The Relationship of the Terra MODIS Fire Product and Anthropogenic Features in the Central Siberian Landscape

Abstract

Fires are a common occurrence in the Siberian boreal forest. The thermal anomalies product of the Terra/Moderate Resolution Imaging Spectroradiometer (MODIS) product suite is designed to detect thermal anomalies (i.e., hotspots or fires) on the Earth's surface. Many, but not all, of the hot spots detected by MODIS in Siberia are wild fires. Agricultural burning and industrial activities also contribute. Using MODIS data from the years 2001, 2002, and 2003 along with a geographical information system (GIS), the type, extent, and duration of hot spots were examined. In addition, high correlations were found between the number of fires and proximity to human activities. Different types of land-cover thermal anomalies were found to have a strong, positive correlation with some anthropogenic features, such as roads, human settlements, and mineral industry locations. The agricultural (r2 = 0.95) and the forest (r2 = 0.81) thermal anomalies had the highest positive correlation with proximity to roads. The correlation was stronger between burned forests and roads (r22001 = 0.81, r22002 = 0.90, r22003 = 0.88) than between any forested land and roads (r 2 = 0.52). The results indicate that forest fires tend to occur near agricultural fires (r22001 = 0.93, r 22002 = 0.87, r 22003 = 0.94). Anthropogenic feature and land-cover thermal anomaly (LCTA) relationships tend to be stronger in a high fire year (2003) than in a low fire year (2001). This alone does imply causation, but might be an indicator of natural and anthropogenic factors acting together to shape where and when fires are burning. These findings have important implications for carbon and climate modelers wishing to use MODIS products to quantify and predict carbon storage and climate change.