Impact Assessment Of Flood Disaster By Integrating Multi-source Remote Sensing Data

Rainstorm and flood disasters have the characteristics of sudden and wide range,which are easy to cause serious damage to buildings,farmland and ecological vegetation.Therefore,it is particularly important to obtain the spatial distribution information of rainstorm and flood disasters.With the development of remote sensing technology,the limitations of traditional ground monitoring of disaster information have been broken through,and dynamic and continuous monitoring can be carried out,with the characteristics of large amount of information obtained and high efficiency.However,there are still the following shortcomings in using remote sensing data to assess flood and waterlogging disasters: the evaluation object has certain limitations,and it is mostly carried out within the scope of impact on human economic activity areas;The assessment data is relatively single,and the acquisition of flood information and post disaster assessment are not comprehensive enough;The research and prediction of integrating multiple remote sensing data on the spatial scale is less.Therefore,this paper takes the "July 20" rainstorm and flood disaster area in Henan Province as the research area,integrates land use/cover data,water system data,remote sensing data and statistical yearbook data,and carries out the assessment of rainstorm and flood disaster in human economic activity areas and natural vegetation growth areas at macro and micro scales through data analysis,fusion and modeling.The main research contents and conclusions are as follows:(1)Research on flood disaster assessment methods.Based on the characteristics of flood disasters,this study selected two directions: economic and ecological vegetation to construct an evaluation index system.The fusion of nighttime lighting data and optical remote sensing data was discussed,and it was found that the human settlements index overcorrected the saturation of the immediate urban core area,while the normalized difference urban index did not consider the significant correlation between human economic activities and land use types.Further,this study proposes the fusion of land cover data and a normalized difference index to generate a normalized difference index of land cover,which is used to identify population aggregation points scattered on cultivated land,forest land,and grassland,refine the spatial distribution of population,and provide spatial details within the city,providing data and theoretical support for the subsequent prediction of population and economic disaster levels.(2)Social impact assessment of flood disasters.The suitability evaluation of population and GDP models was conducted through relevant models,and the results showed that the polynomial model was optimal.Therefore,using the land cover normalized urban index and GDP and population to achieve spatial modeling,a spatial density map with a spatial resolution of 10 meters was obtained for the GDP and population of Zhengzhou,Xinxiang,Anyang,and Hebi cities.The randomly selected resident population of 14 streets in the research area of four cities in Henan Province was used as validation data,The results show that the overall error range is controlled between-0.3 and 0.3;By creating an observation time series for each pixel in the study area for modeling and decomposition,and predicting each pixel,the results showed that TNL did not significantly change in most regions compared to internationally observed NPP-VIIRS-DNB images.However,after the flood,TNL brightness significantly decreased at 80% and 90% confidence levels,while recovering slowly;Using the normalized difference urban index ratio before and after the disaster to calculate the degree of power failure and delineate the risk zone,the results show that the highest affected area and economic proportion of Zhengzhou City is Xingyang City,and the highest affected population proportion is Weihui City;By overlaying the distribution map of land use types in high-risk and medium risk areas,it can be seen that high-risk areas are mainly distributed in built-up areas,a small portion of cultivated land,and woodland areas.Due to the lack of data on night lighting data in areas without lighting,the assessment of vegetation and human activity areas in areas without lighting is limited,and the observed disaster situation of farmland and forest land in each area is lower than the actual disaster area,which will be detrimental to the overall grasp of disaster information and the rush to harvest and plant natural vegetation.Therefore,it is necessary to carry out corresponding assessments at the level of natural vegetation.(3)Natural vegetation level impact assessment of flood disasters.In order to reduce the impact of seasonality,periodicity,and data quality,this study introduced the difference between the average change rate of Enhanced Vegetation Index(EVI)in2019 and 2020 over the same period of time to assess the degree of vegetation damage in 2021.The results showed that the total decline in EVI in Hebi City was the most significant,and the recovery of some EVIs in the main urban area of Zhengzhou City was the best,mainly due to the fact that the vegetation in this area was mainly dominated by shrubs and trees Located in a plain area with relatively developed water systems;By calculating the vegetation coverage recovery rate in the study area,the results show that more than half of the damaged areas have a recovery rate of over 0.75.Due to factors such as terrain,vegetation type,altitude,and mountain slope,secondary disasters are prone to occur after flood disasters,resulting in 4.84% of the damaged areas not being restored,but the vegetation coverage continues to decline.Therefore,timely replanting or emergency harvesting should be carried out;By overlaying areas with poor or poor vegetation restoration with different land use types,it is concluded that MODIS-EVI has a better evaluation effect on cultivated land,grassland,forest land,and other vegetation types.Compared with NDUI monitoring,the MODIS-EVI monitoring errors between the predicted cultivated disaster area and the actual disaster area in Zhengzhou City,Anyang City,Hebi City,and Xinxiang City are-6.40%,7.44%,-10.06%,and 27.80%,respectively,The accuracy is much higher than NDUI’s land use type disaster damage prediction.