Land cover change in semi-arid areas, in the context of climate change, can significantly affect the carbon sequestration potential in these fragile ecosystems. Semi-arid Inner Mongolia, P.R.C, is experiencing climate change with associated land cover/use change that includes an increase in irrigated agriculture and population growth. Land cover change was monitored at the regional level through the use of MODIS derived remote sensing products along with meteorological and carbon exchange data obtained from five EC flux towers across a variety of ecosystem types that include natural and disturbed grassland, shrublands as well as land use types such as croplands and dune stabilizing poplar stands. Firstly, I quantified the land cover/use change at the regional and biome levels in the context of landscape fragmentation and the possible consequences. Secondly, I used flux data to validate intra-annual dynamics of satellite derived GPP in different ecosystem types using existing GPP models. In addition I developed and validated a new model, the modified vegetation photosynthesis model.
Finally, I derived predictive models of plant species diversity, developed specifically for semi-arid grasslands which used improved vegetation and water indices.
The major finding of this study were: (1) increasing portions of dominant grassland shrubland and barren cover within the decade points toward a water stressed landscape that is becoming more homogenous and is corroborated by a decrease in proportions of rare cover types. The rapid increase in socio-economic growth leading to a growing population base is described by increasing cohesion and aggregation of urban/built-up patches as well as an increasing number of patches and interspersion of cropland land use. (2) variance in GPP and water content indices were the two most important variables for predicting species richness in Inner Mongolia, while MODIS-derived vegetation and water content indices were selected as significant independent variables for specific biome type. The predictive power of models improved greatly when the region was stratified by biome and life-form type and especially when the anthropogenically modified cover types such as croplands were excluded from the model. (3) The intrannual dynamics of satellite derived GPP vegetation photosynthesis model (VPM) and modified VPM (MVPM) models were validated by flux towers at five ecosystem types across semi-arid Inner Mongolia. Though not as computationally intensive like most process-based models, MVPM offers an advantage over VPM by being independent of any ground measured meteorological data. MVPM provides a cost effective method of predicting GPP, especially at remote study sites which lack the required infrastructure to set up EC flux towers. While there was reasonable agreement between the observed GPPtower and predicted GPPVPM and GPPMVPM indicating the potential of these models for modeling of GPP in semi-arid ecosystems, there was some uncertainty in the predictive ability of these models, attributed to different sources of error.