The group is broadly interested in the understanding and model simulation of climate change and variability. Our publications cover subjects such as anthropogenic changes in hydroclimate and atmospheric circulation, global and regional uncertainties in climate projections and high-resolution climate modeling. One of the current research projects focuses on tropical air-sea coupling. The goal is to quantify various coupling feedback processes and to build a simple and practical framework for modeling air-sea interactions. We are also working on the connection between transient climate sensitivity and regional ocean heat uptake.
Tropical Air-sea Interaction
|Point-wise regression coefficients of monthly precipitation anomalies onto monthly SST anomalies from a comprehensive coupled climate model (top) and a simple linear model derived from uncoupled atmosphere-only models (bottom).|
The interaction between the atmosphere and the ocean is inherently a two-way coupling. On the one hand, sea surface temperature (SST) anomalies are known to induce large variations in the atmosphere, particularly in the tropics. On the other hand, atmospheric variability can also be internally driven by atmospheric intrinsic dynamics, which in turn influences the ocean. Because the atmospheric intrinsic variability is essentially chaotic, it cannot be extracted from a fully coupled system. The inability to separate the ocean-induced and atmospheric intrinsic variability has hindered the quantitative understanding of air-sea interactions and has created a fundamental challenge in interpreting the observed air-sea relationships.
The group has developed a new framework able to cleanly separate and quantify the ocean-induced and atmospheric intrinsic variability. This was accomplished by using uncoupled atmosphere-only simulations forced with different underlying SSTs. By enabling the quantification of individual processes in air-sea coupling, the new framework could be an important step toward the development of process-oriented diagnostics for understanding the observed and simulated air-sea relationships.
Find out more in He et. al. [2018, Journal of Climate].
Subtropical Precipitation Changes
|Time evolution of global mean surface warming (left) and zonal mean precipitation changes (right) from abrupt 4xCO2 simulations. Note the slow global warming and the rapid precipitation decline at approximately 30o latitude.|
The subtropics encompass many of the world's driest human habitats, and climate models ubiquitously project a future decline in subtropical precipitation. This precipitation decline has been widely attributed to either the increase in atmospheric moisture or the poleward expansion of the Hadley cell - both mechanisms are associated with the slow warming of the global surface temperature. However, by abruptly increasing CO2, we found that the subtropical precipitation decline forms much faster than the global surface warming and is therefore unrelated to the previously proposed mechanisms. Then what actually causes the subtropical precipitation decline?
Find out more in He and Soden [2017, Nature Climate Change].
Climate Sensitivity and Ocean Circulation
|Time evolution of global mean surface warming from 1%CO2 simulations branching off from a pre-industrial control simulation (1860-start, blue) and a present-day control simulation (1990-start, red) and the difference between the two (black).|
Climate models show large uncertainty in the simulation of transient climate sensitivity. Much of this uncertainty could result from biases in the simulation of the base climate. However, because models differ in multiple ways, it is often difficult to narrow the uncertainty to specific dynamical processes. In this study, we compared two simulations with the same model but in which CO2 is increased from either a pre-industrial or a present-day control simulation. This allowed different base climate ocean circulations that are representative of those in current climate models to be imposed upon a single model. As a result, the model projects different transient climate sensitivities that are comparable to the multi-model spread. What can we learn from these simulations?
Find out more in He et al. [2016, Journal of Climate].
Tropical Circulation Changes
|Percentage changes in tropical mean precipitation (P), surface moisture (q) and convection (M) from fully coupled (top), uncoupled CO2 only (bottom left) and uncoupled warming only (bottom right) simulations. Note the negative changes from M in both CO2 only and warming only simulations.|
As the climate warms, the tropical overturning circulation slows down. It is commonly believed that the circulation weakening is largely associated with the warming. However, by increasing CO2 in an uncoupled simulation with fixed surface temperature, we found that substantial weakening also occurs. So what exactly causes the weakening of tropical circulation?
Reginal Climate Changes and Sea Surface Warming Patterns
|Land precipitation change from a fully coupled simulation (left) and an uncoupled simulation with only a spatially uniform warming (right). Note the similarity between the two.|
Studies found that the pattern of anthropogenic changes in sea surface temperature (SST) plays a dominant role in circulation and precipitation changes over tropical oceans. It is expected that this pattern of SST changes should also have large influences on climate changes over land. This expectation is based on the fact that natural variations in tropical SST (e.g., ENSO) often have substantial remote climatic impact. However, we found that land climate changes are actually quite similar with and without the projected pattern of anthropogenic SST changes. Why does climate respond differently to natural and anthropogenic variations in SST patterns?
Find out more in He et al. [2014, Geophysical Research Letters].