TOPOCLIM-EWS: Topological Insights for Hydroclimatic Extremes and Early Warnings Signals

Objective

TOPOCLIM-EWS aims to develop a next-generation framework for detecting early warning signals of hydroclimatic extremes by applying and advancing Topological Data Analysis (TDA) to climate time series and spatiotemporal datasets. The project addresses a pressing need in the Mediterranean basin—one of the world’s most vulnerable regions—where existing methods struggle to capture the complex and nonlinear dynamics that drive droughts and floods. The fellowship introduces methodological innovations, including the development of two-parameter CROCKER plots to analyze the joint evolution of climate variables and detect bifurcations, and the use of the Mapper algorithm to reveal multi-scale structural patterns and regime shifts hidden to conventional approaches. Raw climate observations will be transformed into robust topological descriptors, with a strong emphasis on interoperability: these features will be designed to integrate seamlessly with established machine learning frameworks, ensuring enhanced predictive pipelines and extended forecast lead times. The research is structured around a synergistic transatlantic collaboration. The outgoing phase at Michigan State University (USA) provides expertise in applied topology and methodological innovation, while the return phase at Middle East Technical University (Türkiye, Horizon Europe Associated Country) offers a strong application environment in climate science, hydroclimatic datasets, and stakeholder engagement. By bridging mathematical theory and applied climatology, TOPOCLIM-EWS will deliver novel, open, and transferable early-warning indicators that can strengthen operational systems, supporting Horizon Europe objectives on climate resilience, disaster risk reduction, and science-based solutions for societal challenges.