Application of GBGWR Model in Spatial Distribution Simulation of Precipitation

doi:10.13474/j.cnki.11-2246.2018.0071

Abstract

Abstract:

In recent years,most parts of China were repeatedly attacked by natural disasters,such as floods and droughts.Therefore,it is of great significance to carry out spatial modeling of heavy flood and arid areas.Aiming at the problem that the model recognition and parameter estimation of the geographic weighted regression (GWR) model are easy to be affected by the observed anomalies,a Bayesian geographic weighted regression (GBGWR) method based on Gibbs sampling is developed to solve the problem.The Markov Chain Monte Carlo Bayesian method based on Gibbs sampling is used to estimate the parameters of the GWR model,and the anomalous data of the observed values are reduced by the smoothing function.Finally,the precipitation data of 35 observation sites in Hunan province from 1985 to 2015 were simulated.The experimental results show that the performance of this method is 19.8% higher than the GWR model,and is 8.2% higher than the BGWR model.It is proved that the method can effectively reduce the influence of the abnormal value and the "weak data" on the regression result.More realistic simulation of the spatial distribution of precipitation in Hunan Province.

Key words: Gibbs sampling, geographic weighted regression, BGWR, Hunan province, precipitation simulation

CLC Number:

P237

DENG Yue, LIU Yang, LIU Jiping, XU Shenghua, LUO An. Application of GBGWR Model in Spatial Distribution Simulation of Precipitation[J]. 测绘通报, 2018, 0(3): 32-37,42.

References

[1] 傅抱璞.地形和海拔高度对降水的影响[J].地理学报,1992(4):302-314.
[2] MARQUÍNEZ J,LASTRA J,GARCÍA P.Estimation Models for Precipitation in Mountainous Regions:The Use of GIS and Multivariate Analysis[J].Journal of Hydrology,2003,270(1-2):1-11.
[3] NAOUM S,TSANIS I K.A Multiple Linear Regression GIS Module Using Spatial Variables to Model Orographic Rainfall[J].Journal of Hydroinformatics,2004,6(1):39-56.
[4] BONI G,PARODI A,SICCARDI F.A New Parsimonious Methodology of Mapping the Spatial Variability of Annual Maximum Rainfall in Mountainous Environments[J].Journal of Hydrometeorology,2007,9(3):492-506.
[5] 彭晓芬,黄甫则,周汝良.云南省年均降雨量空间插值模拟方法比较[J].西南林业大学学报,2010,30(5):25-28.
[6] 张书文,李占甫.时间序列分析在预测年降水量中的应用[J].水资源研究,1996(4):7-11.
[7] 白江涛,白建军,王磊,等.基于GIS的关中——陕南地区降雨量空间插值分析[J].安徽农业科学,2011,39(33):20872-20876.
[8] 何红艳,郭志华,肖文发,等.利用GIS和多变量分析估算青藏高原月降水[J].生态学报,2005,25(11):2933-2938.
[9] DELBARI M,AFRASIAB P,JAHANI S,et al.Spatial Interpolation of Monthly and Annual Rainfall in Northeast of Iran[J].Meteorology & Atmospheric Physics,2013,122(1-2):103-113.
[10] SHEN Z,CHEN L,LIAO Q,et al.Impact of Spatial Rainfall Variability on Hydrology and Nonpoint Source Pollution Modeling[J].Journal of Hydrology,2012,s 472-473(24):205-215.
[11] 覃文忠,王建梅,刘妙龙.地理加权回归分析空间数据的空间非平稳性[J].辽宁师范大学学报(自然科学版),2005,28(4):476-479.
[12] BRUNSDON C.Geographically Weighted Reg Ression:A Method for Exploring Spatial Nonstation Arity[J].Geographical Analysis,1996,28(4):281-298.
[13] 武文娟,徐京华,时进,等.基于GWR的四川省医院床位数时空分布及其影响因素研究[J].测绘通报,2016(4):49-53.
[14] LESAGE J P.A Spatial Econometric Examination of China's Economic Growth[J].Geographic Information Sciences,1999,5(2):143-153.
[15] LESAGE J P.A Family of Geographically Weighted Regression Models[C]//Advances in Spatial Econometrics.Berlin Heidelberg:Springer,2004:241-264.
[16] SUBEDI N.Local Modeling of Tree Crown Area Using Bayesian Geographically Weighted Regression[D].[S.l.]:State University of New York College of Environmental Science and Forestry,2012.
[17] 崔长彬,姜石良,张正河.河北县域经济影响因素的空间差异分析——基于贝叶斯地理加权回归方法[J].经济地理,2012,32(2):39-45.
[18] 丁刚.基于BGWR模型的区域创新能力建设中政府效能测评[J].技术经济,2013,32(4):20-26.
[19] 叶金桃,王俊超,冯光柳,等.降雨型滑坡联合监测及预警预报系统研究[J].测绘通报,2012(2):31-34.