EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Bias-Corrected ESM–Driven Rainfall Downscaling via Tweedie Compound Poisson–Gamma and LASSO for Local Climate Projection

fatkhurokhman fauzi, Anik Djuraidah, Anang Kurnia, Rochdi Wasono

Malaysian Journal of Fundamental and Applied Sciences · 31.08.2026 · 22 (4) · 731-748

DOI: 10.11113/mjfas.v22n4.4832 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Malaysian Journal of Fundamental and Applied Sciences, online erschienen am 31.08.2026. Verfasst von fatkhurokhman fauzi und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Climate change has increased rainfall variability and intensified the need for reliable local rainfall projections, particularly in agricultural regions such as West Java. Earth System Models (ESMs) provide useful climate change information but remain limited by coarse spatial resolution and systematic bias when applied to local-scale rainfall analysis. This study proposes an ESM-driven rainfall downscaling framework by integrating Tweedie Compound Poisson–Gamma with Least Absolute Shrinkage and Selection Operator (TCPG-LASSO), climate imprint-based resolution enhancement, and ISIMIP bias correction. Three ESMs, namely BNU-ESM, MIROC-ESM, and MIROC-ESM-CHEM, were evaluated using rainfall station data in West Java. Two modelling scenarios were compared: TCPG-LASSO without rescaling and bias correction and TCPG-LASSO with ESM rescaling and Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) bias correction. Model performance was assessed using Taylor diagrams and time series plots. The results show that West Java rainfall follows a unimodal seasonal pattern and has a non-negative, positively skewed distribution, supporting the use of the TCPG framework. LASSO effectively selected a limited number of relevant ESM grids and reduced multicollinearity. The corrected TCPG-LASSO model provided better agreement with observed rainfall than the uncorrected model, with an average correlation of about 0.7. Among the evaluated ESMs, BNU-ESM showed the best performance and was used for six-month Standardized Precipitation Index (SPI-6) projection during 2021–2030. Across the 805 valid station-month SPI-6 values from June 2021 to December 2030, normal conditions dominated, with 583 occurrences. Dry and wet anomalies were also projected, including 85 moderately dry, 15 severely dry, 2 extremely dry, 55 moderately wet, 47 very wet, and 18 extremely wet occurrences. These findings demonstrate that the proposed framework can support drought monitoring, agricultural planning, and hydrometeorological disaster mitigation in West Java.

Zitieren

APA

fauzi, F., Djuraidah, A., Kurnia, A., & Wasono, R. (2026). Bias-Corrected ESM–Driven Rainfall Downscaling via Tweedie Compound Poisson–Gamma and LASSO for Local Climate Projection. Malaysian Journal of Fundamental and Applied Sciences, 22(4), 731-748. https://doi.org/10.11113/mjfas.v22n4.4832

Vancouver

fauzi F, Djuraidah A, Kurnia A, Wasono R. Bias-Corrected ESM–Driven Rainfall Downscaling via Tweedie Compound Poisson–Gamma and LASSO for Local Climate Projection. Malaysian Journal of Fundamental and Applied Sciences. 2026;22(4):731-748. doi:10.11113/mjfas.v22n4.4832

BibTeX
@article{fauzi2026,
  author = {fauzi, fatkhurokhman and Djuraidah, Anik and Kurnia, Anang and Wasono, Rochdi},
  title = {Bias-Corrected ESM–Driven Rainfall Downscaling via Tweedie Compound Poisson–Gamma and LASSO for Local Climate Projection},
  journal = {Malaysian Journal of Fundamental and Applied Sciences},
  year = {2026},
  volume = {22},
  number = {4},
  pages = {731--748},
  doi = {10.11113/mjfas.v22n4.4832}
}
Kostenlos starten →