Comparison of Spatial Interpolation Methods Based on Rain Gauges for Annual Precipitation on the Tibetan Plateau
Xiaoke Zhang1,2, Xuyang Lu2,3, Xiaodan Wang2
More details
Hide details
1School of Public Administration, Hohai University, Nanjing 210098, China
2Key Laboratory of Mountain Surface Processes and Ecological Regulation, Institute of Mountain Hazards
and Environment, CAS, Chengdu 610041, China
3Xainza Alpine Steppe and Wetland Ecosystem Observation and Experiment Station, CAS, Xainza 853100, China
Publish date: 2016-05-25
Submission date: 2014-11-26
Final revision date: 2015-09-08
Acceptance date: 2016-02-16
Pol. J. Environ. Stud. 2016;25(3):1339–1345
Accurate precipitation data are of great importance for environmental applications. Interpolation methods are usually applied to afford spatially distributed precipitation data. However, due to the scarcity of rain gauges, different spatial interpolation methods may result in deviations from the real spatial distribution of precipitation. In this study, three different interpolation methods were investigated with regard to their suitability for producing a spatial precipitation distribution on China’s Tibetan Plateau. Precipitation data from 39 rain gauges were spatially interpolated using ordinary kriging, cokriging with covariates as elevation (Cok-elevation), and cokriging with covariates as tropical rainfall measuring mission (Cok-TRMM). The results showed that the mean absolute error (MAE), mean relative error (MRE), and root mean square error (RMSE) for Cok-TRMM amounted to 103.85 mm, 0.32, and 134.50 mm, respectively. These numbers were lower than the fi gures for ordinary kriging (MAE 111.01 mm, MRE 0.34, RMSE 144.86 mm) and Cok-elevation (MAE 111.43 mm, MRE 0.34, RMSE 144.35 mm). In addition, the correlation coefficient between observed and predicted values of Cok-TRMM (r2 = 0.53) was higher than that for ordinary kriging (r2 = 0.46) and Cok-elevation (r2 = 0.46). Our results demonstrate that Cok-TRMM is more effective at producing a spatial precipitation distribution on the Tibetan Plateau and can serve as a new spatial interpolation method for precipitation in data-scarce regions.