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

Accueil du site → Doctorat → Pays-Bas → Human - wildlife interface in African Savanna : quantifying landscape fragmentation for predicting wildlife distribution and human - wildlife conflicts

University of Twente (2012)

Human - wildlife interface in African Savanna : quantifying landscape fragmentation for predicting wildlife distribution and human - wildlife conflicts

Pittiglio Claudia

Titre : Human - wildlife interface in African Savanna : quantifying landscape fragmentation for predicting wildlife distribution and human - wildlife conflicts

Auteur : Pittiglio Claudia

Etablissement de soutenance : University of Twente

Grade : Doctor University of Twente 2012

Descriptif
Landscape fragmentation is one of the main determinants of biodiversity loss in east African rangelands. It affects wildlife species by reducing the size of suitable contiguous habitat and by increasing the level of isolation between habitat patches. Also it appears to increase the occurrence of human-wildlife conflicts by increasing the human-wildlife interface. Existing approaches, like the ’patch mosaic approach’, are limited to quantify landscape fragmentation in gradient landscapes because they are based on the discrete classification of the environment into land cover classes. They quantify fragmentation by measuring size, connectivity and spatial arrangement of suitable habitat patches. In gradient landscapes, like savanna, where clear boundaries cannot be drawn among vegetation classes, this method is not accurate and may fail to explain wildlife distribution. Alternative models based on continuous environmental variation are still relatively new.
The main objectives of this research are to : * quantify landscape fragmentation (dominant scale, intensity and distribution of block sizes) of cultural and gradient landscapes in spatial and temporal dimension by developing a new approach (the continuum model) based on continuous environmental variation and comparing it with existing approaches. The continuum model is based on wavelets transform which is able to assess multi scales of temporal and spatial series of images of different resolution. Its performance is tested against that of the ’patch mosaic approach’. * analyze the relationship between the distribution (and density) of wild herbivore species of different body size and patterns of landscape fragmentation in east African rangelands. Are highly fragmented areas related to lower density of wild herbivore species ? (which species, which pattern and at what scale ?). Are highly fragmented areas related to lower species richness (of different body size) ? * relate human-wildlife conflicts with patterns of landscape fragmentation in East African rangelands. Are human wildlife conflicts higher in highly fragmented areas ?
The research is currently developing, testing and validating the continuum model in a European landscape (Andalucía, Spain) in order to : 1) apply the model to different landscapes (gradients of natural vegetation and mosaics of agricultural fields) ; 2) make use of a large dataset of images of high spatial and temporal resolution ; 3) compare the performance of the continuum model with that of the patch mosaic approach in an European environment. The continuum model will be applied to a study area in Tanzania (in Tarangire-Manyara ecosystem - Arusha region) in order to analyze and explain patterns of landscape fragmentation and wildlife distribution of different body size (like, elephants, zebra, impala) and human-wildlife conflicts in savanna.

Présentation

Version intégrale (22,8 Mb )

Page publiée le 2 février 2018