Comparison of random forests and support vector machine for real-time radar-derived rainfall forecasting

Abstract This study aims to compare two machine learning techniques, random forests (RF) and support vector machine (SVM), for real-time radar-derived rainfall forecasting. The real-time radar-derived rainfall forecasting models use the present grid-based radar-derived rainfall as the output variable and use antecedent grid-based radar-derived rainfall, grid position (longitude and latitude) and elevation as the input variables to forecast 1- to 3-h ahead rainfalls for all grids in a catchment. Grid-based radar-derived rainfalls of six typhoon events during 2012–2015 in three reservoir catchments of Taiwan are collected for model training and verifying. Two kinds of forecasting models are constructed and compared, which are single-mode forecasting model (SMFM) and multiple-mode forecasting model (MMFM) based on RF and SVM. The SMFM uses the same model for 1- to 3-h ahead rainfall forecasting; the MMFM uses three different models for 1- to 3-h ahead forecasting. According to forecasting performances, it reveals that the SMFMs give better performances than MMFMs and both SVM-based and RF-based SMFMs show satisfactory performances for 1-h ahead forecasting. However, for 2- and 3-h ahead forecasting, it is found that the RF-based SMFM underestimates the observed radar-derived rainfalls in most cases and the SVM-based SMFM can give better performances than RF-based SMFM.

[1]  G. A. Corzo Perez,et al.  Self-Learning Cellular Automata for Forecasting Precipitation from Radar Images , 2013 .

[2]  George C. Craig,et al.  Blending a probabilistic nowcasting method with a high‐resolution numerical weather prediction ensemble for convective precipitation forecasts , 2012 .

[3]  Ian Cluckie,et al.  An experiment of rainfall prediction over the odra catchment by combining weather radar and a numerical weather model , 2006 .

[4]  Ping-Feng Pai,et al.  Potential assessment of the support vector regression technique in rainfall forecasting , 2007 .

[5]  Shie-Yui Liong,et al.  FLOOD STAGE FORECASTING WITH SUPPORT VECTOR MACHINES 1 , 2002 .

[6]  José Augusto Baranauskas,et al.  How Many Trees in a Random Forest? , 2012, MLDM.

[7]  Shie-Yui Liong,et al.  Flow categorization model for improving forecasting , 2005 .

[8]  Junfei Chen,et al.  Statistical Uncertainty Estimation Using Random Forests and Its Application to Drought Forecast , 2012 .

[9]  H. Hirose,et al.  Comparison of artificially intelligent methods in short term rainfall forecast , 2010, 2010 13th International Conference on Computer and Information Technology (ICCIT).

[10]  Gwo-Fong Lin,et al.  The very short-term rainfall forecasting for a mountainous watershed by means of an ensemble numerical weather prediction system in Taiwan , 2017 .

[11]  Subimal Ghosh,et al.  SVM-PGSL coupled approach for statistical downscaling to predict rainfall from GCM output , 2010 .

[12]  Zbynek Sokol,et al.  Nowcasting of 1-h precipitation using radar and NWP data , 2006 .

[13]  Pao-Shan Yu,et al.  Statistical downscaling of daily precipitation using support vector machines and multivariate analysis , 2010 .

[14]  Thanh-Tung Nguyen,et al.  An L1-Regression Random Forests Method for Forecasting of Hoa Binh Reservoir's Incoming Flow , 2015, 2015 Seventh International Conference on Knowledge and Systems Engineering (KSE).

[15]  Xiaohong Chen,et al.  Flood hazard risk assessment model based on random forest , 2015 .

[16]  Gwo-Fong Lin,et al.  Ensemble forecasting of typhoon rainfall and floods over a mountainous watershed in Taiwan , 2013 .

[17]  Mohammad Mahadi Hasan,et al.  Correcting bias in radar Z – R relationships due to uncertainty in point rain gauge networks , 2014 .

[18]  Gwo-Fong Lin,et al.  Effective typhoon characteristics and their effects on hourly reservoir inflow forecasting , 2010 .

[19]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[20]  Rajib Maity,et al.  Potential of support vector regression for prediction of monthly streamflow using endogenous property , 2010 .

[21]  P. Mujumdar,et al.  A comparison of three methods for downscaling daily precipitation in the Punjab region , 2011 .

[22]  Martin L. Hazelton,et al.  Predicting the Onset of Australian Winter Rainfall by Nonlinear Classification , 2005 .

[23]  Shie-Yui Liong,et al.  Forecasting of hydrologic time series with ridge regression in feature space , 2007 .

[24]  Gwo-Fong Lin,et al.  Support vector machine-based models for hourly reservoir inflow forecasting during typhoon-warning periods , 2009 .

[25]  Pao-Shan Yu,et al.  Comparison of neural network architectures and inputs for radar rainfall adjustment for typhoon events , 2011 .

[26]  I-Fan Chang,et al.  Support vector regression for real-time flood stage forecasting , 2006 .

[27]  Antanas Verikas,et al.  Mining data with random forests: A survey and results of new tests , 2011, Pattern Recognit..

[28]  Auroop R. Ganguly,et al.  Distributed Quantitative Precipitation Forecasting Using Information from Radar and Numerical Weather Prediction Models , 2003 .

[29]  Kuolin Hsu,et al.  Quantitative Precipitation Nowcasting: A Lagrangian Pixel-Based Approach , 2012 .

[30]  Mark Junjie Li,et al.  Forecasting Time Series Water Levels on Mekong River Using Machine Learning Models , 2015, 2015 Seventh International Conference on Knowledge and Systems Engineering (KSE).

[31]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[32]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[33]  Kurt Hornik,et al.  The support vector machine under test , 2003, Neurocomputing.

[34]  Pao-Shan Yu,et al.  Pruning of support vector networks on flood forecasting , 2007 .

[35]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[36]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[37]  Gwo-Fong Lin,et al.  Effective forecasting of hourly typhoon rainfall using support vector machines , 2009 .

[38]  B. Golding.  Nimrod: a system for generating automated very short range forecasts , 1998 .

[39]  Gwo-Fong Lin,et al.  Typhoon flood forecasting using integrated two-stage Support Vector Machine approach , 2013 .

[40]  S. Sorooshian,et al.  Watershed rainfall forecasting using neuro-fuzzy networks with the assimilation of multi-sensor information , 2014 .

[41]  Kuolin Hsu,et al.  Short-term quantitative precipitation forecasting using an object-based approach , 2013 .