The dry and windy climate and low ground cover in spring in the black soil region of Northeast China make the soil strongly affected by wind erosion,which seriously threatens the food security and ecological security ...The dry and windy climate and low ground cover in spring in the black soil region of Northeast China make the soil strongly affected by wind erosion,which seriously threatens the food security and ecological security of this region.In this paper,based on the daily observation data of 124 meteorological stations in study area from 1961 to 2020,seasonal and monthly wind erosion climate factor(C)in spring(March to May)were calculated by using the method proposed by the Food and Agriculture Organization of the United Nations(FAO),the wind erosion characterization in spring were systematically analyzed based on C by various statistical analysis methods.The results showed that in the past 60 years,spring wind erosion climate factor(CSp)and monthly C of the whole region and each province(region)all showed highly significant decreasing trend,but they began to show rebounded trend in the middle or late 2000s.CSp of the study area showed a significant upward trend since 2008 with an increase of 4.59(10a)^(-1).The main contributors to this upward trend are the changes of C in March and in April.For the four provinces(regions),CSp in Heilongjiang,Jilin,Liaoning and eastern Inner Mongolia all showed rebounded since 2008,2011,2008 and 2009,respectively.The rebounded trend of CSp in eastern Inner Mongolia was the most obvious with a tendency rate of 11.27(10a)^(-1),and its mutation occurred after 1984.The rebound trend of CSp in Heilongjiang Province takes the second place,with a trend rate of 4.72(10a)^(-1),but there’s no obvious time mutation characteristics.The spatial characteristics of CSpand monthly C are similar,showing decreasing characteristics centered on the typical black soil belt of Northeast China.Compared with 1961-1990,in the period from 1991 to 2020,the proportion of high value areas(CSp>35,monthly C>10)has decreased to varying degrees,while the proportion of low value areas(CSp≤10,monthly C≤4)has increased.The trends of seasonal and monthly C in 82.2%~87.7%of the stations show significant decreases at 95%confidence level.CSp is closely related to wind speed at 2m height,temperature difference,minimum temperature and precipitation in the same period,of which the correlation between CSp and wind speed is the strongest,indicating that the main control factor for CSp in the study area is wind speed,but the impact of the change of temperature and precipitation on CSp cannot be ignored.展开更多
为进一步降低样本成本并加快模型收敛速度,提出基于探索和开发的指数加权算法(exponential-weight algorithm for exploration and exploitation,EXP3)和增量微调卷积神经网络(fine-tuning convolutional neural networks,FCNN)的入侵...为进一步降低样本成本并加快模型收敛速度,提出基于探索和开发的指数加权算法(exponential-weight algorithm for exploration and exploitation,EXP3)和增量微调卷积神经网络(fine-tuning convolutional neural networks,FCNN)的入侵检测系统(EXP3-FCNN)。利用EXP3算法自适应选择最佳主动学习策略,代替单一的主动学习算法,提高样本质量;利用增量微调卷积神经网络提取流量数据更深层次的特征;使用AWID数据集作为实验数据。实验结果表明,该方案在保证模型精确度、召回率等性能指标的基础上,降低了样本成本,提高了模型的收敛效率。展开更多
基金supported by the Open Research Fund of Innovation and Open Laboratory of Eco-meteorology in Northeast China,China Meteorological Administration(stqx2019zd02)Heilongjiang Meteorological Science and Technology Research Project(HQGG202004)Heilongjiang Provincial Natural Science Foundation of China(LH2020C105)。
文摘The dry and windy climate and low ground cover in spring in the black soil region of Northeast China make the soil strongly affected by wind erosion,which seriously threatens the food security and ecological security of this region.In this paper,based on the daily observation data of 124 meteorological stations in study area from 1961 to 2020,seasonal and monthly wind erosion climate factor(C)in spring(March to May)were calculated by using the method proposed by the Food and Agriculture Organization of the United Nations(FAO),the wind erosion characterization in spring were systematically analyzed based on C by various statistical analysis methods.The results showed that in the past 60 years,spring wind erosion climate factor(CSp)and monthly C of the whole region and each province(region)all showed highly significant decreasing trend,but they began to show rebounded trend in the middle or late 2000s.CSp of the study area showed a significant upward trend since 2008 with an increase of 4.59(10a)^(-1).The main contributors to this upward trend are the changes of C in March and in April.For the four provinces(regions),CSp in Heilongjiang,Jilin,Liaoning and eastern Inner Mongolia all showed rebounded since 2008,2011,2008 and 2009,respectively.The rebounded trend of CSp in eastern Inner Mongolia was the most obvious with a tendency rate of 11.27(10a)^(-1),and its mutation occurred after 1984.The rebound trend of CSp in Heilongjiang Province takes the second place,with a trend rate of 4.72(10a)^(-1),but there’s no obvious time mutation characteristics.The spatial characteristics of CSpand monthly C are similar,showing decreasing characteristics centered on the typical black soil belt of Northeast China.Compared with 1961-1990,in the period from 1991 to 2020,the proportion of high value areas(CSp>35,monthly C>10)has decreased to varying degrees,while the proportion of low value areas(CSp≤10,monthly C≤4)has increased.The trends of seasonal and monthly C in 82.2%~87.7%of the stations show significant decreases at 95%confidence level.CSp is closely related to wind speed at 2m height,temperature difference,minimum temperature and precipitation in the same period,of which the correlation between CSp and wind speed is the strongest,indicating that the main control factor for CSp in the study area is wind speed,but the impact of the change of temperature and precipitation on CSp cannot be ignored.
文摘为进一步降低样本成本并加快模型收敛速度,提出基于探索和开发的指数加权算法(exponential-weight algorithm for exploration and exploitation,EXP3)和增量微调卷积神经网络(fine-tuning convolutional neural networks,FCNN)的入侵检测系统(EXP3-FCNN)。利用EXP3算法自适应选择最佳主动学习策略,代替单一的主动学习算法,提高样本质量;利用增量微调卷积神经网络提取流量数据更深层次的特征;使用AWID数据集作为实验数据。实验结果表明,该方案在保证模型精确度、召回率等性能指标的基础上,降低了样本成本,提高了模型的收敛效率。
