Dynamical systems for remote validation of very high-resolution ocean models

TL;DR

Problem: validating very high-resolution coastal ocean models is hampered by scarce in situ data. Approach: a framework that couples ROMS outputs with satellite imagery through dynamical-systems constructs, using Lagrangian Descriptors with the function $M(\vec{x}_0,t_0,\tau)$ to identify coherent transport features and a ground-truth area metric $\mathcal{A}$ to quantify fit. The method uses Sentinel-2/3 data corrected by ACOLITE to reveal surface patterns (POM/DOM/Chlorophyll) and compares them to model-derived patterns under different boundary and turbulence settings, including a pullback attractor perspective for time-dependent forcing. Key finding: the best overall configuration is the mixed radiation-nudging boundary with sponge (C6, exp2), with average $\mathcal{A}$ indicating substantial but variable agreement; results highlight sensitivity to bathymetry and boundary treatment. Significance: provides a practical, scalable route to calibrate coastal models with Earth observation, with immediate relevance to pollution dispersion, search and rescue, and coastal management.

Abstract

This paper presents and investigates a novel methodology for validating high-resolution ocean models using satellite imagery. High-resolution ocean models provide detailed information in coastal areas where other available data products are too coarse. Models are usually fitted by comparing them with observations; However, accessing in situ data in all small coastal areas is not feasible, as in situ observations are scarce and obtained through dedicated ships or instruments in limited and selected regions. Our work aims to use alternative remote sensing information to overcome this challenge. The approach involves establishing connections between the satellite observations and the outcomes of various computational experiments carried out using the Regional Ocean Modeling System (ROMS), which allows the selection of different parameters to run the ocean model. These choices are not fully determined a priori and each one produces distinct outputs, which are then linked to the images through dynamical systems objects. By defining a performance index, we are able to quantify which experiment provides a better representation of the local ocean state.

Figures (8)

• Figure 1: Bathymetry of the Rafina area in the domain $L_0$. The color intensity corresponds to the depth of the sea floor, with brighter colors indicating greater depths. a) NAVIONICS source data; b) GEBCO source data.
• Figure 2: Vertical discretization at longitude 24.15 $^\circ$E. The green color represents the bathymetry, while the lines indicate the different $\sigma-$levels based on the stretching function utilized.
• Figure 3: The figure depicts the WRF domain, with the wind speed intensity at 10 meters above the ground visualized in the background. The visualization corresponds to the specific date and time of November 2, 2019, at 12:00:00. Arrows are employed to indicate the direction of the wind.
• Figure 4: Sentinel 2 images on the 22nd of November 2019 in the Rafina Port area showing convoluted shapes. (a) Quasi-true RGB color image;(b) turbidity image obtained from ACOLITE algorithms.
• Figure 5: Sentinel 3A and B images in the Rafina Port area showing convoluted shapes of chlorophyll. (a) on the 22th of November 2019, (b) on the 23th of November 2019, (c) on the 25th of November, (d) on the 16th of December, (e) on the 17th of December, and (f) on the 18th of December. Red rectangles mark the region where very high-resolution simulations have been performed.