Method for detection and identification of lightning strokes likely to cause wildfires

Lightning and wildfires are natural phenomena prevalent in the Southern African sub-continent. The endemic vegetation evolved to survive these fires. In the past, lightning was the most common trigger of fires, but currently anthropogenic causes account for most fires in occupied areas. Despite this, lightning remains critical, as fires in remote and uninhabited locations are mostly initiated by lightning. These fires have both a significant ecological and socioeconomic impact, as such fires are typically detected at an advanced stage. These fires are difficult to extinguish due to rugged terrain and low accessibility, resulting in large burnt areas. The detection, location and prediction of lightning strokes that are probable to ignite fires is important to protect lives, property as well as ecologically sensitive areas. This process is illustrated by various recent case studies. Of particular concern are smouldering underground fires that are ignited by lightning and may result in fires that occur days, or even weeks later after the initiation event. Methods for detecting and extinguishing these fires have been devised. More than 90% of fires, are started by strokes of positive polarity. Conversely, around 90% of cloud-to-ground strokes are of negative polarity, therefore positive strokes are the focus of the detection process and result in a relatively small analysis set. Lightning stoke data can be combined with satellite detection of vegetation hydration levels to objectively quantify the risk of wildfires. The aim of this research is to provide early detection and location of lightning strokes that are associated with a high probability of causing wildfires. Here, the technique is introduced with limited data as a discussion paper and is being developed and verified with additional data in ongoing further work.