Research On Data Griding Of Three-Dimensional Oceanic Element Fields Based On Argo Data

The Argo profile data set has been developed rapidly in recent years as the main global quasireal-time Marine hydrological observation data.It provides a scientific reference for the researchers to study the structure of oceanic temperature salt,analyze the spatial and temporal variation characteristics and interrelationship of various hydrological factors,reveal the structure of ocean temperature and the three-dimensional characteristics of flow field,and explore the marine environment and the climate change.Ocean phenomena are continuous and complex in both space and time.But sparse discrete and randomly distributed Argo profile data is difficult to provide a dynamic visual description of the intuitive,continuous hydrological elements and ocean process information.Therefore,it is meaningful to study the evolution law of Marine environment and provide effective data support for engineering application by studying method of constructing three-dimensional grid data of ocean element field.The current Argo plan is still in its development stage.Because the number of buoys is insufficient and the observation time is inconsistent,and the sampling location is uncertain caused by drifting,the profile datasets have sparse and scattered characteristics both in time and space,which is difficult for gridding the data of the Marine element field.In view of the above situation,considering the spatio-temporal correlation of ocean elements and combining spatio-temporal Kriging interpolation technology,the technique of constructing three-dimensional grid data of oceanic element field by Argo profile data is studied and practiced in this paper.The main content and achievements of the thesis are as follows:(1)For the efficient neighborhood search capability in spatio-temporal interpolation,a single data organization method based on JSON format is proposed according to the semi-structured feature of single Argo profile data file.Based on the spatio-temporal distribution characteristics of ocean buoy,the database is managed,and the search efficiency of neighborhood sampling points is improved through coding index.(2)The shortcomings of spatio-temporal Kriging interpolation and the inadaptability of application in sparse distributed Argo profile data.are analyzed in detail.The spatio-temporal Kriging improvement method is proposed in this paper,and the error effect of time variation on the fitting of spatial variation function is controlled.Finally,the interpolation of spatiotemporal variation function is constructed by integrating model,and the result of cross validation proves that the method effectively improves the precision.(3)The Argo three-dimensional grid data of oceanic emperatures is constructed using the improved spatio-temporal Kriging interpolation.First,the data types and pretreatment methods involved in the construction process are introduced.Then,the method of constructing grid data of subsurface sea temperature field based on "partition-average" and the fusion idea of surface temperature field grid data based on AVHRR remote sensing data are proposed.The analysis of spatial and temporal distribution characteristics and quantitative analysis of other data shows that the method and results of grid data construction in this paper are reasonable and reliable.