The Spatial-Temporal Characteristics of Evapotranspiration of China’s Terrestrial Ecosystems during 1991～2000

The spatial-temporal dynamics of ecosystems’evapotranspiration (ET) are important factors in climate change and terrestrial carbon and water cycles, closely relating to ecosystem net primary productivity (NPP). On the basis of multi-year NOAA-AVHRR remote sensing data, meteorology data (including temperature, precipitation, humidity and radiation), soil data and other ancillary data, a modified ecosystem productivity process model, Boreal Ecosystem Productivity Simulator (BEPS), are developed to estimate the evaptranspiration (ET) of terrestrial ecosystems in China at 10-km spatial resolution on different temporal scales. The estimated results are used to develop the spatial-temporal variations of ET and their responses to climate changes. Results show that in 1990s, the annual average ET nationwide increased at a mean rate of 442.55 mm/a, with maximum annual ET of 475.91mm/a in 1998, when average temperature and total precipitation reached peaks and minimum value of 425.59mm/a in 1992, when precipitation hit the lowest point. Annual ET were positively correlated with temperature and precipitation, and the correlation between ET and precipitation (R2=0.950, P0.05, n=10) was more significant than that between ET and temperature (R2=0.399, P0.05, n=10), indicating that precipitation may be the main factor influncing the interannnual variations of terrestrial ecosystem in China. Annual average ET exhibited a distinguishable pattern, increasing from northwest to southeast. The distribution characteristics of ET are analyzed on the basis of climate zones or land cover types. It indicates that the ET of broad-leaved forests is generally higher than that of need-leaved forests, while ET of evergreen forests is also higher than that of deciduous forests. Also, the evaporation and transpiration of different types of land cover are compared. Furthermore, the seasonal variation of ET shows unimodal pattern, and interannual variation of ET can be associated with land cover and climate condition. Although annual ET demonstrates an increase in the analysed period, the spatial distribution of interannual trend indicats a higher degree of spatial heterogeneity, coupled with regional climate variations and extreme climate events.