考虑降水时间相关性的地面观测-雷达-卫星遥感逐时降水融合方法研究

阮惠华; 张钧民; 许剑辉; 戴晓爱

doi:10.16032/j.issn.1004-4965.2023.028

考虑降水时间相关性的地面观测-雷达-卫星遥感逐时降水融合方法研究

doi: 10.16032/j.issn.1004-4965.2023.028

阮惠华¹,
张钧民^{2, 3},
许剑辉^2, ,,
戴晓爱³

1.
广东省气象探测数据中心，广东广州 510641
2.
广东省科学院广州地理研究所/广东省遥感与地理信息应用重点实验室/广东省地理空间信息技术与应用公共实验室，广东广州 510070
3.
成都理工大学地球科学学院，四川成都 610059

基金项目:

广东省省级科技计划项目 2018B020207012

国家自然科学基金 41901371

详细信息

通讯作者:
许剑辉，男，广东省人，副研究员，主要从事多源遥感数据时空融合与同化研究。E-mail：xujianhui306@gdas.ac.cn

中图分类号: P413
计量
- 文章访问数: 119
- HTML全文浏览量: 9
- PDF下载量: 23
- 被引次数: 0
出版历程
- 收稿日期: 2021-10-08
- 修回日期: 2023-01-08
- 网络出版日期: 2023-09-11
- 刊出日期: 2023-06-20

AN XGBOOST-BASED GEOSTATISTICAL DATA FUSION METHOD FOR INTEGRATING HOURLY GAUGE-RADAR-SATELLITE PRECIPITATION DATA BY CONSIDERING THE TEMPORAL CORRELATION CHARACTERISTICS OF PRECIPITATION

1.
Guangdong Meteorological Observation Data Center, Guangzhou 510641, China
2.
Key Laboratory of Guangdong for Utilization of Remote Sensing and Geographical Information System/Guangdong Open Laboratory of Geospatial Information Technology and Application/Guangzhou Institute of Geography, Guangdong Academy of Sciences, Guangzhou 510070, China
3.
School of Earth Sciences, Chengdu University of Technology, Chengdu 610059, China

摘要

摘要: 以广东省北部山区2018年汛期强降水时段4月23—28日龙舟水、5月7—11日龙舟水和9月16—17日台风“山竹”3次典型暴雨过程的逐时降水为研究对象，研究基于XGBoost算法与地统计学理论的地面观测-前两个时刻逐时降水-雷达-卫星遥感的多源逐时降水融合模型，充分考虑相邻时刻降水的时间相关性，得到空间分辨率为1 km的逐时降水融合数据。此外，分别利用XGBoost与随机森林(RF)算法进行不考虑降水时间相关性的地面观测-雷达-卫星遥感逐时降水融合对比试验，并对试验结果进行精度评价。结果表明：(1) 在3次暴雨过程中，三种融合模型的逐时降水融合结果具有类似的空间分布；(2) 与XGBoost和RF逐时降水融合结果相比，融合了降水时间相关性的逐时降水融合结果在不同暴雨过程的准确性均有明显改进，3次暴雨过程的决定系数(R²)平均提高了7.89%和23.27%；(3) XGBoost逐时降水融合模型的精度整体上优于RF逐时降水融合模型，3次暴雨过程的R²分别提高了7.1%、4.3%和31.4%。
- 逐时降水 /
- 多源数据融合 /
- 时间相关性 /
- XGBoost算法 /
- 地统计学
Abstract: In this study, an hourly precipitation data fusion model was proposed based on the framework of XGBoost-based machine learning algorithm and geostatistical theory. The model integrated the multi-source hourly precipitation data, including gauge, radar, Integrated Multi-satellite Retrievals for GPM (IMERG), Global Satellite Mapping of Precipitation (GSMaP), and gridded precipitation data. The gridded precipitation data recorded in the previous two hours were analyzed to summarize the temporal correlation characteristics of precipitation in the proposed hourly precipitation data fusion model. The 1 km IMERG and GSMaP precipitation data were downscaled by using the area-to-point Kriging method. The proposed hourly precipitation data fusion model was built using hourly precipitation data from 200 rain gauges and tested with three regional rainstorm events occurred during April 23 to 28, May 7 to 11, and September 16 to 17, 2018 in northern Guangdong. The results were compared with the results estimated from the XGBoost and Random Forest (RF) algorithms without considering the temporal correlation characteristics of precipitation. The results showed that: (1) For the three regional rainstorm events, the three models reported similar spatial distribution of hourly precipitation. (2) Compared with the XGBoost and RF models, the proposed model that integrated the temporal correlation of hourly precipitation data could significantly improve the accuracy of estimated precipitation data. (3) The XGBoost model showed better performance than the RF model in capturing the nonlinear relationship between gauge precipitation data and independent variables.
- hourly precipitation /
- multi-source precipitation data fusion /
- temporal correlation characteristics /
- eXtreme Gradient Boosting /
- geostatistics

HTML全文

图 1 研究区域及气象站点分布

a.广东省边界范围；b.研究区域及气象站点分布。

下载: 全尺寸图片幻灯片

图 2 考虑时间相关性的多源逐时降水融合方法技术流程图

下载: 全尺寸图片幻灯片

图 3 2018年过程Ⅰ(a)、过程Ⅱ(b)和过程Ⅲ(c)三种方案的逐时降水融合与站点降水观测比较

下载: 全尺寸图片幻灯片

图 4 不同融合方法的逐时降水融合的R²(a)、RMSE(b)和MAE(c)箱图

下载: 全尺寸图片幻灯片

图 5 暴雨过程Ⅰ逐时降水融合R²(a)、RMSE(b)、MAE(c)精度评价

下载: 全尺寸图片幻灯片

图 6 暴雨过程Ⅱ逐时降水融合R²(a)、RMSE(b)、MAE(c)精度评价

下载: 全尺寸图片幻灯片

图 7 暴雨过程Ⅲ的逐时降水融合R²(a)、RMSE(b)、MAE(c)精度评价

下载: 全尺寸图片幻灯片

图 8 不同融合方法的累计降水融合结果空间分布图

a.暴雨过程Ⅰ；b.暴雨过程Ⅱ；c.暴雨过程Ⅲ。

下载: 全尺寸图片幻灯片

表 1 相关数据来源及介绍

数据	变量名称	数据描述	数据来源
气象站降水	Gauge Precipitation	时间分辨率为1h	广东省气象局(http://data.cma.cn/wa)
雷达降水	Radar	时间分辨率为6 min, 空间分辨率1 km	广东省气象局(http://data.cma.cn/wa)
IMERG降水	IMERG	时间分辨率为0.5 h, 空间分辨率为0.1°	https://gpm.nasa.gov/data/imerg
GSMaP降水	GSMaP	时间分辨率为1 h, 空间分辨率为0.1°	https://sharaku.eorc.jaxa.jp/GSMaP/
海拔	DEM	ALOS DSM, 空间分辨率为30 m	https://www.eorc.jaxa.jp/ALOS/en/aw3d30/index.htm
与海岸线距离	Coastline	空间分辨率为1 km

下载: 导出CSV

表 2 同降水融合方案R²表现最好的时刻占总暴雨过程的比例情况

过程	过程Ⅰ	过程Ⅱ	过程Ⅲ
方案Ⅰ	53.82%	55.08%	68.23%
方案Ⅱ	23.09%	23.22%	22.71%
方案Ⅲ	23.09%	21.70%	9.06%

下载: 导出CSV

参考文献(35)

[1]	师春香, 潘旸, 谷军霞, 等. 多源气象数据融合格点实况产品研制进展[J]. 气象学报, 2019, 77(4): 774-783.
[2]	SONG Y, LIU H, WANG X, et al. Numerical simulation of the impact of urban non-uniformity on precipitation[J]. Adv Atmos Sci, 2016, 33(6): 783-793.
[3]	MIN X, YANG C, DONG N. Merging satellite and gauge rainfalls for flood forecasting of two catchments under different climate conditions[J]. Water, 2020, 12(3): 802.
[4]	BRUNETTI M T, MELILLO M, GARIANO S L, et al. Satellite rainfall products outperform ground observations for landslide prediction in India[J]. Hydrol Earth Syst Sci, 2021, 25(6): 3 267-3 279.
[5]	FOEHN A, HERNANDEZ J G, SCHAEFLI B, et al. Spatial interpolation of precipitation from multiple rain gauge networks and weather radar data for operational applications in Alpine catchments[J]. J Hydrology, 2018, 563: 1 092-1 110.
[6]	CHAO L, ZHANG K, LI Z, ZHU Y, et al. Geographically weighted regression based methods for merging satellite and gauge precipitation[J]. J Hydrology, 2018, 558: 275-289
[7]	CHEN Y, HUANG J, SHENG S, et al. A new downscaling-integration framework for high-resolution monthly precipitation estimates: Combining rain gauge observations, satellite-derived precipitation data and geographical ancillary data[J]. Remote Sens Environ, 2018, 214: 154-172.
[8]	VERDIN A, RAJAGOPALAN B, KLEIBER W, et al. A Bayesian kriging approach for blending satellite and ground precipitation observations[J]. Water Resour Res, 2015, 51(2): 908-921.
[9]	YANG P, NG T L. Fast bayesian regression Kriging method for real‐time merging of radar, rain gauge, and crowdsourced rainfall data[J]. Water Resour Res, 2019, 55(4): 3 194-3 214.
[10]	SUN R, YUAN H, YANG Y. Using multiple satellite-gauge merged precipitation products ensemble for hydrologic uncertainty analysis over the Huaihe River basin[J]. J Hydrology, 2018, 566: 406-420.
[11]	RAHMAN K U, SHANG S, SHAHID M, et al. Application of Dynamic Clustered Bayesian Model Averaging(DCBA)algorithm for merging multi-satellite precipitation products over Pakistan[J]. J Hydrometeorology, 2020, 21(1): 17-37.
[12]	潘旸, 沈艳, 宇婧婧, 等. 基于贝叶斯融合方法的高分辨率地面-卫星-雷达三源降水融合试验[J]. 气象学报, 2015, 73(1): 177-186.
[13]	卢新玉, 刘艳, 王秀琴, 等. 新疆地区多源降水融合试验[J]. 干旱区研究, 2020, 37(5): 1 223-1 232.
[14]	潘旸, 沈艳, 宇婧婧, 等. 基于最优插值方法分析的中国区域地面观测与卫星反演逐时降水融合试验[J]. 气象学报, 2012, 70(6): 1 381-1 389.
[15]	罗聪, 张华龙, 曾沁, 等. 多模式融合的广东网格定量降水预报方法的研发与评估[J]. 气象, 2021, 47(5): 539-549.
[16]	朱岩, 翟丹华, 吴志鹏, 等. 基于Xgboost算法的短时强降水预报方法[J]. 气象科技, 2021, 49(3): 406-418.
[17]	HEN H, CHANDRASEKAR V, CIFELLI R, et al. A machine learning system for precipitation estimation using satellite and ground radar network observations[J]. IEEE Trans Geosci Remote Sens, 2019, 58(2): 982-994.
[18]	BAEZ-VILLANUEVA O M, ZAMBRANO-BIGIARINI M, BECK H, MCNAMARA I, THINH N X. RF-MEP: A novel Random Forest method for merging gridded precipitation products and ground-based measurements[J]. Remote Sens Environ, 2020, 239: 111606.
[19]	ZHANG L, LI X, ZHENG D, et al. Merging multiple satellite-based precipitation products and gauge observations using a novel double machine learning approach[J]. J Hydrology, 2021, 594: 125969.
[20]	潘旸, 谷军霞, 宇婧婧, 等. 中国区域高分辨率多源降水观测产品的融合方法试验[J]. 气象学报, 2018, 76(5): 755-766.
[21]	YU J, LI X F, LEWIS E, et al. UKGrsHP: a UK high-resolution gauge-radar-satellite merged hourly precipitation analysis dataset[J]. Clim Dyn, 2020, 54(5): 2 919-2 940.
[22]	吴薇, 杜冰, 黄晓龙, 等. 四川区域融合降水产品的质量评估[J]. 高原山地气象研究, 2019, 39(2): 76-81.
[23]	田凤云, 吴志勇, 徐征光, 等. 黄土高原典型流域网格化逐时降水融合产品CMPA V1.0的精度分析[J]. 水电能源科学, 2019, 37(2): 5-9.
[24]	RAHMAN K U, SHANG S, ZOHAIB M. Assessment of merged satellite precipitation datasets in monitoring meteorological drought over Pakistan[J]. Remote Sens, 2021, 13(9): 1 662.
[25]	张钧民, 阮惠华, 许剑辉, 等. 基于XGBoost的多源降水数据融合方法研究[J]. 热带地理, 2021, 41(4): 845-856.
[26]	O S, FOELSCHE U, KIRCHENGAST G, et al. Evaluation of GPM IMERG Early, Late, and Final rainfall estimates using WegenerNet gauge data in southeastern Austria[J]. Hydrol Earth Syst Sci, 2017, 21(12): 6 559-6 572.
[27]	KYRIAKIDIS P C. A Geostatistical Framework for Area-to-Point Spatial Interpolation[J]. Geographical Analysis, 2004, 36(3): 259-289.
[28]	USHIO T, SASASHIGE K, KUBOTA T, et al. A Kalman Filter Approach to the Global Satellite Mapping of Precipitation(GSMaP)from combined passive microwave and infrared radiometric data[J]. J Meteor Soc Japan Ser Ⅱ, 2009, 87: 137-151.
[29]	HU M, HUANG Y. Atakrig: An R package for multivariate area-to-area and area-to-point kriging predictions[J]. Comput & Geosci, 2020, 139: 104471.
[30]	CHEN T, GUESTRIN C. XGBoost: A Scalable Tree Boosting System[C]/Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. 2016: 785-794.
[31]	MENG C, MO X, LIU S, HU S. Extensive evaluation of IMERG precipitation for both liquid and solid in Yellow River source region[J]. Atmos Res, 2021, 256: 105570.
[32]	陈晓宏, 钟睿达, 王兆礼, 等. 新一代GPM IMERG卫星遥感降水数据在中国南方地区的精度及水文效用评估[J]. 水利学报, 2017, 48(10): 1 147-1 156.
[33]	李彦妮, 黄昌, 庞国伟. 全球降雨计划GSMaP与IMERG卫星降雨产品在陕西地区的精度评估[J]. 干旱区地理, 2021: 1-17.
[34]	WU H, YANG Q, LIU J, WANG G. A spatiotemporal deep fusion model for merging satellite and gauge precipitation in China[J]. J Hydrology, 2020, 584: 124664.
[35]	SHARIFI E, SAGHAFIAN B, STEINACKER R. Downscaling satellite precipitation estimates with multiple linear regression, artificial neural networks, and spline interpolation techniques[J]. J Geophy Res Atmos, 2019, 124(2): 789-805.