论文部分内容阅读
选择在塔里木河下游普遍生长的胡杨林(Populus euphratica)为供试植被(3种不同大小胡杨),以大气温度、太阳净辐射、大气相对湿度、冠层顶风速、地下水位和胡杨树茎横截面积等6个影响因子作为影响胡杨林蒸腾量的自变量,基于最小二乘法建立了多元线性回归模型与非线性模型相结合的多元非线性回归耦合模型,并应用模型对地下变水位条件下塔里木河下游河岸胡杨林的耗水过程分别进行了时尺度和日尺度上的模拟研究。结果表明:在日内变化(时尺度)方面,大气温度、相对湿度、辐射、地下水位和树茎横截面积等5个因子是影响胡杨林蒸腾量的主要因素,在不同地下水位条件(Hg=1.0 m、1.2 m、2.5 m、3.0 m)条件下,胡杨蒸腾量观测与模拟的平均确定系数分别为0.69、0.87、0.82和0.88;而在日均变化(日尺度)方面,大气温度、地下水位和树茎横截面积等3个因子是影响胡杨林蒸腾量的主要因素,胡杨林的蒸腾量观测值与模拟值表现出较好的相关性,其确定系数与决定系数分别为R=0.73、R2=0.532,平均相对误差为19.6%,其显著性水平均通过p=0.05,表现出较好的拟合性。总之,模拟结果与试验观测结果比较吻合,该回归耦合模型具有使用简便、影响因子易测定,能够更好刻画植被腾发量的复杂非线性特性,为干旱区自然植被耗水量的估算提供了计算方法和科学依据。
Populus euphratica, a species commonly growing in the lower reaches of the Tarim River, was selected as the test vegetation (populus euphratica) with three different sizes. Atmospheric temperature, net solar radiation, relative atmospheric humidity, canopy top wind speed, Sectional area and other six influencing factors as independent variables affecting the transpiration of Populus euphratica, a multivariate nonlinear regression coupled model was established based on the least-squares method, which combined multivariate linear regression model with nonlinear model. Based on the model, The water consumption process of Populus euphratica forest in the lower reaches of the river was simulated respectively on the time scale and the day scale. The results showed that there are five factors affecting atmospheric transpiration, such as atmospheric temperature, relative humidity, radiation, groundwater level and stem cross-sectional area, m, 1.2 m, 2.5 m and 3.0 m), the average coefficient of determination and simulation of Populus euphratica were 0.69,0.87,0.82 and 0.88, respectively; while in the average daily change (daily scale), the average temperature, groundwater level And stem cross-sectional area were the main factors that affected the transpiration of Populus euphratica. The observed values of transpiration of Populus euphratica showed good correlation with the simulated values, and the coefficients of determination and the coefficients of determination were R = 0.73 and R2 = 0.532, the average relative error was 19.6%, the significance level of which passed p = 0.05, showing a good fit. In conclusion, the simulation results are in good agreement with the experimental observations. The regression coupled model has the advantages of easy to use and easily measurable influencing factors, and can better describe the complex nonlinear characteristics of vegetation evapotranspiration, which provides a basis for estimating the water consumption of natural vegetation in arid areas Methods and scientific basis.