| Jiangsu Province has the most tidal flat resources in China.There are huge radiant sandbars developed offshore.The land shore tidal flats covered by it have frequently changing material sources and strong hydrodynamic conditions.It is an important part of Jiangsu tidal flats with important research value.However,due to the irradiated sandbar's land bank length of about 200 km and the large north-south span,the hydrodynamic environment in different regions is significantly different.Subject to the harsh field survey conditions of muddy tidal flats,there are few studies on the characteristics of the tidal flats on the entire radiant sandbar.The application of remote sensing technology makes up for the limitations of traditional investigation methods and promotes the study of tidal beach sediment distribution and transport laws,but the inversion accuracy of existing methods needs to be improved.Based on this,the irradiated sandbar land bank was selected as the research area,and field sampling and multispectral remote sensing image data were combined,and combined with principal component analysis(PCA,Principal Components Analysis)and wavelet neural network model(WNN,Wavelet Neural Network)to obtain the spatial distribution characteristics and vertical transport trend of the surface sediment information of the tidal flat on the shore of the radial sandbar in 2019.Further select the central coast of Jiangsu(Doulong Port-Wanggang River Estuary)with frequent erosion and siltation changes and development intensity as the typical area,and select multi-period remote sensing images of 2002,2008,2014 and 2019 for four years,combined with the sediments of the past Data,and carried out long-term tide flat surface sedimentary characteristics change analysis.And verify the possible error effect when using the established model to invert sediment information in different years.With a view to extending the remote sensing application of tidal beach sediment in coastal zone and providing the method support and reference basis for rational exploitation and protection of tidal beach resources,the main research programs are as follows:(1)The PCA of image combined with WNN modeling is an effective method for remote sensing inversion of surface sediment characteristics and particle size transport trends in silty tidal flats.The accuracy verification results of the method in the study area and the typical area show that: the mean absolute error of the model test group of sediment size parameters(average distribution,sorting coefficient,skewness)of thestudy area is respectively below 0.41?,0.24,0.32,respectively.The average relative error mean of average particle size is the smallest,the sorting coefficient is the second,the skewness is the largest,it can be seen that the inversion accuracy decreases gradually with the increase of order of particle size parameter,but the variation range of variation coefficient of three particle size parameters remains stable.Among the three sediment component contents(clay,silt,sand),the average absolute error,average relative error,and maximum coefficient of variation were 5.89%,19.33%,and 21.49%,respectively,indicating that the overall difference between the inversion value and the measured value It is stable and the inversion effect of sediment components is better.In the sedimentary particle size parameters of typical areas,the average particle size inversion results are still the most ideal,and the sorting coefficient and skew inversion results are poor;among the sediment components,the inversion results of sand and silt components are the most ideal.The inversion results of clay are poor.(2)The average particle size of the surface sediments on the tidal flat on the radiant sand bar is mostly between 4 and 5?,and the sorting coefficients are mostly in the poorly sorted state,and the skewness is mainly positive and extreme positive.The content of sand component increased from land to sea and the content of silt component decreased,and the clay content generally did not exceed 10%.The sediment types are mainly sandy silt and silt sand.Factors such as hydrodynamic conditions and the effects of human activities are more consistent.(3)For the four subdivisions on the shores of Radiant Sandbar,the average particle size of most of the sediments in the north wing and the inner margin area is between 4 and 5 ?,and the sorting coefficient is mostly in the case of poor sorting.The type of sediments is mainly sandy silt,and the particle size migration trend of the north wing sediments is mostly northerly.At the intersection of the two tidal wave systems in the inner margin area,there is a clear migration from the two sides to the converging point.The particle size of the central part of the sedimentation trend is relatively coarse in the four regions.The overall classification is poor.The skewness is still more obvious from the landward to the ocean pole.Among the sediment types,the content of silty sand is the most,and the proportion of sediment types is also higher.The trend of sediment particle size migration is also complicated due to the impact of human activities such as port construction and fish farming.The south wing sediments are in four areas.At a relatively fine level,the sorting coefficients are all inthe case of poor sorting.Most of the skewness values are in the case of positive biases.Most of the sediment types are silt,and the sediment particle size migration trend is except for the tidal flat offshore.Northwest most of the outside Northeast.(4)Since 2002,the average length of the surface sediments in the tidal flats has been around 4? since 2002,and the average size of the sediments has been slightly refined.Poor state,and the sorting coefficient value slightly increased;the average of skewed state belongs to the positive or extremely positive state,the proportion of sediments in extreme positive state increases,and the extreme positive state is more obvious.The types of sediment are mainly silty sand and sandy silt,and the sandy silt gradually surpassed the silty sand,occupying a dominant position in the distribution.It also indicates that there is a trend of refinement in sediments,which is consistent with the trend of average particle size change in sediments.Most of the sediments are in the trend of northward particle size migration,which is consistent with the local hydrodynamic conditions,mainly due to the transport effect of damp residual flow.(5)The 2019 sediment inversion model is applied to the sediment information extraction in different years.From the comparison of the inversion effect,the model application can reflect the sediment situation in the current year with certain accuracy.Among the three particle size parameters,the model of average particle size has the best reusability effect.The maximum values of average absolute error,average relative error,and coefficient of variation are: 1.28?,32.17%,and 32.86%;sorting coefficient and skewness.The model's reusability is slightly worse.According to the integrated results of sediment type inversion,with the increase of years,the reusability accuracy of the model shows a downward trend,and the sandy silt sediment type has the best inversion effect.Factors affecting the reusability accuracy of the model include the consistency of remote sensing data sources,the similarity of time series,and the representativeness of sample collection.In the application of remote sensing models,attention should be paid to the normalization of multi-source remote sensing images,and the representativeness of samples should be paid to improve the accuracy of model reuse. |
PDF Full Text Request