Spatial Pattern Study Of CO₂ Fluxes Of Haibei Alpine Meadow On The Qinghai-Tibet Plateau Based On GIS And RS

The productivity of terrestrial ecosystem is a key part of carbon cycle in global change research. Long-term measurement of CO₂, H₂O and hear fluxes between natural vegetation and atmosphere have facilitated the research on carbon cycle in terrestrial ecosystems and its response to global climatic changes. Although these observed datum have highly temporal continuity, but these are still point measurements with limited fetch area. Quantitative extrapolations and scaling-up of these tower measurements are now urgently needed for carbon cycle research. Achieving these quantitative regional estimates will require the integration of remote sensing, geographical information systems, and mathematical modeling. This study aimed to achieve a preliminary discuss of the flux-NDVI relationship which facilitate the scaling-up of site-specific CO₂ flux measurements to regional levels in the Typical Alpine Meadow area on the Qinghai-Tibet Plateau.(1) Spatial database of the Tibet Plateau: 16-day composite NDVI, 16-day total precipitation, and 16-day average PAR. NDVI data were acquired from the website "Earth Observing System Data Gateway" of NOAA and NASA. A further process had been conducted for excluding abnormal data of 16-day composite NDVI. Data of 16-day precipitation and PAR were gained from China Meteorological Administration base on daily measurements. AUSPLIN model was used for spatial interpolation of NDVI and PAR.(2) Model development: Based on the flux data of alpine shrub meadow at Haibei on northeast of Tibet Plateau, the seasonality of components of the flux data were analyzed. These components include gross primary production, GPP; ecosystem respiration, Re; and net ecosystem exchange, NEE. The relationships between flux components and Normalized Difference Vegetation Index (NDVI) were also analyzed. Thus, we formed the foundation of semi-empirical statistical model. We established the statistical models for this site with the flux components, NDVI and other impact factors (precipitation, photosynthetically active radiation), and we validated the model at the site. The results of validation indicate that the model we established can well simulate the characteristics of flux components.(3) Spatial Clustering: Based on the 1:1 million China Rangeland Resources map, NDVI time-series in growing-season were used for spatial clustering. Particular areas, with similar seasonal dynamics of NDVI of Haibei flux tower in the typical alpine meadow region, were extractive as the areas for flux scaling.(4) Results: ArcGIS software system was used for simulating the flux components in the specific scaling-up areas. Annual average GPP is 100 gCO₂/m² and the annual total GPP was 5.51 TgC in 2003. The simulated result indicated that the region was a week source of CO₂, the regional average discharges was -12.39 g CO₂/m², and the annual total NEE was -0.68 TgC in 2003.