Skip to main content
ARS Home » Research » Publications at this Location » Publication #240406
Title: Deriving satellite crop rotation maps for distributed modeling of water quality in the Choptank River watershed

Author
item Hively, Wells - Dean
item Sadeghi, Ali
item Lang, Megan
item BANDARU, VARAPRASAD - University Of Maryland
item KIM, ANDREW - University Of Maryland
item McCarty, Gregory

Submitted to: Soil Science Society of America Annual Meeting
Publication Type: Abstract Only
Publication Acceptance Date: 7/15/2009
Publication Date: 11/1/2009
Citation: Hively, W.D., Sadeghi, A.M., Lang, M.W., Bandaru, V., Kim, A., Mccarty, G.W. 2009. Deriving satellite crop rotation maps for distributed modeling of water quality in the Choptank River watershed [abstract]. Soil Science Society of America Annual Meeting. 2009 CDROM.

Interpretive Summary:

Technical Abstract: The Choptank River watershed Conservation Effects Assessment Project (CEAP) has now accumulated a four-year set of quarterly SPOT satellite images that we are using to monitor agricultural practices on farms within the study area on Maryland’s Eastern Shore. The imagery is also useful for developing high resolution distributed input datasets for water quality modeling. Models such as AnnAGNPS and SWAT are often run on input data layers that derive topographic features from high resolution landscape maps (such as LIDAR), but land use coverage is often derived only from aggregated county-level statistics that are translated into landscape maps in a semi-random fashion. Through image classification we are able to identify crop types in the landscape (maize; soybean; wheat/soybean double crop; vegetated fallow; bare fallow) with an initial classification accuracy of 77%. As the project advances, methods will be developed to improve classification accuracy and map classifications to known field boundaries. Resulting geospatial coverages can be used to provide realistic distributed input data layers that are expected to improve the accuracy of water quality models through the incorporation of spatially-precise landscape metrics particular to individual crop types, such as erodibility and application and loss of nutrients and pesticides. We are also using the data to investigate regional patterns of crop rotation and the temporal niches that farmers find suitable for growing winter cover crops.