Scaling up from field to region for wind erosion prediction using a field-scale wind erosion model and GIS

Factors that affect wind erosion such as surface vegetative and other cover , soil properties and surface roughness usually change spatially and tempor ally at the field-scale to produce important field-scale variations in wind erosion. Accurate estimation of wind erosion when scaling up from fields t o regions, while maintaining meaningful field-scale process details, remain s a challenge. The objectives of this study were to evaluate the feasibilit y of using a field-scale wind erosion model with a geographic information s ystem (GIS) to scale up to regional levels and to quantify the differences in wind erosion estimates produced by different scales of soil mapping used as a data layer in the model. A GIS was used in combination with the revis ed wind erosion equation (RWEQ), a field-scale wind erosion model, to estim ate wind erosion for two 50 km(2) areas. Landsat Thematic Mapper satellite imagery from 1993 with 30 m resolution was used as a base map. The GIS data base layers included land use, soils, and other features such as roads. The major land use was agricultural fields. Data on 1993 crop management for s elected fields of each crop type were collected from local government agenc y offices and used to 'train' the computer to classify land areas by crop a nd type of irrigation (agroecosystem) using commercially available software . The land area of the agricultural land uses was overestimated by 6.5% in one region (Lubbock County, TX, USA) and underestimated by about 21% in an adjacent region (Terry County, TX, USA). The total estimated wind erosion p otential for Terry County was about four times that estimated for adjacent Lubbock County. The difference in potential erosion among the counties was attributed to regional differences in surface soil texture. Ln a comparison of different soil map scales in Terry County, the generalised soil map had over 20% more of the land area and over 15% greater erosion potential in l oamy sand soils than did the detailed soil map. As a result, the wind erosi on potential determined using the generalised soil map was about 26% greate r than the erosion potential estimated by using the detailed soil map in Te rry County. This study demonstrates the feasibility of scaling up from fiel ds to regions to estimate wind erosion potential by coupling a field-scale wind erosion model with GIS and identifies possible sources of error with t his approach. Published by Elsevier Science B.V.