Correlation Between Sea Surface Temperature And Water Depth Based On Remote Sensing

China is a large maritime country,and the strategy of maritime power is one of the important strategies of China's overall development goal in the 21 st century.In the field of marine research,water depth research and sea surface temperature research play an important role: water depth information can reflect the marine environment,seabed topography,and play an important role in regional economic and military activities;Sea surface temperature is an important environmental parameter reflecting the state of ocean thermal radiation.Due to the large heat capacity of seawater,the small change of its temperature will have a great impact on the climate.However,water depth inversion requires measured water depth data or measured water optical parameters,and sea surface temperature inversion requires atmospheric profile data and complex formula calculation.According to the heat conduction theory,the sea surface temperature is affected by the atmosphere and the interior of the ocean.At the same time,the temperature gradually decreases with the increase of seawater depth.Therefore,the research hypothesis of this paper is put forward: there is a certain correlation between sea surface temperature and water depth.With the increase of water depth,sea surface temperature gradually decreases.The sea surface temperature value can be derived from the water depth value or the water depth value can be derived from the sea surface temperature value.According to the problems of small number and low depth of existing water depth measurement points,this paper puts forward the extrapolation experiment of water depth measurement points and verifies the accuracy.Then,according to the measured points and extrapolation points,the statistical correlation model,BP neural network model and support vector regression model are constructed;Then,the sea surface temperature inversion is carried out according to the radiative transfer model,and compared with the water depth inversion results.The research points are extracted for the construction and evaluation of the relationship model between the two.The main research results are as follows:(1)The correlation between offshore distance and water depth reaches 0.781,a nd the constructed offshore distance inversion water depth model R2 reaches 0.760.The water depth inversion model constructed by others is used as the standard of accuracy evaluation,and the correlation between them reaches 0.788.This shows that in an area with simple and regular terrain like Hainan Province,the relationship model between offshore distance and water depth has certain application value,but there is still a large optimization space,such as the optimal spacing between extrapolated points and measured points.(2)Among the three water depth inversion models,BP neural network model has the highest accuracy,and R2 reaches 0.939.The R2 of the support vector regression model which performs well in the small sample problem is 0.9017.It is speculated that there is still optimization space because the best parameters are not found.(3)The next sea surface temperature inversion is due to the large error in the11 th band of Landsat 8.At present,the more mature and widely used temperature inversion method for Landsat 8 image is still the single band algorithm.In the single band,the radiative transfer equation method is simple and accurate.Therefore,the radiative transfer model is selected for sea surface temperature inversion.After comparing the inversion results with the sea temperature,it is found that there is a certain correlation between the two inversion results.(4)The water depth value and sea surface temperature value are extracted from160 research points.The results show that the correlation between the two images in January 2020 is-0.958,the correlation between the images in April 2020 is-0.361,the correlation between the images in July 2020 is-0.829,and the correlation between the images in February 2021 is-0.755.Then the relationship model between the two is constructed: the model R2 in January 2020 is 0.928,the model R2 in April is 0.166,the model R2 in July is 0.722,and the model R2 in February 2021 is 0.687.(5)The results show that there is a certain correlation between sea surface temperature and water depth,but due to the insufficient data in this paper,the correlation between the two images in April 2020 and February 2021 is low.If there are abundant measured water depth data and atmospheric profile data,the correlation between them is expected to reach a new height.In this study,after screening and evaluating the accuracy of three water depth inversion methods,BP neural network model is selected,and the correlation analysis is carried out with the inversion results of sea surface temperature calculated by radiative transfer equation.It is concluded that there is a negative correlation between water depth and sea surface temperature,and the derivation between them can be carried out through the model.It provides a new feasible scheme for future marine research.