WHAT IS FDM? Fire Diffusion Models (FDM) is a forest fire simulation and forecasting software developed in Visual Basic for Windows 32-bit Operating System.

The software has been developed starting from the United States Department of Agriculture (USDA) mathematical models, effective for point fire forecasting and simulation. To extend USDA models to region-wide models, we have built and implemented a real-time propagation algorithm based on graph theory. We have also identified four fire descriptors to produce risk maps useful to forecast fire risk levels.

FIRE FORECASTING Basically the objective of any fire forecast is supporting the fire management process. During fire event it is important to have immediate prediction of its behavior, in order to optimize the number and type of fire resources to use, reduce the risk to lose human life and so on. The forecast is realized with an approach that consists in three main steps:

1. Acquiring data that describe the fire environment in the interested area (as morphological features, fuels, fuel humidity, wind speed and wind direction)

2. Computing two main descriptors: rate of spread and fire intensity

3. Estimating other mathematical descriptors, as distance of spread, height of flames, area and perimeter fired in a given time. Finally identifying conditions causing crowing and spotting events.

WHY FDM? Managing and optimizing human and material resources in case of fire. FDM can be used to optimize decisional processes during forest fire events. After the first alarm is given, Forest Fire Prevention Service personnel does not usually have adequate tools to predict where the fire is going to spread, and which is its actual intensity. It is important to predict the fire behavior directly in reception site, to optimize the use of material and human resources to oppose that specific fire event. The software provides precious information without having in input additional data from special systems, as meteorological stations, sensors, video cameras and so on.

Preventing Fire. It is possible to product several scenarios changing environmental start conditions, like meteorological and fuel conditions, fire injection point and so on. In this way it is possible to quantify the fire risk level for each scenario identified, especially for the most frequent. Understanding the fire behave or identifying the area with the higher fire risk, allows you to proactively prevent the fire, placing barriers (as road, rock and water zones) in the appropriate place and scheduling forest and site maintenance. Further applications could be identified in the ability to understand where injecting fire in order to modify the forest configuration or its distribution on a specific area.

Training Forest Fire Prevention Service Personnel. Based on the previous applications, the software is a valid tool when training personnel working at Forest Fire Prevention Service.

WHAT FDM DOES Having in input territorial data and environmental conditions, FDM generates two classes of data: fire simulation data and dynamic risk data.

Fire simulation provides real-time fire evolution, on video, using a dedicated algorithm that uses the minimum number of step to complete. The simulator can be also used to identify fired zones after a given interval of time; usually associable to the time needed to start opposing the fire.

Data can be also elaborated to produce fire risk maps.

Output data are stored as standard format file: every FDM output image is a standard Bitmap format file (BMP), while log, simulation and data files are in plain text format files.

Input Data

FDM requires in input morphological data, simple environmental data and the injection point. The four morphological data files (altitude, fuels, slope and exposition) are stored in a proprietary binary format, in order to have the faster access - Using that format, the software need less than 15 second to read 90.000 point’s data using a 100 MHz Pentium, while in a relational database format it requires more than 5 minutes. Figures 1 and 2 show how the software displays territorial data file.

Meteorological data (like wind speed and direction, humidity, temperature, height of this data) are provided by the user using the GUI, as well as the fire injection point.

Fig 1 Height File for Florence zone

Fig 2 Fuel File for the Florence Zone

Output Data:

The main result is the event description obtained using five mathematical descriptors:

1. Rate of Spread: it is the rate (meters/minutes) of the fireline. This variable allows computing perimeter and distance of fire, giving in input the injection point and the projection time.
2. Fireline Intensity: this is the quantity of heat (KW/m) released in the unit of time for a meter of fire. It allows to estimate efforts needed to oppose the fire
3. Heat per Unit Area: is the quantity of heat (KJ/mq) released in the unit of area; it depends on the fuel type and its characteristic moisture
4. Reaction Intensity: is the quantity of heat (KW/mq) released in the unit of time for a unit of area
5. Flame Length: is the height of flames, measured at the bed of the fuel.

The most important output provided by FDM is the fire simulation for a projection time and four fire risk maps.

The fire simulation is an on video real-time simulation, based on runtime data. The simulation uses a special propagation algorithm that completes with a number of iterations equal to the number of point used to describe the region (grid points) During the iteration you have both partial and definitive simulation of the fire, meaning that time of propagation for elaborated points is the definitive time. So that it is possible to stop the simulation without changing elaboration results.

Runtime Simulation

 The Simulator menu generates:

1) One data file containing the propagation time for each point of the grid, that is the time needed by the fire to reach that point, starting from the injection point.

2) The online simulation, that is runtime on video drawing of fired and firing points. In the Altitude map the fireline is painted in red, while points already fired are in black color.

3) One ASCII log file, after the simulation stop, having ground coordinates and propagation time for each point of the grid. This is the main interface to integrate FDM output with other software.

4) One set of proprietary data file. This is a special format file used by FDM to retrieve simulation data for further elaborations

There is a second group of output composed by two kinds of risk maps:

• Static Risk Map: it is obtained evaluating the heat produced by each fuel in the ground; the map provides the estimation of the static risk level for the zone
• Dynamic Risk Map: obtained using static risk map and meteorological and topographical information at a given time. Due to the fact these maps correlates the morphological characteristics of the processing area with the meteorological conditions, they provide a tool to determine “how much” the ground is predisposed to the fire under that conditions

Once dynamic data file or simulation data file was generated, FDM enables the Output menu. A very friendly GUI drives the user to watch and analyze output data.

The Simulation menu launches a window presenting the contours of fired area as function of the time, using the simulator output data files. It is possible to show screens previously obtained and stored.

4 FDM Dynamic Risk Maps

 The Risk Maps menu show screens for the four dynamic risk maps, using 16 colors or 16 red tonalities.

The Complete Risk menu is enabled when each of Dynamic Risk Maps is stored and available on FDM directories, and they will be painted in a single window together with Fuels and Altitude maps. In this way the user will have a powerful interface to understand fire risk and its features.

Archiving different simulation for different environment conditions, it is possible to figure out several fire scenarios and their anti-fire strategies

READY TO USE FDM requires standard Intel 386 or higher processor with 32-bit Microsoft Windows Operating System.

Very often scientific software requires high-line computers, with high-speed CPUs, due to complex computing and image rendering processes. FDM uses a propagation model very light and efficient, with optimized data file format for the best performances. Finally FDM is a standard Windows 32-bit application that doesn’t require any kind of system and program administration.

CONCLUSIONS For the fire forest problems, FDM is a valid tool to easily produce qualitative and quantitative considerations, helping a lot when forecasting fire scenarios and anti-fire strategies.

The software has “a certification” for the following points

1. Accuracy of USDA models, developed by personnel at the USDA Forest Service, Intermountain Fire Sciences Laboratory, in Missoula, MT.

2. Accuracy of mathematical models used to extend USDA point models to region-aware models

3. Excellent development due to a deep know-how on Object Oriented technology, and mathematical model code writing.

INFO For information contact info@fabiolongo.com