Evaluating the necessity of post-processing techniques on d4PDF data for extreme climate assessment

Area of Science:

• Climate Science
• Hydrology
• Environmental Modeling

Background:

• Global extreme precipitation events are increasing in frequency and severity.
• Existing climate models often have coarse resolution, leading to significant bias errors in climate change assessments.
• Accurate simulation of extreme climatic flood events is crucial for policy decisions.

Purpose of the Study:

• To adopt post-processing techniques (interpolation and bias correction) on the d4PDF model database.
• To improve the simulation accuracy of extreme climatic flood events in the Chao Phraya River Basin under future climate scenarios (+4 K).
• To enhance the robustness of climate change impact assessments.

Main Methods:

• Applied gradient plus inverse distance squared interpolation due to limited rain gauges.
• Adjusted bias correction methods with monthly and seasonal periods.
• Utilized gamma distribution combined with generalized Pareto distribution for rainy season bias correction, and gamma distribution for dry season.

Main Results:

• The proposed post-processing method effectively simulated extreme rainfall events, extended dry periods, and intensified rainfall during the rainy season.
• Post-processed d4PDF trends showed significant variations in annual rainfall across different sea surface temperature patterns and ensemble members.
• Highest annual rainfall recorded at 4,450 mm/year (Nan River) and lowest at 710 mm/year (lower Chao Phraya River Basin).

Conclusions:

• Post-processing techniques are vital for improving the accuracy and robustness of climate model outputs, particularly the d4PDF model.
• The study highlights the need for further investigation into rainfall pattern variances among ensembles for future climate change impact studies.
• Findings provide novel insights for more reliable climate change impact assessments in river basins.