Development of probability density functions for future South American rainfall.

Area of Science:

• Climate science
• Hydrology
• Ecology

Background:

• Accurate projections of future rainfall are crucial for understanding climate change impacts.
• General Circulation Models (GCMs) offer insights but have inherent biases and varying performance.
• South America's diverse climate necessitates region-specific rainfall assessments.

Purpose of the Study:

• To estimate probability density functions (PDFs) for future rainfall in five South American regions.
• To systematically weight predictions from 24 Coupled Model Intercomparison Project Phase 3 (CMIP3) GCMs based on performance.
• To assess the impact of model weighting on rainfall PDFs and subsequent forest biomass simulations.

Main Methods:

• Weighting GCM predictions using Bayes' theorem to sequentially update model performance.
• Rating GCMs based on their ability to reproduce inter-annual variability in seasonal rainfall.
• Applying derived model weightings to simulate forest biomass using the Lund-Potsdam-Jena Dynamic Global Vegetation Model (LPJmL).

Main Results:

• GCM performance rankings varied significantly across different seasons and regions.
• No single GCM consistently outperformed others in all assessed areas.
• Differential model weighting led to notable shifts in the estimated rainfall PDFs for specific regions and seasons.

Conclusions:

• A weighted ensemble approach improves rainfall PDF estimation compared to unweighted models.
• Region- and season-specific model weightings are essential for accurate climate projections.
• The derived weightings provide a robust basis for simulating climate-driven ecological changes, such as forest biomass dynamics.