Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → États-Unis → 2010 → ESTIMATING EVAPOTRANSPIRATION USING REMOTE SENSING : A HYBRID APPROACH BETWEEN MODIS DERIVED ENHANCED VEGETATION INDEX, BOWEN RATIO SYSTEM, AND GROUND BASED MICRO-METEOROLOGICAL DATA

Wright State University (2010)

ESTIMATING EVAPOTRANSPIRATION USING REMOTE SENSING : A HYBRID APPROACH BETWEEN MODIS DERIVED ENHANCED VEGETATION INDEX, BOWEN RATIO SYSTEM, AND GROUND BASED MICRO-METEOROLOGICAL DATA

Chatterjee, Sumantra

Titre : ESTIMATING EVAPOTRANSPIRATION USING REMOTE SENSING : A HYBRID APPROACH BETWEEN MODIS DERIVED ENHANCED VEGETATION INDEX, BOWEN RATIO SYSTEM, AND GROUND BASED MICRO-METEOROLOGICAL DATA.

Auteur : Chatterjee, Sumantra

Université de soutenance : Wright State University

Grade : Doctor of Philosophy (PhD), Environmental Sciences , 2010.

Résumé
We investigated water loss by evapotranspiration (ET) from the Palo Verde Irrigation District (PVID) and the Cibola National Wildlife Refuge (CNWR) in southern California bordering the Colorado River collaborating with the United States Bureau of Reclamation (U.S.B.R.). We developed an empirical model to estimate ET for the entire PVID using satellite derived MODIS enhanced vegetation index (EVI), and ground based measurements of solar radiation and vapor pressure. We compared our predictions with U.S.B.R. estimates through statistical cross validation and showed they agree with an error less than 8%. We tested the same model for an alfalfa field inside PVID to check its applicability at a smaller spatial scale. We showed that the same model developed for PVID is the best model for estimating ET for the alfalfa field. We collected data from three Bowen ratio energy balance (BREB) towers installed in the invasive saltcedar (Tamarix spp) dominated riparian zone in the CNWR and a fourth tower in the alfalfa field in PVID. The riparian sites were selected according to different densities of vegetation. We collected data from these sites at various intervals during the period between June 2006 to November 2008. We reduced the errors associated with the Bowen ratio data using statistical procedures taking into account occasional instrument failures and problems inherent in the BREB method. Our results were consistent with vegetation density and estimates from MODIS EVI images. To estimate ET for larger patches of mixed vegetation we modified the crop coefficient equation and represented it in terms of EVI. Using this approach, we scaled the alfalfa field data to the entire PVID and compared the results with U.S.B.R. (2001-2007) estimates. We predicted ET well within the acceptable range established in the literature. We empirically developed ET models for the riparian tower sites to provide accurate point scale ET estimation and scaled for the entire riparian region in CNWR with our modified crop coefficient approach. We investigated the assumptions upon which the Bowen ratio equation is developed. In the presence of turbulence some of the assumptions may not be valid, and the final data may require correction factors.

Mots-clés : Evapotranspiration ; Remote Sensing ; Enhanced Vegetation Index ; Bowen Ratio

Présentation

Version intégrale (4,74 MB)

Page publiée le 26 avril 2011, mise à jour le 26 décembre 2018