Skip to main content
ARS Home » Research » Publications at this Location » Publication #95911
Title: CLASSIFICATION TECHNIQUES FOR MAPPING BIOPHYSICAL PARAMETERS IN THE U.S. SOUTHERN GREAT PLAINS

Author
item Doraiswamy, Paul
item STERN, ALAN - SSAI INC., LANHAM, MD
item COOK, PAUL - USDA/NASS

Submitted to: IEEE IGARSS Annual Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 8/28/1998
Publication Date: N/A
Citation: N/A

Interpretive Summary: The identification and mapping of vegetation cover types is an important step toward characterization of hydrological balance over a watershed area. The objective of this research was to develop a land use classification for a watershed in Central Oklahoma and to map the vegetation characteristics. The study was part of an NASA-USDA collaborative research project to map regional soil moisture in the U.S. Southern Great Plains. Landsat satellite imagery was the primary source of data used for the classification. Areal photography and systematic ground truth data were acquired to develop the land use classification and retrieval of vegetation parameters. The classification procedure developed was applied to a 60 X 45 km area located in the southern tip of the study area and included the Little Washita Watershed. The vegetation cover types were dominated by grasslands (Forage and Pasture), about 60%, and included winter wheat (15%), trees (10%), alfalfa (5%), and summer crops (corn and legume) about 3%. The results were verified with ground surveys and the level of accuracy was acceptable.

Technical Abstract: The U.S. Department of Agriculture is interested in assessing crop acreage at the small scale for precise acreage assessment. This research was conducted as part of a 1997 NASA-USDA collaborative research project for regional mapping of soil moisture in the U.S. Southern Great Plains. The study area was primarily in Central Oklahoma where the land cover was a mixture of agricultural crops and various grasses. The study was conducted from mid-June to mid-July when a transition of winter to summer crops occurred. Three rows of Landsat images along a path between Kansas and Texas were the main data source. Areal photography and two site visits (March and June) provided ground truth data from selected areas to develop spectral signatures for nine major vegetation classes. Spectral signatures for the classes were then extended to the entire region. The classification procedure developed was applied to a 60 X 45 km area located in the southern tip of the study area and included the Little Washita Watershed. A validation of the classification results was accomplished using part of the ground truth data not used for developing the signature classes. The vegetation cover types were dominated by grasslands (Forage and Pasture), about 60%, and included winter wheat (15%), trees (10%), alfalfa (5%), and summer crops (corn and legume) about 3%. This research successfully validated a classification technique to map vegetation in a very mixed land use area. The image processing technique used temporal satellite imagery to circumvent cloud cover problems and successfully identify land use patterns.