Cascading Climate Tipping Points and Compound Extremes: AI-Driven Earth System Digital Twins for Anticipatory Governance
Peter Makieu
*
Department of Agribusiness Management, School of Agriculture and Food Sciences, Njala University, Sierra Leone and School of Environmental Engineering, Suzhou University of Science and Technology, Jiangsu Province, China.
Mohamed Yansaneh
Department of Agribusiness Management, School of Agriculture and Food Sciences, Njala University, Sierra Leone and School of Environmental Engineering, Suzhou University of Science and Technology, Jiangsu Province, China.
Mohamed Jalloh
Department of Agribusiness Management, School of Agriculture and Food Sciences, Njala University, Sierra Leone and School of Environmental Engineering, Suzhou University of Science and Technology, Jiangsu Province, China.
Edison D. Dartue
School of Environmental Engineering, Suzhou University of Science and Technology, Jiangsu Province, China.
Fatmata Dankay Kamara
School of Environmental Engineering, Suzhou University of Science and Technology, Jiangsu Province, China.
Mitchell Vampelt
School of Environmental Engineering, Suzhou University of Science and Technology, Jiangsu Province, China.
*Author to whom correspondence should be addressed.
Abstract
Cascading climate tipping points and compound extreme events pose urgent and under-characterized threats to Earth system stability. Multiple hazards occurring simultaneously or in rapid succession can reduce effective tipping thresholds by 15–30% compared to single-stressor exposures, narrowing safety margins implied by international temperature targets. Traditional Earth System Models (ESMs) are limited in capturing nonlinear feedbacks and threshold-crossing dynamics. AI-enabled Earth System Digital Twins (ESDTs), data-assimilative, high-resolution, and continuously updated, offer transformative potential to anticipate these risks in near real time.
This systematic review, conducted following PRISMA 2020 guidelines and registered in PROSPERO, synthesizes 287 peer-reviewed studies (2015–2024) across four thematic pillars: (i) cascading climate tipping points, (ii) compound extremes as tipping accelerators, (iii) AI-driven ESDTs, and (iv) anticipatory governance frameworks. At least five tipping elements—including the Greenland Ice Sheet, West Antarctic Ice Sheet, and tropical coral reefs—approach critical thresholds within the Paris Agreement target range, with compound extremes amplifying risks nonlinearly. Current ESDTs lack explicit tipping cascade modules, interpretability standards, and equitable access.
We introduce CASCADE-AI (Cascading Assessment of Systemic Climate Abrupt Dynamics through Artificial Intelligence), linking tipping cascade characterization, compound extreme assessment, physics-AI hybrid digital twin architecture, and governance interfaces. Two critical integration gaps are identified: between compound extreme attribution and tipping characterization, and between ESDT outputs and governance application. Ten prioritized research actions are proposed across near (2025–2028), medium (2028–2032), and long-term (2032–2040+) horizons.
Keywords: Compound extreme events, physics-informed neural networks, CASCADE-AI framework, nonlinear climate dynamics, early warning signals, decision-making under deep uncertainty, climate risk management