| Two spatial scale study areas are selected in this paper, the large scale is China and typical province, the small scale is Daxinganling Mountains. Using fire statistics, meteorological data and climate change scenarios data of different time scales to study spatial and temporal patterns of forest fires, response of forest fires to climate change. In this paper, we studied the suitability and sensitivity of FWI in China, studied the trend of meteorological factors, FWI system components, fire occurrence and FWI spatial pattern in 2010-2100 under SRES A2, B2 climate change scenarios. Build fire occurrence probability models of Daxinganling Mountains, and given a quantitative method to determine fire season.Freaky snow and ice storms have plagued southern China since mid-January to mid-February 2008, this long time snow and ice damaged forest ecosystem severely, and also greatly impacted forest fire occurrence and firefighters'safety. We studied forest fire hotpot distribution from satellite image, forest fire occurrence, firefighters injured and dead number, the impact of weather on fire occurrence, and assessed the impact of snow and ice damage on fire occurrence and firefighting safety in Hunan Province. We used MODIS 1B dataset, ground truth data and plant distribution data to classify the fuel into Broadleaf Forest, Masson Pine Forest, Chinese Fir Forest, Mixed forest, Other Coniferous Forest, and Bamboo Forest. We classified the damage level of forest based on ground truth and remote sensing analysis, and analyzed the increased or decreased amount of different fuel. We classified the damage level into 4 classes based on the difference of NDVI change before and after the snow damage.(1)The starting value of FWI system has great impact on the results of FWI system components, the impact days of starting value are influenced by physical properties, time lag days, and size of different components which represented. The minimum days that starting value influenced is FFMC and the ISI, the impact days are 39, the longest is DC, more than 366 days.(2)The suitability and sensitivity of FWI in different provinces in China is different, the most applicable provinces are Yunnan province, Heilongjiang, Inner Mongolia Autonomous Region, for Hunan, Zhejiang, Jiangxi, Fujian, Hubei, Anhui, Jiangsu Provinces, FWI are not suitable. Most of the provinces which are sensitive to FWI are the Southwest Forestry provinces, including Yunnan, Guizhou and Guangxi, followed by Heilongjiang, Jilin, Liaoning, Inner Mongolia, Sichuan and other provinces, the high sensitivity means that small changes of FWI will have a great impact on fire occurrence.(3)Forest fire number of China has obvious fluctuation, the main cycle is 8.29a, and the second cycle is 29a. In recent years, the lightning fires, foreign fires showed an increasing trend, distribution of fire casualties in different provinces has the same pattern with slope of China, the complexity of the terrain is the main factor to cause different casualties level in different provinces.(4)2010-2100, forest fires in China has the tendency to increase, forest fire number will increase more than 70% from 2010s to 2090s under A2 and B2 scenarios in Heilongjiang province. In Yunnan province, the increase percentage is great different in A2 and B2 scenarios. Forest fire number will increase 26.29% under A2 scenario, and forest fire number will increase 60.02% under B2 scenario. The overall growth rate FWI in Heilongjiang is greater than in Yunnan province. The FWI has the fastest growth rate in May and June. In Yunan province, the fastest growth rate of FWI is from January to May under B2 scenario and the fastest growth rate of FWI are January and February under B2 scenario.(5)There are two obvious cycles in Daxinganling Mountains, the main cycle is 17a and the sub-cycle is 6a. The fire cycle of Daxinganling is 236a. The spatial distribution of lightning fires and human caused fires is aggregation distribution, and human caused fires have more intense aggregation degree. Most of lightning fires distribute between 49o-54onorth latitude and 120o-127o east longitude. Human caused fires distribute along railway and roads, the southeast part of Daxinganling Mountains has the most intense aggregation degree. The aggregation degree of human caused fires and lightning fires is correlated with study scale, and the aggregation degree will change along the change of study scale. Altitude is correlated with the distribution of human caused fires and lightning fires, the lightning fires distribute in higher altitude than human caused fires as a whole. The near distance to roads, railways and residential places is one of the indicators to measure the ability of human activities. With the increasing of distance, the human caused fire number reached the maximum value at 1606m, and then fire number decreased fast with the distance becomes longer.(6)In Daxinganling Mountains, fire season days increased, and fire restricted season days has more significant tendency to increase. The difference of the first day of fire season and fire restricted season decreased obviously. Fire number after 1987 decreased dramatically, and fire number increased from 2000, and the increased mount is not equal in different month. The increased total fire number in each year after 2000 is due to the increasd fire number in June and August; especially the fire number has the fastest increasing rate in August. The percentage of lightning fire of total fire number increased greatly.(7)In Daxinganling Mountains, the total fire occurrence probability increased 95.66%, human caused fire occurrence probability increased 71.07%, lightning fire occurrence probability increased 292.98%, and big fire occurrence probability increased 118.05% from 2010s to 2090s under the A2 scenario. The total fire occurrence probability increased 41.42%, human caused fire occurrence probability increased 33.11%, lightning fire occurrence probability increased 84.18%, and big fire occurrence probability increased 34.56% from 2010s to 2090s under the B2 scenario. The lightning fire has the maximum increased probability both in these two climate change scenarios, the lightning fires will become more severe in the future.(8)61.00% hotspots distribute in the area damaged by snow and ice. Fire number and burned area is extremely high in March 2008, fire number is 3097, burned area is 23227.68ha, fire number is over the total of 1999 to 2007 in March, accounts for 120.65% of the total fire number of 1999-2007 in March, is 10.86 times of the average fire number of 1999-2007 in March. Burned area accounts for 88.40% of the total burned area of 1999-2007 in March, and are 4.69 times of the average burned area of 1999-2007 in March. The number of injured and dead fighters is 40, accounts for 72.73% of the total number of 1999-2007 in March, and is 6.56 times of the average number of injured and dead of 1999-2007 in March. The abnormal increase of fire number, burned area and the number of injured and dead go beyond the limit of weather change impact on forest fire occurrence. The number of injured and dead has significant correlation with fire number and burned area. Air temperature and rainfall does not have significant impact on fire number and burned area. The extreme increase of fire number, burned area and the number of injured and dead are impacted by weather, human activities under the background of great change of fuel accumulation, flammability, and fuel structure. The area of class I is 400.07×104ha, account for 41.68% of total area, the area of class II is 403.95×104ha, account for 41.93% of total area, class III is 100.90×104ha, account for 10.42% of total area, and class IV is 57.76×104ha, account for 5.96% of total area. After the damage the surface fuel loading increased dramatically, the increased amount is different for different fuel and different damage level. The maximal increased times of surface fuel is 32.81, and the maximal fuel loading is 142.82t/ha.
|
PDF Full Text Request