CSTR:11738.11.ncdc.isnow.db2716.2023
10.12072/ncdc.isnow.db2716.2023
This data set contains the thematic map of 1:500000 flood disaster types in Xinjiang from 1984 to 2020 and its corresponding mapping data. The thematic map is in jpg format, the data is raster data (tif), and the geographic coordinate system is WGS_ In 1984, the projection was the isoproduct conic projection of Asia North Alberts Equal Area Conic, with a spatial resolution of 500m.
The grid value value represents the flood risk level, which can be used to quantitatively describe the snowmelt flood caused by the rapid melting of ice and snow caused by the high temperature in spring and the rainstorm flood caused by the heavy rainfall in summer and autumn, and the flood risk level and its spatial characteristics caused by these two different causes. This thematic map provides the spatial reference data of snowmelt flood, rainstorm flood and its mixed flood disaster risk level for regional flood control and disaster reduction macro-management. Adopt fully open sharing.
The annual daily precipitation in Xinjiang is ≥ 25mm, the extreme number of days; The extreme value of annual daily precipitation in Xinjiang; The maximum monthly precipitation for many years in Xinjiang is described in the annex.
This thematic map is composed of the "thematic map of 1:500000 snowmelt flood disaster risk level in Xinjiang region" and the "thematic map of 1:500000 storm flood disaster risk level in Xinjiang region", which are superimposed by GIS grid.
Among them, the 1:500000 snowmelt flood risk level thematic map in Xinjiang region is based on the snowmelt flood disaster and its associated natural, social and risk basic information for many years (1984-2020) in Xinjiang, and the model is built by GIS and random forest (RF) methods and completed by GIS mapping. The grid value value is 1. This pixel is the snowmelt flood disaster prone area, which is a high-risk area, and 0 is a risk-free area. The thematic map of risk level of 1:500000 rainstorm and flood disaster in Xinjiang region is based on the basic information of rainstorm and flood disaster and its associated nature, society and risk in Xinjiang for many years (1984-2020). The model is built by GIS and random forest (RF) methods and completed by GIS mapping. The grid value is 1. This pixel is the rainstorm and flood risk-free area, 2 is the low-risk area, 3 is the general risk area, and 4 is the key risk area.
Taking the regional rainstorm and flood disaster in Xinjiang as an example, the basic principle of RF is to assume that there are N rainstorm and flood risk levels and M characteristic factors. First, the bootstrap method is used to randomly select samples from the sample set to build a new training sample set. The samples that are not sampled each time constitute out-of-bag (OOB) data as the test sample set.
Secondly, feature factors are randomly selected from M feature factors. Each classification tree branches from one of the Mtry feature factors with the most classification ability, and each tree grows to the maximum without any pruning. The above process was repeated several times to generate Ntree classification tree.
Finally, the generated Ntree classification tree is aggregated to obtain a random forest, and the mode of the classification result is used as the classification result of the random forest, that is, the risk level of rainstorm and flood. Here, the R language randomForest package is used to build the RF assessment model of rainstorm and flood disaster in Xinjiang. The technical core is to determine the optimal Ntree and Mtry.
The thematic map of flood risk level caused by different causes in Xinjiang region is obtained through processing. The specific processing includes projection conversion, range standardization, GIS and RF calculation simulation, accuracy verification and GIS mapping.
This work is licensed under
CC BY 4.0 (Creative Commons Attribution 4.0 International License).
