The Predict Model And Dynamic Variation Of Litter Moisture In Nanwenghe Natural Reserve

This paper is instructed by the fuel moisture machine, a constantly monitoring of 1-h time-lag stick moisture of five stands (Meadow in wetland, Larix gmelinii, Quercus mongolica Fischer, South-facing slope of Larix gmelinii-Betula platyphylla and North-facing slope of Larix gmelinii-Betula platyphylla) in NanWenghe Natural Reserve was conducted during the fire season period of spring and autumn in 2015. Combined with the method of sampling-drying on the moisture of litter, the author aims to obtain a brand new to method test the moisture by the machine instead of sampling-drying. A research about the dynamic variation of meteorological factors as well as the fuel moisture in different types of forest was made to make an analysis of the correlation between different meteorological factors and fuel moisture. In the meanwhile, the paper took the method of meteorological element regression to construct a predicted model for multiple liner regression on the litter in dissimilar types of forest.As a result, fuel moisture became significant different and went up and down within 24 hours. The duration for going up and down will change with the season. In spring, the duration of fuel moisture lose will be longer than that in autumn. Fuel moisture showed hysteresis on the influence of meteorological element. In spring, the hysteresis will be 3 hours and only 2 hours in autumn. While the fuel moisture machine used in the paper could make a continuous monitoring of 1-h time-lag stick moisture. But there are some differences on the moisture between the litter and the stick. On the basis of the correlation of moisture between 1-h time-lag stick and litter, this paper made a predictive model for moisture of litter instructed by the stick. The MAE of the model was between 14.71% and 20.75%. The average was 17.22%, MRE was between 17.11% and 30.37%, the average was 21.77%. This model could account for 59.4% to 90.0% of the variation of fuel moisture. By the analysis of the correlation of moisture among different meteorological factors and litter, it could be concluded that the moisture of local litter was highly influenced by the rainy and air humidity. Air humidity showed small influence on the litter moisture. A sound result was found to establish local litter moisture prediction through the method of meteorological element regression. The predictive model MAE was between 16.21% to 22.92% and the average was 5.95%, MER was between 28.31% and 35.25% and the average was 31.13%. This model could account for 52.0% to 71.0% of the variation of fuel moisture. The results further provides data base and reference methods on research litter moisture content and spatial distribution pattern of simulation where in Daxing’an Mountain, the research has a very important practical significance.