Attributing Extremes in Weather

This project investigated the predictive potential of climate indices in explaining extreme weather outcomes across multiple meteorological variables, such as temperature, wind, and humidity. A climate index is a quantitative measure that summarizes key aspects of the climate system, such as sea surface temperature anomalies or atmospheric pressure patterns, and is often used to study and predict large-scale climate phenomena (e.g., El Niño-Southern Oscillation, North Atlantic Oscillation). The goal was to determine whether these indices could serve as reliable predictors for extreme weather events, contributing to improved long-range weather forecasting and climate research.

Technologies and Tools

• Duration: January 2023 – June 2024 (6 Months)
• Technologies and Tools: Python, Bash, Scipy (Pearson correlations), Matplotlib (visualizations), Iris (correlation significance visualizations), Jupyter Notebooks
• Data Sources: ERA5 reanalysis weather data, open-source climate indices from climexp.knmi.nl

Key Objectives

1. Correlation Analysis:
   • Analyzed correlations between open-source climate indices (from climexp.knmi.nl) and ERA5 reanalysis weather data across various lead times.
   • Investigated the ability of climate indices to explain extreme weather events, defined as weather variables exceeding their standard deviation at a select value.
2. Geo-Spatial Considerations:
   • Accounted for multiple testing and spatial dependencies in the data to ensure statistically robust results.
   • Developed tools for large-scale spatiotemporal analysis, enabling the visualization and interpretation of correlations across different regions.
3. Predictive Potential:
   • Assessed the utility of climate indices in predicting meteorological parameters, contributing to advancements in long-range weather forecasting and global food security.

Results and Insights

The project involved analyzing large-scale climate datasets to identify potential relationships between climate indices and extreme weather events. Statistical methods, including Pearson correlations, were used to quantify these relationships, while geo-spatial tools were employed to account for spatial dependencies and multiple testing. Visualizations were created to interpret the results and communicate findings effectively.

The findings from this investigation contributed to a broader understanding of the relationships between climate indices and extreme weather outcomes. By developing tools for spatiotemporal analysis and exploring the predictive potential of climate indices, the project supported ongoing efforts to improve long-range weather forecasting and enhance global food security. The methodologies and insights gained can be adapted for future research on climate variability and its impacts.

Conclusion

This project explored the use of climate indices as predictors for extreme weather events, employing robust statistical methods and geo-spatial analysis to investigate potential relationships. While the results provided valuable insights, the complexity of spatiotemporal correlations leaves room for further exploration and interpretation. The tools and methodologies developed during the project contribute to ongoing climate research and provide a foundation for future investigations into the predictive potential of climate indices.