Wanniarachchi et al. (2025) Enhancing the effectiveness of satellite precipitation products with topographic and seasonal bias correction

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-12-01
• Authors: Susantha Wanniarachchi, Ranjan Sarukkalige, Hapu Arachchige Prasantha Hapuarachchi, Pattiyage I. A. Gomes, Upaka Rathnayake
• DOI: 10.1016/j.jhydrol.2025.134688

Research Groups

• School of Civil and Mechanical Engineering, Curtin University, Western Australia, Australia
• Department of Civil Engineering, Sri Lanka Institute of Information Technology, Malabe, Sri Lanka
• The Bureau of Meteorology, Australia
• Department of Civil Engineering and Construction, Atlantic Technological University, Sligo, Ireland

Short Summary

This study introduces the Heavy Rain Peak Adjustment (HRPA) method for satellite precipitation bias correction, comparing its effectiveness against the Seasonal Autoregressive Integrated Moving Average (SARIMA) model. The HRPA method significantly enhances the accuracy of Global Satellite Mapping of Precipitation-Near-Real-Time (GSMaP-NRT) data, particularly for heavy precipitation events and at lower elevations, outperforming SARIMA in reducing errors and improving correlation with observed data.

Objective

• To introduce and evaluate the Heavy Rain Peak Adjustment (HRPA) method alongside the Seasonal Autoregressive Integrated Moving Average (SARIMA) model for satellite precipitation bias correction.
• To assess the applicability of hourly resolution Global Satellite Mapping of Precipitation-Near-Real-Time (GSMaP-NRT) data with topographic and seasonal bias correction in a flood-prone river region in Australia.

Study Configuration

• Spatial Scale: Ovens River region, Australia, covering an area of 6239 km². Analysis conducted at 31 rain gauge stations and using 0.1° × 0.1° grid resolution for satellite data.
• Temporal Scale: Hourly precipitation records from 2007 to 2017 (11 years).

Methodology and Data

• Models used: Heavy Rain Peak Adjustment (HRPA) method, Seasonal Autoregressive Integrated Moving Average (SARIMA) model.
• Data sources:
  • Global Satellite Mapping of Precipitation-Near-Real-Time (GSMaP-NRT) data (Japan Aerospace Exploration Agency).
  • Hourly precipitation records from 31 rain gauges (Bureau of Meteorology, Australia).
  • Geographic data: Geofabric National dataset (version 3.3), 1-second SRTM-derived Digital Elevation Model (DEM) from Geoscience Australia.

Main Results

• The mean residual between observed and HRPA-corrected GSMaP-NRT precipitation was -0.02 mm, indicating substantial accuracy improvement.
• The HRPA method consistently outperformed SARIMA across all objective functions and elevation classes. For HRPA, average performance metrics included a linear regression R² of 0.911, NSE (log) of -0.847, and RMSE of 0.628.
• HRPA showed superior performance, particularly at lower elevations, while SARIMA struggled significantly at higher elevations, often yielding negative R², NSE, NSE(log), and KGE values.
• For intermediate-elevation station 82162, HRPA demonstrated a 96% increase in NSE, a 107% improvement in KGE, and a 64% reduction in RMSE compared to SARIMA.
• Autocorrelation and partial autocorrelation plots for hilly terrain stations showed significant wave-like patterns at 24-hour and 48-hour lags, indicating greater uncertainty and systematic influence of past residuals in satellite precipitation estimates over complex terrain.
• Precipitation bias correction factors (Fc,i) for heavy precipitation events (>10 mm/h) varied: approximately 5.3 for high-flow/low-flow winter events, 5.3 to 9.4 for summer events, and 1.8 to 2.5 for intermediate-flow events.

Contributions

• Introduces and validates the novel Heavy Rain Peak Adjustment (HRPA) method, demonstrating its superior effectiveness over the SARIMA model for correcting biases in satellite precipitation data.
• Provides a practical and robust approach for enhancing the reliability of near-real-time satellite precipitation products, particularly for heavy rainfall events and in regions with complex topography.
• Addresses the critical challenge of underestimation of heavy precipitation and topographic effects in satellite-based estimates, which is crucial for accurate flood forecasting in data-scarce areas.
• Offers a method that can be used to predict actual precipitation from GSMaP-NRT data with further refinements, aiding in near-real-time flood peak estimation.

Funding

The authors acknowledge support from the Bureau of Meteorology of Australia and the Japan Aerospace Exploration Agency for providing weather data. No specific funding projects or programs were explicitly listed.

Citation

@article{Wanniarachchi2025Enhancing,
  author = {Wanniarachchi, Susantha and Sarukkalige, Ranjan and Hapuarachchi, Hapu Arachchige Prasantha and Gomes, Pattiyage I. A. and Rathnayake, Upaka},
  title = {Enhancing the effectiveness of satellite precipitation products with topographic and seasonal bias correction},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134688},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134688}
}