Submitted to: Proceedings of the World Water and Environmental Resources Congress
Publication Type: Proceedings
Publication Acceptance Date: February 15, 2006
Publication Date: May 20, 2006
Citation: Gowda, P., Peters, R.T., Howell, T.A. 2006. Mapping contrasting tillage practices in the Texas Panhandle with Landsat Thematic Mapper (TM) data. In: Proceedings of the World Water and Environmental Resources Congress. Examining the Confluence of Environmental and Water Concerns, May 21-25, 2006, Omaha, Nebraska. 2006 CDROM. Interpretive Summary: Environmental models require information on tillage management practices to predict water holding capacity, evapotranspiration, carbon sequestration, and soil losses due to wind and water erosion from agricultural lands. Collecting this information can be time-consuming, labor intensive, costly and can involve destructive sampling. Moreover, field data are limited because they provide point, rather than area information. Remote sensing techniques show promise in providing such spatial data over a large area in a time and cost-effective manner. In this study, we evaluated a set of Landsat TM based logistic regression models for their ability to identify tillage practices in Ochiltree County during 2005 planting season. Results indicate that these models are promising for the rapid collection of tillage information over larger areas in the Texas Panhandle
Technical Abstract: Tillage information is crucial in environmental modeling as it has a direct impact on soil erosion and water holding capacity of agricultural soils. A remote sensing approach is promising for the rapid collection of tillage information on individual fields over large areas. In this study, six Thematic Mapper (TM)-based logistic regression models proposed by van Deventer et al (1997) were used to distinguish conventional and conservation tillage practices in Ochiltree County located in the Texas panhandle. Accuracy assessments of tillage maps derived from Landsat 5 TM data were made using field data collected during the 2005 planting season. Logistic regression models were easy to use, cost and time effective, and produced reasonably accurate tillage maps. The “percent correct” and kappa (k) values varied from 61-83% and 0.02-0.73, respectively, with best values for logistic regression models that use TM bands 1, 3 and 5 images. This approach is promising for the rapid collection of tillage information on individual fields over large areas.