Kilometer-Scale E3SM Land Model Simulation over North America
Dali Wang, Peter Schwartz, Fengming Yuan, Danial Ricciuto, Peter Thornton, Shih-Chieh Kao, Michele Thornton, Chen Wang, Kathryn Mohror, Qinglei Cao, Jayesh Krishnar, Danqing Wu · 2025
Read Kilometer-Scale E3SM Land Model Simulation over North America by Dali Wang, Peter Schwartz, Fengming Yuan, Danial Ricciuto, Peter Thornton, Shih-Chieh Kao, Michele Thornton, Chen Wang, Kathryn Mohror, Qinglei Cao, Jayesh Krishnar, Danqing Wu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-11141.txt.
Introduction
Wang and colleagues report in 2025 the first full kilometer-scale runs of the E3SM land model over North America. The simulation covers over twenty million land cells at one-kilometer resolution. It ran on more than a hundred thousand processor cores. That is hundreds of times larger than earlier efforts of this kind. Scaling tests showed strong efficiency.
Fine land detail helps model local geography and extreme weather, which matter for energy and ecosystem planning. The model shares the wider E3SM infrastructure and a toolkit called KiloCraft. It can zoom from a single cell to the whole continent.
Catalog record
| CREATOR | Dali Wang, Peter Schwartz, Fengming Yuan, Danial Ricciuto, Peter Thornton, Shih-Chieh Kao, Michele Thornton, Chen Wang, Kathryn Mohror, Qinglei Cao, Jayesh Krishnar, Danqing Wu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-11141
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Kilometer-Scale E3SM Land Model Simulation over North America Dali Wang, Peter Schwartz, Fengming Yuan, Danial Ricciuto, Peter Thornton, Shih-Chieh Kao, Michele Thornton, Chen Wang, Kathryn Mohror, Qinglei Cao, Jayesh Krishnar, Danqing Wu Abstract The development of a kilometer-scale E3SM Land Model (km-scale ELM) is an integral part of the E3SM project, which seeks to advance energy-related Earth system science research with state-of-the-art modeling and simulation capabilities on exascale computing systems. Through the utilization of high-fidelity data products, such as atmospheric forcing and soil properties, the km-scale ELM plays a critical role in accurately modeling geographical characteristics and extreme weather occurrences. The model is vital for enhancing our comprehension and prediction of climate patterns, as well as their effects on ecosystems and human activities. This study showcases the first set of full-capability, km-scale ELM simulations over various computational domains, including simulations encompassing 21.6 million land gridcells, reflecting approximately 21.5 million square kilometers of North America at a 1 km × \times 1 km resolution. We present the largest km-scale ELM simulation using up to 100,800 CPU cores across 2,400 nodes. This continental-scale simulation is 300 times larger than any previous studies, and the computational resources used are about 400 times larger than those used in prior efforts. Both strong and weak scaling tests have been conducted, revealing exceptional performance efficiency and resource utilization. The km-scale ELM uses the common E3SM modeling infrastructure and a general data toolkit known as KiloCraft. Consequently, it can be readily adapted for both fully-coupled E3SM simulations and data-driven simulations over specific areas, ranging from a single gridcell to the entire North America. I Introduction Earth System Models (ESMs) play a vital role in advancing our understanding of the Earth’s climate system and its response to natural and human-induced changes. These models incorporate various components such as the atmosphere, oceans, and land surface, along with their dynamic interactions, to provide valuable insights into past, present, and future climate conditions. ESMs contribute to climate research, policy-making, and public awareness, enabling us to make informed decisions and take proactive measures to address the challenges posed by climate change. The development and application of ESMs require significant computational resources and expertise. The Energy Exascale Earth System Model (E3SM), funded by the US Department of Energy (DOE), was developed to align with the nation’s scientific and the DOE’s mission objectives. E3SM distinguishes itself from many other ESMs by emphasizing energy-related research. This includes investigating the effects of climate change on energy systems, such as water availability, severe weather events, and sea level rise [ 1 ] . E3SM’s key advantages include its high-resolution capability, allowing for better representation of smaller-scale processes like regional weather patterns and ocean currents. The model is optimized for advanced exascale computers, enabling complex simulations that would be computationally prohibitive on older systems, leading to more detailed and accurate predictions [ 2 ] . E3SM Land Model (ELM) [ 3 ] provides detailed representations of vegetation, soil moisture, surface temperature, and land-use changes. ELM interacts with other E3SM components, such as the atmosphere and oceans, to gain insights into how these elements influence each other and the overall Earth system. Key processes modeled by ELM include energy and water exchange, carbon cycling, vegetation dynamics, and human activities like agriculture and deforestation. ELM is crucial for understanding surface conditions that influence atmospheric circulation and climate patterns, making them valuable tools for studying climate change impacts, land-use change scenarios, and carbon sequestration potential. I-A Related work The majority of ESM simulations are conducted on coarse grids larger than 0.5 \times 0.5 degrees [ 4 ] . The Energy Exascale Earth System Model version 1 (E3SMv1) simulates at two different horizontal resolutions: a standard resolution with approximately 100 km (around 1 degree) grid spacing for land and atmosphere, and a high-resolution with 25 km (0.25 degree) grid spacing for land and atmosphere. The high-resolution version allows for a detailed representation of ocean eddies and atmospheric storms, which are vital for understanding planetary circulation and their impacts on human activity [ 2 ] . E3SM is currently developing kilometer-scale models that fully utilize Exascale computers to provide detailed representations of environmental features. The development of kilometer(km)-scale ELM model is being carried out in conjunction with other E3SM components aimed at exascale computing, such as the kilometer(km)-scale Atmosphere and Ocean models [ 5 , 6 ] The ELM software comprises half a million lines of code, featuring highly specialized data types containing around 2,000 globally accessible, multidimensional arrays, and over 1,000 subroutines [ 7 , 8 ] . In recent years, there have been advancements in creating an ultrahigh-resolution ELM (uELM) simulation that leverages exascale computing for accurate land simulations on continental and global scales. For instance, a new computational model and framework for uELM simulation on hybrid architectures of exascale computers have been introduced [ 9 , 10 ] . Various strategies for porting uELM have been established, alongside the development of a functional unit test (FUT) framework [ 11 , 12 ] . These uELM simulations were carried out on a small domain with fewer than 10,000 grid cells, utilizing small-scale HPC clusters with only a few nodes (less than 5). An ELM simulation at a 1 km by 1 km resolution over a small domain, containing fewer than 10,000 grid cells, were reported in 2022. However, that study employed a simplified ELM model (with a satellite phenology submodel and a reduced number of subgrid components) to understand the sub-grid topographic effects on land-atmosphere interactions over mountainous areas [ 13 ] . With the availability of high-fidelity data products, such as the forcing and soil properties [ 14 , 15 , 16 ] , km-scale ELM can deliver high spatial resolution beneficial for accurately simulating geographical features and extreme weather events. However, the inherent computational challenges pose a significant barrier to their implementation. Specifically, the increased grid density demands advanced computing resources and storage capacities to handle the vast data volumes. I-B Km-scale ELM simulation To effectively tackle the computational challenges associated with km-scale ELM simulations, we have developed methods that optimize the model configuration and enhance data handling and processing capabilities. The study delivered an unprecedented ELM simulation capability, with its contributions described as follows: 1. The study conducted the largest km-scale ELM simulation to date. This simulation, which operated at full ELM capacity using a common E3SM modeling framework, covered a continental scale and was 300 times larger than any previous study reported. 2. The study presented the largest km-scale ELM simulation using up to 100,800 CPU cores across 2,400 nodes. The study archived outstanding scaling performance by utilizing computational resources that were about 400 times larger than those used in prior efforts. 3. This study developed a data toolkit to support ELM simulation at various scales, ranging from a single gridcell to the entire North America domain of 21.5 million gridcells. This study also created the first 1 km by 1 km dataset (including domain, surface properties, and forcing) for ELM simulations across North America. II Method The following four sections outline our method: model configuration, input data generation, numerical experiments to assess the ELM simulation readiness, and results validation and performance profiling. II-A Data-driven, km-scale ELM