Quantification of aboveground rangeland productivity and anthropogenic degradation on the Arabian Peninsula using Landsat imagery and field inventory data

Abstract

The productivity of semi-arid rangelands on the Arabian Peninsula is spatially and temporally highly variable, and increasing grazing pressure as well as the likely effects of climatic change further threatens vegetation resources. Using the Al Jabal al Akhdar mountains in northern Oman as an example, our objectives were to analyse the availability and spatial distribution of aboveground net primary production (ANPP) and the extent and causes of vegetation changes during the last decades with a remote sensing approach. A combination of destructive and non-destructive biomass measurements by life-form specific allometric equations was used to identify the ANPP of the ground vegetation (<50 cm) and the leaf and twig biomass of phanerophytes. The ANPP differed significantly among the life forms and the different plant communities, and the biomass of the sparsely vegetated ground was more than 50 times lower (mean = 0.22 t DM ha(-1)) than the biomass of phanerophytes (mean = 12.3 t DM ha(-1)). Among the different vegetation indices calculated NDVI proved to be the best predictor for rangeland biomass. Temporal trend analysis of Landsat satellite images from 1986 to 2009 was conducted using a pixel-based least square regression with the annual maximum Normalized Differenced Vegetation Index (NDVImax) as a dependent variable. Additionally, linear relationships of NDVImax and annual rainfall along the time series were calculated. The extent of human-induced changes was analysed using the residual trends method. A strongly significant negative biomass trend detected for 83% of the study area reflected a decrease in annual rainfall but even without clear evidence of deforestation of trees and shrubs, human-induced vegetation degradation due to settlement activities were also important. (C) 2010 Elsevier Inc. All rights reserved.