Remotely sensed biophysical variables of wheat as indicators for crop dynamics modelling

International audience

英语. The objective of the work is the calibration, validation and simulation of STICS model on wheat commercial plots using the remote sensing time series to map the variability of crop biomass and yield. Sentinel time series provide the Green Area Index (GAI-BV), field phenological observations and soil clay and organic matter for calibration; the model is validated on three-year field extensive campaigns in terms of nutrients, biomass and yield. Abstract. Crop dynamics is one of the key factors in crop simulations. It is parameterized as a function of intervals between key successive development stages. These stages are described as growth stages rather than organogenetic stages and correspond to changes in trophic or morphological strategy of crop that influence the evolution of leaf area index and grain filling. Crop models require biophysical variables (BV) including Green Area Index (GAI) as an input to monitor crop growth. Remote sensing-based BV allow to obtain spatially representative information about an agrosystem at different spatial scales-local to regional. A consistent long time series of GAI can be derived by combining reflectance data from Landsat OLI and Sentinel with artificial neural networks (ANN). This derived data was able to show key information on the GAI curve that are related to different growth stages. The GAI curve behavior is related to crops biomass and the photosynthetically active organs. These stages can also be described by coupling field phenological observations with GAI curve peaks. This coupled model can be useful in applying crop models over large areas with limited data. In this work, the GAI curve long time series coupled with field phenological observations was applied to determine the development stages of wheat. The experimental data was obtained from commercial agricultural fields of wheat located in the Central Plateau of the Iberian Peninsula, Spain, that was grown under different nutritional conditions. The Simulateur mulTIdisciplinaire pour les Cultures Standard (STICS) crop model used on this analysis. The development stages were used to validate crop parameterization that was obtained from previously modelled wheat growing conditions in south France and to compare between their time scales.