Research On Remote Sensing Retrieval Of Particle Size Distribution In The Coastal Waters Of Qinhuangdao Based On OLCI Images

Particle size distribution(PSD)is a common parameter used to describe the optical scattering of marine particulate matter in oceanography,representing the variation characteristics of particle quantity concentration with particle size.It is an important index for studying the properties of particulate matter.Based on the remote sensing data of the Sentinel-3 OLCI satellite,thesis studied the retrieval of the particle size distribution in the coastal waters of Qinhuangdao.It provided a new direction for understanding and predicting the functional types of the phytoplankton community,ocean carbon cycle,and water optical properties.The PSD slope is a parameter that characterizes the particle size distribution.This study was based on the in situ measured suspended particulate matter data,discussing the response relationship between the optical characteristics and PSD slope.Then,the empirical relationship between spectral slope η of backscatter and PSD slope ξ was established.On this basis,the empirical model and semi-analytical model of PSD slope were constructed.Furthermore,the optimal region inversion model was also determined with model accuracy verified.Based on OLCI image data,the PSD slope distribution map was generated.At the same time,the temporal and spatial variation of PSD slope was studied with chlorophyll a and total suspended matter.The specific research results were as follows:(1)To get accurate information on particle size distribution,the effect of different reference diameters on particle size distribution was studied in the thesis.By studying the effect of different reference diameters on the fitting precision of power law model and the PSD slope.It was found that the reference diameters should be larger than 22.7 μm,and 101 μm could be used as the best reference diameters to obtain the accurate size distribution information.(2)In the thesis,combined with the Sentinel-3 OLCI center band,the empirical model and semi-analytical model of PSD slope were successfully constructed.The best band ratio of the empirical model was Rrs(442)/Rrs(560),the root mean square error(RMSE)and the mean absolute percentage error(MAPE)were 0.30 and 6.66%,respectively.Rrs(754)and Rrs(779)were selected as the best input band of the semi-analytical model,and RMSE and MAPE were0.74 and 17.45%,respectively.The inversion accuracy of the empirical model and the semianalytical model was analyzed by using the OLCI image data corresponding to the verifying station.The empirical model results were consistent with the in-situ PSD slope,and the semianalytical model results were high.The RMSE were 0.59 and 1.29,and the MAPE were 8.66%and 24.87%,respectively.Finally,the empirical model inversion accuracy was higher than the semi-analytical model.(3)Based on the empirical model established in the thesis,the PSD slope map in the coastal waters of Qinhuangdao was generated from OLCI images.Overall,the slope of PSD presented a ladder distribution with the distance from shore.The PSD slope decreased with the distance from shore and increased with the distance from land.The PSD slope was greater in June than in September.In order to research the variation characteristics of PSD slope,the inversion range was extended to the whole Bohai Sea.The PSD slope in summer,winter,spring,and autumn were 3.85,3.73,3.58,and 3.52,respectively.Explain the PSD slope in combination with chlorophyll a and total suspended particulate matter concentration.From the inversion results,in spring,summer,and autumn,the regions with the lower concentration of chlorophyll-a and the total suspended matter appeared to have higher PSD slope values,and there was good covariation among them.The winter co-variation was not significant.