Sabine Chabrillat, Thomas Schmid, Robert Milewski, Paula Escribano, Monica Garcia, Eyal BenDor, Stephane Guillaso, Marta Pelayo, Andres Reyes, Veronica Sobejano Paz, Marcos Jimenez Michavila. 2018. Mapping Crop Variability Related to Soil Quality and Crop Stress Within Rainfed Mediterranean Agroecosystems Using Hyperspectral Data. 9th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (whispers), Workshop on Hyperspectral Image and Signal Processing, 1-5.
Cultivation and land use practices have a long history within the Mediterranean region exploiting soils as a natural resource. The soils are an essential factor contributing to agricultural production of rainfed crops such as cereals, olive groves and vineyards. Inadequate management is endangering soil quality and productivity, and in turn crop quality and productivity are affected. The main objective of this project is to map soil and crop variability related to crop stress and land management within a Mediterranean environment based on hyperspectral data within the visible, near-infrared, short-wave infrared as well as thermal infrared (0.4-12 mu m). For this, we use CASI and AHS hyperspectral imagery acquired during the growing season within the Camarena agricultural area in central Spain, characterized by Mediterranean climate, extended agricultural rainfed uses, mostly evolved soils, and erosion features associated to contrasting soil horizons. Simultaneous to the airborne campaign, an intensive field campaign took place for the characterization of soil and crop variability including chemical and biophysical variables, soil degradation stages, and crop production in selected test sites. In this paper, we focus on the optical VNIR-SWIR spectral domain and present project objectives, selected field and airborne data, and preliminary analyses that show, in this Mediterranean agroecosystem affected by soil degradation, the strong influence of soil quality on crop variability and production based on hyperspectral imagery and yield data.
