[1]池 钦.基于随机森林的降雨预报季节性分析[J].大众科技,2022,24(10):17-20.
　Seasonal Analysis of Rainfall Forecast Based on Random Forest[J].Popular Science & Technology,2022,24(10):17-20.
点击复制

基于随机森林的降雨预报季节性分析()
分享到：

《大众科技》[ISSN:1008-1151/CN:45-1235/N]

卷:
24
期数:
2022年10
页码:
17-20
栏目:
信息技术与通信
出版日期:
2022-10-20

文章信息/Info

Title:
Seasonal Analysis of Rainfall Forecast Based on Random Forest
作者:
池 钦　
（安徽理工大学空间信息与测绘工程学院，安徽 淮南 232001）
关键词:
GPT3; 随机森林; PWV; 短临降雨; 季节性
Keywords:
GPT3 random forest PWV short term and imminent rainfall seasonal
文献标志码:
A
摘要:
全球导航卫星系统（Global Navigation Satellite Systems，GNSS）能够以高精度和高时间分辨率有效地反演大气可降水量（precipitable water vapor，PWV）。GNSS衍生的PWV可用于反映强对流天气过程中的水汽变化。通过研究PWV、气象参数与降雨的相关系可以帮助研究人员利用随机森林模型进行降雨预报。但缺少测站位置的气象参数限制了PWV的进一步应用。因此，文章利用GPT3模型得到经验气象参数帮助GNSS反演PWV，并利用wuh2测站建立随机森林降雨预报模型，研究季节性对预报效果的影响。结果表明，在7月—9月的预报效果是最好的，达到了93%以上，1月—3月的效果是最差的，但也在75%以上。在今后的研究中，可以针对不同季度改变建模策略，来提高预报的精度。
Abstract:
Global Navigation Satellite Systems (GNSS) can effectively retrieve the precise water vapor (PWV) with high accuracy and high time resolution. The PWV derived from GNSS can be used to reflect the changes of water vapor in the process of severe convective weather. By studying the correlation between PWV, meteorological parameters and rainfall, we can help us to use the random forest model to forecast rainfall. However, the lack of meteorological parameters at the station location limits the further application of PWV. Therefore, this paper uses the empirical meteorological parameters obtained by GPT3 model to help GNSS retrieve PWV, and uses wuh2 station to establish a random forest rainfall prediction model to study the influence of seasonality on the prediction effect. The results show that the forecast effect from July to September is the best, reaching more than 93%, and the forecast effect from January to March is the worst, but also more than 75%. In the future research, the modeling strategy can be changed according to different seasons to improve the accuracy of prediction.

参考文献/References:

[1] 施闯，张卫星，曹云昌，等. 基于北斗/GNSS 的中国－中南半岛地区大气水汽气候特征及同降水的相关分析[J]. 测绘学报，2020，49(9): 1112-1119. [2] 王勇，刘备，刘严萍，等. 基于小波变换的GPS水汽与气象要素相关性分析[J]. 大地测量与地球动力学，2017，37(7): 721-725. [3] 李黎，宋越，周嘉陵. 利用小波变换对暴雨过程中GNSS气象要素的初步探索[J]. 大地测量与地球动力学，2020，40(3): 225-230. [4] Wang H, Asefa T, Sarkar A. A novel non-homogeneous hidden Markov model for simulating and predicting monthly rainfall[J]. Theoretical and Applied Climatology, 2021, 143(1): 627-638. [5] Shou K J, Lin J F. Evaluation of the extreme rainfall predictions and their impact on landslide susceptibility in a sub-catchment scale[J]. Engineering Geology, 2020, 265: 105434. [6] Li G, Chang W, Yang H. A novel combined prediction model for monthly mean precipitation with error correction strategy[J]. IEEE Access, 2020, 8: 141432-141445.

备注/Memo

备注/Memo:
【收稿日期】2022-06-16 【作者简介】池钦（1998－），男，浙江瑞安人，安徽理工大学空间信息与测绘工程学院在读硕士研究生，研究方向为GNSS水汽反演。
更新日期/Last Update: 2022-11-29