湿度码预测江西典型地表细小死可燃物含水率适用性分析

Analysis on the applicability of moisture code to predict the moisture content of surface fine dead fuel on Jiangxi province,China

【目的】分析加拿大火险等级系统中湿度码预测江西典型地表细小死可燃物含水率的适用性,为该地区含水率研究提供基础数据以及更好的理解湿度码在我国的适用性。【方法】以江西省南昌市茶园山林场典型林分马尾松林(MWS)、杉木林(SM)、柳杉林(LS)和毛竹林(MZ)下地表细小死可燃物为研究对象,在2018年10月23日—2019年1月20日(江西省防火期)以非破坏性采样方法对其含水率动态变化进行监测,并同步监测气象要素,利用Spearman偏相关分析地表细小死可燃物含水率动态变化与气象要素和湿度码的相关性,选择气象要素回归法和湿度码回归法建立地表细小死可燃物含水率预测模型,并进行模型外推分析,得到预测模型精度和外推能力,分析湿度码预测江西典型地表细小死可燃物含水率的适用性。【结果】监测共进行90 d,主要得到以下结果:1)地表细小死可燃物含水率动态变化主要受相对湿度和降雨的影响,此外,马尾松还受空气温度极显著影响,柳杉和杉木含水率与风速呈极显著负相关;2)地表细小死可燃物含水率与细小可燃物湿度码、腐殖质码和干旱码都呈极显著负相关,且相关性逐渐降低;3)直接使用细小可燃物湿度码得到的含水率值与所有地表细小死可燃物含水率实测值均有极显著差异,平均绝对误差和相对误差范围分别为43.717%~80.377%和49.576%~151.698%;4)气象要素回归模型的平均绝对误差和平均相对误差范围分别为20.395%~40.765%和37.417%~59.175%,湿度码回归模型的平均绝对误差和平均相对误差范围分别为19.366%~34.372%和33.895%~47.258%,湿度码回归模型误差显著低于气象要素回归模型;5)气象要素回归模型和湿度码回归模型的外推绝对误差均值分别为50.244%和45.824%,湿度码回归模型外推能力要优于气象要素回归模型。【结论】直接使用湿度码预测江西典型地表细小死可燃物含水率并不适用,选择回归法对其校正,虽然较气象要素回归法显著提高了预测精度,但依旧无法满足火险需要。在今后研究中,还需对湿度码中关键参数:平衡含水率和失水系数等进行校正,提高基于湿度码预测地表细小死可燃物含水率的精度。

【Objective】To analyze the applicability of moisture code in Canadian fire risk rating system to predict moisture content of typical surface fine dead fuel in Jiangxi province,so as to provide basic data for moisture content research in this area and better understand the applicability of moisture code in China.【Method】Taking the typical forest stands of Chayuanshan forest farm in Nanchang city of Jiangxi province as the research object,the dynamic change of moisture content was monitored by non-destructive sampling method from October 23,2018 to January 20,2019(Fire prevention period in Jiangxi province),and meteorological factors were monitored simultaneously.Spearman partial correlation was used to analyze the correlation between dynamic change of moisture content and meteorological factors and moisture code.Meteorological factor regression method and moisture code regression method were selected to establish the prediction model of moisture content.The accuracy and extrapolation ability of the prediction model were obtained by extrapolation analysis,and the applicability of moisture code to predict the moisture content of typical surface fine dead fuel in Jiangxi province was analyzed.【Result】The monitoring was conducted for a total of 90 days,and the main results were as follows:1)the dynamic changes of moisture content of fine dead fuel is mainly affected by relative humidity and rainfall.In addition,Pinus massoniana is also significantly affected by air temperature and the Crytomeroa fotunei and Cunninghamia lanceolate is negatively correlated with wind speed.2)the moisture content was significantly negatively correlated with fine fuel moisture code,duff moisture code and drought code,and the correlation gradually decreased.3)the moisture content value by directly using the fine fuel moisture code is extremely different from the measured values of the measured moisture content,and the average absolute error and relative error range are 43.717%-80.377%and 49.576%-151.698%respectively.4)The average absolute error and average relative error range of meteorological factor regression model were 20.395%-40.765%and 37.417%-59.175%,respectively.The average absolute error and average relative error range of moisture code regression model were 19.366%-34.372%and 33.895%-47.258%,respectively.5)The mean absolute errors of the extrapolation of the meteorological factor regression model and the moisture code regression model are 50.244%and 45.824%,respectively.The extrapolation ability of the moisture code regression model is better than that of the meteorological factor regression model.【Conclusion】It is not suitable to directly use moisture code to predict moisture content of typica surface fine dead fuel of Jiangxi Province.Regression method is used to correct the moisture content of fuel,although the prediction accuracy is significantly improved compared with the meteorological factor regression method,it still can not meet the fire danger system requirements.In the future research,the key parameters of moisture code,such as equilibrium moisture content and drying coefficient,should be corrected to improve the accuracy of moisture content prediction of fine dead fuel based on moisture code.