Using satellite imagery to asses trends in soil and crop productivity across landscapes

Abstract

Measuring different indicators of production and soil health over the long term will help build a picture of soil health and productivity across the landscape. This paper examines the potential contribution of satellite imagery to this area. This investigation undertook a very long time series analysis of Landsat imagery (approximately 40 years) and MODIS imagery (approximately 10 years). Novel datasets and approaches were used to assess areas based on land use history and land cover condition. Spring Normalised Difference Vegetation Index (NDVI), land cover maps based on NDVI thresholds, annual cumulative NDVI and fractional ground cover (FGC) were used to identify trends in vegetation cover change at a landscape scale, and their relationship with factors such as land use intensification history, geomorphology, rainfall, and land use. This work has improved the broad, baseline understanding of production variation across the landscape, while also providing a practical demonstration of the integration of a range of disparate data sources.