Skip to main content
ARS Home » Research » Publications at this Location » Publication #111642
Title: LAND COVER CLASSIFICATION AND NDVI ANALYSIS FOR THE SGP99 EXPERIMENT AREA

Author
item OLDAK, ANNA - UNIV OF WARSAW, POLAND
item REYNOLDS, CURT - USDA, FAS
item Jackson, Thomas

Submitted to: Sring Meeting American Geographical Union Washington DC
Publication Type: Abstract Only
Publication Acceptance Date: 4/5/2000
Publication Date: N/A
Citation: N/A

Interpretive Summary: None.

Technical Abstract: One of the goals of the Southern Great Plains 1999 (SGP99) Hydrology Experiment is to obtain information of soil moisture pattern on the regional scale. The soil moisture algorithm, based on remote sensing measurements, requires land cover information as one of the inputs into the model. Vegetation cover strongly affects the microwave emission from the soil, and it is important in soil moisture investigation and analysis. A land cover study was conducted between July 8-21, 1999. For this period, and prior to it, Landsat TM scenes were collected. A ground truth survey was designed based on aerial photographs which indicated 16 main land cover categories. During the experiment, there was some harvesting of wheat, alfalfa, and corn. Land cover categories were selected to follow the changes in vegetation cover by distinguishing such classes as bare soil, bare soil with wheat stubble, and harvested fields with growing weeds (bare esoil with green vegetation). Also, special attention was paid to different types of pasture. Two types of pasture were distinguished: 1) grass areas that are grazed and used as regular pastures and 2) grass areas left idle. The proposed list of 16 land cover categories reflects all the main types of vegetation of the experimental period. Land cover classification was performed based on those images that had the least cloud coverage. From the available Landsat 5 and Landsat 7 TM images, four images were used: March 9, May 12, July 15, and July 23, 1999. The classification results were analyzed using confusion and separability matrices. The Normalized Difference Vegetation Index was calculated in order to follow the changes in biomass during the experimental period.