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

Accueil du site → Doctorat → États-Unis → 1999 → A multiscale analysis and model of vegetation change in a semiarid landscape

Arizona State University (1999)

A multiscale analysis and model of vegetation change in a semiarid landscape

Francis, Joyce Marie

Titre : A multiscale analysis and model of vegetation change in a semiarid landscape

Auteur : Francis, Joyce Marie

Université de soutenance : Arizona State University

Grade : Doctor of Philosophy (PhD) 1999

The slopes of the San Francisco Peaks of northern Arizona provide a steep environmental gradient that can be used to investigate the effects of changing ecological conditions on spatial pattern and distribution of vegetation. I use this gradient to : (1) examine the changes in spatial pattern of vegetation along an semiarid environmental gradient ; (2) characterize the relationship between spatial scale and pattern across this landscape ; and, (3) investigate potential changes in vegetation distribution along the gradient in response to climatic change. Lacunarity analysis was used to determine the spatial pattern of trees and shrubs at five sites along the gradient. The intra- and interspecific associations were determined using join count statistics coupled with a Monte Carlo simulation. Tree stems approach a random distribution at all sites. Canopies and biomass displayed increasing levels of aggregation with increasing moisture availability. Join count analysis revealed that although Pinus edulis and Juniperous monosperma are not segregated at any of the sites, they are segregated at the edges of their range from shrub species and from Pinus ponderosa. Data from a high resolution thematic mapper simulator (NS001) were aggregated and combined with Landsat thematic mapper data to provide a continuum of grain sizes from 5 m to 30 m on a side. The spatial pattern of each image was analyzed to explore the effects of grain size on various patch and landscape level metrics. Finer grained images appeared to be more fragmented than coarse grained images. The metrics varied smoothly as a function of grain size and were fitted to nonlinear models. These models failed to accurately predict the metrics for a second, independent landscape but did display similar scaling patterns for both landscapes. The effects of climate change may be most drastic along environmental gradients. A fine scale model of changes in vegetation distribution in response to climate change was developed using a remotely sensed vegetation map coupled with output from a microclimate model in a geographic information system. The model demonstrates that a predicted 2°C rise in mean annual temperature may result in major shifts in vegetation in this region.


Page publiée le 11 novembre 2013, mise à jour le 5 janvier 2017