Skip to main content
U.S. flag

An official website of the United States government

Publication Date
28 July 2017

Simulating the Global Water Cycle with A High Spatial-Resolution Climate Model

Print / PDF
Powerpoint Slide
Image
(top left) Annual mean hi-res ACME v0.3 precipitation rate, (top right) difference between hi-res and low-res model precipitation, (bottom left) difference between hi-res model and GPCP precipitation, and (bottom right) difference between CMIP5 AMIP multi-model mean and GPCP precipitation.
Science

Scientists from the Lawrence Livermore National Laboratory and the Oak Ridge National Laboratory examined simulations of the observable present-day climate to diagnose how the version 0.3 of the Department of Energy’s new climate model represented the global water cycle. Comparing with the best available observational datasets, they determined that while the model captured most fundamental aspects of the global water cycle, the model also produced long-standing biases, which were only partially improved with increasing resolution.

Impact

This study contributes to the growing number of studies that evaluate the impact of increasing the resolution on a climate model’s ability to better represent the water cycle. It also provides a new framework to diagnose why the precipitation is increasingly produced by the large-scale physics scheme, rather than the convective scheme that simulates sub-grid scale cloud motions.

Summary

The U.S. Department of Energy is developing a new high-resolution climate model under the Accelerated Climate Modeling for Energy (ACME) project. One of the its driving questions is, “What are the processes and factors governing precipitation and water cycle today, and how will precipitation evolve over the next 40 years?” This study assessed how well the version 0.3 of the ACME model is able to represent the present-day atmospheric hydrologic cycle and examined how increasing the horizontal resolution from a grid spacing of approximately 100 km to 25 km changes the representation of the global water cycle. This is relevant given that previous studies have reported differing results regarding the impact of horizontal resolution on the water cycle. The model was evaluated using the best available observational estimates, and the diagnosis found several biases in the model, which are shared by other state-of-the-art climate models, namely a global mean precipitation rate that is too high, light rain that occurs too frequently, and an atmospheric residence time of water that is too short. Increasing the resolution does not improve those biases but improves the frequency of heavy precipitation events and shifts the precipitation produced by the convective physics scheme to that produced by the large-scale physics scheme. The study provides a basis on which to compare subsequent versions of the model and provides a reminder of building a body of literature for different models so that we can get a sense for which behaviors are common across all climate models and which are model-dependent.

Point of Contact
Christopher Terai
Institution(s)
Lawrence Livermore National Laboratory (LLNL)
Funding Program Area(s)
Publication