Can weather generation capture precipitation patterns across different climates, spatial scales and under data...

Area of Science:

• Earth Sciences
• Ecology
• Climate Research

Background:

• Stochastic weather generators create long-term weather data essential for scientific research.
• Previous studies on these generators were limited in scope, scale, and accessibility.
• Multi-site algorithms were often not publicly available, hindering reproducibility.

Purpose of the Study:

• To evaluate the performance of the TripleM multi-site precipitation generator.
• To assess generator performance across diverse US climatic regions.
• To investigate the impact of gauge network extent and density on generator accuracy.

Main Methods:

• Tested the TripleM precipitation generator in three distinct US climate zones.
• Resampled observational data to analyze network extent and density effects.
• Evaluated performance using nine statistical metrics, focusing on precipitation variability and spell lengths.

Main Results:

• TripleM performed more accurately in wet temperate climates than in drier regions.
• Larger spatial scales showed performance degradation due to increased climate heterogeneity.
• Network density was identified as a key factor for applicability in data-scarce areas.

Conclusions:

• The TripleM generator's effectiveness is climate-dependent and influenced by spatial scale.
• Network density significantly impacts performance, providing guidance for data-poor contexts.
• This research enhances understanding of stochastic weather generator limitations and applications.