Research On Dynamic And Prediction Model Of Dead Fuel Moisture Content Of Typical Stands In Great Xing’ An Mountains

The Great Xing’an Mountains of Heilongjiang Province is the largest forest district with most severely damaged by forest fires in China. It is very important to predict surface dead fuel moisture content accurately and effectively in this region and significant for the forest fireproof working to improve the accuracy of the prediction of fuel moisture content.Measurements were made on moisture content of fuels (Larix gmelinii, Betula platyphylla and Mixed forest of Larix gmelinii-Betula platyphylla) under different air temperature and humidity in the laboratory. This analysis was conducted to determine the effects of air temperature and relative humidity on the equilibrium moisture content (EMC) and time-lag. The Nelson and Simard methods were select to establish the fuel moisture content prediction models in this paper through this research.The aim of this paper was to predict surface dead fuel moisture content by using field data which were obtained from seven typical forest types in Xilinji Forestry Bureau and Nanwenghe Reserve of Great Xing’an Mountains:mixed forest of Populus davidiana-Betula platyphylla、Pinus sylvestris Linn、Larix gmelinii、Betula platyphylla、mixed forest of Larix gmelinii-Betula platyphylla、Quercus mongolica Meadow in wetland. Dynamic of surface dead fuel moisture content and meteorological elements were studied by using ecological factors which include different aspects (north-facing slope and south-facing slope)、different slope-position (high slope-position、mid slope-position and low slope-position) and different seasons (spring and autumn)..The accuracies of three daily-based prediction methods (Nelson, Simard, meteorological elements regression) were studied with different factors. The fuel moisture content prediction models of ten stand types were established. The accuracies were assessed by the analysis of extrapolation errors. and the influences of different elements on the predictive accuracy of the model were analyzed. The hourly-based fuel moisture content prediction models (Larix gmelinii, Betula platyphylla and mixed forest of Larix gmelinii-Betula platyphylla) were established to analyze the effect of models on smaller time scales for the accuracy. The followings are the conclusions:1、Air temperature was negatively correlated with EMC and time-lag, whereas relative humidity was positively correlated with EMC and time-lag. The performance of Nelson and Simard model which could apply in the prediction model establishing in this paper.2、Geographical heterogeneity in litter fuel moisture content was strong in different condition. The fuel moisture content of Larix gmelinii on north-facing slope were maximum and Pinus sylvestris Linn on south-facing slope were minimum. There was significant correlation between different stand fuel moisture content and different meteorological elements. The response of fuel moisture content for the meteorological elements showed obvious time- lag.3、Fuel moisture content prediction models:(1)In general the predictive accuracy of Simard model was best and meteorological element regression model were worst. (2) The predictive accuracy of mixed forest of Populus davidiana-Betula platyphylla models were the best, and the Meadow in wetland models were the worst. (3) The spring models were the lowest and got a best modeling effect. (4) The models of south-facing slope got the best effect. (5) On south-facing slope, the predictive accuracy of models decreased with the increase of slope-position, the accuracy of low slope-position model was best. The variation in north-facing slope models was contrary to south-facing slope models.4、Compare with the daily-based prediction model, the hourly-based model got a better modeling effect, and could be able to better meet the requirements of the prediction accuracy of surface dead fuel in fire behavior forecast.5、The models based on the data from the experiment, is more accurate than the models of meteorological data from Mohe Station and Jiagedaqi District Station, which were farther. The meteorological data in experimental site could better reflect that how fuel moisture content changed really.In summary, meteorological elements、aspect、slope-position、forest type、time-scale of prediction model affected the dynamic of fuel moisture content and model accuracy. Three methods, Nelson、Simard and meteorological elements regression could predict surface dead fuel moisture content for typical stands in Great Xing’an Mountains. Thus,it is necessary to establish different models according to the various forest type and micro terrain changes.