Modeling and Data Assimilation

Forecasters rely on sophisticated computer models to provide an estimate of what the weather will look like tomorrow, next week, and beyond. With this information people can plan what to wear, what roads to avoid, or how to prepare for severe weather. Decision makers such as water managers and emergency planners can also use model information to decide what steps should be taken to prepare for extremes, such as water scarcity and floods.

To keep communities safe and to support decision makers, the accuracy and predictive capability of forecast models needs to be continually improved through research and technological advances.

What We Do

PSL's Modeling and Data Assimilation Division advances understanding and predictions of climate and weather using the Unified Forecast System (UFS) — a community-based framework for modeling Earth Systems — across a wide range of timescales.

This includes working with the community of developers that contribute to the UFS, including other NOAA contributors, to improve the forecast model and data assimilation system, as well as developing diagnostic tools and supporting reanalysis/reforecast datasets.

We have a strong focus on “Research to Operations” and work closely with the National Weather Service's National Centers for Environmental Prediction to transition research innovations into operational forecast systems.

Research Activities

Advance

Data Assimilation

We develop advanced methods for data assimilation - the process of combining real-world observational data with computer models to improve their accuracy - that seek to optimally merge past forecasts with new observations to estimate the state of the Earth System, either now or in the past. Our work interconnects several different Earth System component models to produce our best estimate of historical conditions, with a focus on developing a strongly-coupled approach where observations in one component can influence the conditions in a different component. We also develop machine learning models to emulate weather forecasts using the latest artificial intelligence technologies.

Improve

Model Accuracy

We explore advanced methods to represent physical processes in models, including ways to represent their uncertainty.

Create

Historical Reanalyses

Our researchers generate historical reanalyses using the Unified Forecast System to provide an essential climate record and to initialize “re-forecasts” that are used to correct errors in real-time predictions. These datasets can be used to study historical extreme events, and to produce historical forecasts necessary for calibration - the process of making current models more accurate and reliable when predicting future weather events.

This is shared

Featured Projects

NOAA NASA Joint Archive (NNJA) of Observations for Earth System Reanalysis

NOAA UFS Replay

UFS Weather Model

UFS Research to Operations

Publication Highlights

Integration of emerging data-driven models into the NOAA research to operation pipeline for numerical weather prediction. Sergey Frolov (PSL), et al. Bull. Amer. Meteor. Soc.
Learn more

Updates to NOAA’s Unified Forecast System’s cumulus convection parameterization scheme between GFSv16 and GFSv17. Lisa Bengtsson (PSL) and J. Han Wea. Forecasting.
Learn more

Nondescript image of weather models on computer screens. Credit: Adobe Stock

Division Chief

Philip Pegion

About | Email

Division Staff List