Climate scientists collaborate on a new high-resolution Earth model – E3SM, which can predict climate trends 50 to 100 years into the future. The simulations developed using E3SM can capture the most sophisticated climate dynamics, from heat transportation through the ocean (advection) to the creation of storms in the atmosphere.
In January, several leading science and research organizations reported 2019 as the second warmest year on record, while July went down as the hottest month ever recorded. Modeling, the climate by using earth models helps in aiding our understanding and predictability of climate dynamics and behavior. This information helps in providing essential information to better inform decisions at the local, national and regional levels, such as hazard management, water resource management, agriculture, and urban planning.
Earth system and Earth Models
Earth system refers to the interaction between hydrosphere, atmosphere, and cryosphere. Earth system modeling helps in simulating all physical, chemical, biological aspects of the Earth system including; oceans, ice, land surface, climate, and biogeochemical cycles. Earth models are the most important tool used to predict, comprehend and reconstruct climate change, The global environment is a complex and ever-evolving system that requires Earth models to decode changes happening in the Earth’s systems, including; the influence of human activities on Earth and to analyze possible future changes and challenges related to it.
Energy Exascale Earth System Model (E3SM)
Seven US Department of Energy (DOE) laboratories including Argonne National Laboratory (ANL) and other research institutes worked in collaboration to advance a high-resolution version of Energy Exascale Earth System Model (E3SM). Using this high-resolution earth model, scientists are trying to predict climatic dynamics and future climatic trends into the next century. By using E3SM the researchers hope to help reduce the effects of extreme climate on energy, water resources, infrastructure, and agriculture. The simulations developed using E3SM project can capture the most detailed dynamics of climate-generating behavior, including energy transport through oceans and storm formation in the atmosphere.
E3SM project and predicting the future of climate
Robert L. Jacbos, a computational climate scientist at ANL said “Climate change can have huge consequences on our need and capability to produce energy, manage water resources and predict impacts on agriculture. Therefore the DOE wants a prediction model that can describe and decode climate changes with enough detail to help decision-makers.”
“E3SM is an Earth model designed to simulate how the combinations of ocean currents, temperature, winds, land surface, and rainfall/snowfall patterns can influence regional climate and built infrastructure on a local and global scale,” explained the lead scientist at ANL. “Moreover, having the capability to predict changes in the climate and the response of the hydrological cycle to the increase in carbon dioxide is significant in planning for our future.”
Infrastructure and ecosystems along our coasts are susceptible to sea-level rise caused, partly by rapid melting of glaciers, and extreme weather events. For example, 2019’s historically heavy rainfalls and flood damage in the central and southern states of America, and 2019’s hot, dry conditions in Alaska, California, and Australia resulting in massive wildfires.
To study how Earth systems work in relation to each other the E3SM Earth model splits the world into thousands of interdependent grid cells featuring oceans, land, sea ice, river and atmosphere. These grids were used in low-resolution as well as in high-resolution configuration. The grid spacing for atmosphere and land was defined to be 25 km for high-resolution and 110 km for low-resolution. Meanwhile, grids for the ocean and sea ice amounted to 8 to 16 km for high-resolution and 30 to 60 km for low resolution.
“The Earth is modeled as a group of cells with 25 kilometers between grid centers horizontally or a quarter of a degree of latitude resolution,” says Azamat Mametjanov, an application performance engineer and a postdoctoral researcher in the Laboratory for Advanced Numerical Simulations at Argonne. “Historically, spatial resolution has been much lower, about 100 kilometers. So we’ve increased the resolution by a factor of four in each direction. We are starting to better resolve the phenomena that energy industries worry about most — extreme weather.”
Scientists believe that E3SM’s high-resolution capabilities will help to resolve geophysical features such as hurricanes and mountain snowpacks that are ambiguous in other models. The most innovative inclusion in the E3SM project was sea surface temperature and sea ice in the North Atlantic Ocean, e.g. Labrador Sea.
“Labrador sea is an important region of the ocean that requires an accurate accounting of air and water flow,” says Dr. Jacbos. “Lower-resolution models tend to depict too much sea ice coverage in the region. This additional sea ice cools the atmosphere above it and degrades our predictions in that area and also downstream. Increasing the resolution of the Earth model also helped resolve the issue with ocean currents in the Labrador Sea, as the results coincided with satellite and ship observations.”
Another important characteristic of the Earth model E3SM is its ability to run over multiple decades. While some models can run at very high resolutions the downside to those models is the time frame of the simulation, which is 5 to 10 years at most. E3SM uses ultra-fast DOE supercomputers, which aided in running the 25-km E3SM simulation to a course of 50 years. The scientists look forward to running a 100 years simulation at a time, as 2100 is considered to be a standard end date used in climate simulation and modeling to predict future climatic trends.
E3SM, computation and future implications
Running an Earth model like E3SM is not an easy task as it requires high resolution, longer time sequences and is a complex process in itself. For each of the cell related to the atmosphere, researchers ran dozens of algebraic operations corresponding to meteorological processes like wind speed, moisture, temperature, atmospheric pressure, the amount of localized heating contributed by sunlight and condensation, etc. The scientists were required to perform these calculations and computation for thousands of times a day. Adding more resolution results in slower computation and a lot of time to run and check the results. The 50-year simulation pursued by the scientists using the E3SM project took about 12 months in real-time to run.
Another variable for which climate scientists must adjust their Earth models is called forcing, which is the natural and anthropogenic drivers stabilizing or changing the climate. The main forcing on the climate system is the sun, which stays relatively constant but during the 20th century, increments in other external factors have been observed, i.e. carbon dioxide, aerosols from volcanoes and sea sprays.
For this first simulation, the team was not so much probing a specific stretch of time as working on the model’s stability, therefore a forcing that represents conditions during the 1950s was selected. The date was a compromise between preindustrial conditions used in low-resolution simulations and the onset of the more dramatic anthropogenic greenhouse gas emissions and warming that would come to a head in this century.
One of the main objectives of the project has been to increase the performance of E3SM on DOE supercomputers. But as computer architectures change with an eye toward exascale computing, the next steps for the project include porting the models to GPUs.
“As the resolution increases using exascale machines, it will become possible to use E3SM to resolve droughts and hurricane trends, which develop over multiple years.” There is still a gap between weather models and climate models, and using E3SM, we are trying to close that gap,” says Dr. Mametjanov.
Energy Exascale Earth System Model (E3SM) is a state of the science earth system modeling, prediction, simulation project that addresses climate and earth system research by using the US Department of Energy (DOE)’s advance computers. The main scientific objectives of Earth model “E3SM ” include; water cycle, cryosphere-ocean system, biogeochemistry and energy, and improving the performance of E3SM model on DOE computers.
New version of Earth model captures detailed climate dynamics
John Spizzirri | November 28, 2020 | Argonne National Laboratory, USA.
The DOE E3SM Coupled Model Version 1: Description and Results at High Resolution
Journal of Advances in Modeling Earth Systems