search

CIGLR-ai-lab / GreatLakes-TempSensors

Collaborative repository for optimizing the placement of temperature sensors in the Great Lakes using the DeepSensor machine learning framework. Aiming to enhance the quantitative understanding of surface temperature variability for better environmental monitoring and decision-making.
MIT License
0 stars 0 forks source link

Gain a contextual understanding of FVCOM data and NARR forcing #16

Closed DaniJonesOcean closed 3 months ago

DaniJonesOcean commented 3 months ago

Issue Purpose: Acquire a foundational understanding of the FVCOM dataset, which is forced by NARR atmospheric reanalysis data, understand its relevance for our study, and carry out some basic data exploration and visualization.

Background: FVCOM provides a detailed simulation of lake circulation, which is critical for applications like our. We aim to use FVCOM outputs, which have been forced with NARR data, as a potential training dataset for DeepSensor. A basic comprehension of the model, its outputs, and the manner in which atmospheric forcing is applied, will be useful for integrating this data into our work.

Tasks:

  1. Contextual Research:

    • Review the provided FVCOM and NARR project pages and paper to understand the model characteristics and the dataset's origin.
    • Summarize how FVCOM works and the role of NARR data in its forcing mechanism.

  2. Preliminary Data Exploration:

    • Access the data here on U-M HPC: /nfs/turbo/seas-dannes/SST-sensor-placement-input/FVCOM/SST_splits
    • Perform a basic visual inspection using xarray. Plot the spatial patterns, variability over time, and anything else that you find interesting and potentially relevant.

Deliverable: Please write a few brief comments on this issue on the following:

Expectations: This does not have to be an in-depth analysis but rather should provide enough insight to inform future research steps with DeepSensor.

eredding02 commented 3 months ago

@DaniJonesOcean I am unable to access the David Cannon article

DaniJonesOcean commented 3 months ago

@eredding02 PDF sent via email

DaniJonesOcean commented 3 months ago

@eredding02 Apologies for the sudden change, but I'm suggesting that we pause activity on this issue to focus on the GLSEA/GLSEA3 and buoy datasets. Those datasets might be sufficient for what we want to do for DeepSensor, and I don't want to spread our efforts too thinly across too many tasks.

We might return to this issue later, but for now I've un-assigned it and have moved it back to the "In prep" column, which can act as a bit of a holding area.

Note: To mirror what I say in my email signature, I do not expect you to work outside of your working hours. My working hours may be different than yours, as you can see by the timing of this comment!

DaniJonesOcean commented 3 months ago

Closing for now to get it off the issue list and board. We can return to it again later if needed.