• 研究论文 •

### 以小时为步长的大兴安岭典型林分地表死可燃物含水率模型预测及外推精度

1. 1东北林业大学林学院, 哈尔滨 150040;
2中国林业科学研究院森林生态环境与保护研究所, 北京 100091;
3沈阳农业大学林学院, 沈阳 110866;
4吉林省林业勘察设计研究院, 长春 130022;
5国家林业局森林病虫害防治总站, 沈阳 110034;
6哈尔滨市林业科学研究院, 哈尔滨 150028
• 收稿日期:2018-05-10 修回日期:2018-09-19 出版日期:2018-12-20 发布日期:2018-12-20
• 作者简介:于宏洲, 男, 1983年生, 博士, 讲师. 主要从事森林防火与GIS研究. E-mail: yhz-163@163.com
• 基金资助:
本文由国家自然科学基金项目(31700575,31870644,31800609)、国家重点研发计划项目(2017YFD0600106-2)和沈阳农业大学引进人才启动经费项目(2015年度)资助

### Prediction models and the extrapolation effects for water content of surface dead fuels in the typical stand of the Great Xing’an Mountains of China by one-hour time step

YU Hong-zhou1,2, SHU Li-fu2, DENG Ji-feng3*, YANG Guang1, LIANG Qi4, LI Jing-hao5, ZHU Hang-yong6

1. 1College of Forestry, Northeast Forestry University, Harbin 150040, China;
2Research Institute of Forest Ecology, Environment and Protection, Chinese Academy of Forestry, Beijing 100091, China;
3College of Forestry, Shenyang Agricultural University, Shenyang 110866, China;
4Jilin Academy of Forestry Survey and Design, Changchun 130022, China;
5Forest Pest Control Station of the State Forestry Administration, Shenyang 110034, China;
6Harbin Forestry Academy, Harbin 150028, China
• Received:2018-05-10 Revised:2018-09-19 Online:2018-12-20 Published:2018-12-20
• Supported by:
This work was supported by the National Natural Science Foundation of China (31700575, 31870644, 31800609), the National Key Research and Development Plan (2017YFD0600106-2), and the Shenyang Agricultural University Startup Foundation for Introduced Talents (2015).

Abstract: The water content of surface dead fuels is one of the most important indicators for forecasting fire danger and fire behaviors. We employed the timelag equilibrium water content methods (i.e. Nelson and Simard models) and the meteorological variable regression method to continuously measure the water content of surface dead fuels by one-hour time step from September to October in 2010 under Populus davidiana + Betula platyphylla, Picea koraiensis and the cutover lands (Pinus sylvestris var. mongolica + Betula platyphylla) with different canopy densities in Pangu Forestry Bureau, the Great Xing’an Mountains, Heilongjiang Province, China. We established prediction models and obtained prediction errors. The models were also used to extrapolate the water contents of surface dead fuels under other forest stands and the extrapolation accuracy was analyzed. The results showed that the mean absolute error, the mean relative error and the mean square error root of Nelson model (0.0154, 0.104 and 0.0226) were lower than those of Simard model (0.0185, 0.117 and 0.0256). In terms of extrapolation effects, the mean absolute error, the mean relative error and the mean square error root of meteorological variable regression method (0.0410, 0.0300 and 0.0740) were lower than those of Simard model (0.610, 0.492 and 0.846), but they were higher than those of Nelson model (0.034, 0.021 and 0.0660). Such results indicated that the timelag equilibrium moisture content method by one-hour time step, especially Nelson model, was sui-table for the forest stands in the Great Xing’an Mountains. Although extrapolation could not reduce the prediction errors, it could help improve the prediction accuracy and the efficiency of the present models applied to different forest stands or in a larger scale. The modeling and extrapolation errors were closely related to species identity and canopy densities, thus the appropriate timelag equilibrium moisture content methods should be selected according to different forest stands and locations.