Hotspot Area Extraction And Movement Pattern Mining Of Ocean Mesoscale Eddy

The ocean accounts for 71% of the total area of the earth and is rich in mineral resources and 80% of the earth’s biological resources.Changes in ocean ecology play a key role in the global energy cycle as well as climate change.Ocean mesoscale eddies are an important oceanic phenomenon and carry great energy,being the main bearer of ocean energy and heat transfer,and also play an important role in mixing and transporting the vertical and horizontal distribution of ocean resources and nutrients.Therefore,it is important to extract the hot spot areas of ocean mesoscale eddies to study their effects on sea surface temperature,fishery resources and fishery distribution.It is important to explore the movement patterns of mesoscale eddy trajectories,and to provide a reference basis for the research direction of tracking and movement direction prediction of ocean mesoscale eddies.At present,the main objects of hot spot area extraction are urban traffic,residential travel,and ship tracks,and there are fewer studies on the extraction of hot spot areas of ocean mesoscale eddies,and there are even fewer studies on mining hot spot areas of mesoscale eddies carrying huge energy,while the magnitude of energy carried by ocean mesoscale eddies is influenced by factors such as amplitude,radius,and average shift speed,and the traditional clustering methods are not effective for mesoscale eddies with multiple features.The traditional clustering methods are not effective in extracting the hotspots of the mesoscale eddies with large energy content.There are various forms of oceanic mesoscale eddy trajectories,such as straight,circling and slewing,etc.The methods for mining the movement patterns of mesoscale eddy trajectories are mainly based on traditional clustering algorithms,and the traditional trajectory clustering methods can only mine the vague and unclear mesoscale eddy movement patterns,and most of the existing algorithms only consider the spatial location of the moving target,ignoring its temporal properties and its own physical characteristics,and do not well The fine-grained movement characteristics of the target are not well explored.This paper takes the ocean mesoscale eddy trajectory as the research object,extracts the significant hot spot area of the mesoscale eddy affecting the marine environmental elements,mines the trajectory movement pattern,and proposes a multi-feature density peak-based clustering method(MF-CFSFDP)and an improved density space-based clustering method(2DB-SCAN)in combination with the characteristics of the ocean mesoscale eddy.The main research contents and results of the paper are as follows.(1)For the existing research objectives,most of the traditional trajectory clustering methods only consider the spatial location information without fully considering the attributes of the trajectory such as time and shift speed,which makes the final clustering results cannot reflect the representative hot spot area distribution and cannot meet the practical application requirements.In this paper,a multi-feature density peak clustering method(MF-CFSFDP)is proposed in Chapter 2.Based on the CFSFDP clustering method,three features of amplitude,radius and average shift speed are added in the clustering process,and the Euclidean distance is used to measure the similarity between mesoscale vortex trajectory points,which solves the problem that the CFSFDP clustering is only applicable to single-feature moving targets,and the clustering method in this chapter The clustering method in this chapter is more applicable to hotspot region mining of multi-feature moving targets.(2)The purpose of ocean mesoscale eddy trajectory clustering is to discover the characteristics of ocean mesoscale eddy movement patterns in the study area,so as to further explore the movement patterns of mesoscale eddy trajectories.The density-based spatial clustering method and the density-based spatio-temporal clustering method are two common clustering methods used in the application of ocean mesoscale eddy trajectory clustering nowadays,and the results of these two clustering methods do not well characterize the specific movement trend of mesoscale eddies.In the process of trajectory clustering of ocean mesoscale eddies,in order to better represent the spatial distribution and moving direction of mesoscale eddies,this paper proposes a densitybased vector trajectory spatial clustering method,which integrates the moving direction characteristics of ocean mesoscale eddies trajectories,improves the line segment spatial distance to be calculated in the clustering process into vector line segment spatial distance,improves the method of calculating the angular distance,and can better The improved method of calculating angular distance can better cluster mesoscale eddies with different moving directions into different clusters.The experimental results show that the improved method of angular distance can better represent the spatial distribution and moving directions of mesoscale eddies in the process of trajectory clustering of ocean mesoscale eddies.