基于天气预报估算参考作物蒸发蒸腾量的预测模型比较

Real-time ET_0 Forecasting by Synoptial Report

参考作物蒸发蒸腾量(ET0)是计算作物需水量和进行灌溉预报的基础要素。利用天气预报可测因子和Penman Monteith公式计算ET0,分别建立多元线性回归模型(MLR)和自适应神经模糊推理系统模型(ANFIS),两种方法的估算值与Penman Monteith公式计算值没有明显的差异,自适应神经模糊推理系统预测值相对于多元线性回归模型具有整体吻合度好,相关性高。两种预测模型的输入项完全可以从当前短期气象预报中获得,程序运行操作简单,具有实用价值,为实时灌溉预报提供了理论基础。

The estimation of evapotranspiration from vegetated surfaces is a basic tool to compute water balances and to estimate water availability and requirements. In this research establishing the multivariate linear regression mode （MLR） and the auto-adapted nerve fuzzy reasoning system model （ANFIS） by synoptial report and ET0 （calculated with Penman Monteith Formula）. The results indicated that the two methods and the Penman Monteith Formula did not have significant difference for the estimation. ANFIS -ET0 （ET0 simulated with ANFIS model） had higher accuracy than MLR -ET0 （ET0 calculated with multivariate linear regression model）, and the forecasting result was highly dependent on the input factor. The input items could be obtained from the current short-range weather forecasting information. The operation of the program is simple, and the ANFIS model has practical value in the real-time irrigation forecast system.