| What's New | Approach | Tools | Research Projects | Events | Publications | People | Contact |
| Introduction | EVIS | MURI | BIS Portal | MV | Glider | OOS | DMAC | EDSS | Feedback Planning |
ANNOUNCEMENT: The Human and Social Dynamics program at NSF notified UW that the proposal on communicating weather forecast uncertainty will be funded in FY08. This will be a joint UW project with participants from Statistics, Atmospheric Science, Psychology, and APL (EIS department). This effort will extend work of the current Multidisciplinary University Research Initiative (MURI) project on visualizing uncertain data.
 |
 |
 |
MURI Research
Researchers are developing methods and systems to help weather forecasters deal with uncertainty in numerical weather prediction models. In this MURI, an OSD-funded project, researchers are developing a web system and interface that combines mesoscale uncertainty information in funnel fashion that aids in understanding and utility.
MURI researchers span four University of Washington departments: Applied Physics Laboratory, Statistics, Atmospheric Sciences, and Psychology. The team's goals include:
- Development of methods for evaluating uncertainty of mesoscale meteorological model prediction
- Improvement of statistical methods for dealing with uncertainty;
- Understanding how forecasters incorporate uncertainty in their forecasts;
- Development of methods of integration and visualization of multi-source information from model output, observations, and expert knowledge
|
|
|
MURI Products:
MUM
The Model Uncertainty Monitor (MUM) processes and displays a variety of uncertainty information related to Numerical Weather Prediction (NWP) systems. To accommodate the goal structure observed in the cognitive task analysis, information is presented in three different time windows: past, present, and future. The past window presents information on the accuracy of previous NWP predictions. As such, it encourages forecasters to use models that are performing best in the specific circumstances and to use evaluation techniques tailored to the situation. The present window provides information on the differences between the current analyses (the NWP model's estimate of the present state of the weather) and observations. The future window provides access to a variety of probabilistic forecast information for specific parameters, such as precipitation, temperature, or wind speed.
The MUM provides the forecaster with a holistic approach to assessing NWP. It centralizes the relevant information for the forecasting task in a single interface that meets the needs of the naval forecaster and is easy to understand. Yet it leaves the final decision about the optimal product or predication up to the forecaster.
The source of weather data is from the Penn State University / National Center for Atmospheric Research mesoscale atmospheric model, known as MM5, specifically the operational implementation of UW's version of the MM5 designed for navigation and wind speed and direction.
ProbCast
 |
Another MURI product is the ProbCast or Probability Forecast tool which aims to bring probabilistic weather forecasting information for the MM5 domain to the non-scientific public. This application brings a focused handful of useful visualizations to the user with a minimum of complexity and unnecessary detail about the underlying science. At the same time ProbCast provides forecasts of temperature, wind, and precipitation. The nominal user of this product is a non-scientific member of the agricultural or power industry, or members of other communities for whom weather has a financial consequence. The product should be usable and useful even without a perfect understanding of the cost/loss structure of decisions made based on weather forecasts. User feedback is solicited by several survey questions, the answers to which can be submitted through an automated form.