摘要
受季风气候等影响,传统T_(m)模型在中国沿海地区的精度难以满足导航定位等领域的发展需求,亟须建立一种高精度的区域性T_(m)模型。通过利用随机森林(Random Forest,RF)的方法,对基于该地区探空数据所构建的非线性T_(m)模型(F-T_(m))进一步精化,选取4个特征值(气压P、地表温度T_(s)、水汽压e_(s)、比湿s)作为输入因子,对F-T_(m)模型的偏差进行预测改正,构建了RFF-T_(m)模型。通过时空特性分析表明,该模型在沿海地区具有较好的适用性。
Due to the influence of monsoon climate,the accuracy of traditional T_(m) model in China's coastal areas is difficult to meet the development needs of navigation and positioning,so it is urgent to establish a high-precision regional T_(m) model.By using Random Forest(RF)method,the nonlinear T_(m) model(F-T_(m))constructed based on radiosonde data in this area is further refined.Four characteristic values(pressure P,surface temperature T_(s),vapor pressure ES and specific humidity S)are selected as input factors to predict and correct the deviations of F-T_(m) model.RFF-T_(m) model is constructed.The spatio-temporal analysis shows that the model has good applicability in coastal areas.
作者
万庆同
Wan Qingtong(Anhui Urban Construction Design and Research Institute Co.,Ltd.,Hefei 230000,China)
出处
《城市勘测》
2022年第3期84-86,共3页
Urban Geotechnical Investigation & Surveying
基金
国家自然科学基金项目(41864002)。
