1. Data‐Driven Predictions of Peak Warming Under Rapid Decarbonization.
- Author
-
Diffenbaugh, Noah S. and Barnes, Elizabeth A.
- Subjects
- *
CLIMATE change models , *GLOBAL warming , *CARBON emissions , *CONVOLUTIONAL neural networks ,PARIS Agreement (2016) - Abstract
The severe impacts associated with recent record‐setting annual global temperatures elevate the need to accurately predict the hottest conditions that could occur even if the most ambitious decarbonization goals are achieved. We use convolutional neural networks (CNNs) to predict peak global warming from recent observed temperature maps and future cumulative CO2 emissions. For the SSP1‐1.9 decarbonization scenario there is >99% probability that mean global warming exceeds 1.5°C, approximately even odds that it reaches 2°C, and ∼90% probability that the hottest year globally exceeds 2023 by at least 0.5°C. Further, for the SSP2‐4.5 decarbonization scenario, there is >90% probability that the hottest annual global temperature anomaly is twice the 2023 anomaly. That our framework makes highly accurate out‐of‐sample predictions of the hottest historical year provides confidence in the predicted future probabilities, suggesting substantial risks from the extreme local conditions that are likely to result from globally hot years during rapid decarbonization. Plain Language Summary: Calendar year 2023 was the hottest year on record globally, reaching ∼1.5°C above the pre‐industrial. Many national, sub‐national and non‐state actors have articulated ambitious decarbonization goals to stabilize the global temperature. However, the intensifying impacts as individual years have approached 1.5°C have heightened the need to more accurately predict not just the mean warming but also the hottest years that could occur even in the context of rapid decarbonization. We train neural networks on an ensemble of global climate models and then use historical observations as input to the trained networks, thus constraining the uncertainty in climate model projections by using the current state of the climate system as the basis for a truly out‐of‐sample prediction. For the historical period, we find that, despite a wide range of climate sensitivities across global climate models, the neural networks make highly accurate predictions of the hottest historical year when given observed climate patterns as out‐of‐sample inputs. Predicting future warming under different cumulative emissions, we find that even if net‐zero emissions are achieved mid‐century, mean warming is virtually certain to exceed 1.5°C and has even odds of reaching 2°C, with high likelihood of individual years that are at least 0.5°C hotter than 2023. Key Points: We train CNNs to predict peak global warming given the map of recent annual temperatures and total additional CO2 emissionsEven if net‐zero emissions are reached mid‐century, mean warming is virtually certain to exceed 1.5°C, with even odds of 2°CThere is high likelihood of individual years that are at least 0.5°C hotter than 2023 even in the most ambitious decarbonization scenario [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF