Fire weather and lightning fires in a changing climate

点击翻译按钮获取中文摘要

Wildland fire is a common occurrence in Canada with 8000 fires burning 2 million hectares on average every year during the past decade. Most of this area burned results from stand-renewing high-intensity crown fires. These high intensity fires can have significant impact on communities; for example, the fires in Slave Lake in 2011 and Kelowna in 2003 sustained damages measured in 100s of millions of dollars. This research will focus on three related studies examining fire weather, fuel moisture, lightning, fire occurrence prediction and climate change. The objectives of these studies are to: 1) develop a blended method using weather stations, radar and lightning activity to estimate total storm precipitation across the landscape through the interpolation of weather stations observations, the level of lightning activity ( as a proxy for precipitation) and precipitation radar data, 2) develop and determine the accuracy of a 1-3 day lightning forecast model for Canada, and 3) develop a 1-3 day lightning-caused fire occurrence prediction system for Canada and then use this prediction system along with General Circulations Models of future climate to estimate the lightning and the lightning – caused fire occurrence for the rest of the 21st century. One of the largest challenges with the fire weather is determining the weather conditions between weather stations and in particular the location and amount of precipitation. During the fire season much of the precipitation is convective in nature; showers and thundershowers that are highly variable in space and time. Fuel moisture in the Canadian Fire Weather Index System is determined by observations of temperature, precipitation, wind speed and relative humidity. At this time, we are addressing only precipitation because it is the source of the largest uncertainty when calculating fuel moisture and fire danger across the landscape. We will develop improved precipitation estimates from the weather observations, radar and lightning activity. Lightning-caused fire occurrence prediction models have been available for years and most models use observed lightning activity from the last 24 hours and observed fuel moisture to estimate fire starts. Forecasts for the next day or two are possible but these forecasts are for hold-over fires (fires that smoulder underground but don’t progress to flaming combustion until the weather conditions are conducive) based on the observed lightning and forecasted fuel moisture derived from weather forecasts. This current approach is limited as it does not consider additional lightning in the fire occurrence prediction. A three day lightning forecast model will be developed for Canada. A lightning-caused fire occurrence prediction system will be developed for Canada. Using this fire occurrence prediction system and the three day lightning forecast model along with General Circulation Model data we will model lightning and lighting-caused fire occurrence for the rest of the 21st century. The proposed work will lead to advances in fire research that will result in meaningful changes in day-to-day fire management activities that will help protect Canadians and our communities.