Associated Projects - University of Leicester

Assessing fires in tropical peatland: A combined Modis/DMC and Landsat approach

The focus of Jasper Beston’s MSc study (2007/8) was on a tropical peatland region in the southeast of Kalimantan, Borneo.  Jasper used MODIS hotspot data counts and burn scars created using DMC and Landsat data to create a relationship which can be used to calculate a value of the average burn area per hotspot (AR).  He investigated and used two methods. The first is referred to as the regional method, and examines the study area as a whole.  The second method is referred to as the within-region method, and creates a relationship for the burn scars and fire hotspots at a smaller scale. Two different levels of hotspot detection confidence were also examined, in order to ascertain which confidence levels best describe fires in a tropical peatland.  Jasper found that the burn scars were best described utilising all hotspot detection confidence levels.  Results for the estimation of (AR) were found to vary between 48.0ha and 57.7ha, and so the largest of these values was used, in order to present a worst-case scenario for burn emissions detection.  This result was then compared to other estimates in the literature produced for the value of AR, and further potential directions of research are examined.

This image shows the correspondence of active fires and burned area for a region called Block C in central Kalimantan in Borneo. Burnt area are shown in brown. Active fires are shown in red which has been made transparent so that regions of correspondence can be seen.

Web Fire Mapper information plus modis data (http://modis-fire.umd.edu/data.asp) were subsetted and reprojected into UTM format using the Modis Image Processer developed in house on the SPLINT SUN workstation in F55b.

Staff involvement: Kevin Tansey