FLARE Project Overview

The FLARE project is one of the major collaborative efforts between Dr. Figueiredo’s group at OSU and Drs. Carey and Thomas at Virginia Tech. FLARE (Forecasting Lake And Reservoir Ecosystems) creates open-source software for flexible, scalable, robust, and near-real time iterative ecological forecasts in lakes and reservoirs.

FLARE Entities

These are the entities involved in the process:

1. Sensors: Gather environmental data using sensors at and around the lakes and reservoirs;
2. Data Loggers: Store the sensor data and periodically push them to the corresponding edge gateway;
3. Edge Gateways: Stage the sensor data alongside system logs in local Git repositories;
   3.1. LoRa Gateways: A gateway in a remote position at FCR with no direct access to the internet connects to another gateway via LoRa connection in order to get access to the Internet.
4. SUNP Computer: In Lake Sunapee, rather than sensors and gateways, SUNP runs our software on a Windows operated computer to fetch data from the buoy.
5. Git Repositories: Data is transferred from edge gateways to cloud storage using Git, staging data and logs on local Git repos on the gateways and commit them to remote Git repos on GitHub;
   5.1. Git Repository Backups: Automatic backup of all FLARE related repositories are done using Bash scripts on a virtual machine deployed on NSF Jetstream2 resources.
6. GitHub Actions: Automate the workflows for data pre-processing, running forecasts, and publishing the forecast results;
7. S3 Cloud Storage: Responsible for storing the input, intermediary, and output data of FLARE forecasts.
8. Virtual Machines: A couple of Ubuntu Server virtual machines on Jetstream2 infrastructure acts as compute node for some services including a node responsible for repository backups and a node that acts as a self-hosted GitHub runner for FLARE.
9. FLARE Website: FLARE website address is https://flare-forecast.org/.

Maintenance and Troubleshooting

🚀 Maintenance and Troubleshooting

An Overview of the Physical and Virtual Nodes

Forecasting Workflow Overview