Abstract
Understanding how marine heatwaves (MHWs) unfold in space and time under anthropogenic climate change is key to anticipate future impacts on ecosystems and society. Yet, our knowledge of the spatiotemporal dynamics of MHWs is very limited. Here, I combine network theory with topological data analysis and event synchronization to high-resolution satellite data and to a set of Earth System Model simulations to reveal the dynamical organization of complex MHW networks. The analysis reveals that MHWs have already crossed a tipping point separating highly synchronized preindustrial MHWs from the more extreme, but less coherent warming events we experience today. This loose spatiotemporal organization persists under a reduced RCP 2.6 emission scenario, whereas a second abrupt transition towards a permanent state of highly synchronized MHWs is foreseen by 2075 under a business-as-usual RCP 8.5 scenario. These results highlight the risks of abrupt ocean transitions, which may dramatically affect marine life and humanity by eroding valuable time for adaptation to climate change.
Highlights
Understanding how marine heatwaves (MHWs) unfold in space and time under anthropogenic climate change is key to anticipate future impacts on ecosystems and society
I have used a novel approach based on Topological Data Analysis (TDA) and the Mapper algorithm to implement a network analysis of observed and simulated MHWs under Historical, RCP 2.6 and RCP 8.5 emission scenarios
This condition of low connectivity characterizes the network of observed MHWs and persists throughout the simulation period in the RCP 2.6 network
Summary
Understanding how marine heatwaves (MHWs) unfold in space and time under anthropogenic climate change is key to anticipate future impacts on ecosystems and society. The full network encodes both spatial and temporal information of MHW events in a low-dimensional compressed representation produced by the clustering of timeframes within nodes. This approach allows for Scientific Reports | (2021) 11:1739. Great flexibility to explore spatiotemporal dynamics and transitions[22] and complements methodologies used to construct functional climate networks where nodes correspond to individual spatial locations[18] Because nodes encode both spatial and temporal information of MHW occurrences, prominent events can be extracted from TDA-based networks and their characteristics (e.g., duration, spatial extent) can be visualized as maps and time series (Supplementary Fig. 1). To navigate through the networks and to interactively explore their spatial and temporal properties, I developed a web tool that is available at: calcoloecologia.biologia.unipi.it:3838/ MHW_App) based on R 4.0.229
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
