基于GPR、CatBoost、XGBoost三种模型预测江西地区水面蒸发量

Prediction of pan evaporation of Jiangxi Province using GPR,CatBoost and XGBoost models

水面蒸发量是水资源规划与管理、农业灌溉设计和水文模拟等方面的基础数据,它是水量平衡计算中的关键要素。为了提高水面蒸发量的预测精度,选用了3种经验模型和3种学习机模型预测江西地区水面蒸发量,3种学习机模型包括GPR模型、XGBoost模型和Cat Boost模型。依据江西地区2001-2015年16个气象站的逐日气象资料,如最高(低)气温、全球太阳辐射、地外太阳辐射、相对湿度和风速,构建10种不同的输入参数,通过对4种统计指标(R2、RMSE、MBE、MAE)的大小进行评估来评价模型的模拟精度。结果表明:当气象资料充足时,推荐CatBoost 10模型为江西地区水面蒸发量的预测模型,该模型在验证期的R2、RMSE、MBE、MAE值分别为0.744、0.842、0.006、0.633 mm/d;在输入组合相同的条件下,3种学习机模型的模拟精度均优于相应的经验模型。通过研究对比提高了江西地区水面蒸发量模型预测的精度。

Pan evaporation provides basic data for the planning and management of water resources,the design of agricultural irrigation and hydrological modeling. It is a key element in the calculation of water balance. In order to improve the accuracy of pan evaporation modeling,three empirical models and three learning machine models were used to predict the pan evaporation in Jiangxi Province,including GPR,XGBoost and CatBoost models. According to the meteorological data of 16 meteorological stations in Jiangxi,such as maximum/minimum temperature,global solar radiation,extra-terrestrial solar radiation,relative humidity,wind speed,10 different input parameters were constructed,and four statistical indicators were adopted( R2,RMSE,MBE,MAE) to evaluate the performance of the models. The statistical results show that when meteorological data is sufficient,the CatBoost 10 model is recommended as the predictive model for pan evaporation in Jiangxi Province. The values of R2,RMSE,MBE,MAE in verification period are 0. 744,0. 842 mm/d,0. 006 mm/d,0. 633 mm/d,respectively. When the input combination is the same,the accuracy of the three learning machine models is better than their corresponding empirical models. This improved the model accuracy of predicting pan evaporation in Jiangxi.