Hi! Thank you very much for your interest! And thanks for raising your concerns as we are still trying to improve how we explain what we are doing. Let me try to explain the different tasks (from my perspective). I would appreciate any feedback on my explanation because I would like to include this write-up within the documentation. Dataset Available In general, we're working with 4 types of datasets: Sea Surface Height (SSH) - $\eta$ Sea Surface Temperature (SST) - $T$ Discretized AlongTrack SSH Observations - $\eta_{obs}$ Native AlongTrack Observations - $\eta_{atrack}$

This is my aggregation of ML resources over the years. If you google machine learning or deep learning you’re likely find a lot of resources. But in the age of ChatGPT and consultants, we as a society have recognized that we need to filter all of this information to get more personalized recommendations based in our needs. This is my attempt to find the best resources based on my own personal experience. this is targeted at applied machine for geosciences. So I am assuming someone trained in physical sciences with training in the scientific method and at least a mediocre level of programming experience. Fortunately, machine learning is quite general enough to learn alone because there are many toy problems to motivate the algorithms. However, I believe that geoscience problems have many many characteristics that make it hard to apply ML out of the box. So I will include resources that can perhaps bridge the gap between the two. My Criteria I have a set criteria which I try to maximize in order to have a relatively balanced a smorgasbord of resources. In other words, I want a range of resources that cater to different learning styles based on things that I personally think are important for applying ML. Breadth vs Depth. Some resources cover a wide range of topics while others dive deep into a few subjects.

Project Statement (TODO) My Conceptual Notes Notation Inverse Problems Dynamical Systems Markov Models Optimal Interpolation Kalman Filter Ensemble Kalman Filter

A differentiable JAX-powered Shallow Water Model Motivation A (flexible) differentiable code for dynamical systems For testing ideas on parameter inference, closures, model errors and state estimation. Key requirement : embedding time-stepping in the learning. Concept : a common interface for different use cases in climate science Why Jax?

Important Links Mathy Bits - The mathy bits related specifically to this paper. (Updating as I Go along) Literature Review - The literature review specific to this paper (Updating as I go along) Data Details - The literature review specific to this paper (Updating as I go along) Results Details - Some preliminary results. My Personal GP Model Zoo - A Website I did during my PhD of all things Gaussian process with modern tools! Data Challenge II - Based Purely on Observations Data Challenge I - Model Simulations and Observations Simulations

Day I Functions: To-Do [ ] Spreadsheet for Datasets Goals Safe Goal