Statistical Downscaling for Climate Change Studies: Comparing TCPG-GLM and Gradient-Boosted TCPG Models for Local-Scale Rainfall in Purwakarta
Article Sidebar
Main Article Content
Abstract
Accurate rainfall projections are vital for climate adaptation and water resource planning in tropical regions, where rainfall patterns are highly variable and zero-inflated. This study evaluates the performance of two statistical downscaling approaches based on the Tweedie Compound Poisson-Gamma (TCPG) distribution for monthly rainfall modeling in Purwakarta, Indonesia: the generalized linear model with Tweedie distribution (GLM-Tweedie) and the nonlinear Tweedie distribution boosting model (TDboost). Using observational rainfall data from five stations and predictor variables from three Earth System Models (BNU-ESM, MIROC-ESM, and MIROC-ESM-CHEM), model performance was assessed using root mean squared error of prediction (RMSEP) and Pearson correlation coefficient. The results show that TDboost generally outperformed GLM-Tweedie in terms of predictive accuracy and temporal consistency. At Cibukamanah station, TDboost achieved the lowest RMSEP under the MIROC-ESM-CHEM scenario (90.63 mm), while its highest correlation was obtained under the BNU-ESM scenario (r = 0.88). Similarly, at Cisomang station, TDboost under the MIROC-ESM-CHEM scenario achieved an RMSEP of 109.75 mm and a correlation of r = 0.78, indicating improved predictive performance compared with GLM-Tweedie. These findings suggest that distribution-aware boosting techniques can better capture the nonlinear and zero-inflated characteristics of tropical rainfall. The results also indicate that the best-performing ESM may depend on the evaluation metric used, with MIROC-ESM-CHEM providing lower prediction errors at some stations and BNU-ESM showing strong temporal correlation when combined with TDboost. Overall, this study highlights the importance of combining appropriate statistical downscaling models with suitable ESM predictors for improving local-scale rainfall prediction in complex tropical environments.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
References
Scaife, A. A.; Ferranti, L.; Alves, O.; Athanasiadis, P.; Baehr, J.; Dequé, M.; Dippe, T.; Dunstone, N.; Fereday, D.; Gudgel, R. G.; et al. Tropical Rainfall Predictions from Multiple Seasonal Forecast Systems. Int. J. Climatol. 2019, 39(2), 974–988. https://doi.org/10.1002/joc.5855
Hendrawan, I. G.; Asai, K.; Triwahyuni, A.; Lestari, D. V. The Interannual Rainfall Variability in Indonesia Corresponding to El Niño Southern Oscillation and Indian Ocean Dipole. Acta Oceanol. Sin. 2019, 38(7), 57–66. https://doi.org/10.1007/s13131-019-1457-1
Fauzi, F.; Kuswanto, H.; Atok, R. M.; Salamah, M. Statistical Downscaling and Bias Correction of the Earth System Models Outputs for Future Climate Projection under Solar Geoengineering: Case Study Indonesia. J. Mech. Eng. 2021, 10(1), 33–49.
Shabu, S. L. J.; Refonaa, J.; Devi, D.; Aishwarya, D.; Babu, K. K.; Reddy, K. P. Rainfall Prediction Using Machine Learning Techniques. AIP Conf. Proc. 2024, 2850(1), 030004. https://doi.org/10.1063/5.0208435
Fauzi, F.; Kuswanto, H.; Atok, R. M. Bias Correction and Statistical Downscaling of Earth System Models Using Quantile Delta Mapping (QDM) and Bias Correction Constructed Analogues with Quantile Mapping Reordering (BCCAQ). In J. Phys.: Conf. Ser.; 2020; 1538, p 012050. https://doi.org/10.1088/1742-6596/1538/1/012050
Soleh, A. M.; Wigena, A. H.; Djuraidah, A.; Saefuddin, A. Statistical Downscaling to Predict Monthly Rainfall Using Linear Regression with L1 Regularization (LASSO). Appl. Math. Sci. 2015, 9(105–108), 5361–5369. https://doi.org/10.12988/ams.2015.56434
Rachmawati, R. N.; Sungkawa, I.; Rahayu, A. Extreme Rainfall Prediction Using Bayesian Quantile Regression in Statistical Downscaling Modeling. Procedia Comput. Sci. 2019, 157, 406–413. https://doi.org/10.1016/j.procs.2019.08.232
Sadeghfam, S.; Khatibi, R.; Moradian, T. Statistical Downscaling of Precipitation Using Inclusive Multiple Modelling (IMM) at Two Levels. J. Water Clim. Change 2021, 12(7), 3373–3387.
Zhang, X.; Shen, M.; Chen, J.; Homan, J. W.; Busteed, P. R. Evaluation of Statistical Downscaling Methods for Simulating Daily Precipitation Distribution, Frequency, and Temporal Sequence. Trans. ASABE 2021, 64(3), 771–784. https://doi.org/10.13031/trans.14097
Kuswanto, H.; Kravitz, B.; Miftahurrohmah, B.; Fauzi, F.; Sopahaluwaken, A.; Moore, J. Impact of Solar Geoengineering on Temperatures over the Indonesian Maritime Continent. Int. J. Climatol. 2021, 41, 1–20. https://doi.org/10.1002/joc.7391
Dewanti, D.; Djuraidah, A.; Sartono, B.; Sopaheluwakan, A. Bias Correction and Ensemble Techniques in Statistical Downscaling Model for Rainfall Prediction Using Tweedie-LASSO in West Java, Indonesia. J. Agrometeorol. 2024, 26(3), 324–330. https://doi.org/10.54386/jam.v26i3.2614
Adam, F. F.; Kurnia, A.; Purnaba, I. G.; Mangku, I. W.; Soleh, A. M. Tweedie Compound Poisson Model with First Order Autoregressive Time Random Effect. Commun. Math. Biol. Neurosci. 2020, 2020, 1–16. https://doi.org/10.28919/cmbn/4960
Dunn, P. K. Occurrence and Quantity of Precipitation Can Be Modelled Simultaneously. Int. J. Climatol. 2004, 24(10), 1231–1239. https://doi.org/10.1002/joc.1063
Hayati, M.; Permatasari, R. Comparison of Generalized Linear Model between Gamma and Tweedie Compound Response for Rainfall Prediction in Lampung Province. Asian J. Probab. Stat. 2024, 26(1), 41–49. https://doi.org/10.9734/ajpas/2024/v26i1583
Hayati, M.; Wigena, A. H.; Djuraidah, A.; Kurnia, A. A New Approach to Statistical Downscaling Using Tweedie Compound Poisson Gamma Response and Lasso Regularization. Commun. Math. Biol. Neurosci. 2021, 2021. https://doi.org/10.28919/cmbn/5936
Halder, A.; Mohammed, S.; Dey, D. K. Bayesian Variable Selection in Double Generalized Linear Tweedie Spatial Process Models. N. Engl. J. Stat. Data Sci. 2023, 187–199. https://doi.org/10.51387/23-nejsds37
Iranzad, R.; Liu, X.; Chaovalitwongse, W. A.; Hippe, D.; Wang, S.; Han, J.; Thammasorn, P.; Duan, C.; Zeng, J.; Bowen, S. Gradient Boosted Trees for Spatial Data and Its Application to Medical Imaging Data. IISE Trans. Healthc. Syst. Eng. 2022, 12(3), 165–179. https://doi.org/10.1080/24725579.2021.1995536
Zhou, H.; Qian, W.; Yang, Y. Tweedie Gradient Boosting for Extremely Unbalanced Zero-Inflated Data. Commun. Stat. Simul. Comput. 2020, 1–23. https://doi.org/10.1080/03610918.2020.1772302
Yang, Y.; Qian, W.; Zou, H. Insurance Premium Prediction via Gradient Tree-Boosted Tweedie Compound Poisson Models. J. Bus. Econ. Stat. 2018, 36(3), 456–470. https://doi.org/10.1080/07350015.2016.1200981
Watanabe, S.; Hajima, T.; Sudo, K.; Nagashima, T.; Takemura, T.; Okajima, H.; Nozawa, T.; Kawase, H.; Abe, M.; Yokohata, T.; et al. MIROC-ESM 2010: Model Description and Basic Results of CMIP5-20c3m Experiments. Geosci. Model Dev. 2011, 4(4), 845–872. https://doi.org/10.5194/gmd-4-845-2011
Ji, D.; Wang, L.; Feng, J.; Wu, Q.; Cheng, H.; Zhang, Q.; Yang, J.; Dong, W.; Dai, Y.; Gong, D.; et al. Description and Basic Evaluation of Beijing Normal University Earth System Model (BNU-ESM) Version 1. Geosci. Model Dev. 2014, 7(5), 2039–2064. https://doi.org/10.5194/gmd-7-2039-2014
Nur’utami, M. N.; Hidayat, R. Influences of IOD and ENSO to Indonesian Rainfall Variability: Role of Atmosphere–Ocean Interaction in the Indo-Pacific Sector. Procedia Environ. Sci. 2016, 33, 196–203. https://doi.org/10.1016/j.proenv.2016.03.070
Hidayat, R.; Ando, K.; Masumoto, Y.; Luo, J. J. Interannual Variability of Rainfall over Indonesia: Impacts of ENSO and IOD and Their Predictability. IOP Conf. Ser.: Earth Environ. Sci. 2016, 31(1), 012043. https://doi.org/10.1088/1755-1315/31/1/012043
Dunn, P. K.; Smyth, G. K. Generalized Linear Models with Examples in R, 2nd ed.; Springer: New York, 2018. https://doi.org/10.1007/978-1-4419-0118-7
Dunn, P. K.; Smyth, G. K. Series Evaluation of Tweedie Exponential Dispersion Model Densities. Stat. Comput. 2005, 15(4), 267–280. https://doi.org/10.1007/s11222-005-4070-y
Monego, V. S.; Anochi, J. A.; de Campos Velho, H. F. South America Seasonal Precipitation Prediction by Gradient-Boosting Machine-Learning Approach. Atmosphere 2022, 13(2), 243. https://doi.org/10.3390/atmos13020243
Anwar, M. T.; Winarno, E.; Hadikurniawati, W.; Novita, M. Rainfall Prediction Using Extreme Gradient Boosting. In Journal of Physics: Conference Series; IOP Publishing Ltd, 2021; 1869, p 012078. https://doi.org/10.1088/1742-6596/1869/1/012078
Dunn, P. K.; Smyth, G. K. Series Evaluation of Tweedie Exponential Dispersion Model Densities. Stat. Comput. 2005, 15(4), 267–280. https://doi.org/10.1007/s11222-005-4070-y
Dunn, P. K.; Smyth, G. K. Generalized Linear Models with Examples in R, 2nd ed.; Springer: New York, 2018. https://doi.org/10.1007/978-1-4419-0118-7
Hidayat, R.; Ando, K.; Masumoto, Y.; Luo, J. J. Interannual Variability of Rainfall over Indonesia: Impacts of ENSO and IOD and Their Predictability. In IOP Conference Series: Earth and Environmental Science; 2016; 31, p 012043. https://doi.org/10.1088/1755-1315/31/1/012043
Nur’utami, M. N.; Hidayat, R. Influences of IOD and ENSO to Indonesian Rainfall Variability: Role of Atmosphere–Ocean Interaction in the Indo-Pacific Sector. Procedia Environ. Sci. 2016, 33, 196–203. https://doi.org/10.1016/j.proenv.2016.03.070
Ji, D.; Wang, L.; Feng, J.; Wu, Q.; Cheng, H.; Zhang, Q.; Yang, J.; Dong, W.; Dai, Y.; Gong, D.; et al. Description and Basic Evaluation of Beijing Normal University Earth System Model (BNU-ESM) Version 1. Geosci. Model Dev. 2014, 7(5), 2039–2064.
https://doi.org/10.5194/gmd-7-2039-2014
Watanabe, M.; Kakuta, H. Retinoid X Receptor Antagonists. Int. J. Mol. Sci. 2018, 19(8), 2354. https://doi.org/10.3390/ijms19082354
Yang, Y.; Qian, W.; Zou, H. Insurance Premium Prediction via Gradient Tree-Boosted Tweedie Compound Poisson Models. J. Bus. Econ. Stat. 2018, 36(3), 456–470. https://doi.org/10.1080/07350015.2016.1200981
Zhou, H.; Qian, W.; Yang, Y. Tweedie Gradient Boosting for Extremely Unbalanced Zero-Inflated Data. Commun. Stat. Simul. Comput. 2020, 1–23. https://doi.org/10.1080/03610918.2020.1772302
Iranzad, R.; Liu, X.; Chaovalitwongse, W. A.; Hippe, D.; Wang, S.; Han, J.; Thammasorn, P.; Duan, C.; Zeng, J.; Bowen, S. Gradient Boosted Trees for Spatial Data and Its Application to Medical Imaging Data. IISE Trans. Healthc. Syst. Eng. 2022, 12(3), 165–179. https://doi.org/10.1080/24725579.2021.1995536
Halder, A.; Mohammed, S.; Dey, D. K. Bayesian Variable Selection in Double Generalized Linear Tweedie Spatial Process Models. N. Engl. J. Stat. Data Sci. 2023, 187–199. https://doi.org/10.51387/23-nejsds37
Hayati, M.; Wigena, A. H.; Djuraidah, A.; Kurnia, A. A New Approach to Statistical Downscaling Using Tweedie Compound Poisson Gamma Response and Lasso Regularization. Commun. Math. Biol. Neurosci. 2021, 2021, Article ID 2021. https://doi.org/10.28919/cmbn/5936
Hayati, M.; Permatasari, R. Comparison of Generalized Linear Model between Gamma and Tweedie Compound Response for Rainfall Prediction in Lampung Province. Asian J. Probab. Stat. 2024, 26(1), 41–49. https://doi.org/10.9734/ajpas/2024/v26i1583
Dunn, P. K. Occurrence and Quantity of Precipitation Can Be Modelled Simultaneously. Int. J. Climatol. 2004, 24(10), 1231–1239. https://doi.org/10.1002/joc.1063
Adam, F. F.; Kurnia, A.; Purnaba, I. G.; Mangku, I. W.; Soleh, A. M. Tweedie Compound Poisson Model with First Order Autoregressive Time Random Effect. Commun. Math. Biol. Neurosci. 2020, 2020, 1–16. https://doi.org/10.28919/cmbn/4960
Dewanti, D.; Djuraidah, A.; Sartono, B.; Sopaheluwakan, A. Bias Correction and Ensemble Techniques in Statistical Downscaling Model for Rainfall Prediction Using Tweedie-LASSO in West Java, Indonesia. J. Agrometeorol. 2024, 26(3), 324–330. https://doi.org/10.54386/jam.v26i3.2614
Kuswanto, H.; Kravitz, B.; Miftahurrohmah, B.; Fauzi, F.; Sopahaluwaken, A.; Moore, J. Impact of Solar Geoengineering on Temperatures over the Indonesian Maritime Continent. Int. J. Climatol. 2021, 41, 1–20. https://doi.org/10.1002/joc.7391
Zhang, X.; Shen, M.; Chen, J.; Homan, J. W.; Busteed, P. R. Evaluation of Statistical Downscaling Methods for Simulating Daily Precipitation Distribution, Frequency, and Temporal Sequence. Trans. ASABE 2021, 64(3), 771–784. https://doi.org/10.13031/trans.14097
Sadeghfam, S.; Khatibi, R.; Moradian, T. Statistical Downscaling of Precipitation Using Inclusive Multiple Modelling (IMM) at Two Levels. J. Water Clim. Change 2021, 12(7), 3373–3387.
Rachmawati, R. N.; Sungkawa, I.; Rahayu, A. Extreme Rainfall Prediction Using Bayesian Quantile Regression in Statistical Downscaling Modeling. Procedia Comput. Sci. 2019, 157, 406–413. https://doi.org/10.1016/j.procs.2019.08.232
Soleh, A. M.; Wigena, A. H.; Djuraidah, A.; Saefuddin, A. Statistical Downscaling to Predict Monthly Rainfall Using Linear Regression with L1 Regularization (LASSO). Appl. Math. Sci. 2015, 9(105–108), 5361–5369. https://doi.org/10.12988/ams.2015.56434
Fauzi, F.; Kuswanto, H.; Atok, R. M. Bias Correction and Statistical Downscaling of Earth System Models Using Quantile Delta Mapping (QDM) and Bias Correction Constructed Analogues with Quantile Mapping Reordering (BCCAQ). In Journal of Physics: Conference Series; 2020; 1538, p 012050. https://doi.org/10.1088/1742-6596/1538/1/012050
Shabu, S. L. J.; Refonaa, J.; Devi, D.; Aishwarya, D.; Babu, K. K.; Reddy, K. P. Rainfall Prediction Using Machine Learning Techniques. AIP Conf. Proc. 2024, 2850(1), 030004. https://doi.org/10.1063/5.0208435
Fauzi, F.; Kuswanto, H.; Atok, R. M.; Salamah, M. Statistical Downscaling and Bias Correction of the Earth System Models Outputs for Future Climate Projection under Solar Geoengineering: Case Study Indonesia. J. Mech. Eng. 2021, 10(1), 33–49.
