Welcome to

Data Dynamic Simulation for Disaster Management

Data Driven Numerical Simulation of Wildfires
Coupled Atmosphere-Wildfire Modeling

We are working on advanced mathematical modeling and information technology in the area of Dynamic Data Driven Application Systems, to help fight and manage wildfires. Here is project summary.

Frequently Asked Questions

Why is this project important? What issue does it address?

The purpose of the project is to provide a continously updated prediction how the fire will move in the next few hours or days and to predict the likely effect of firefighting strategies. Such technology in the field might save lives and property both in fighting fires and by better control of prescribed burns.

Where, when, and how do you see the system being used in the field?

We are not there yet, though we of course hope that our results will be used in the field in the future.

What is the project's current status? What's next?

The mathematical model of fire and weather was developed at NCAR earlier, based on the Clark-Hall atmospheric code. The fire model is based on propagation of the fireline curve in the normal direction and exponential decay of fuel. We have now developed a new code that implements this validated fire model using a level set method. The new code is suitable for data assimilation (modification of its state by new data). The new fire code has been coupled with WRF, which is a de-facto standard in weather modeling, and it is distributed for research and education purposes. The data assimilation component is not ready for public distribution yet.

We are developing mathematical and statistical methods to update predictions by injecting data into a model continously and we have tested them on simplified mathematical models. Our new morphing filter, which combines correction to position with correction to intensity of physical fields is particularly promising. We work on information techology to manage the different kinds of data that will be coming in the weather/fire application and adding software components for injecting the data into the existing software. We are also developing a new fire model based on reaction-diffusion partial differential equations.

Next comes reanalysis of an actual wildfire from observed data and injecting data into a weather-fire model that will run continuously on our local computers for the mountains near Boulder.