Prediction, Assimilation & Risk Communication Section

Overview

The Prediction, Assimilation and Risk Communication Section studies advanced techniques for data assimilation and probabilistic weather forecasts, and the intersections between weather hazards and society. Section activities span a wide range of high-resolution applications using both WRF and MPAS as forecast models, assimilation of a variety atmospheric observations, from in-situ measurements to remotely sensed satellite and radar data, and frequent real-time prediction experiments. We also utilize social science and interdisciplinary approaches to study weather and climate risk communication and decision making, along with the factors that influence these processes. A present focus is to design and demonstrate a convection-permitting, US-wide, probabilistic prediction system that is initialized with cycling data assimilation at O (1 km) resolution.

Research

Community Data Assimilation

WRFDA: We develop, facilitate community contributions, and support the WRF data assimilation (WRFDA) system. WRFDA is built within the WRF software framework and offers several advanced DA algorithms (e.g., 3D/4D-Var, Hybrid-3D/4DEnVar) together with assimilation capabilities for a variety of observations from conventional, satellite, and radar observing networks. We also develop and maintain WRFPlus, which includes tangent linear and adjoint versions of the WRF model and is applied in WRFDA-4DVAR or forecast sensitivity to analysis and observations.

WRF/DART & MPAS/DART: In collaboration with the Data Assimilation Research Section within CISL, we develop and support interfaces for WRF and MPAS to the Data Assimilation Research Testbed (DART). WRF/DART and MPAS/DART employ the ensemble Kalman filter provided by DART. WRF/DART is widely used for data-assimilation research at universities and US national labs, and is the basis for experimental prediction systems such as the NCAR real-time ensemble.

Weather, Risks & Decisions

Weather Risks and Decisions in Society (WRaDS): Using social sciences and interdisciplinary approaches and methods, we build fundamental understanding of how weather-related risks intersect with society, including how they factor into decision making and influence outcomes. Our work includes studies with members of the public, forecasters, public officials, broadcast meteorologists, and other stakeholders. We also help build capacity for weather-society research and practice in the broader community.

Communicating Hazard Information in the Modern Environment (CHIME): We are integrating social sciences with hazard prediction research to investigate how dynamic weather risks and risk information interact with society in the modern information environment. This includes building understanding about how evolving weather forecasts and warnings, societal vulnerabilities, and information flow influence protective decisions as hazardous weather approaches and arrives.

Real-Time Predictions

NCAR real-time ensemble: Using WRF and DART, we have implemented a real-time high-resolution ensemble prediction system that provides 6-hourly, 15-km ensemble analyses and once daily, 10-member, 3-km ensemble forecasts over the conterminous United States. This complete prediction system has operated continuously since April 2015, demonstrating the viability of a continuously cycled mesoscale ensemble analysis to initialize convection-permitting ensemble forecasts. We employ the real-time system as a baseline in research related to ensemble data assimilation, ensemble prediction, and verification, and as a testbed for visualization of ensemble information.

WRFDA real-time forecast/analysis system: We have also implemented a 6-hourly, 15-km analysis system using WRFDA's hybrid ensemble-variational capabilities and the ensemble of 6-hour forecasts from the NCAR real-time ensemble. The system assimilates conventional observations and AMSU-A radiances and makes 48-hour forecasts every 6 hours.

STEP Hydromet Experiment: PARC participates in the STEP Hydromet Experiment, which serves each June, July, and August as a real-time testbed for integrated hydrometeorological prediction systems on the basin scale and covers the Colorado Front Range of the Rocky Mountains and the adjacent high plains. Using WRF and WRFDA, we provide hourly updated, high-resolution forecasts based on assimilation of radar radial velocity and reflectivity.

People

MMM Staff

Other Labs

Affiliate Scientists

• Dale Barker
• Lance Bosart