. The most current data is available free for download along with extensive metadata, FAQs, and other detailed technical information at the following website: /Research_and_Science/Cropland/SARS1a.php. NASS developed both the CropScape and VegScape web services in cooperation with the Center for Spatial Information Science and Systems, George Mason University, Fairfax, VA.
- Why was the Cropland Data Layer created?
The purpose of the Cropland Data Layer Program is to use satellite imagery to provide acreage estimates to the Agricultural Statistics Board for major commodities and to produce digital, crop-specific, categorized geo-referenced output products.
- What differences can be expected when comparing CropScape pixel counts and official NASS statistics for counties, ASD, and states?
There will be differences between CropScape and official NASS estimates when comparing acreage statistics at the state, district, and county levels. Statistics generated by CropScape are dependent upon pixel counting. Pixel counting is usually downward biased when compared to the official estimates. Counting pixels and multiplying by the area of each pixel will result in biased area estimates and should be considered raw numbers needing bias correction. Official crop acreage estimates at the state and county level are available at QuickStats.
Here are a list of references discussing the subject matter of pixel counting and estimation:
a) Gallego F.J., 2004, Remote sensing and land cover area estimation. International Journal of Remote Sensing. Vol. 25, n. 15, pp. 3019-3047.
b) European Commission, Joint Research Centre, MARS; Best practices for crop area estimation with Remote Sensing - Section 4.1.1.
c) Czaplewski, R. L. (1992). Misclassification bias in areal estimates. Photogrammetric Engineering and Remote Sensing, 58, 189-192.
- Can I download the entire country in one file through CropScape?
Zoom to the national scale map, choose the year that you want to download and click on the "Download Defined Area of Interest Data" button on the toolbar. Respond "yes" to the download confirmation question. The downloadable file will be a Winzip compressed file containing the CDL in a GeoTIFF (TIF) file format. As an alternative to CropScape, a national CDL mosaic in an Erdas Imagine IMG file format is available for download on the SARS National Download webpage.
- Are color legends available for the Cropland Data Layers?
The following downloadable jpeg files are color legends by year for the Continental United States CDLs:
US_2022_CDL_legend.jpg
US_2021_CDL_legend.jpg
US_2020_CDL_legend.jpg
US_2019_CDL_legend.jpg
US_2018_CDL_legend.jpg
US_2017_CDL_legend.jpg
US_2016_CDL_legend.jpg
US_2015_CDL_legend.jpg
US_2014_CDL_legend.jpg
US_2013_CDL_legend.jpg
US_2012_CDL_legend.jpg
US_2011_CDL_legend.jpg
US_2010_CDL_legend.jpg
US_2009_CDL_legend.jpg
US_2008_CDL_legend.jpg
- Is there a way to view the Cropland Data Layer data on CropScape with just one or two commodities shown at the national, state, district and/or county levels?
To view the Cropland Data Layer data on CropScape with just one or two commodities shown at the national, state, district and/or county levels: 1) Select "Define Area of Interest by State/ASD/County" or "Define Area of Interest" or "Import Area of Interest" from the top toolbar 2) Once you have an area of interest (AOI) defined select the "Area of Interest Statistics" from the top toolbar 3) Choose the commodity of interest from the popup, you can choose one or many 4) Export the selected crop(s) for mapping to create a graphic containing only the selected crop(s) in your defined AOI.
- Has someone compiled all of the exported CDL attribute and raw pixel count acres tables by year and state into a master spreadsheet?
2022_CDL_Histogram_Summary.xlsx
2021_CDL_Histogram_Summary.xlsx
2020_CDL_Histogram_Summary.xlsx
2019_CDL_Histogram_Summary.xlsx
2018_CDL_Histogram_Summary.xlsx
2017_CDL_Histogram_Summary.xlsx
2016_CDL_Histogram_Summary.xlsx
2015_CDL_Histogram_Summary.xlsx
2014_CDL_Histogram_Summary.xlsx
2013_CDL_Histogram_Summary.xlsx
2012_CDL_Histogram_Summary.xlsx
2011_CDL_Histogram_Summary.xlsx
2010_CDL_Histogram_Summary.xlsx
2009_CDL_Histogram_Summary.xlsx
2008_CDL_Histogram_Summary.xlsx
- Is more detailed accuracy assessment information available than what is provided in the metadata? Can you provide the full accuracy assessment error/confusion matrices for all states?
The strength and emphasis of the CDL is crop-specific land cover categories. The accuracy of the CDL non-agricultural land cover classes is entirely dependent upon the USGS, National Land Cover Database (NLCD). Thus, the USDA, NASS recommends that users consider the NLCD for studies involving non-agricultural land cover.
The training and validation data used to create and accuracy assess the CDL has traditionally been based on ground truth data that is buffered inward 30 meters. This was done 1) because satellite imagery (as well as the polygon reference data) in the past was not georeferenced to the same precision as now (i.e. everything "stacked" less perfectly), 2) to eliminate from training spectrally-mixed pixels at land cover boundaries, and 3) to be spatially conservative during the era when coarser 56 meter AWiFS satellite imagery was incorporated. Ultimately, all of these scenarios created "blurry" edge pixels through the seasonal time series which it was found if ignored from training in the classification helped improve the quality of CDL. However, the accuracy assessment portion of the analysis also used buffered data meaning those same edge pixels were not assessed fully with the rest of the classification. This would be inconsequential if those edge pixels were similar in nature to the rest of the scene but they are not as they tend to be more difficult to classify correctly. Thus, the accuracy assessments as have been presented are inflated somewhat.
Beginning with the 2016 CDLs we published both the traditional "buffered" accuracy metrics and the new "unbuffered" accuracy assessments. The purpose of publishing both versions is to provide a benchmark for users interested in comparing the different validation methods. For the 2017 CDL season we are now only publishing the unbuffered accuracy assessments within the official metadata files and offer the full "unbuffered" error matrices for download on this FAQs webpage. We plan to continue producing these unbuffered accuracy assessments for future CDLs. However, there are no plans to create these unbuffered accuracy assessments for past years. It should be noted that accuracy assessment is challenging and the CDL group has always strived to provide robust metrics of usability to the land cover community. This admission of modestly inflated accuracy measures does not render past assessments useless. They were all done consistently so comparison across years and/or states is still valid. Yet, by now providing both scenarios for 2016 gives guidance on the bias.
The full error matrices are included in the downloadable links below.
CDL_2022_accuracy_assessments.zip
CDL_2021_accuracy_assessments.zip
CDL_2020_accuracy_assessments.zip
CDL_2019_accuracy_assessments.zip
CDL_2018_accuracy_assessments.zip
CDL_2017_accuracy_assessments.zip
CDL_2016_accuracy_assessments.zip
CDL_2015_accuracy_assessments.zip
CDL_2014_accuracy_assessments.zip
CDL_2013_accuracy_assessments.zip
CDL_2012_accuracy_assessments.zip
CDL_2011_accuracy_assessments.zip
CDL_2010_accuracy_assessments.zip
CDL_2009_accuracy_assessments.zip
CDL_2008_accuracy_assessments.zip
- What are the "Crop Mask Layer" or "Cultivated Layer" and the "Crop Frequency Data Layers" shown on the CropScape website and is the data available for download?
The CropScape website offers a "Cultivated Layer" or "Crop Mask Layer". It is based on Cropland Data Layers from the most recent five years of CDL data and is updated annually. An Erdas Imagine Spatial Model is used to create the Cultivated Layer. The processing logic is as follows. If a pixel is identified as cultivated in at least two out of the five years of CDL data then it is assigned to the 'Cultivated' category. The exception is that all pixels identified as cultivated in the most recent year are assigned to the 'Cultivated' category regardless of whether or not they were cultivated in the previous four years of CDL data.
The Crop Frequency Layers identify crop specific planting frequency and are based on land cover information derived from every year of available CDL data beginning with the 2008 CDL, the first year of full Continental U.S. coverage. There are currently four individual crop frequency data layers that represent four major crops: corn, cotton, soybeans, and wheat. Please be aware that there is overlap between these layers where double-cropping of these four crop types exists.
The Cultivated Layer and Crop Frequency Data Layers with accompanying metadata detailing the methodology are available for download at /Research_and_Science/Cropland/Release/.
- Where can I find metadata for the data offered on CropScape?
Extensive metadata files are available by state and year in HTM and XML file formats at the following website: (/Research_and_Science/Cropland/metadata/meta.php). As of early 2020 the metadata uses the FGDC-STD-001-1998 standards. We expect to soon transition to the ISO standards.
- What is the preferred citation for the Cropland Data Layer and CropScape?
USDA National Agricultural Statistics Service Cropland Data Layer. {YEAR}. Published crop-specific data layer [Online]. Available at https://nassgeodata.gmu.edu/CropScape/ (accessed {DATE}; verified {DATE}). USDA-NASS, Washington, DC.
- To whom do I address concerns about CropScape or the Cropland Data Layer?
Distribution issues can be directed to the NASS Customer Service Hotline at 1-800-727-9540. Content and technical questions can be directed to the NASS Spatial Analysis Research Section (SARS) at SM.NASS.RDD.GIB@usda.gov.
- I have downloaded a user-defined subset of the Cropland Data Layer from CropScape. I am using ArcGIS to view my download, but there is no histogram information in the attribute table. How do I generate the statistics for my downloaded area in ArcGIS?
To create a pixel "count" field in the attribute table of the downloaded CDL use the "Build Raster Attribute Table" Function in ESRI ArcGIS. In ESRI ArcGIS Version 9.3 and 10 this function is located at ArcToolbox > Data Management Tools > Raster > Raster Properties > Build Raster Attribute Table. Specify the downloaded CDL tif file as the Input Raster and accept all other defaults and click OK. After it has run successfully, a new "Count" data field is added to the attribute table. Count represents a raw pixel count. To calculate acreage multiply the count by the square meters conversion factor which is dependent upon the CDL pixel size. The conversion factor for 30 meter pixels is 0.222394. The conversion factor for 56 meter pixels is 0.774922.
- How do I add category names to a downloaded .tif image in ESRI ArcGIS?
If your downloaded CDL .tif file does not contain category names, then you can add them using the following instructions. Download the file: generic_cdl_attributes.tif.vat.dbf. This generic file contains all possible CDL colors and category names. As long as the .tif file and the .tif.vat.dbf file have the same file name, then the category names will load automatically in ArcMap. So, change the file name (not extension) of the generic_cdl_attributes.tif.vat.dbf to match the file name of the downloaded CDL .tif file. Then add the .tif file as a layer in ArcMap. The category names will display in the Table of Contents window.
Example 1 - If the downloaded .tif file is: _NASS_DATA_CACHE_CDL_2011_clip_20110307142903_862761787.tif Change the generic_cdl_attributes.tif.vat.dbf file name to: _NASS_DATA_CACHE_CDL_2011_clip_20110307142903_862761787.tif.vat.dbf
Example 2 – If you renamed the downloaded .tif file to MyCDL.tif, then rename the generic_cdl_attributes.tif.vat.dbf file name to MyCDL.tif.vat.dbf.
- I am using Erdas Imagine to view the data that I downloaded from CropScape, but there is no histogram information. How do I build statistics in Erdas Imagine?
To generate statistics in Erdas Imagine, go to Tools > Image Information then click on "Compute the statistics." If you have the TIF file open in a Viewer, then you will have to close it and reopen the TIF file. Now when you view the attribute information, there should be a Histogram column, which represents the pixel count per category. To calculate acreage multiply the count by the square meters conversion factor which is dependent upon the CDL pixel size. The conversion factor for 30 meter pixels is 0.222394. The conversion factor for 56 meter pixels is 0.774922.
- How do I add category names and/or colors to my downloaded CDL data in Erdas Imagine?
You first must build statistics for the TIF file as outlined in the question above. To add category names, open the TIF in a Viewer and select Raster > Attributes. In the Raster Attribute Editor select Edit > Add Class Names. This new "Class Names" column can be populated manually or you can download this prepared DAT file located at: /Research_and_Science/Cropland/docs/cdl_class_names.zip. Unzip the CDL names and colors file and save the DAT files to your computer. Then in the Raster Attribute Editor highlight the "Class Names" column by left-clicking on the header of the Class Names column. Next, right-click on the Class Names column and select the Import option. Specify the CDL_Names DAT file as the file to import and this will add all possible CDL class names to your TIF attribute table. You can add colors by importing the CDL_Colors DAT file similar to the steps used to add the class names.
- How does CropScape calculate acreage? Is it an ArcGIS Server geoprocessing task?
The pixel counting algorithm used by CropScape is straightforward. An area of interest (AOI) is defined by an enclosed boundary. The AOI is then rasterized and the pixels that fall within this AOI are counted. ArcGIS server is not used in this application. Please refer to Question 61 or 64 on this FAQs webpage for details on how acreage calculated on CropScape will compare to official NASS acreage estimates.
- What are CropScape's query limits?
CropScape allows users to analyze and interact with areas less than 2,000,000 square kilometers. However, users can download the entire national CDL by year by following the instructions in this FAQ question in this FAQ question (Click Here) which could then be used to perform analysis using their own GIS or image processing software.
- What should I do if a CropScape CDL download repeatedly times out?
If you are using Microsoft Internet Explorer, try changing the default timeout value (greater than 100000ms if your network connection is slow). Here is a link detailing how to change the default timeout value: http://support.microsoft.com/kb/813827. You could also try using a web browser other than Internet Explorer, such as Firefox or Google Chrome.
- What should I do when I get a CropScape error downloading certain states or when I get a pop up window that says scripts are running on this website which may make my computer unresponsive?
This issue is caused by security controls of Internet Explorer when rendering a state with a large boundary file, which can take a long time depending on your internet connection speed. There are three possible solutions:
Solution 1: If you get a window that says "...Do you want to abort the script?", click the "NO" button to continue.
Solution 2: Follow the technical support at http://support.microsoft.com/kb/175500 to download a patch and fix this problem automatically.
Solution 3: Try using another browser, such as Firefox, Safari, or Google Chrome.
- Is there a developer's guide for CropScape?
The George Mason University group responsible for creating the CropScape website offer an online developer's guide available at https://nassgeodata.gmu.edu/CropScape/devhelp/help.html.
- Is REST API or REST Endpoint available in CropScape?
CropScape provides Web geoprocessing services for invocations or workflow in other web geospatial applications. A CropScape Developer Guide is available at https://nassgeodata.gmu.edu/CropScape/devhelp/help.html. There are 7 operations, including GetCDLStat, GetCDLImage, GetCDLComp, GetCDLFile, GetCDLValue, ExtractCDLByValues, and GetCDLPDF. These Web services support HTTP GET/KVP POST/XML and SOAP encoding.
- Is the CDL available as a web mapping service (WMS)?
The WMS is implemented and is available to the general public. It is OGC standard compliant. The CDL can be served as a data layer from the user's application. If you receive an error when adding to ArcCatalog that reads "Could not add the specified data object to the map. Failed to open raster dataset" then try changing the WMS default version to 1.0.0 when you add the WCS.
- Is it possible to filter and display just a single crop or group of crops using the Cropscape WMS layer in a web map? Is it possible to modify the web request to filter for a particular crop?
The CropScape/CDL data are in raster format. Raster data layer is one layer with all crop types in the same layer. Therefore, WMS map has only one layer. You cannot use layerDefs and SLD in WMS to extract individual crop land cover. If you want to extract the individual crop cover, you have to download the CDL data and then extract crop type by specific attribute value(s).
- I'm having trouble finding the address of the service for use in my GIS application. Could you please direct me to a link that will work with ArcGIS?
Please refer to the Developer’s Guide. To add all CDL data layers to ArcGIS use: https://nassgeodata.gmu.edu/CropScapeService/wms_cdlall.cgi. In the Developer’s Guide there are many examples for using CropScape web services. Use EPSG:5070 (CONUS Albers) in the GetMap Request if you would like the CDL in its native projection.
- When adding the CDL WMS to ArcGIS Explorer or ArcCatalog, I get an error message saying "Invalid Format for Exception Parameter at WMS 1.3.0". How do I correct this?
The CDL WMS can be accessed at https://nassgeodata.gmu.edu/CropScapeService/wms_cdlall.cgi. However, the default WMS GetCapabilities request in ArcGIS Explorer or ArcCatalog is version 1.3.0 with EXCEPTIONS=application/vnd.ogc.se_xml, which is not a standard format (XML, INIMAGE or BLANK). Try changing the WMS version from default 1.3.0 to 1.1.0 when adding to the WMS server. For convenience, ArcGIS users can access CDL WMS at https://nassgeodata.gmu.edu/CropScapeService/wms_cdlall.cgi. Another option for ArcGIS users is to set the default WMS version back to 1.1.0 as shown in these screenshots: updated screenshots available soon.
- In order to write a tool in Python that accesses the CropScape/NASS server to download files for your year and area of interest (AOI), is there a way to programmatically upload the AOI in order to utilize the GET Request function?
The area of interest must be a compressed ESRI Shapefile. The .shp, .shx, .dbf, and .prj files must all be compressed with no subdirectories in a single ZIP file. Next, you will need to publish your ZIP file to a website URL (internal or external). The use the example below where "UUUU" represents your URL and "YYYY" is the year of interest.
https://nassgeodata.gmu.edu/axis2/services/CDLService/GetCDLFile?year=YYYY&aoiURL=UUUU
If you receive a "HTTP Error 500", then check that you are using a static link rather than dynamic. If you are using BOX then this link may help with creating a static link (https://community.box.com/t5/Archive-Forum/How-to-mass-download-Static-Share-Links/td-p/11973).
- Are the data on CropScape available to be consumed in a mash up in the ArcExplorer web viewer?
ArcGIS Explorer Online supports map, image and feature services from ArcGIS Server (http://doc.arcgis.com/en/arcgis-online/create-maps/add-layers.htm). For ArcGIS Explorer, the CDL could be added as one of its data layers via WMS. You can access the CDL WMS at https://nassgeodata.gmu.edu/CropScapeService/wms_cdlall.cgi.
You can learn how to add a legend of a WMS layer in ArcGIS Explorer Desktop at http://webhelp.esri.com/arcgisexplorer/900/en/legend_window.htm.
Currently, we do not define the legend as a sublayer of CDL layers in our CDL WMS, so the legend for your selected CDL layers can not be displayed like the guide suggests. But, you could try to access the legend by sending a request: https://nassgeodata.gmu.edu/CropScapeService/wms_cdlall.cgi?version=1.0.0&service=wms&request=
getlegendgraphic&layer=cdl_2009&format=image/png.
You can generate a kml file for the CDL data of the area of interest in CropScape, then add the kml file as one layer in ArcGIS Explorer Desktop, then right click data layer, and select "Show Popup", and the legend will be shown in the popup window.
- What should I do if I get an error message upon startup of CropScape stating "To perform all operations successfully, please download and install Adobe Flash Player plug-in in your browser"?
CropScape is optimized for use with Adobe Flash player. Please upgrade your Adobe Flash player at http://www.adobe.com/.
- Can the CropScape data be exported to a KML or KMZ file for use in Google Earth, Bing Maps, or to your local hard drive?
CropScape data can be exported to a KML format that is downloaded to your local drive. This KML file can then be used in Google Earth if desired. Instructions for how to download data and specify the KML format are detailed in the Section 3.e.ii of the Help hyperlink in the upper righthand corner of the CropScape webpage. For Bing Maps you will need to write JavaScript code using the Bing Map API to add a KML layer, please follow the instructions at: http://msdn.microsoft.com/en-us/library/cc316942.aspx.
- Is it possible to create a CropScape PDF map with only one specific category or group of categories shown?
CropScape currently does not have the capability to create PDF maps with only one specific crop or group of crops shown. However, the user can export the actual CDL data with only a single crop or subset of crops in a Geotiff (TIF) format using the CropScape "Area of Interest Statistics" tool. That downloaded data can then be used to create a more polished PDF map using ESRI ArcGIS software. Below are the procedures:
1. Select your area of interest;
2. Use the "Area of Interest Statistics" tool to calculate statistics;
3. Check the box of the crop type(s) you wish to display in the pop-up statistical result window;
4. Click "Export the selected crop(s) for mapping";
5. Click the "Download" button to download the resulting image in a Geotiff (TIF) file format;
6. Load the downloaded TIF file in ArcGIS where you can then add additional data, such as boundaries and/or legends, to create your own map that can then be exported to a PDF file.
- What is the recommended projection for the "Import Area of Interest" option?
The recommended projection for your imported AOI data is Albers Equal Area Conic Projection (EPS:5070).
- To whom do I address concerns about CropScape or the Cropland Data Layer?
Distribution issues can be directed to the NASS Customer Service Hotline at 1-800-727-9540. Content and technical questions can be directed to the NASS Spatial Analysis Research Section (SARS) at SM.NASS.RDD.GIB@usda.gov.
- Where can I find metadata about the Cropland Data Layer?
Extensive metadata records are available by state and year at the following webpage: (/Research_and_Science/Cropland/metadata/meta.php). The metadata uses the FGDC-STD-001-1998 standards. We expect to soon transition to ISO standards.
- Who created the CDL datasets and how has the CDL Program changed over time?
Originally, field preparation and digitizing work were performed in NASS Field Offices and the remote sensing analysis performed by the Spatial Analysis Research Section (SARS) of NASS. However, in 1997 SARS entered into a data sharing partnership with USDA's Foreign Agricultural Service and USDA's Farm Service Agency. The agreement provided access to Landsat 5 coverage in the states selected for the project by SARS. The first states covered with the data sharing partnership were Arkansas, North Dakota and South Dakota. Improvements in hardware along with software enhancements made program expansion possible for the 1999 growing season. NASS Research Development Division solicited additional states to find outside cooperators/partners to provide an analyst and hardware to perform duties associated with the Acreage Estimation Program. The Illinois and Mississippi State Field Offices were able to obtain partnership agreements with external State/Federal Agencies.
For crop year 2000, the states of Iowa and Indiana were added to the Program. North Dakota was able to obtain a partner for the 2000 crop year cooperatively with North Dakota State University (NDSU) through an EPA water quality grant for 5 years. Indiana was added to the program for crop year 2000 also, but as a regional type center where the ground data collection and digitization was performed at the Indiana State Office, and the acreage estimation was performed at the Illinois State Office.
For crop year 2001, the Missouri southeastern boot heel area was added to the program. All boot heel digitizing was performed by the Missouri Ag Statistics Service, and image processing duties were performed by the Arkansas Ag Statistical Service. Nebraska and Wisconsin were added as pilot states, where all digitizing was performed by the Nebraska and Wisconsin Ag Statistics Service offices respectively, and image processing functions were performed by SARS. Maryland/Delaware were also added as a pilot program where digitizing was done by the University of Maryland Mid-Atlantic RESAC group, and image processing was performed by the SARS group.
For crop year 2002, Nebraska expanded to full state coverage, and Wisconsin expanded to full state coverage in 2003. In 2002, a ten state Mid-Atlantic based Cropland Data Layer product was sponsored in part by a NASA/Raytheon/Synergy Project through Towson University, with the digitizing and image analysis performed under contract by NASS. The Mid-Atlantic CDL products were based on the 2002 June Agricultural Survey and the Agriculture Coverage Evaluation Survey (ACES) that coincided with the 2002 Agricultural Census.
For crop year 2004, the IRS Resourcesat-1 AWiFS sensor was used over Nebraska, Indiana and Arkansas to perform acreage analysis. The AR, IN and NE CDL's were released with both Landsat TM classifications as well as AWiFS classifications. The AWiFS sensor has 56 meter spatial resolution, and five day repeat coverage. The best possible scene dates taken during the month of August 2004 were used to create the AWiFS CDL products. A cooperative partnership between University of Maryland Department of Geography and SARS helped process the Louisiana 2004 CDL.
A Florida CDL for 2004 was released in February of 2007 using Landsat 5/7 imagery. The Florida CDL was the first CDL created exclusively with See5, and it was the first usage of the segmentation based gap filled Landsat-7 SLC-off imagery. It included the first usage of the Farm Service Agency/Common Land Unit and the Florida Citrus Grove layer for ground truth training.
For crop year 2005, the Idaho Cropland Data Layer was created with a cooperative partnership between Utah State University, the United Potato Growers of Idaho and NASS. This partnership produced both a Landsat TM and Resourcesat-1 AWiFS classification over the Idaho Snake River Plain. The 2005 Midwestern CDL update contained new AWiFS classifications and a revised Wisconsin TM based classification. The new AWiFS classifications cover Nebraska and North Dakota. The Wisconsin revision was performed under contract for the Wisconsin State, Bureau of Environmental and Occupational Health and Department of Health and Family Services. The 2005 Mississippi Delta Region was classified using the regression tree classifier See5.0 available from www.rulequest.com over the 2001 NLCD defined mapping Zone 45 https://www.mrlc.gov/ for the States of Arkansas, Louisiana and Missouri. The Zone 45 classification results from See5.0 were overlaid on top of the Arkansas, Louisiana and Missouri bootheel, resulting in an accurate ag classification and an enhanced non-ag land cover classification leveraging results from the 2001 NLCD products. The traditional pixel based PEDITOR classification covers the remaining parts of these states.
The 2006 Delta/Midwestern/Pacific Northwest CDL products covered eleven states: AR, IL, IN, IA, LA, MO, MS, NE, ND, WA, WI. Illinois and Indiana were processed with Peditor. The remaining States were processed using See5 decision tree software. The Mississippi Delta CDL and the remaining Midwestern and Prairie States were processed exclusively with See5 using the FSA Common Land Unit for ground truth. The 2006 Washington CDL does have a smoothing algorithm applied to remove pixel scatter. This is the only state and year of CDL that has any level of post-classification smoothing.
The 2007 CDL product became operational in NASS delivering for the first time in-season acreage estimates for the October 2007 Crop Report across all speculative corn and soybean states. Twenty-one states total (AR, CA, IL, IN, IA, ID, KS, LA, MI, MN, MO, MS, MT, ND, NE, OH, OK, OR, SD, WA, WI) were processed into CDL's. Additionally, new CDL's were created for crop year 2006 for KS, MN, MO, OH, OK, SD. Michigan State University/Land Policy Institute entered into a cooperative partnership with SARS and obtained funding to provide an image analyst to process Michigan.
The 2008 crop year is the first year that the entire Continental United States is covered by the CDL. Real-time CDL acreage estimates were produced for the June Ag Survey for winter wheat, the August Crop Report and the October Crop Report for corn and soybeans. The 2008 CDL was reprocessed and released on 12/11/2017. The 2008 and 2009 Cropland Data Layers (CDL) for the entire Continental United States have been reprocessed and re-released at a 30 meter spatial resolution. The move from 56m to 30m resolution was made possible with the inclusion of Landsat 5 Thematic Mapper data, which was not freely available during the initial processing period. Additionally, the reprocessing effort used more complete Farm Service Agency administrative data for training and accuracy assessing the classifications.
The 2009 crop year again offers coverage for the entire continental United States and produced real-time CDL acreage estimates for the June Ag Survey and September Small Grain Summary for winter wheat, the August, September and October Crop Reports for corn, soybeans, rice and cotton. 2009 was the inaugural year for coverage of the Continental United States. The original 2009 product was released at 56 meters resolution. The 2008 CDL was reprocessed and released on 12/11/2017. The 2008 and 2009 Cropland Data Layers (CDL) for the entire Continental United States have been reprocessed and re-released at a 30 meter spatial resolution. The move from 56m to 30m resolution was made possible with the inclusion of Landsat 5 Thematic Mapper data, which was not freely available during the initial processing period. Additionally, the reprocessing effort used more complete Farm Service Agency administrative data for training and accuracy assessing the classifications.
The 2010 CDL product was released the first week of January 2011 co-incident with the release of the new CropScape web service. The 2010 product utilized Landsat TM/ETM+ and AWiFS imagery for production of a 30m product covering the Continental United States.
The 2011 CDL product was released January 31, 2012. The 2011 product utilized Deimos-1, UK-DMC 2, Landsat TM/ETM+, and AWiFS imagery for production of a 30m product covering the Continental United States. Coincident with the release of the 2011 product, the entire historical CDL catalog was re-released with minor category code and class name revisions. These revisions were done to eliminate redundant or unused categories. Please view the 2011 crosswalk document for a detailed listing of the revisions.
The 2012 CDL product was released January 31, 2013. The 2012 product utilized Deimos-1, UK-DMC 2, and Landsat TM/ETM+ imagery for production of a 30m product covering the Continental United States.
The 2013 CDL product was released January 31, 2014. The 2013 product utilized Deimos-1, UK-DMC 2, and Landsat 8 imagery for production of a 30m product covering the Continental United States. Coincident with the release of the 2013 product, the entire historical CDL catalog was re-released with minor category code and class name revisions. These revisions were done to eliminate redundant or unused categories. Please view the 2013 crosswalk document for a detailed listing of the revisions.
The 2014 CDL product was released February 2, 2015. The 2014 CDL product utilized Deimos-1, UK-DMC 2, and Landsat 8 imagery for production of a 30m product covering the Continental United States.
The 2015 CDL product was released February 12, 2016. The 2015 CDL product utilized Deimos-1, UK-DMC 2, and Landsat 8 imagery for production of a 30m product covering the Continental United States.
The 2016 CDL product was released February 3, 2017. The 2016 CDL product utilized Deimos-1, UK-DMC 2, and Landsat 8 imagery for production of a 30m product covering the Continental United States. Beginning with the 2016 CDL season we are creating CDL accuracy assessments using unbuffered validation data. These "unbuffered" accuracy metrics will now reflect the accuracy of field edges which have not been represented previously. This admission of modestly inflated accuracy measures does not render past assessments useless. By providing both buffered and unbuffered validation scenarios for 2016 gives guidance on the bias. There are no plans to create unbuffered accuracy assessments for prior CDL seasons.
The 2017 CDL product was released January 26, 2018. The 2017 CDL product utilized satellite imagery from the Landsat 8 OLI/TIRS sensor, the Disaster Monitoring Constellation (DMC) DEIMOS-1 and UK2, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 sensors collected during the current growing season. The spatial resolution is 30 meters covering the Continental United States.
The 2018 CDL product was released February 15, 2019. The 2018 CDL product utilized satellite imagery from the Landsat 8 OLI/TIRS sensor, the Disaster Monitoring Constellation (DMC) DEIMOS-1 and UK2, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 sensors collected during the current growing season. The spatial resolution is 30 meters covering the Continental United States. A new CDL category was added in 2018, Code 215 - Avocados.
The 2019 CDL product was released February 5, 2020. The 2019 CDL product utilized satellite imagery from the Landsat 8 OLI/TIRS sensor, the Disaster Monitoring Constellation (DMC) DEIMOS-1 and UK2, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 sensors collected during the current growing season. The spatial resolution is 30 meters covering the Continental United States. The 2007-2018 CDLs used FSA CLU data for agricultural training with a 30 meter inward buffer applied. The inward buffering removes spectrally mixed field edge pixels from the land cover classifier. Starting with the 2019 CDL products, the inward buffer has been reduced from 30 meters to 15 meters. The result is a noticeable increase in crop identification at field borders which impacts the CDL, the Cultivated Layer, and Crop Frequency Layers. The newly published USGS NLCD 2016 was used as training for the non-agricultural component of the 2019 CDL. A new CDL category was added in 2019, Code 228 - Double-Cropped Triticale/Corn.
The 2020 CDL product was released February 1, 2021. The 2020 CDL product utilized satellite imagery from the Landsat 8 OLI/TIRS sensor, the Disaster Monitoring Constellation (DMC) DEIMOS-1, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 sensors collected during the current growing season. The spatial resolution is 30 meters covering the Continental United States. There were no new CDL categories added in 2020.
The 2021 CDL product was released February 14, 2022. The 2021 CDL product utilized satellite imagery from the Landsat 8 OLI/TIRS sensor, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 sensors collected during the current growing season. The spatial resolution is 30 meters covering the Continental United States. There were no new CDL categories added in 2021.
The 2022 CDL product was released January 30, 2023. The 2022 CDL product utilized satellite imagery from the Landsat 8 and Landsat 9 OLI/TIRS sensors, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2A and -2B sensors collected during the current growing season. The spatial resolution is 30 meters covering the Continental United States. There were no new CDL categories added or changes in processing methodology in 2022.
- Why was the Cropland Data Layer created?
The purpose of the Cropland Data Layer Program is to use satellite imagery to provide acreage estimates to the Agricultural Statistics Board for major commodities and to produce digital, crop-specific, categorized geo-referenced output products.
- Where can I obtain the Cropland Data Layer (CDL) and what is the cost?
The entire archive of CDL products are available free at CroplandCROS and the USDA NRCS Geospatial Data Gateway. The most current year of CDL data is also available for download at https://www.nass.usda.gov/Research_and_Science/Cropland/Release/index.php
- Can I redistribute the CDL?
Yes, the NASS Cropland Data Layer has no copyright restrictions. The CDL is considered public domain and free to redistribute. However, NASS would appreciate acknowledgment for the usage of our CDL product. The preferred citation is as follows:
USDA National Agricultural Statistics Service Cropland Data Layer. {YEAR}. Published crop-specific data layer [Online]. Available at https://nassgeodata.gmu.edu/CropScape/ (accessed {DATE}; verified {DATE}). USDA-NASS, Washington, DC.
- Is NASS still providing annual DVD's of the CDL?
The CropScape website is meant to eliminate the need for DVD production. All CDL data products are available at CroplandCROS and the SARS website.
- What projections are used?
CropScape and CDL data use USA Contiguous Albers Equal Area Conic USGS Version with a spheroid of GRS 1980 and datum of NAD83. The downloadable zip files from the SARS website are offered in the native Albers projection. Please notice that this projection has an offset from the standard North America Albers Equal Area Conic in standard parallels. The projection information is as follows (in human-readable OGC WKT):
PROJCS["NAD_1983_Albers",
GEOGCS["NAD83",
DATUM["North_American_Datum_1983",
SPHEROID["GRS 1980",6378137,298.257222101,
AUTHORITY["EPSG","7019"]],
TOWGS84[0,0,0,0,0,0,0],
AUTHORITY["EPSG","6269"]],
PRIMEM["Greenwich",0,
AUTHORITY["EPSG","8901"]],
UNIT["degree",0.0174532925199433,
AUTHORITY["EPSG","9108"]],
AUTHORITY["EPSG","4269"]],
PROJECTION["Albers_Conic_Equal_Area"],
PARAMETER["standard_parallel_1",29.5],
PARAMETER["standard_parallel_2",45.5],
PARAMETER["latitude_of_center",23],
PARAMETER["longitude_of_center",-96],
PARAMETER["false_easting",0],
PARAMETER["false_northing",0],
UNIT["meters",1]]
In order to conform to Geospatial Data Gateway technical specifications, any CDL data downloaded through the Geospatial Data Gateway is re-projected from Albers to the dominant Universal Transverse Mercator (UTM) zone with a spheroid and datum of WGS84. The one exception to the UTM projection is for Wisconsin. Wisconsin is projected using the Wisconsin Transverse Mercator (WTM) projection. This WTM projection is based on the 1991 adjustment to NAD83, and is called WTM83/91. Projection parameters and additional information about WTM83/91 is posted on the DNR website: http://dnr.wi.gov/maps/gis/wtm8391.html
WTM83/91 Parameters
Projection: Transverse Mercator
Scale Factor at Central Meridian: 0.9996
Longitude of Central Meridian: 90 Degrees West (-90 Degrees)
Latitude of Origin: 0 Degrees
False Easting: 520,000
False Northing: -4,480,000
Unit: Meter
Horizontal Datum: NAD83, 1991 Adjustment (aka HPGN or HARN)
- What file format is the CDL published in?
The CDL data is available in a raster-based GeoTIFF (.TIF) file format. The GeoTIFF for a single CDL data layer will have at least three associated files: .tif, .tfw, .aux, and possibly a .vat.dbf and .ovr files. Please keep all associated files in the same directory as the GeoTIFF for proper viewing in ArcGIS or Erdas Imagine. Older CDLs may also be available in an ERDAS Imagine (.img) file format. The Erdas Imagine file will have at least two files associated with it: .img and .rrd and possibly .ige or .rde files.
- How has the methodology used to create the CDL changed over the program's history?
The classification process used to create older CDLs (prior to 2006) was based on a maximum likelihood classifier approach using in-house software. The pre-2006 CDL's relied primarily on satellite imagery from the Landsat TM/ETM satellites which had a 16-day revisit. The in-house software limited the use of only two scenes per classification area. The only source of ground truth was the NASS June Area Survey (JAS). The JAS data is collected by field enumerators so it is quite accurate but is limited in coverage due to the cost and time constraints of such a massive annual field survey. It was also very labor intensive to digitize and label all of the collected JAS field data for use in the classification process. Non-agricultural land cover was based on image analyst interpretations.
Starting in 2006, NASS began utilizing a new satellite sensor, new commercial off-the-shelf software, more extensive training/validation data. The in-house software was phased out in favor of a commercial software suite, which includes Erdas Imagine, ESRI ArcGIS, and Rulequest See5. This improved processing efficiency and, more importantly, allowed for unlimited satellite imagery and ancillary dataset inputs. The new source of agricultural training and validation data became the USDA Farm Service Agency (FSA) Common Land Unit (CLU) Program data which was much more extensive in coverage than the JAS and was in a GIS-ready format. NASS also began using the most current USGS National Land Cover Dataset (NLCD) dataset to train over the non-agricultural domain. The new classification method uses a decision tree classifier.
NASS continues to strive for CDL processing improvements, including our handling of the FSA CLU pre-processing and the searching out and inclusion of additional agricultural training and validation data from other State, Federal, and private industry sources. New satellite sensors are incorporated as they become available. Currently, the CDL Program uses the Landsat 8 and 9 OLI/TIRS sensor, the Disaster Monitoring Constellation (DMC) DEIMOS-1 and UK2, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 A and B sensors. Imagery is downloaded daily throughout the growing season with the objective of obtaining at least one cloud-free usable image every two weeks throughout the growing season.
Please refer to (FAQ Section 4, Question 4) on this FAQs webpage to learn more about how the handling of grass and pasture related categories has evolved over the history of the CDL Program.
- What states have been completed and how many years of data are available?
The CDL program became operational with one state in 1997, Arkansas. CDL coverage prior to 2008 is listed, with 2008 providing the first annual coverage for the Continental United States. Please visit the SARS website for a list of all states and years of available CDL data.
- Do the classifications cover the entire state?
All CDL data for crop year 2007 and newer provide statewide cloud-free coverage. For CDL years 2006 and earlier, please reference the official metadata files to verify the extent of coverage and level of cloud contamination.
- Is the CDL data available during the growing season? Will this data be available anywhere as it is being compiled, or do we have to wait until January for data release?
The CDL is released to the public in late January/early February following the end of the typical US growing season. Prior to the public release the CDL is considered confidential and market sensitive during the growing season and cannot be released until after the official NASS year end area county estimates are published. Furthermore, the CDL is considered preliminary during the growing season and could be misleading to our users, as we continue to receive updated ground truth and satellite imagery throughout the season.
- What data is available through the Geospatial Data Gateway?
When downloading the CDL using the NRCS Geospatial Data Gateway, all available years of CDL production for the requested state are included in a single compressed WinZIP file. Geospatial Data Gateway technical restrictions do not allow us to offer the CDL by individual state/year. The zip file will include all years of CDL data for the requested state in a GeoTIFF (.tif) file format projected in UTM and the accompanying metadata.
- Can you explain how to download the CDL data from the NRCS Geospatial Data Gateway?
Below are instructions for downloading from the NRCS Geospatial Data Gateway (https://datagateway.nrcs.usda.gov/). Recently some users are reporting problems with retrieving all years of CDL data when downloading from the Geospatial Data Gateway. If the instructions below do not work or you do not receive the most recent years of CDL data then try using the 'Direct Data Download' link in the lower right-hand corner of their webpage.
1. Go to the website: https://datagateway.nrcs.usda.gov/
2. Click on the "Get Data" button on the upper right hand side of the page
3. Select the state of interest from the drop down list
4. Select a single county. Your download will include the entire state regardless of what county you select
5. Click on "Submit Selected Counties"
6. Place a checkmark next to "Cropland Data Layer by State" under the "Land Use Land Cover" category. Click continue
7. Select "FTP" from the "Delivery" category. Click continue
8. Fill out all fields marked with a "*"
9. Click continue
10. Review order for accuracy then click "Place Order"
11. Note your order number. You will receive an email with a link to download your order.
- Do I need any special type of GIS or image processing software to view the CDL?
If you already have GIS capability, you should be able to work with the downloadable GeoTIFF files directly in your software. If you do not have software capable of viewing a GeoTIFF (.tif) or Erdas Imagine (.IMG) file formats then we suggest using the basic viewing and GIS functionality available on the CroplandCROS web service.
- What software is used to create the classifications?
Prior to the 2006, the CDL processing was done entirely with in-house maintained software called PEDITOR. The PEDITOR classification was based on a maximum likelihood classifier. Limitations to the software also made it impossible to use more than two satellite scenes per classification, which was a significant hindrance when trying to minimize cloud coverage.
Beginning with the 2006 CDL products, the CDL program phased out the use of PEDITOR and transitioned to a commercial software suite. ERDAS Imagine is used in the pre- and post- processing of all raster-based data. ESRI ArcGIS is used to prepare the vector-based training and validation data. Rulequest See5.0 is used to create a decision tree based classifier. The MRLC NLCD Mapping Tool is used to apply the See5.0 decision-tree via ERDAS Imagine.
- What satellite sensors are used to create the CDL?
The CDL Program uses medium spatial resolution (30 meter) satellite imagery. Currently, it is too costly to use higher resolution satellites to perform crop acreage estimation over large areas. The current CDL Program uses the Landsat 8 and 9 OLI/TIRS sensor, the Disaster Monitoring Constellation (DMC) DEIMOS-1 and UK2, the ISRO ResourceSat-2 LISS-3, and the ESA SENTINEL-2 A and B sensors. Imagery is downloaded daily with the objective of obtaining at least one cloud-free usable image every two weeks throughout the growing season.
Older CDL products (prior to 2008) relied on Landsat 4/5/7 and the IRS-P6 Resourcesat-1 sensor. The official metadata contains a detailed listing of the inputs for each state and year.
- How statistically accurate are the classifications?
Detailed accuracy assessment tables are published within the official metadata files. Generally, the large area row crops have producer accuracies ranging from mid 80% to mid 90%. The full error matrices used to create the accuracy assessment information contained within the metadata files is available for download in Question 11 and 54 of this FAQs webpage. NOTE ABOUT THE UNBUFFERED VALIDATION ACCURACY TABLES BEGINNING IN 2016: The training and validation data used to create and accuracy assess the CDL has traditionally been based on ground truth data that is buffered inward 30 meters. This was done 1) because satellite imagery (as well as the polygon reference data) in the past was not georeferenced to the same precision as now (i.e. everything "stacked" less perfectly), 2) to eliminate from training spectrally-mixed pixels at land cover boundaries, and 3) to be spatially conservative during the era when coarser 56 meter AWiFS satellite imagery was incorporated. Ultimately, all of these scenarios created "blurry" edge pixels through the seasonal time series which it was found if ignored from training in the classification helped improve the quality of CDL. However, the accuracy assessment portion of the analysis also used buffered data meaning those same edge pixels were not assessed fully with the rest of the classification. This would be inconsequential if those edge pixels were similar in nature to the rest of the scene but they are not- they tend to be more difficult to classify correctly. Thus, the accuracy assessments as have been presented are inflated somewhat. Beginning with the 2016 CDL season we are creating CDL accuracy assessments using unbuffered validation data. These "unbuffered" accuracy metrics will now reflect the accuracy of field edges which have not been represented previously. Beginning with the 2016 CDLs we published both the traditional "buffered" accuracy metrics and the new "unbuffered" accuracy assessments. The purpose of publishing both versions is to provide a benchmark for users interested in comparing the different validation methods. For the 2017 CDL season we are now only publishing the unbuffered accuracy assessments within the official metadata files and offer the full "unbuffered" error matrices for download on the FAQs webpage. We plan to continue producing these unbuffered accuracy assessments for future CDLs. However, there are no plans to create these unbuffered accuracy assessments for past years. It should be noted that accuracy assessment is challenging and the CDL group has always strived to provide robust metrics of usability to the land cover community. This admission of modestly inflated accuracy measures does not render past assessments useless. They were all done consistently so comparison across years and/or states is still valid. Yet, by now providing both scenarios for 2016 gives guidance on the bias.
- What are the geo-positional errors or spatial accuracies associated with the CDL?
Prior to 2006, the Landsat TM/ETM categorized images were co-registered to MDA/EarthSat Inc's ortho-rectified GeoCover Stock Mosaic images using automated block correlation techniques. The resulting correlations were applied to each categorized image and then added to a master image or mosaic using NASS' in-house software, PEDITOR. The GeoCover Stock Mosaics are within 50 meters root mean squared error overall.
Newer Cropland Data Layers (2006 to current) retain the spatial attributes of the input imagery. The AWiFS imagery has a positional accuracy of 60 meters at the circular error at the 90 percent confidence level (CE90). CE90 is a standard metric often used for horizontal accuracy in map products and can be interpreted as 90% of well-defined points tested must fall within a certain radial AWiFS distance. The Landsat 4/5/8 imagery is obtained via download from the USGS Global Visualization Viewer (Glovis) website (http://glovis.usgs.gov/). Please reference the metadata on the Glovis website for each Landsat scene for positional accuracy. The majority of the Landsat data is available at Level 1T (precision and terrain corrected). The DEIMOS-1 and DMC-UK 2 imagery used in the production of the Cropland Data Layer is orthorectified to a radial root mean square error (RMSE) of approximately 10 meters. More information about the geo-positional accuracy of the ESA Sentinel-2 imagery can be found at https://sentinel.esa.int/. More information about the geo-positional accuracy of the ISRO ResourceSat-2 LISS-3 imagery can be found on page 104 of the ResourceSat2 Handbook at http://www.euromap.de/download/R2_data_user_handbook.pdf.
- Are color legends available for the Cropland Data Layers?
The following downloadable jpeg files are color legends by year for the Continental United States CDLs:
US_2022_CDL_legend.jpg
US_2021_CDL_legend.jpg
US_2020_CDL_legend.jpg
US_2019_CDL_legend.jpg
US_2018_CDL_legend.jpg
US_2017_CDL_legend.jpg
US_2016_CDL_legend.jpg
US_2015_CDL_legend.jpg
US_2014_CDL_legend.jpg
US_2013_CDL_legend.jpg
US_2012_CDL_legend.jpg
US_2011_CDL_legend.jpg
US_2010_CDL_legend.jpg
US_2009_CDL_legend.jpg
US_2008_CDL_legend.jpg
- Has someone compiled all of the exported CDL attribute tables by year and state into a master spreadsheet?
2022_CDL_Histogram_Summary.xlsx
2021_CDL_Histogram_Summary.xlsx
2020_CDL_Histogram_Summary.xlsx
2019_CDL_Histogram_Summary.xlsx
2018_CDL_Histogram_Summary.xlsx
2017_CDL_Histogram_Summary.xlsx
2016_CDL_Histogram_Summary.xlsx
2015_CDL_Histogram_Summary.xlsx
2014_CDL_Histogram_Summary.xlsx
2013_CDL_Histogram_Summary.xlsx
2012_CDL_Histogram_Summary.xlsx
2011_CDL_Histogram_Summary.xlsx
2010_CDL_Histogram_Summary.xlsx
2009_CDL_Histogram_Summary.xlsx
2008_CDL_Histogram_Summary.xlsx
- Has someone summarized CropScape/CDL acreage for all land cover categories at the county-level for the Continental United States?
Below is a link to a WinZIP file containing CSV spreadsheets that summarize the pixel counts and acreage for all US counties for each land cover category for the CDL beginning with the year 2007 to current. These are raw pixel counts and are not official NASS estimates.
The yearly CDL pixel counts and acreage county summaries are available at County_Pixel_Count.zip.
- Is more detailed accuracy assessment information available than what is provided in the metadata? Can you provide the full accuracy assessment error/confusion matrices for all states?
The strength and emphasis of the CDL is crop-specific land cover categories. The accuracy of the CDL non-agricultural land cover classes is entirely dependent upon the USGS, National Land Cover Database (NLCD). Thus, the USDA, NASS recommends that users consider the NLCD for studies involving non-agricultural land cover.
The training and validation data used to create and accuracy assess the CDL has traditionally been based on ground truth data that is buffered inward 30 meters. This was done 1) because satellite imagery (as well as the polygon reference data) in the past was not georeferenced to the same precision as now (i.e. everything "stacked" less perfectly), 2) to eliminate from training spectrally-mixed pixels at land cover boundaries, and 3) to be spatially conservative during the era when coarser 56 meter AWiFS satellite imagery was incorporated. Ultimately, all of these scenarios created "blurry" edge pixels through the seasonal time series which it was found if ignored from training in the classification helped improve the quality of CDL. However, the accuracy assessment portion of the analysis also used buffered data meaning those same edge pixels were not assessed fully with the rest of the classification. This would be inconsequential if those edge pixels were similar in nature to the rest of the scene but they are not as they tend to be more difficult to classify correctly. Thus, the accuracy assessments as have been presented are inflated somewhat.
Beginning with the 2016 CDL season we are creating CDL accuracy assessments using unbuffered validation data. These "unbuffered" accuracy metrics will now reflect the accuracy of field edges which have not been represented previously. This admission of modestly inflated accuracy measures does not render past assessments useless. By providing both buffered and unbuffered validation scenarios for 2016 gives guidance on the bias. There are no plans to create unbuffered accuracy assessments for prior CDL seasons.
The full error matrices are included in the downloadable links below.
CDL_2022_accuracy_assessments.zip
CDL_2021_accuracy_assessments.zip
CDL_2020_accuracy_assessments.zip
CDL_2019_accuracy_assessments.zip
CDL_2018_accuracy_assessments.zip
CDL_2017_accuracy_assessments.zip
CDL_2016_accuracy_assessments.zip
CDL_2015_accuracy_assessments.zip
CDL_2014_accuracy_assessments.zip
CDL_2013_accuracy_assessments.zip
CDL_2012_accuracy_assessments.zip
CDL_2011_accuracy_assessments.zip
CDL_2010_accuracy_assessments.zip
CDL_2009_accuracy_assessments.zip
CDL_2008_accuracy_assessments.zip
- What is the preferred citation for the Cropland Data Layer and CropScape?
USDA National Agricultural Statistics Service Cropland Data Layer. {YEAR}. Published crop-specific data layer [Online]. Available at https://nassgeodata.gmu.edu/CropScape/ (accessed {DATE}; verified {DATE}). USDA-NASS, Washington, DC.
- To whom do I address concerns about CropScape or the Cropland Data Layer?
Distribution issues can be directed to the NASS Customer Service Hotline at 1-800-727-9540. Content and technical questions can be directed to the NASS Spatial Analysis Research Section (SARS) at SM.NASS.RDD.GIB@usda.gov.
- Is the CropScape/Cropland Data Layer available in a shapefile format?
We do not offer the data in a vector format, such as shapefile. The Cropscape/CDL data can be downloaded in a raster-based GeoTIFF file format and used in most common GIS software. In ESRI ArcGIS you would most likely require the 'Spatial Analyst' extension to perform any in-depth GIS applications using the GeoTIFF file. And any common image processing software, such Erdas Imagine, ENVI or PCI, should be able to perform basic image processing/GIS applications using the GeoTIFF file. This type of pixel-based data does not lend itself to being converted to vector since the resulting polygon file would be enormous. Depending on the size of area you are studying it is technically possible to convert Cropscape data to a shapefile, but it would have to be a rather small area such as a single county or smaller.
If you do convert the data to a shapefile format and want to add the CropScape/CDL category names in ESRI ArcGIS, then start by downloading this spreadsheet file cdl_codes_names.xlsx that lists all CDL codes and category names. Open this file and change the "Code" column header to match the name of the attribute field in your newly created shapefile. Then open both the excel file and the shapefile in ArcGIS. Right click on the shapefile in the ArcGIS Table of Contents and do a JOIN on the commonly named "code" attribute field. You can then right click on the shapefile and use Data > Export Data to save a new shapefile with the category names added.
- I have downloaded CropScape data or National CDL Mosaic data from the SARS website and when trying to uncompress the data it appears to be huge - on the order of 6.0 PB (Petabytes). What software should I use to view the contents of the compressed ZIP file?
Several users have reported having trouble uncompressing and viewing their downloaded CropScape/CDL data. In every case this was caused by the user trying to use Windows Explorer to view or extract the contents of the zip file. This issue is resolved by using actual WinZIP software (www.winzip.com) to unzip the contents of the zip file. WinRAR or 7-Zip software should also uncompress the downloaded zip file properly.
- Are more detailed CDL category definitions available?
The AGRICULTURAL CATEGORIES are based on data from the Farm Service Agency (FSA) Common Land Unit (CLU) Program. Thus, all crop specific categories are determined by the FSA CLU/578 Program which offers detailed documentation at the following website: https://www.fsa.usda.gov/programs-and-services/laws-and-regulations/handbooks/index. The online manual titled 2-CP contains much of the crop information used for CDL purposes, especially the section "Exhibit 10 2003 and Subsequent Year Crops Reported on FSA-578."
There are hundreds of potential FSA crop types and thousands of other variables that we have done our best to crosswalk for CDL purposes. This Microsoft Excel spreadsheet FSA-to-CDL_Crosswalk details our current crosswalking and can be used to determine exactly what FSA crop types appear within grouped CDL categories, such as "Other Crops" (CDL code 44), "Misc Vegs and Fruit" (CDL code 47), "Herbs" (CDL code 57), "Other Tree Crops" (CDL code 71), and "Greens" (CDL code 219).
As for the NON-AGRICULTURAL CATEGORIES in the CDL, we sample non-ag training and validation from the USGS National Land Cover Database (NLCD). The NLCD legend with category definitions is available at: https://www.mrlc.gov/data/legends/national-land-cover-database-2016-nlcd2016-legend. In the CDL we have added 100 to their code numbers (i.e. NLCD code 11 "Open Water" is code 111 in the CDL). The NLCD Cultivated Crops category is ignored for CDL purposes. We have also made the decision to merge NLCD "Grassland/Herbaceous" and NLCD "Pasture/Hay" into a single CDL category called "Grassland/Pasture" (CDL code 176).
- How do I determine how much area of a certain crop(s) is grown within a defined area of interest (AOI) and then create a graphic of only that crop(s)?
A user can summarize the area of a certain crop within a certain radius using the CropScape online tools. This can be done at the state, county level, agricultural statistics district (ASD), or region. It is also possible to define your own Area of Interest (AOI) either by using the drawing tools or importing your own shapefile. Note that the recommended projection for your imported AOI data is Albers Equal Area Conic Projection (EPS:5070). Here are the basic steps:
1. define an Area of Interest (AOI) - you can do this by state/ASD/county or use the drawing tools to create your own AOI or the Import Area of Interest to use your own shapefile
2. after defining your AOI - click on the 'Area of Interest Statistics' button on the toolbar at the top of the page
3. this brings up an acreage and pixel count summary table
4. if you would like to download a graphic of the area for a single crop or group of crops then you can place a checkmark next to the crop(s) you are interested in
5. click on the 'Export selected crop(s) for mapping' button
6. then click the 'Download' button and specify where to save the file
Users should be aware of the potential limitations of acreage summaries that are based on only pixel counting. Most land cover classification datasets will contain some level of counting bias (typically downward). Pixel counting is usually downward biased when compared to the official estimates. Counting pixels and multiplying by the area of each pixel will result in biased area estimates and should be considered raw numbers needing bias correction. Official crop acreage estimates at the state and county level are available at QuickStats.
Here are a list of references discussing the subject matter of pixel counting and estimation:
a) Gallego F.J., 2004, Remote sensing and land cover area estimation. International Journal of Remote Sensing. Vol. 25, n. 15, pp. 3019-3047.
b) European Commission, Joint Research Centre, MARS; Best practices for crop area estimation with Remote Sensing - Section 4.1.1.
c) Czaplewski, R. L. (1992). Misclassification bias in areal estimates. Photogrammetric Engineering and Remote Sensing, 58, 189-192.
- Does the CDL differentiate between grassland types such as urban grassland, pastures used for grazing, and other grass-related land cover types?
Unfortunately, the pasture and grass-related land cover categories have traditionally had very low classification accuracy in the CDL. Moderate spatial and spectral resolution satellite imagery is not ideal for separating grassy land use types, such as urban open space versus pasture for grazing versus CRP grass. To further complicate the matter, the pasture and grass-related categories were not always classified definitionally consistent from state to state or year to year. In an effort to eliminate user confusion and category inconsistencies the 1997-2013 CDLs were recoded and re-released in January 2014 to better represent pasture and grass-related categories. A new category named Grass/Pasture (code 176) collapses the following historical CDL categories: Pasture/Grass (code 62), Grassland Herbaceous (code 171), and Pasture/Hay (code 181). We continue to search for program enhancements and ancillary datasets that may help improve the identification of grassland and pasture categories within the CDL. We recommend users consider using the USGS NLCD (https://www.mrlc.gov/) for research involving non-agricultural categories and grassland/pasture categories.
- How do I add class names and/or histogram values to the GeoTIFF file when viewing the CDL in ESRI ArcGIS Software?
If your downloaded CDL .tif file does not contain category names, then you can add them using the following instructions. Download the file: generic_cdl_attributes.tif.vat.dbf. This generic file contains all possible CDL colors and category names. As long as the .tif file and the .tif.vat.dbf file have the same file name, then the category names will load automatically in ArcMap. So, change the file name (not extension) of the generic_cdl_attributes.tif.vat.dbf to match the file name of the downloaded CDL .tif file. Then add the .tif file as a layer in ArcMap. The category names will display in the Table of Contents window.
Example 1 - If the downloaded .tif file is: _NASS_DATA_CACHE_CDL_2011_clip_20110307142903_862761787.tif Change the generic_cdl_attributes.tif.vat.dbf file name to: _NASS_DATA_CACHE_CDL_2011_clip_20110307142903_862761787.tif.vat.dbf
Example 2 - If you renamed the downloaded .tif file to MyCDL.tif, then rename the generic_cdl_attributes.tif.vat.dbf file name to MyCDL.tif.vat.dbf.
To create a pixel "count" field in the attribute table of the downloaded CDL use the "Build Raster Attribute Table" Function in ESRI ArcGIS. In ESRI ArcGIS Version 9.3 and 10 this function is located at ArcToolbox > Data Management Tools > Raster > Raster Properties > Build Raster Attribute Table. Specify the downloaded CDL tif file as the Input Raster and accept all other defaults and click OK. After it has run successfully, a new "Count" data field is added to the attribute table. Count represents a raw pixel count. To calculate acreage multiply the count by the square meters conversion factor which is dependent upon the CDL pixel size. The conversion factor for 30 meter pixels is 0.222394. The conversion factor for 56 meter pixels is 0.774922.
- Do the classifications added to county and state level match the official NASS estimates?
There will be differences between CropScape and official NASS estimates when comparing acreage statistics at the state, district, and county levels. Statistics generated by CropScape are dependent upon pixel counting. Pixel counting is usually downward biased when compared to the official estimates. Counting pixels and multiplying by the area of each pixel will result in biased area estimates and should be considered raw numbers needing bias correction. Official crop acreage estimates at the state and county level are available at QuickStats.
Here are a list of references discussing the subject matter of pixel counting and estimation:
a) Gallego F.J., 2004, Remote sensing and land cover area estimation. International Journal of Remote Sensing. Vol. 25, n. 15, pp. 3019-3047.
b) European Commission, Joint Research Centre, MARS; Best practices for crop area estimation with Remote Sensing - Section 4.1.1.
c) Czaplewski, R. L. (1992). Misclassification bias in areal estimates. Photogrammetric Engineering and Remote Sensing, 58, 189-192.
- How are fields with multiple crop types planted in the same season handled in the Cropland Data Layer, such as late season cover crops or winter wheat followed by soybeans?
The primary focus of the Cropland Data Layer (CDL) is on large area summer crops. The Farm Service Agency CLU data is our primary source of agricultural training data for the CDL classifier. We depend on the data that the farmer reports on their FSA/CLU signup forms. The ground truth is prepared to show whether a single/double crop was planted in a particular field. For example, a winter wheat field planted in the Fall of 2009 will be identified in the 2010 CDL, as we consider the time of harvest as the current year of production. If the field is multi-use during a given year, for example winter wheat (ww) followed by soybeans (sb), then a double cropping situation exists and the category for that given field will be ww/sb, and is indicated as such in the legend. If a field is only sb during that year, then it will be identified as sb only. Therefore, all major crop rotations/patterns are captured with this method and are consider mutually exclusive for a given pixel/field. We do not monitor the fruit and vegetable winter industry (i.e., Florida/California), as we focus primarily on the large area summer crops and are not equipped to monitor triple or quad cropping practices. Please reference the official metadata for a complete list of all possible CDL categories, including valid double-crop categories.
- NASS says this is a Cropland data layer product, what about the areas that are not agriculturally intensive?
The strength of the CDL is in its agricultural classifications. The major crop types for a CDL state will normally have a classification accuracy of 85% to 95%.
Prior to 2006, the field level training data was collected solely through the June Agricultural Survey (JAS). The JAS is an annual national survey of randomly selected areas of land. The selected areas are targeted toward cultivated parts of each state based on its area frame. Our enumerators are given questionnaires to ask the farmers what, where, when and how much are they planting. Our surveys focus on cropland, but the enumerators record all land covers within the sampled area of land whether it is cropland or not. NASS uses broad land use categories to define land that is not under cultivation, including; non-agricultural, pasture/rangeland, waste, woods, and farmstead making it difficult to know what specific type of land use/cover actually is on the ground. Thus, non-agricultural land cover contained within the 2005 and older CDL products were based solely on an individual analyst's interpretation.
Newer CDLs (2006 to current) use agricultural training and validation data provided by the FSA CLU Program. The FSA CLU data does not contain much, if any, non-agricultural data. The only source of non-ag training available at the scale required to meet the needs of the CDL Program is the USGS National Land Cover Dataset (NLCD). We sample the non-ag categories of the NLCD proportionate to the available FSA CLU data for a state and include this in the CDL classification process. Thus, the accuracy of the non-agricultural land cover classes within the Cropland Data Layer are entirely dependent upon the NLCD. We recommend that users consider the NLCD for studies involving non-agricultural land cover.
The FSA CLU data does contain a small amount of non-agricultural data and this non-ag FSA data was used in the classification process in early versions of the CDL. Thus, there are some CDL states that may have multiple categories for the same non-ag land cover type, such as category 87 (FSA-sampled wetland) and category 190 and 195 (NLCD-sampled wetlands). This is should only be an issue in the 2006 and 2007 CDL products. Beginning in 2008, the use of the FSA CLU non-ag for classification training was discontinued. Beginning with the 2013 CDL, the use of FSA-sampled grasses and pasture (code 62) was discontinued.
- Have the attribute names, codes, and/or colors changed over the history of the program?
All category codes, class names and legend colors are standardized and consistent for all states and all years of the Cropland Data Layer Program.
The 1997-2013 CDLs were recoded and re-released January 31, 2014 to better represent pasture and grass-related categories. A new category named Grass/Pasture (code 176) collapses the following historical CDL categories: Pasture/Grass (code 62), Grassland Herbaceous (code 171), and Pasture/Hay (code 181). This was done to eliminate confusion among these similar land cover types which were not always classified definitionally consistent from state to state or year to year and frequently had poor classification accuracies. Please view the 2013 crosswalk document for a detailed listing of the revisions.
This follows the recoding of the entire CDL archive in January 2012 to better align the historical CDLs with the current product. These revisions were done to eliminate redundant and/or unused categories. The majority of the changes apply to the non-agricultural domain. Please view the 2011 crosswalk document for a detailed listing of the revisions.
- Problems opening the CDL within ENVI?
Some users have reported issues with viewing the downloadable CropScape/CDL GeoTIFF (.TIF) files when using ENVI software. Be sure to keep all associated files (.tif, .tfw, and .aux) together in the same directory. ENVI version 4.4/ENVI Zoom and newer should open the CDL without issue.
- How can I create a legend for the CDL using ESRI ArcGIS software?
Step-by-step instructions are provided on how to create a CDL legend using ArcGIS at: CDL_Create_Legend.pdf.
- Is any smoothing or filtering applied to the CDL?
In general, no smoothing or filtering is done to the final CDL classification. However, there have been exceptions in the past. The original 2006 CDL products did contain a small level of smoothing, but in March of 2009 all but one of the 2006 CDL products with were re-released with no smoothing. The one exception is the 2006 Washington CDL which still contains the smoothing. Smoothing has also been applied to cranberries in the 2008, 2009, 2010 and 2011 New England States and to oranges in 2008, 2009 and 2010 Florida. Please refer to the "Processing Description" Section of the official metadata files to find out if any smoothing was applied to a particular state or year.
- What other spatial data are available from NASS?
Please visit the Charts and Maps link on the USDA NASS website for information about other spatial datasets. There is also a "Geo Spatial Data" Section located on the right side of that webpage that contains links to more geospatial data. Of particular note are CropScape, VegScape, land use stratification by state, and Crop Progress charts.
The Agricultural Statistics Districts (ASD) for the entire U.S. are available in ESRI shapefile format ASD shapefile. An ASD is defined as a contiguous group of counties having relatively similar agricultural characteristics. The ASD's used by NASS usually divide each state into as many as nine Agricultural Statistics Districts to make data comparison easier. Each district is more homogeneous with respect to agriculture than the state as a whole. The following link provides national State, ASD, and county codes in tabular .csv format asds2009.csv. Lists of state, ASD, and county can also be found at the following link: county_list.txt.
- What other remote sensing/GIS publications/reports has SARS released in the public domain?
Many NASS research reports are available online at the following website: /Education_and_Outreach/Reports,_Presentations_and_Conferences/Reports_by_Date/. Many SARS conference presentations are available online at: /Education_and_Outreach/Reports,_Presentations_and_Conferences/Presentations/.There is also an online link to a collection of papers by other agencies, universities and private industry that make reference to the CDL available at: "Other CDL Citations".
- To whom do I address concerns about CropScape or the Cropland Data Layer?
Distribution issues can be directed to the NASS Customer Service Hotline at 1-800-727-9540. Content and technical questions can be directed to the NASS Spatial Analysis Research Section (SARS) at SM.NASS.RDD.GIB@usda.gov.
