Assessing Deoxygenation in Gulf of Mexico through Interpolation and 3D Modeling

Abstract

Deoxygenation poses a significant global concern, affecting oceans and marine reliant industries. The Gulf of Mexico stands out as an area with persistent hypoxia and yearly eutrophication events, resulting in reduced biodiversity, shifts in species distributions, and diminished fishery resources. Focusing on the central Gulf of Mexico 200 miles south of Louisiana, using World Ocean Database oxygen measurements over the year 2022, this study strives to shed light on methods used for predicting DO at various depths as well as the use of 3D modeling and Bathymetry to create advanced visuals on ocean-based elevation data. By applying Empirical Bayesian Interpolation in 3D, different transformations of the data are explored, and continuous surfaces of predicted DO levels at various depths are yielded. Models are assessed using cross-validation, semivariograms, and statistical performance measures. Among the three models tested, the highest performing model exhibited the lowest average standard error and mean error and with no applied data transformation or vertical trend removals. Exploring local Kriging models at different locations and standard errors, revealed larger standard errors at locations further away from known data points Arrived at confirmation of no hypoxic conditions in this Gulf location. 3D rendering through exportation of prediction surfaces as multidimensional voxel layers is exercised. Vertical and horizontal angled sections show the predicted DO measurements as intersections of different slopes, as well as isosurfaces visualizing depths showing the same DO values. These advanced 3D renderings exhibited that predicted DO concentrations are at their lowest between 66 and 1,000 meters.