COMBINING REMOTE SENSING AND PLANT GROWTH MODELING TO DESCRIBE THE CARBON AND WATER BUDGET OF SEMI-ARID GRASSLAND

Authors: NOUVELLON, Y - CIRAD MONTPELLIER FRANCE; SEEN, D - DIRAD MONTPELLIER FRANCE; BEGUE, A - CIRAD MONTPELLIER FRANCE; RAMBAL, S - CEFE-CNRS MONTPELLIER FR; Moran, Mary; Qi, Jiaguo; CHEHBOUNI, A - IMADES HERMOSILLO MEXICO; KERR, Y - CESBIO TOULOUSE FRANCE

Submitted to: International Geoscience and Remote Sensing Symposium Proceedings
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/15/1998
Publication Date: N/A
Citation: N/A

Interpretive Summary: To properly manage extensive rangeland regions, it is necessary to have regional information on rangeland extent and general health. An approach was developed to combine a grassland growth simulation model with images obtained from orbiting satellite sensors to map rangeland plant and soil conditions. By combining the model with regional images, it was possible to obtain both temporal and spatial information about rangeland health. The results were encouraging because the simulated estimates of grassland biomass and soil moisture compared well with measured values at several sites in southeastern Arizona. Ultimately, this coupled modeling/remote sensing approach should result in accurate, timely information on rangeland health that can be used to make intelligent management decisions at the regional scale.

Technical Abstract: In this paper we investigate the opportunity of coupling a vegetation growth model developed for semi-arid perennial grasslands with a soil/vegetation reflectance model in order to use remote sensing data to improve the model simulations. The vegetation functioning model developed for this purpose has been validated in southeastern Arizona and northeastern Sonora on several semi-arid grassland sites. The assimilation of radiometric data into the shortgrass prairie ecosystem model is based on an interative numerical procedure that recalibrates the combined model until model simulations match radiometric observations. For this purpose, a prior sensitivity analysis was carried out for the vegetation growth model to determine the most important input parameters or initial conditions on which to base the recalibration procedure. The results obtained and the potential of such an approach are discussed.