New Mexico Supercomputing Challenge

"I Smell Smoke..." Otero County Fire Escape

Team: 8

School: Artesia High

Area of Science: Forestry, Logistics

Interim: <b>“I smell smoke... Otero County Fire Escape” </b>


Cloudcroft, a heavily forested area with isolation from other towns and limited exits, is in constant fear of fire in their forest in times of drought. What does the town have planned in the case a fire does appear? What does the Public Safety Departments (hereafter PSD) do to prepare the town for forest fires? And lastly, what do the citizens of Cloudcroft do to protect themselves? These are all questions we will attempt to answer and put to the test in our model this year.


In order to answer the above questions, we will ask the PSD their plans for these scenarios, to date, how often they must put their plans to work, and how the results are. We will then talk to citizens, chosen at random, about their personal preparation for forest fires. Knowing that Cloudcroft is also an area of fickle weather and surprise rain storms, it will be necessary to incorporate these factors into our model.
In the modeling phase, we will use an agent-based approach with heterogeneous agents. The agent behaviors will reflect our research into real- world behavior of people in emergency situations. In dealing with the actual fire within the model itself, we will apply stochastic cellular automata to our model. Utilizing this, we will make our representation of a Cloudcroft forest fire more like an actual forest fire.
We plan to implement our model in the modeling program NetLogo. Our model last year also used NetLogo, and we are incorporating certain aspects of that model into this year's, such as the basic rules restricting agents from moving through invalid regions. (E.g. Structures, etc.) In order to validate our model, we'll present the output of the basic scenarios to the representatives of the PSD to see how close it matches what they would expect from a real world fire.
After that we will conduct numerous experiments using repeated simulations of our scenarios in order to verify our model, tweaking the code of our model as needed. We will then begin analyzing our results to find the most likely outcomes.


Creating a forest fire model and an evacuation model is no small task. We have gathered a sizable amount of research regarding the fractal distribution of fire spread and the likely causes of fire, as well as the set of existing fire models that are capable of simulating fires. In addition, we have read a few articles about other fire model makers who have found a real world application. One such model particularly interested us.
In Ithaca, New York, a team of geologists and computer scientists created a model simulating matches and trees. The model, centered in Yellowstone National Park, showed one of the reasons for the large Yellowstone Fire of 1988: the accumulation of dead trees from previously suppressed fires. The researchers discovered that the frequency and distribution of small and medium fires can be used to assess the risk of larger fires, much in the same way as small tremors are used to assess the risk of larger earthquakes. courtesy of
Several models have been developed over the years to simulate fire growth and spread as it relates to the ecosystem. BEHAVE Plus Fire Modeling System is the one most frequently used to assess and predict fire behavior in wild land fuels, such as forests. The most recent version is more user-friendly and has many graphic interfaces, including tree density and weather situations, which will be germane to our model.
It was discovered in our research that humans have very little to do with the spread of fires. Humans are only responsible for starting fires and stopping them. In addition, fires are affected by fuel, weather, and topography- recall that fires tend to burn upwards. The potential for the beginning of a blaze in an area such as Cloudcroft could bring devastating consequences if ignored.


The tactics applied in last year’s model will be put to good use in this model. Because of the various challenges, such as the lack of exits, the higher risk for fire, and the lack of territory knowledge, the model will need to be implemented much more often, and it will serve as an instrument on the panel of the PSD. A trip to Cloudcroft to interview citizens and the PSD will be underway by the beginning of 2013. Lastly, we will develop potential cases and scenarios for future implementation.
Computer Simulation of Matches and Trees Can Be Used to Predict and Prevent Large Forest Fires, Cornell Geologists Say. Courtesy of
Behave Plus fire modeling System, version 5.0. Patricia L. Andrews. September 2009. PDF.
How to Generate and Interpret Fire Characteristics Charts for Surface and Crown Fire Behavior. Patricia L. Andrews. USDA March 2011. PDF.

Team Members:

  Devon Mullins
  Keenan Flynn
  Jason Aguilar

Sponsoring Teacher: r gaylor

Mail the entire Team

For questions about the Supercomputing Challenge, a 501(c)3 organization, contact us at: consult @

New Mexico Supercomputing Challenge, Inc.
Post Office Box 30102
Albuquerque, New Mexico 87190
(505) 667-2864

Supercomputing Challenge Board of Directors
Board page listing meetings and agendas
If you have volunteered for the Challenge, please fill out our In Kind form.