The Pattern of Climate Change

As the Earth warms due to anthropogenic emissions, some regions warm more than others. This includes well-known and understood phenomena such as polar amplified warming and enhanced warming over land. However, how the pattern of sea-surface temperatures (SSTs) will change is more uncertain. One reason for this uncertainty is that the pattern of SST change observed over the past 40-60 years is quite different from what most climate models project would happen due to anthropogenic forcing over this time period, with more warming in the West Pacific and Indian Ocean but less warming or even regional cooling in parts of the East Pacific. In contrast, models suggest that the East Pacific should be a region of amplified warming, and our work has shown that models are unable to reproduce the observed SST trend pattern, even when accounting for decadal climate variability (Download Wills et al. 2022). The discrepancy in the recent warming pattern between models and observations has strengthened the Walker circulation and modified regional climate around the Pacific basin, and it has also reduced the effective climate sensitivity and global warming rate over this time period by increasing low-cloud cover (Download Armour et al. 2024). Our group’s ongoing research aims to understand the contributions of greenhouse-gas, anthropogenic-aerosol, and volcanic forcing and internal variability to observed SST pattern trends and to understand how climate models can be improved to better capture these trends.

Our group’s work to understand the contributions to the observed recent warming pattern builds in particular on our development of pattern recognition methods to isolate the forced response in observations (Download Wills et al. 2020). This work is discussed in more detail in the Climate Data Science page, including information about the Forced Component Estimation Statistical Methods Intercomparison Project (ForceSMIP).

Our group’s work to understand how to improve climate models to better capture observed trends focuses on how long-standing biases in the mean-state of climate models, such as the double-ITCZ bias, cold-tongue bias, and biases in tropical evaporation, influence their simulation of SST trends. We also investigate how increasing climate model resolution can alleviate these long-standing mean-state biases, as discussed further in the High-Resolution Climate Modeling page.