Land Surface Temperature Anomalies and Fire Occurrence
The Relationship between Land Surface Temperature Anomalies and Fire Occurrence in Cariboo Region in 2017
Introduction
Wildfires pose an increasing threat to ecosystems worldwide and put public safety in danger, especially for people living in the area close to forests. Wildfire plays a significant role in not only affecting forest regeneration and succession, resulting in a great loss of forest resources, but also threatening safety of humans and properties. Negatively affecting the ecosystem at a global scale, forest fires alter the biogeochemical cycle and net carbon balance, disturb forest structure and lead to long-terms changes in soil properties. With the great attention paid to climate change and fires, both national and international forums have attached increasing importance to preventive measures of wildfire suppression.
The vegetation will reduce stomatal conductance in reaction to increasing water stress conditions through transpiration regulation mechanism, leading to an increasing canopy temperature. Moderate Resolution Imaging Spectroradiometer (MODIS) land surface reflectance can detect the anomalies with respect to temperature and land surface temperature (LST) is often derived to analyze vegetation stress conditions. Since temperature is a main controller of fuel moisture content that associated with fire ignition, spread and other fire behavior, LST is expected to have some relationship with wildfires. This study will take fuel types and fuel reaction to the temperature into consideration as well.
Fire events happened in Cariboo Region, BC in 2017
Methods
Results and Discussion
LST anomaly value on different days prior to fire events. LST_DOY1, LST_DOY2 and LST_DOY3 represent 1,2,3 days prior to fire events respectively. LSTDOY represents the days for fire occurrence.
Evaluating the effect of anomaly values prior to fire events
We can see from the boxplot that as the fire day approaching, the median of LST anomaly is growing, which indicates the temperature is getting warmer. Through the ANOVA test, a small p value suggests that a significant relationship is found between the anomaly value before and on the fire events day.
LST Anomalies VS. Fire Frequency
Through comparing the expected and observed fire distribution. We can see that fires are more likely to happen when the temperature is hotter than average.
It is likely for us to predict fire occurrence based on daily LST observations when we see an increasing trend of anomaly, and the predicted value falls into a range that fire occurs frequently.
Evaluation on the Impact of Fuel Types
There is no significant difference between fuel type groups in terms of slope of change in LST anomaly. The relationship is not obvious in this study may be caused by limited data obtained. But we can still see some characteristics such as the fuel type matted grass is more associated with negative slope of change and this suggests that grass needs less heat to trigger fire.
Future Work
The remote sensing technique has higher spatial resolution that complementary to typical fire danger rating systems based on meteorological data. We hope the combination can facilitate local forest fire management planning and optimization of resource allocation for fire prevention geographically. What needs to be mentioned is that the approach in this study is observational, further studies are expected to investigate in more regions to test the predictive capability of the developed approach. Moreover, factors relating to topography, weather and fuels are worth to be integrated into analysis.
