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

Accueil du site → Doctorat → Allemagne → 2005 → Entwicklung eines fernerkundungsgestützten Modells zur Erfassung von pflanzlicher Biomasse in NW-Namibia

Rheinischen Friedrich-Wihelm-Universität zu Bonn (2005)

Entwicklung eines fernerkundungsgestützten Modells zur Erfassung von pflanzlicher Biomasse in NW-Namibia

Richters, Jochen

Titre : Entwicklung eines fernerkundungsgestützten Modells zur Erfassung von pflanzlicher Biomasse in NW-Namibia

Development of a remote sensing based plant biomass model in NW-Namibia

Auteur : Richters, Jochen

Université de soutenance : Rheinischen Friedrich-Wihelm-Universität zu Bonn

Grade : Dr. rer. nat. 2005

The doctoral thesis byJochen Richters focuses on the development of the remote sensing basedmodel "RBM Kaokoveld" which aims at the assessment of plant biomass. Thework has been carried out at the University of Bonn within the projectB7 of the Sonderforschungsbereich (SFB) 389 ACACIA (University of Cologne).The investigation area, the Kaokoveld in north-western Namibia, is characterizedby semiarid Colophospermum mopane savannas and arid semi-deserts,ranging along a precipitation gradient from the highlands of the GreatEscarpment in NE to the downward slope of the Namib Desert in the SW ofthe investigation area. The aim of this work is the precise assessment of spatiotemporal biomass inan ecologically fragile savanna system. This is accomplished by calculatingthe photosynthetic radiation absorbed by the vegetation and the efficiency bywhich the absorbed energy is transformed into biomass. The knowledge of thespatiotemporal productivity is important as regards to both detecting degradationof vegetation under the impact of pastoral nomadic cattle grazing and the evaluationwithin the context of global change. Whereas such an approach has been already used in earlier studies, the proposedregional model has two major advantages : First, with a spatial resolution of1 by 1 km2 and a lowest temporal resolution of 10 days, only freeavailable datasets are used to derive plant biomass. Second, most of the datainput for the model comes from the TERRA EOS-1 MODIS system, which providesaccurately defined data products. This eases the development of the application.In addition to the improvement of the database, the "RBM Kaokoveld" representsa physically based model that takes into account topographic and pedologic impactsas well as recent climate factors for the assessment of plant biomass. The modellingwas performed using ENVI/IDL and allows the calculation of decadal time steps.The major output of the model is the spatially distributed produced plant biomassfor every time step. Additionally, the accumulated biomass over the whole modellingtimeframe (usually one year) is calculated. The results of the "RBM Kaokoveld" have been thoroughly verified and comparedto results of competing models. This evaluation underlines both correspondenceto general findings of plant productivity and the good performance of the RBMconcept. Furthermore, a large array of ground truth data from several fieldtrips shows the "groundedness" of the model. As a whole, the "RBM Kaokoveld" provides for a powerful tool to generate usefuldata of spatiotemporal productivity at a regional scale. In the future, thesedatasets may be used for the modelling of an ecologically oriented nomadic pasturebalance to identify overgrazing and degradation as well as for the assessmentof regional carbon sinks in the context of a global climate change


Version intégrale (5 Mb)

Page publiée le 19 mars 2008, mise à jour le 11 novembre 2021