攀西林火时空分布格局与气候影响因素的空间异质性研究

基于四川攀西地区2004—2020年防火期林火灾情数据和气象数据，对攀西林火的时间分布特征、空间分布规律、主要气候影响因子分布趋势与格局进行挖掘；利用ArcGIS格网法，结合双变量空间自相关分析方法，探讨气候影响因子对攀西林火发生的影响机制。结果表明： 2004—2020年攀西地区林火发生次数随时间发展呈现上升趋势，其中2004—2014年间，火灾次数和过火面积波动增加；2015—2020年火灾次数显著下降，但过火面积呈增加趋势，攀西地区林火主要发生在河谷盆地和丘陵地带，且多集中在丘陵的边缘地带。攀枝花市是林火频发区，凉山州次之。不同级别林火对不同时期气候影响因子的响应存在差异，火险期最高气温、降水、相对湿度及连续无降水日数的相关性显著，而非火险期连续无降水日数和平均风速的相关性较高，各气象要素与林火的逐月滞后相关也存在差异，表明前期气温、降水通过影响地面可燃物进而影响林火发生，持续性的干旱极有可能引发火灾。空间相关显示，林火次数与气温、蒸发、连续无雨日数的高高值主要分布在攀西地区的河谷盆地及丘陵区域，降水、相对湿度与林火的低高值聚集在北部的山原区，平均风速、干旱指数与林火的低高值和低低值聚集在西部山原区，表明攀西地区河谷盆地及丘陵区域的高温干旱使林火呈现聚集效应；而攀西区域南部的山地区，由于海拔较高，人为活动少火源较少，虽然是高温干旱区，也未出现林火聚集效应。

Study on Spatio-temporal Distribution and Heterogeneity of Climate Forces of Wildfires in Panxi Region

According to the data of wildfires and meteorological data in Panxi region from 2004 to 2020, this study firstly explored the spatial-temporal pattern of wildfires in Panxi region from temporal variation, spatial distribution, and distribution pattern of the main climate driving factors. Meanwhile, the spatial and temporal correlation between wildfires and meteorological factors in Panxi was studied by using ArcGIS grid method and bivariate spatial auto correlation analysis method. The result were as follow: From 2004 to 2020, the number of wildfires in Panxi region showed an increasing trend. From 2004 to 2014, the number of fires and the burned area fluctuated and increased. From 2015 to 2020, the number of fires decreased significantly, but the burned area showed an increasing trend. From the perspective of spatial distribution, the wildfire in the Panxi area mainly occurred in the valley basins and hilly areas, and are mostly concentrated on the edge of the hills. Panzhihua City was an area with frequent wildfires, followed by Liangshan Prefecture. From the perspective of time series correlation, there were differences in the responses of different levels of wildfires to meteorological factors in different periods. The highest temperature and precipitation in the fire danger period, relative humidity, and the number of consecutive days without precipitation were significantly correlated with the number of wildfires. The number of consecutive days without precipitation and average wind speed in the non-fire danger period were significantly correlated with the number of wildfires. There were also differences between meteorological elements and monthly lag of wildfires. Climate factors had a lag effect on the occurrence of wildfires in Panxi. Temperature and precipitation in early stage affected the occurrence of wildfires by affecting the ground fuel, and continuous drought was very likely to trigger fire. From the spatial correlation, the High−High value of wildfire frequency, air temperature, evaporation and continuous days without rain were mainly distributed in the valley basin and hilly region of Panxi region. The Low−Low value of precipitation, relative humidity and wildfire were distributed in the western Panxi, while the Low−High values were in the northern mountain plain region. The Low−High value of mean wind speed, scPDSI and wildfire were distributed in the western, while the Low−Low value were distributed in northeastern mountain plain. The high temperature and drought in valley basin and hilly area in Panxi region made wildfire showed aggregation effect. However, in the southern mountain area of Panxi region, due to the high altitude, there were few artificial activities and few fire sources. Although it was a high-temperature arid area, there was no wildfire aggregation effect.