NASA John C. Stennis Space Center
Texas Disasters II

[music begins]

>> Meredith: The objective of our project was to assist the Texas Forest Service in mapping and analyzing fuel loads and phenology in Texas grasslands. 

>> Meredith: The study area for our project consisted of 6 counties in the state of Texas: Jack, Palo Pinto, Parker, Stevens, Wise and Young with an emphasis on a 2011 wildfire called the Possum Kingdom Complex. 

>> Meredith: The partners for our project were the Texas Forest Service and the United States Department of Agriculture Forest Service. All of the support from our partners was greatly appreciated. 

>> Meredith: Texas vegetation, especially grasslands and shrub lands, are highly susceptible to seasonal wildfires during prolonged periods of severe drought. In recent years, the risk of severe wildfires has increased due to variable climate conditions and recent urban expansion into wilderness areas. Abnormally wet years can increase grassland vegetation biomass. If a dry year follows, this increased biomass will contribute to increased fuel loads leading to more frequent and intense fires. Therefore there is a need to develop fire risk assessment products from satellite data that can be produced and posted near real time. 

>> Meredith: According to the Texas Forest Service, 80% of wildfires in Texas occur within 2 miles of developed areas. In 2010, the lone star state recorded an above average rainfall year, the following year in 2011 a terrible drought occured which lead to 2011 becoming the worst wildfire season in the hsitory of the state. Roughly, 31,000 wildfires burned 4 million acres and destroyed approximately 3,000 homes. 6 of the 10 largest documented wildfires in thehistory of the state occured in April of 2011.

>> Meredith: Methods for wildfire management have typically relied on models based on Landsat data to produce comprehensive fuel load and fuel type maps at 30 meter resolution. The Texas Forest Service uses MODIS data processed by FORWARN to assess wildfires using 13 fuel models. 

>> Michael: In order to make the fuel maps sensitive to fluctuations in phenology, the Normalized Difference Vegetation Index, or NDVI, was used. NDVI compares the amount of reflected infrared and visible light to quantify the greenness of vegetation.

>> Michael: The nation wide MODIS-derived phenology data set creates an NDVI curve that has 14 layers at 231 m resolution. These 14 layer images contain seven NDVI and corresponding day of the year values. Relative fuel loads and dryness can be estimated by measuring changes in phenology responses over the course of the year.

>> Michael: The 14 layer MODIS NDVI imagery was stacked with the most recent current cloud-free Landsat 8 imagery. A classification was performed on this stack to produce landcover classes that had similar infrared Landsat and phenological characteristics. 

>> Michael: The final classified image showed a concentration similar vegetation in the south-central portion of the study area. From 2000-2014 this area showed exaggerated response in NDVI to both wet and dry years.

>> Michael: If the wet and dry years are back to back, this can lead to more intense and frequent wildfires. These risks will increase as more development occurs in the area. 

[end music and video]