Experimental Stormscale Ensemble

Experimental Stormscale Ensemble
Name	Experimental Stormscale Ensemble
Abbreviation	ESE
Type	atmospheric ensemble forecast system
Developer	National Oceanic and Atmospheric Administration National Weather Service National Centers for Environmental Prediction Storm Prediction Center Weather Prediction Center
Introduced	2010s (experimental)
Purpose	convection-allowing probabilistic forecasting for severe thunderstorms, tornado outbreaks, hail and flash flood events
Resolution	convection-permitting (~1–3 km)
Members	ensemble of high-resolution members (varied)
Status	experimental / research-to-operations transition

Experimental Stormscale Ensemble

The Experimental Stormscale Ensemble is an exploratory, convection-allowing ensemble forecasting system designed to provide probabilistic guidance for high-impact thunderstorm events, such as tornado outbreaks, derecho occurrences, and flash flood episodes. It was developed by research groups within the National Oceanic and Atmospheric Administration, including teams at the National Weather Service, National Centers for Environmental Prediction, and collaborative partners from academic institutions and federal laboratories. The system emphasizes storm-scale dynamics, ensemble spread, and rapid-update cycles to support operational centers like the Storm Prediction Center and regional forecast offices.

Overview

The Ensemble was conceived to bridge deterministic models like the High-Resolution Rapid Refresh and global systems such as the Global Forecast System with probabilistic frameworks exemplified by the European Centre for Medium-Range Weather Forecasts ensemble and the North American Ensemble Forecast System. It integrates tools and concepts from projects including the Distributed Hydrologic Model Intercomparison Project, the Community Earth System Model, and the Cyclone Global Navigation Satellite System research community. Stakeholders include the Office of the Federal Coordinator for Meteorology, the University Corporation for Atmospheric Research, and university groups at Penn State University, University of Oklahoma, Texas A&M University, and Colorado State University. The Ensemble informs decision-making for agencies such as the Federal Emergency Management Agency and utilities like Entergy during convective outbreaks.

Development and Versions

Early prototypes drew on the Weather Research and Forecasting Model and assimilation strategies from the Gridpoint Statistical Interpolation system, with code contributions from the National Center for Atmospheric Research and model physics options influenced by NOAA Geophysical Fluid Dynamics Laboratory parameterizations. Development iterations referenced operational programs like NOAA-20 satellite missions, assimilated data streams from Doppler radar networks including NEXRAD and research radars at Oklahoma Mesonet sites. Versions advanced through collaborations with agencies such as the Department of Energy, the National Science Foundation, and international partners at Met Office and Environment and Climate Change Canada, and were evaluated against datasets from projects like Tropical Rainfall Measuring Mission and GPM.

Methodology and Configuration

The Ensemble uses convection-permitting grids comparable to those employed by High-Resolution Rapid Refresh, with horizontal resolutions on the order of 1–3 km and forecast lead times tailored for mesoscale convective systems akin to cases studied in the VORTEX and VORTEX2 field campaigns. Member generation employed perturbations inspired by techniques used by the European Centre for Medium-Range Weather Forecasts Ensemble Prediction System and the Canadian Ensemble while incorporating stochastic physics schemes similar to those developed at the Met Office and ECMWF. Lateral boundary conditions often originate from the Global Forecast System or reanalyses such as ERA5, with nesting strategies informed by the North American Mesoscale system and regional frameworks promoted by the Hydrometeorological Testbed.

Data Assimilation and Observations

Data assimilation fused observations from NEXRAD radar, Geostationary Operational Environmental Satellite imagery from the GOES series, surface networks including the Automated Surface Observing System, and vertical profiles from radiosonde launches coordinated with University of Alabama in Huntsville and NOAA Air Resources Laboratory campaigns. Assimilation methodologies referenced variational and ensemble approaches like 3D-Var, 4D-Var, and the Ensemble Kalman Filter developed at institutions including Princeton University and Massachusetts Institute of Technology. Additional inputs incorporated storm-scale observations from research programs such as IMPROVE, IHOP, and Hydro-CLASH, and remote-sensing data from platforms including Doppler on Wheels and C-band networks.

Verification and Performance

Verification employed metrics standardized by the World Meteorological Organization and research consortia, comparing skill against benchmarks including the High-Resolution Rapid Refresh and North American Mesoscale forecasts. Probabilistic measures leveraged the Brier score, continuous ranked probability score methodologies advanced at NOAA ESRL, and object-based verification frameworks developed at University of Oklahoma and University of Washington. Case studies included notable events like May 3, 1999 tornado outbreak reanalyses and El Reno tornado assessments, and performance was cross-validated using datasets from SPC Convective Outlook archives and Storm Data reports.

Applications and Operational Use

Operational users at the Storm Prediction Center, National Weather Service field offices, and emergency managers in FEMA Region IV and FEMA Region VI have used experimental outputs to refine convective outlooks, warn lead times, and allocate resources during events similar to Hurricane Harvey-influenced convective episodes. Utilities, transportation agencies such as Federal Aviation Administration, and infrastructure operators at facilities like Entergy and Port Authority of New York and New Jersey have trialed ensemble-informed decision support. Research-to-operations pathways involved partnerships with NOAA's Weather Program Office, the Office of Science and Technology Policy, and international exchanges with Met Éireann and Bureau of Meteorology.

Limitations and Future Work

Limitations include computational cost constraints noted by supercomputing centers like Oak Ridge National Laboratory and the National Center for Atmospheric Research Computational and Information Systems Laboratory, sensitivities to convection-permitting physics highlighted in studies at Colorado State University and University of Illinois, and challenges in assimilating nonconventional data streams tested by Johns Hopkins University and Massachusetts Institute of Technology Lincoln Laboratory. Future work emphasizes coupling with hydrologic models from USGS and National Water Center initiatives, machine-learning post-processing collaborations with Google Research and IBM Research, and integration into operational suites at National Centers for Environmental Prediction and global centers such as ECMWF and the Met Office. Continued field campaigns like VORTEX-SE and cross-agency projects with NSF and DOE aim to expand observation networks and improve ensemble calibration.

Category:Numerical weather prediction