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

Accueil du site → Master → Pays Bas → Investigating remote-sensing based estimations of ecological pasture conditions in Kara-Unkur watershed, Kyrgyzstan

Wageningen University (2016)

Investigating remote-sensing based estimations of ecological pasture conditions in Kara-Unkur watershed, Kyrgyzstan

Masselink, Loes

Titre : Investigating remote-sensing based estimations of ecological pasture conditions in Kara-Unkur watershed, Kyrgyzstan

Auteur : Masselink, Loes

Université de soutenance : Wageningen University

Grade : Master of Science (MS) 2016

Résumé
Kyrgyzstan has a rich nomadic history in which pastures play an important role. However, currently the country is dealing with severe levels of grazing-induced pasture degradation. Decreased biomass, palatability and biodiversity both have and continue to reduce the pastures’ productivity. To improve future pasture productivity by amending the management strategies, up-to-date information regarding the ecological conditions of the pastures is considered of key importance. At the moment, however, there is no simple methodology to detect changing ecological pasture conditions. As remote sensing is presented as a promising cost- and time-effective approach, this research aimed to understand the potential of a remote sensing-based methodology for the detection of changing ecological pasture conditions in Kara-Unkur watershed, Kyrgyzstan. After data collection (both in the field and with remote sensing) the relations between Vegetation Indices (VIs) from Landsat ETM+ images and biomass-, palatability- and species richness field data (N=17) were investigated. Both simple regression and multiple linear regression (MLR) analyses (including terrain attributes) were applied. Subsequently, trends of these three pasture degradation types (biomass, palatability and species richness) were mapped through time series analysis. The results show that biomass is most accurately estimated by a model including Modified Soil Adjusted Vegetation Index (MSAVI) with a slope factor (R2 = 0.65, F = 0.0006). Regarding palatability, a model including Enhanced Vegetation Index (EVI), Near Infrared (NIR) band, Red band and Northness Index was most accurate (R2 = 0.61, F = 0.0160). Species richness was most accurately estimated by a model including Topographic Wetness Index (TWI), Eastness Index and estimated biomass, i.e. the results of biomass modelling (R2 = 0.81, F = 0. 0028). Subsequent trend analyses of all three estimated ecological pasture conditions presented very similar (mainly negative) trend patterns in Jas Jerim pasture area within Kara-Unkur watershed. Despite lacking validation of both models and subsequent trend patterns with additional field data, this study confirms the high potential of a remote-sensing based methodology for detection of changing ecological pasture conditions in Kara-Unkur watershed, Kyrgyzstan

Présentation

Version intégrale (4 Mb)

Page publiée le 5 novembre 2016, mise à jour le 1er janvier 2018