Scientists gain a powerful tool to scrutinize changing weather patterns in the United States

Scientists gain a powerful tool to scrutinize changing weather patterns in the United States

CONUS404 is an unprecedented high-resolution long-term weather dataset

October 6, 2023 – Written by Laura Snyder

An extraordinary new dataset of high-resolution weather simulations spanning more than four decades across the continental United States is now available to the Earth system science community.

This unprecedented resource—which required nearly a year of supercomputing time to create, and is approximately one petabyte in size—provides rich opportunities for scientists and stakeholders interested in how weather patterns are actually shifting as the climate warms, among many other applications. For example, scientists are already using the data to research new techniques to improve long-range forecasting, plan water resource allocation, and better understand the causes and impacts of extreme and rare weather events.

The dataset, known as CONUS404, is the result of a collaboration between the National Center for Atmospheric Research (NCAR) and the United States Geological Survey (USGS).

“To study the rare and extreme climate events that we really care about, we need decades of data and that data must be at high resolution,” said Roy Rasmussen, a senior scientist at the National Research Center who led the project. “With CONUS404, it is possible to study both long-term events that can span many years, such as droughts, and rare events that do not last long but rarely occur, such as severe floods.”

CONUS404 was funded by the US Geological Survey and the US National Science Foundation, an NCAR sponsor.

Filling a gap for water managers

Weather in the United States is relatively well observed by local weather stations, stream gauges, snowpack sensors, radars, weather balloons, satellites, and more. But these observations by themselves cannot always give a clear picture of how weather patterns change over time. This is because the data is often regionally aggregated – with sparse information about conditions in remote and rugged terrain – and unreliable. For example, the accuracy of temperature, humidity, wind, and other important weather data can be affected by device performance and local conditions. And measurements across different monitoring platforms do not always agree.

Because of these factors, scientists rely on meteorological “reanalysis” that combines observations and modeling to create data sets that provide internally consistent weather information at fixed points on a grid across an entire region or globe. These reanalysis products are important tools for scientists. For example, they can be used to check how well climate models can simulate past conditions, a crucial test for determining how well they simulate the future. Scientists also use these products to start or “start” model simulations with real-world conditions.

Although reanalysis products are important, they are generally of low resolution of about 30 kilometers (19 mi) or more between grid points—a distance that is too coarse to capture relatively fine-scale weather events, such as summer thunderstorms, And the local terrain that influences those event events, such as mountain ranges. They are also too coarse to provide meaningful data about rainfall in individual watersheds, important information for water managers. It is this last point that has been a particular frustration for the USGS, which is responsible for collecting and distributing information to the country about water resources, including water flow and groundwater data.

To address this gap, the USGS partnered with NCAR to “downsize” one of the most widely used global reanalysis datasets, called ERA5to create a high-resolution dataset for the contiguous United States (CONUS) using NCAR's Weather Research and Forecasting (WRF) model.

The resulting dataset covers over 40 years (1980-2021) with a 4 km grid spacing – hence the name CONUS404.

Simulating weather covering such a large area over such a long period and with such high accuracy has never been possible before. But several factors have come together over the past decade to make this project feasible, including advances in supercomputing capabilities and weather models. Even with advances in computing, it still takes more than 11 months to complete a simulation on the USGS Denali supercomputing system.

Improvements in WRF over the past few years have corrected many of the problems that arose in previous attempts to run the model at high resolution over CONUS, but for shorter time periods. One problem was that the WRF tended to make the central United States very hot and very dry, which in turn affected the model's ability to accurately simulate thunderstorms in the region. But the updated version of the WRF includes a groundwater module that cools and dampens the area, resulting in a more realistic simulation. The updated version also does a better job of simulating western snowpack, which affects runoff and surface temperatures, and corrects the trend in the model to make winter temperatures too cold in snow-covered areas.

“We are now able to capture the fundamental factors that cause weather in the real world,” Rasmussen said. “We're not perfect, and we're still learning all the time, but the model does an amazing job of getting the historical weather right.”

Data mining

The paper presenting the dataset was published earlier this summer in the journal Bulletin of the American Meteorological SocietyBut many scientists are already looking at the data to answer their research questions. For example, the CONUS404 dataset helped researchers uncover patterns that are now used during historical droughts Improving seasonal drought forecasts In the west.

Scientists are also looking for subtle evidence of changing weather patterns over the past few decades, including one study that identified a shift in the way precipitation falls, from less rain and light rain to more heavy rain. Climate models have long predicted this change, but the low-resolution reanalysis datasets that have existed up to this point have not been detailed enough to identify the change.

Researchers are also analyzing changes in harmful local winds that sometimes accompany storms. Other scientists are looking at extremes in streamflow and whether CONUS404 data can be used as input into crop models to simulate water use and food production.

While the new dataset is just beginning to be exploited, the NCAR-USGS collaboration is already working on the second part of the project: another 40-plus-year weather simulation across the United States, this time in the future. Scientists will use the same methodology, but instead of scaling back to reanalyze what happened in the past, they will use data from NCAR's Community Earth System Model, Version 2, (CESM2) to project what they think conditions will be like in the future. Together, the two datasets will provide more than 80 years of simulation data that will give an unprecedented look at how weather will continue to change as the climate warms.

“Connecting these two pieces — climate and weather modeling at NCAR — is extremely important,” said NCAR scientist Andreas Prien, a co-author of the study. “We have to be smart about how we use climate data to get it to a useful scale.”

CONUS404 data can be accessed free of charge from NCAR Research Data Archive.

About the article

Title: CONUS404: Reanalysis of the NCAR-USGS 4 km long-term regional hydrological climate over CONUS
Authors: Rasmussen RM, Chen F, Liu CH, Ikeda KA, Breen A, Kim J, Schneider T, Dai A, Gotches D, Dagher A, Zhang Y, Jay A, Dudia J, Si He, M. Harold, L. Xue, S. Chen, A. Newman, E. Dougherty, R. Abolafia-Rosenzweig, N. Lybarger, R. VIGER, D. Lesmes, K. Skalak, J. BrakeBill, 1999 (PubMed) Cline D., Dunne K ., Rasmussen K., and Miguez-Macho G. (1999).
magazine: Bulletin of the American Meteorological Society

View all news

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *