A study of the inter-process dependence of sub-grid scale physics parameterizations and process ordering impact.
DOE scientists explored how the order in which physics process parameterizations are handled within a global atmosphere model impact the model climate and climate sensitivity, i.e., how much the model will warm in response to increased greenhouse gas forcing. Through a complete set of physics re-orderings for four physics processes, they were able to show that (a) process order impacts the model predicted climate sensitivity, and (b) there exist two distinct solution regimes which are dependent on the interaction between shallow convection and stratiform cloud processes.
This new understanding suggests that periodic assessment of the physics process order should be an integral part of global atmosphere model development. The study also demonstrates the importance that process order has on model accuracy and predicted climate sensitivity, suggesting that inter-model comparisons would benefit from an understanding of how individual models handle physics process ordering. It also determines a list of best practices which will be helpful for future model design efforts.
Atmospheric models consist of several process parameterizations. The standard modeling approach is to apply each individual parameterization sequentially such that each parameterization feels the effects of all the parameterizations preceding it. This approach is noncommutative, in that changing the parameterization order will change the model solution. Researchers from DOE have recently analyzed the impact of rearranging the physics parameterization order in the Energy Exascale Earth System Model (E3SM). They discovered that the process order impacts model climate and climate sensitivity. The study demonstrates that there is a bifurcation in the modeled climate dependent on the interaction between the shallow convection and stratiform cloud processes. Using these results, it is possible to determine a set of best practices for process ordering that will improve the credibility of future modeling efforts. This study also demonstrated that differences in inter-model studies for climate sensitivity could be attributed to how individual models handle the coupling of physics parameterizations.
Contact(s) (BER PM)
Earth System Modeling
Dr. Peter Caldwell
Atmospheric, Earth, and Energy Division, Lawrence Livermore National Laboratory
This work was performed with support from the BER-Scientific Discovery Through Advance Computing (SciDAC) project titled Multiscale Methods for Accurate, Efficient, and Scale-Aware Models of the Earth System and the Climate Model Development and Validation project ‘‘ACME-SM: A Global Climate Model Software Engineering Surge.”
Donahue, A. S. and P.M. Caldwell. “Impact of physics parameterization ordering in a global atmosphere model.” J. Adv. Model. Earth Syst. 10(2): 481-499 (2018). [DOI: 10.1002/2017MS001067]
SC-23.1 Climate and Environmental Sciences Division, BER
BER supports basic research and scientific user facilities to advance DOE missions in energy and environment. More about BER
Aug 24, 2019
New Approach for Studying How Microbes Influence Their Environment
A diverse group of scientists suggests a common framework and targeting of known microbial processes [more...]
Aug 08, 2019
Nutrient-Hungry Peatland Microbes Reduce Carbon Loss Under Warmer Conditions
Enzyme production in peatlands reduces carbon lost to respiration under future high temperatures. [more...]
Aug 05, 2019
Amazon Forest Response to CO2 Fertilization Dependent on Plant Phosphorus Acquisition
AmazonFACE Model Intercomparison. The Science Plant growth is dependent on the availabi [more...]
Jul 29, 2019
A Slippery Slope: Soil Carbon Destabilization
Carbon gain or loss depends on the balance between competing biological, chemical, and physical reac [more...]
Jul 15, 2019
Field Evaluation of Gas Analyzers for Measuring Ecosystem Fluxes
How gas analyzer type and correction method impact measured fluxes. The Science A side- [more...]
List all highlights (possible long download time)