| The mositrue content of surface litter of forest dircectly affects the occurrence proibability of forest fire and a series of fire behavior characteristics after being ignited.The research on the dynamic change of moisture content and the prediction model is an important part of current forest fire forecast and prediction research.The dynamic change of surface litter is mainly driven by meteorological factors,and predicting moisture content on the hourly scale can reduce the error and improve the prediction accuracy.Comprehensive analysis of the effects of various meteorological factors on the dynamic change of moisture content of surface litter and the establishment of moisture content prediction model in hours can provide methods and available parameters for the construction of forest fire risk prediction system in China.It is of great significance to provide key support technology for the establishment of high-precision prediction technology of moisture content of surface litter and a new generation of forest fire risk prediction system.This study selected typical forests-Mongolian oak(Quercus mongolica)forest and Korean pine(Pinus koraiensis)forest in the temperature forest ecosystem in Northeast China.The litter of Mongolian oak and Korean pine were selected as the research objects.The fuel bed with different compactness similar to the field structure is constructed in the laboratory to simulate the change of moisture content of the litter bed under different air temperature,relative humidity,wind speed and rainfall conditions.The effects of meteorological factors on moisture content of litter bed with different compactness were obtained,and the relevant parameter models were established.During the spring fire prevention period,non-destructive sampling was used in Mongolian oak and Korean pine forests to continuously monitor the moisture content of surface litter for 7 days in an hour step,and a total of 336 sets of data were obtained.The diurnal variation of moisture content and its response to meteorological factors are analyzed.Direct estimation method(Nelson and Simard method)and regression method of meteorological elements are selected to establish prediction model of moisture content of two types of surface litter,and the error and accuracy of model was calculated and analyze.The main results are as follows:(1)Under constant temperature and humidity,the moisture content of oak and pine decreases exponentially with time.The average equilibrium moisture content of litter bed of oak is 14.4%,ranging from 2.1%to 34.9%,and that of pine bed is 11.9%,ranging from 1.8%to 32.7%.The average drying coefficient of litter bed of oak is 0.104 h-1,ranging from 0.021 h-1 to 0.292 h-1,and that of pine bed is 0.106 h-1,ranging from 0.015 h-1 to 0.325 h-1.Equilibrium moisture content of surface litter beds of oak and pine were significantly correlated with air temperature and relative humidity,and the bed compactness has no effect on it.The drying coefficient of two types of litter beds are significantly correlated with air temperature,relative humidity and bed compactness.Nelson and Simard models are selected to fit the equilibrium moisture content of two kinds of litters,and Simard model is better than Nelson model.Based on the bed compactness and relative humidity as classification conditions,the prediction model of the drying coefficient of oak and pine bed in the form of k=aebT was established,the drying coefficient changed exponentially with the temperature change.The MAE range of the prediction model of drying coefficient of the oak bed is 0.003 h-1 to 0.013 h-1,and MRE range is 3.85%to 12.21%.The MAE range of the prediction model of drying coefficient of the pine bed is 0.003 h-1 to 0.011 h-1,and MRE range is 3.09%to 21.89%.(2)At a fixed wind speed,the moisture content of litter bed of oak and pine decreased exponentially with time.When the moisture content of the bed exceeds 35%,the moisture content decreases sharply,which is mainly evaporation.When the moisture content is lower than 35%,the rate of change of moisture content decreases,at this time,the water is mainly diffused.The drying coefficient of litter bed varies from 0.325 h-1 to 1.646 h-1 in different compactness of oak,it increases first and then decreases with the increase of wind speed,reaching the maximum value at 3 m·s-1.The drying coefficient of litter bed varies from 0.319 h-1 to 1.224 h-1 in different compactness of pine,which is positively correlated with wind speed.The drying coefficients of the two types of fuel are doubled from windless to windy.The prediction model of the drying coefficient of k=Caw2+bw+c was established for the litter bed of oak,when the compactness of the litter bed was 0.0092,0.0138 and 0.0184,the MAE of the prediction model of drying coefficient was 0.035 h-1,0.108 h-1 and 0.085 h-1,and the MRE was 3.6%,13.1%and 14.9%,respectively.The prediction model of the drying coefficient of k=aw was established for the litter bed of pine,when the compactness of the litter bed was 0.0158,0.0236 and 0.0315,the MAE of the prediction model of drying coefficient was 0.079 h-1,0.104 h-1 and 0.096 h-1,and the MRE was 13.6%,21.0%and 20.6%,respectively.(3)Under rainfall conditions,the moisture content of litter bed of oak and pine increased logarithmically with time.The litter bed of oak is more easily saturated and its saturated moisture content is higher than that of pine.When the compactness of litter bed of oak is 0.0092,0.0138 and 0.0184,the longest time needed to reach saturated moisture content is 0.89 h,1.00 h and 1.72 h,respectively;when the compactness of litter bed of pine is 0.0158,0.0236 and 0.0315,the longest time needed to reach saturated moisture content is 1.06 h,1.28 h and 2.06 h.Saturation time and saturated moisture content of litter bed of oak are affected by rainfall and bed compactness,while saturation time of the litter bed of pine is affected by rainfall,bed compactness and initial moisture content,saturated moisture content is significantly correlated with compactness of fuel bed and rainfall.The prediction model of saturation time and saturated moisture content for two types of fuels,such as T=aR+b and Ms=aR+b,are established.When the compactness of litter bed of oak is 0.0092,0.0138 and 0.0184,the MAE of saturation time prediction model is 0.025 h,0.018 h and 0.100 h,MRE is 4.0%,2.4%and 8.2%,and the MAE of saturation moisture content prediction model is 12.9%,3.8%and 3.8%,MRE is 3.0%,1.1%and 1.0%,respectively.When the compactness of litter bed of pine is 0.0158,0.0236 and 0.0315,the MAE of saturation time prediction model is 0.073 h,0.114 h and 0.107 h,MRE is 10.7%,14.6%and 9.2%,and the MAE of saturation moisture content prediction model is 10.3%,10.5%and 33.9%,MRE is 2.9%,3.0%and 11.9%,respectively.(4)The air temperature and relative humidity have a stable diurnal variation,the lowest temperature in the early morning,and then begin to rise to 14:00,continue to decline.The relative humidity changes exactly opposite the air temperature,and there is no stable diurnal variation of wind speed and rainfall.The surface fine dead fuel of oak and pine have a strong diurnal variation law,and the two types fuels have the same variation law.In the absence of rainfall,the moisture content of the litter bed rises from early morning to about 8:40 and reaches its maximum value,then begin to decline.At 13:40-15:40,the moisture content is the lowest,then begins to rise again.The random wind in the course of daily change will make a slight change in the diurnal variation,but the overall trend remains unchanged.Air temperature,relative humidity and rainfall have a significant effect on the change of moisture content of surface fine dead fuels.The change of moisture content has a certain lag to the response of meteorological factors.The bigger the gap between collecting time of meteorological factors and monitoring time of moisture content,the worse the correlation.The time lags of the two types of fuels of oak and pine were 4.24 h,16.34 h and 29.25 h,52.32 h,respectively,by Nelson method and Simard method.For the litter bed of oak,the MAE and MRE of the Nelson,Simard and meteorological element regression methods were 11.04%,8.03%,30.29%and 26.67%,6.45%,79.76%,respectively;for the litter bed of pine,the MAE and MRE of the Nelson,Simard and meteorological element regression methods were 2.04%,2.01%,14.61%and 2.10%,5.27%,70.80%,respectively.The Simard method has the best prediction effect,Nelson method takes the second place,and meteorological factor regression method has the worst prediction effect,whether it is oak or pine. |
PDF Full Text Request