Causes of model dry and warm bias over the SGP and their impact on climate projections
Submitter:
Zhang, Minghua — Stony Brook University
Area of research:
General Circulation and Single Column Models/Parameterizations
Journal Reference:
Science
To reduce uncertainties in the prediction of future climate change by numerical models, scientists strive to understand and reduce the model biases in the simulations of the present-day climate relative to observations. By using long-term measurements from the DOE ARM Climate Research Facility, the team identified the cause of a longstanding warm-model bias over the US Southern Great Plains as being due to deficient strong precipitation events that are otherwise followed by shallow convective clouds and less solar radiation to the surface. They also found that models with larger biases tend to project larger warming of future climate in this region, and that if the models are calibrated without the biases in the current climate, the average projected warming will be reduced by 20% and the model spread will be narrowed by 40%.
Impact
Because of the limitation of available computing powers, current climate models cannot explicitly resolve clouds and atmospheric convective systems. Instead, they use parameterizations to calculate the approximate effects of these systems. This study furthers the knowledge on how and why a significant model bias can be formed if the approximation fails to simulate the heavy precipitation events associated with organized convective systems. It advances the potential to narrow the uncertainties in predicting future climate change.
Summary
Climate models show a conspicuous summer warm and dry bias over the Southern Great Plains (SGP). Using results from 19 climate models in CMIP5 (the Coupled Model Intercomparison Project Phase 5), the team found a robust dependence of warm bias on dry bias with the precipitation deficit leading the warm bias over this region. Using long-term measurements from the DOE Atmospheric Radiation Measurement (ARM) Climate Research Facility, the team discovered that the precipitation deficit is associated with the widespread failure of GCMs in capturing strong rainfall events in the summer over the SGP. The team then used the good correlation between the projected warming and the present-day warm bias to empirically correct the future projections of temperature. The corrections substantially narrowed the intermodal spread of the projections and reduced the projected warming by 20% by the end of the 21st century under the RCP8.5 scenarios. With the correction, instead of a sharp decrease, the projected precipitation by the end of the 21st century in this region will likely be neutral or slightly positive.