Background Understanding the spatial patterns of forest ﬁres is of key importance for ﬁre risk management with ecological implications. Fire occurrence, which may result from the presence of an ignition source and the conditions necessary for a fire to spread, is an essential component of fire risk assessment.
Methods The aim of this research was to develop a methodology for analyzing spatial patterns of forest ﬁre danger with a case study of tropical forest fire at Margalla Hills, Islamabad, Pakistan. A geospatial technique was applied to explore inﬂuencing factors including climate, vegetation, topography, human activities, and 299 ﬁre locations. We investigated the spatial extent of burned areas using Landsat data and determined how these factors inﬂuenced the severity rating of ﬁres in these forests. The importance of these factors on forest ﬁres was analyzed and assessed using logistic and stepwise regression methods.
Results The findings showed that as the number of total days since the start of fire has increased, the burned areas increased at a rate of 25.848 ha / day (R 2 = 0.98). The average quarterly mean wind speed, forest density, distance to roads and average quarterly maximum temperature were highly correlated to the daily severity rating of forest fires. Only the average quarterly maximum temperature and forest density affected the size of the burnt areas. Fire maps indicate that 22% of forests are at the high and very high level (> 0.65), 25% at the low level (0.45-0.65), and 53% at the very low level (0.25 – 0.45).
Conclusion Through spatial analysis, it is found that most forest fires happened in less populated areas and at a long distance from roads, but some climatic and human activities could have influenced fire growth. Furthermore, it is demonstrated that geospatial information technique is useful for exploring forest fire and their spatial distribution.