Geospatial assessment of agroecosystem health: development of an integrated index based on catastrophe theory

Assessing agroecosystem health at landscape scale has been rarely reported. A geospatial assessment framework, by integrating remote sensing, geographical information systems, landscape metrics, geostatistics and catastrophe theory, was proposed and applied to characterize the spatial variations of agroecosystem health for a typical region in the eastern coastal agricultural plain, China. After appropriate pretreatments, eleven hierarchically structured indicators, subject to catastrophe models, were aggregated into an integrated index for each 30 m grid cell across the study area. Great spatial variations in agroecosystem health were identified. Areas covered by water bodies and impervious surfaces, under which soils can not function, generally presented low values. Larger slope or higher heavy metal polluted areas and those near roads showed low or moderate values of agroecosystem health index. Concerning the spatial variations of agroecosystems health, patches of approximately 4 km in diameter were identified, within which more homogenous patterns would be expected. Specifically, the spatial variations of agroecosystems health should be resulted from the location-specific coupled influences of the underlying indicators. The integrated agroecosystem health index is believed to be helpful for managers to promote environmental management. The geospatial assessment framework reported in this paper was simple and operational, and has potential to be applicable to other areas with similar conditions at landscape scale.