Analysis Of Factors Affecting Snow Melt And Snow Depth Forecast In Heilongjiang Province

Northeast China is the region with the largest seasonal snow storage in the country.In the context of global warming and particularly significant warming in Heilongjiang Province,the possibility of land erosion and flooding disasters caused by snow melt has increased significantly,and it is especially important to analysis the influencing factors of snow melt and make daily forecasts of snow depth.In this paper,we selected five influencing factors,namely temperature,wind speed,depth,total radiation and black carbon content,and conducted an orthogonal experiment to simulate snow melt indoors using an artificial climate chamber,and analyzed the main influencing factors of snow melt using the experimental data.BP neural network models were developed to predict the snow depth at each station using 12 sets of daily observations of precipitation,air temperature,daily maximum air temperature,daily minimum air temperature,average wind speed,maximum wind speed,sunshine hours,surface air temperature,daily maximum ground temperature,daily minimum ground temperature,average relative humidity,minimum relative humidity and daily snow depth at Mohe,Longjiang,Tieli,Hulin and Harbin stations with spatial typicality from 1991 to 2020.The following conclusions were obtained.(1)The analysis of the orthogonal test data showed that temperature and snow depth had significant effects on snow melting time,and the contribution rates of temperature and snow depth to snow melting time were 47.0% and 43.1% respectively,of these,temperature had the greatest effect on snow melt;the effect of wind speed,total radiation and black carbon content on snow melting time was: wind speed > black carbon content > total radiation,and the contribution rates of wind speed,total radiation and black carbon content to snow melting time were 6.2%,1.8% and 1.9%,respectively.(2)Based on the simulation results of BP neural network model,the 12 indicators of Mohe station with higher weights were daily maximum ground temperature,daily minimum ground temperature,surface temperature and air temperature,with weights of 6.126,5.061,3.404 and 2.269,respectively;the 12 indicators of Longjiang station with higher weights were daily maximum ground temperature,daily minimum ground temperature,surface temperature,and daily minimum air temperature,with weights of3.335,2.866,2.218,and 2.090,respectively;the 12 indicators of Tieli station with higher weights were daily minimum ground temperature,daily maximum ground temperature,surface temperature,and daily minimum air temperature,with weights of1.268,1.064,0.969,and 0.858,respectively;the indicators of the 12 indicators in Hulin station were daily minimum ground temperature,daily maximum ground temperature,average wind speed,and daily maximum air temperature,with indicators of 1.551,1.244,0.941,and 0.939,respectively;the indicators of the 12 indicators in Harbin station were daily minimum temperature,daily minimum ground temperature,daily maximum ground temperature and surface temperature,with weights of 3.278,2.851,2.561 and 2.471,respectively.(3)The daily maximum ground temperature and daily minimum ground temperature accounted for greater weights among the 12 indicators at Mohe,Longjiang,Tieli,Hulin and Harbin stations,with the weights of 6.126,3.335,1.268,1.244,2.561 for the daily maximum ground temperature and 5.061,2.866,1.064,1.551,2.851 for the daily minimum ground temperature,respectively;the index with smaller weight was the maximum wind speed,in which the weights of the maximum wind speed of each station were 0.509,0.587,0.819,0.824 and 1.320 respectively.(4)The daily change of snow depth can also reflect the snow melting situation.The above conclusions together indicated that temperature was the dominant factor affecting snow melting and was also the preferred indicator for snow depth prediction modeling,while wind speed had a smaller effect on snow melting and could be used as a secondary indicator in modeling.This paper mainly clarified the dominant factors of snow melt,provided quantitative data for the influence of various factors on snow melt,and used BP neural network to predict snow depth,which provided some theoretical support for future snow depth prediction in Heilongjiang Province,and also provided suggestions for road safety and flood forecasting in Heilongjiang Province.