基于BP网络的白杨树潜在蒸散量模拟优化研究

Optimization of simulation in poplar potential evapotranspiration based on BP network

【目的】探寻最优的白杨树潜在蒸散量(ET_(0))模拟方法。【方法】基于贵德县2022年1—6月的气象数据,建立了不同算法优化下白杨树ET_(0)的模拟模型,选取151组数据对模型进行训练,选取30组数据对模型进行验证。【结果】影响白杨树ET_(0)的主要因素为平均气温,2 m高处风速,饱和水汽压差,相对湿度,基于不同算法优化后的BP网络模型在模拟白杨树ET_(0)方面的精度优于传统的BP网络模型。【结论】EWOA-BP网络模型的模拟精度最佳,可以作为预测白杨树ET_(0)的首选模型。

【Objective】There are many environmental factors that affect vegetation ET_(0),and it is difficult to obtain some of them.Therefore,it is particularly important to find a simple and feasible method to simulate the response process of vegetation ET_(0)to environmental factors,and to quantitatively analyze the ecological water demand of typical vegetation in alpine region(taking Guide County as an example).【Method】Based on the meteorological data of Guide County from January to June,2022,a BP network model for simulating vegetation ET_(0)under different algorithm optimization was established,151 groups of data were selected to train the model,and the remaining 30 groups of data were tested and verified.【Result】The main factors affecting poplar ET_(0)are average temperature,2 m wind speed,saturated vapor pressure difference,relative humidity,and the correlation from large to small is average temperature,saturated vapor pressure difference,2 m wind speed and relative humidity.BP neural network model optimized by different algorithms is used to simulate poplar ET_(0),and the simulation effect is improved compared with the traditional BP network.【Conclusion】among which EWOA-BP optimization model has the best simulation result,which can be used as the first choice model to predict poplar ET_(0)in alpine region.