Evapotranspiration Quantitative Estimation Research Of Tahe Forest Ecosystem

Forests are the foundation of human survival and development, and it plays a vital role in global ecological environment. It has variety of ecological functions such as absorbing carbon dioxide, preserving water, adjusting climates, breaking wind and fixing the sand, conserving biological diversity, etc., it is the main body of terrestrial ecosystem.In recent years, with the increasing seriousness of the problem such as global warming, carbon sequestration ability of forest has attracted the attention of researchers. Carbon cycle research of forest ecosystem has become an important research area. In the process of influence between global change and carbon cycle, researchers realized that the water cycle and the carbon cycle are interacting and coupling together, and should be modelled at the same time. Numerous studies have shown that the atmospheric precipitation is the only source of water for land, and most of which has evaporated back to the atmosphere, therefore, evapotranspiration counts a significant proportion in the water cycle.Evapotranspiration(ET) is an important parameter of agriculture, meteorology and hydrology research, and an important part of the global hydrological cycle. The traditional method of estimating ET is based on the single point calculation of the observation site, but such method could not capture ET of large-scale area. Due to the ability of distributed hydrological cycle in capturing hydrological parameters, considering the spatial heterogeneity of hydrological process, revealing physical process of hydrological cycle closer to the objective reality, simulating hydrological cycle more accurately, and combining easier with remote sensing and geographic information system, such abilities utilize the unique advantages of various disciplines, conductive to studies in large-scale area, thus, it has become the focus of current water cycle research.In order to achieve better simulation research in carbon cycle, archiving more efficient service and function of forest ecological system, this paper uses the improved DHSVM distributed hydrological model, using Penman-Monteith formulae for evaporation calculation, and estimate daily ET of Tahe area in2007. Data processing includes using TM remote sensing data to obtain leaf area index(LAI) and other surface data. Slope, aspect and other topographic index are obtained via Digital Elevation Model(DEM). Other major inputs of the model include meteorological data, land cover data and soil type data. This paper applies BP neural network to verify the accuracy of the results via building the relationship between daily ET and watershed outlet flow, and establishing a water balance equation in the study watershed to test the model accuracy as well. The followings are main conclusions obtained in this study:(1)It is estimated that the annual total ET of Tahe watershed is234.01mm, annual ET quantity is234.01mm, the ratio of the amount of ET and rainfall (ET/P) is0.64.(2)ET has a significant correlation with seasons. The ET has the highest value in summer and the average daily value is1.56mm, average daily ET in autumn and spring are0.20mm、0.29mm, respectively, and winter has the lowest ET value.(3)The change of ET has a significant correlation with land cover types. In the research area, broad-leaved forests have higher ability of ET than mixed forests, followed by coniferous forests.(4) ET is greatly influenced by meteorological factors. Rainfall and radiation effect the most.(5) This paper applying BP neural network and establishing the water equation balance of the study area, together to verify the adaptability of the model in the study area. The accuracy simulation results of BP neural network are85%or more, the error of water balance equation in the study area is only1.02%, these proves the accuracy of the estimated model results.