Juniper Encroachment and Management in the Intermountain West Relative to Wildfires Utilizing Remote Sensing

One of the most pronounced vegetation changes in recent history is the expansion of junipers (Juniperus spp.) throughout the intermountain west United States. These native species have expanded from their traditional fire-safe habitats into fire-dependent communities as a result of climatic fluctuations, grazing patterns, and wildfire suppression efforts. As junipers expand their range, they begin to dominate plant communities resulting in the recession of shrubs, grasses, and forbs. Land management agencies have a strong commitment to find areas that are vulnerable to juniper encroachment, so that these areas can be studied and more effectively managed. Aiding in this effort, this project used remote sensing to develop two tools that determine fire intensity on a per pixel basis and identify different phases of juniper encroachment, respectively. Landsat 8, representing land cover data was combined with topography information (slope and aspect) in a linear regression model that quantified fire intensity on a per pixel basis, identifying areas that would burn hotter and longer based on fuel type. The overall accuracy of the model was 86% with a kappa coefficient of 0.81. Visual validation using NAIP imagery in comparison with the fuel classification result showed good visual correlation of the fuel model with dense juniper stands. The second output of the project was an image/object based classification tool that uses multispectral imagery and supervised point classification to classify different vegetation types according to the spectral detail of the objects. The goal of the model is to improve phase identification of juniper stands. Initial visual verification with NAIP shows the model to be performing very satisfactorily but is dependent on the spatial resolution of the user fed input imagery. Furnishing land managers with these tools will assist in forecasting areas prone to juniper invasion based upon surrounding seedbanks, as well as, predict the ensuing intensity of fires should ignition occur.