Evaluation And Remote Sensing Inversion Of Soil Quality Saline-alkali Degradation In Coastal Area Of The Yellow River Delta

The land is the basis of human survival and development.For a long time,due to China's large population and limited resources,the per capita occupancy of resources is lower than the world's average level.To maximize the use of land resources,in recent years,human beings have carried out high-intensity development and utilization of land,which leads to the deterioration of land quality,the imbalance of regional ecological environment,and the degradation of soil is more and more serious.As the youngest land in New China,the Yellow River Delta is affected by the dynamic system,its ecological environment is fragile,and its soil degradation is serious.Soil quality saline-alkali degradation is the main form of soil degradation in the Yellow River Delta.Therefore,it is of great significance to scientifically evaluate soil quality saline-alkali degradation in this region and timely grasp the characteristics of soil quality saline-alkali degradation degree and distribution.In this paper,kenli District and Wudi County in the Yellow River Delta were selected as the study areas,and soil samples were collected to obtain soil and vegetation attribute data.Soil salt content,p H value,mineralization of groundwater,groundwater depth,vegetation coverage,five factors as an evaluation index of saline soil degradation,build the evaluation system of the Yellow River delta saline soil degradation,calculate the saline soil degradation index(SADI),in the littoral area of the Yellow River delta saline soil deterioration status evaluation and analysis.On this basis,relevant soil information was extracted from Landsat 8 images,and spectral parameters were constructed and screened by using SPSS,EXCEL and other data processing software,which were combined into multiple variable groups as independent variables of soil quality saline-alkali degradation inversion model.Classical statistical analysis methods(SLR,CR,MLR)and machine learning algorithms(BPNN,SVM,RF)established inversion models.We were using the data of Wudi county to verify the best remote sensing inversion model through comparison and optimization.Furthermore,the best model was applied to the images of Kenli district and Wudi County in the same period of field survey to invert soil quality saline-alkali degradation.Finally,the model was applied to the remote sensing image of Kenli area in the spring of 2020 to realize the dynamic analysis of soil quality saline-alkali degradation.The main conclusions are as follows:(1)The SADI of soil ranged from 0.1220 to 0.7423 in the study area.According to the cluster analysis results of SADI,soil quality saline-alkali degradation grade was divided into three levels:mild saline-alkali degradation(SADI?0.33),moderate saline-alkali degradation(0.330.53).The mild saline-alkali degradation mainly occurred in the southwest and central areas of the reclamation area,accounting for 32.81%of the total area.Wheat,corn,rice and cotton were mainly cultivated.Moderate saline-alkali degradation areas were distributed in the west,south and northwest of Kenli District,accounting for 44.85%of the total area.In addition to cotton and wheat,some fruit trees were planted.Severe saline-alkali degradation areas were distributed in the east coastal tidal flat of the reclamation area,accounting for 22.34%,and the main plants were thatch,p Hragmites australis and suaeda saline-alkali tolerant plants.(2)Based on remote sensing images of the reclamation area,the sensitive bands and sensitive spectral parameters were screened by correlation analysis.The correlation between NDVI,DVI,NDWI,RVI and SADI is greater than 0.74.A variety of soil quality saline-alkali degradation evaluation inversion models were established using six methods:unary linear regression,unary curve regression,multiple linear regression,BP neural network,support vector machine,and random forest,with the sensitive band and sensitive spectral parameters as independent variables.After verification and optimization,with NDVI and RVI as independent variables,the model constructed based on the BP neural network method was the best inversion model.The modeling R~2was 0.819,RMSE was 0.0736,and validation R~2was 0.526.(3)Based on the best inversion model,soil quality saline-alkali degradation in Kenli district and Wudi County was obtained.In kenli district,there was little difference between the inversion results and evaluation results of different soil quality saline-alkali degradation levels:37.67%of the slightly saline-alkali degradation area,40.53%of the moderately saline-alkali degradation area,21.80%of the severely saline-alkali degradation area.The spatial distribution characteristics of each saline-alkali degradation grade were consistent with the evaluation result diagram,showing that the degree of soil quality saline-alkali degradation in some inland areas of southwest China was relatively light,and the degree of soil quality saline-alkali degradation in the eastern coastal area was relatively heavy.Generally,the farther away from the sea,the lighter the degree of soil quality saline-alkali degradation was.The area of slight saline-alkali degradation in Wudi County was 41.75%,the area of moderate saline-alkali degradation was 45.12%,and the area of severe saline-alkali degradation was 13.13%.(4)Dynamic analysis results of soil quality saline-alkali degradation showed that soil quality saline-alkali degradation in 2020 was significantly reduced compared with 2015.The soil area with mild saline-alkali degradation increased by 11.46%,while the soil area with moderate and severe saline-alkali degradation decreased by 2.44%and 9.02%,respectively.Soil salinity degradation in the middle of Kenli district,along the river,and near the mouth of the Yellow River was improved.Explore this article constructed the evaluation system of the littoral area of the Yellow River delta saline soil degradation,and on the basis of constructing the Yellow River delta littoral area soil salinity remote sensing inversion model of degradation,realized the long contrary to saline soil degradation in the region and dynamic analysis,the waterfront saline soil degradation,real-time,accurate monitoring is of great significance.