摘要
通过统计分析理论研究了黑龙江省林火发生规律,并建立了基于气象因子的森林火灾面积预测模型.结果表明:兴安落叶松林区林火主要发生在4—6月和10月,阔叶红松林区林火主要发生在3—6月和10月;利用林火发生当日的平均风速、相对湿度和平均温度的取值范围可知,兴安落叶松林区发生高等级林火概率较大的月份依次为4月、5月和6月,阔叶红松林区则依次为5月、4月和3月.所建模型的平均精度达到63.3%,能够较精确地预测林火发生后林地可能的过火面积.
Based on statistic analysis theory, the occurrence patterns of forest fire in Heilongjiang Province were studied, and the prediction model for forest fire-burnt area was established based on meteorological factors. The results showed that most of forest fires in Larix gmelinii forest region occurred from April to June and in October, and those in broadleaved Korean pine forest region mainly occurred from March to June and in October. By adopting the values of average wind speed, relative humidity and mean temperature, it was predicted that in L. gmelinii forest region, the first three months with greater probability of larger fire occurrence were March, May and June in order, while in broadleaved Korean pine forest region, they were May, March and April. The average precision of the model was 63.3%, suggesting that it could be used to predict the burnt area by forest fires.
出处
《应用生态学报》
CAS
CSCD
北大核心
2007年第12期2705-2709,共5页
Chinese Journal of Applied Ecology
基金
国家自然科学基金项目(30471404)
黑龙江省自然科学基金项目(ZJD04-0102)
黑龙江省科技攻关资助项目(GB05B602).
关键词
多元回归分析
气象因子
林火面积
预测模型
multi-regression analysis
meteorological factor
burnt area
prediction model
