rat 2.1

Resevoir Monitoring using Satellite Remote Sensing

Reservoir Assessment Tool 2.0

The Reservoir Assessment Tool (RAT) Framework was originally developed by Biswas et al. (2021) at SASWE, University of Washington. RAT uses satellite remote sensing data to monitor water surface area and water level changes in artificial reservoirs. It uses this information, along with topographical information (either derived from satellite data, or in-situ topo maps) to estimate the Storage Change (∆S) in the reservoirs. Additionally, RAT models the Inflow (I) and the Evaporation (E) of each reservoir. Finally, RAT uses the modeled I, and E, and estimated ∆S, to estimate the Outflow (O) from reservoirs.

Since the RAT framework developed by Biswas et al. (2021) was a first-of-its-kind open-source reservoir monitoring tool, it is reffered to as the version 1 of RAT, or RAT 1.0. It currently runs for 3 regions - (1) South and South East Asia, (2) Africa, and (3) South America, and can be accessed here.

The Reservoir Assessment Tool (RAT) 2.0 introduces numerous improvements over the RAT 1.0. The improvements in RAT 2.0 are as follows:

1. A robust multi-sensor approach to surface area estimation is implemnted, called, Tiered Multi-Sensor algorithm (Optical and SAR), abbreviated as TMS-OS. It uses a combination of various satellite sensors, such as Sentinel-2, Landsat-8, and Sentinel-1. This allows us to obtain highly accurate surface area estimates even during challenging conditions, such as during high cloud cover.
2. The temporal frequency of surface area, ∆S and Outflow monitoring was increased from a monthly resolution (in RAT 1.0) to 1-5 day frequency. This allows RAT to quantify sub-weekly reservoir operations.
3. In comparison to using only Landsat 7 and Landsat 8 in RAT 1.0, RAT 2.0 uses a total of 4 satellite sensors as of now, and more to be added. RAT 2.0 uses two optical sensors (Sentinel-2 and Landsat-8) to estiamte the surface area of the reservoirs; one SAR sensor (Sentinel-1) to estimate the trends in surface area changes; and one altimeter (Jason-3) to estiamte the water level heights with very high accuracy.
4. The latest version of the VIC hydrological model (VIC 5, Hamman et al. (2018)) and MetSim are used in parallel computation mode. This allows for rapid model runs, and more efficient data storage options in the form of NetCDF files.
5. Evaporation is modeled explicitly using the Penman Combination method, allowing for more control over the parameters used to calculate evaporation.

The RAT 2.0 was developed using the Mekong basin as the test-bed. Hence, RAT 2.0 currently monitors 36 reservoirs in the Mekong Basin, and can be accessed here. A guide on how to interact with the RAT 2.0 interface can be found here.

The RAT 2.0 GitHub repository

This GitHub repository contains the backend code of RAT 2.0, which performs the following tasks:

1. Downloads, Processes and Stores meteorological data for Hydrological modeling.
2. Prepares input data and runs the MetSim meteorological disaggregator, followed by the VIC hydrological model for the Mekong basin.
3. Routes the surface runoff using the VIC Routing model at the reservoir locations to obtain Inflow.
4. Processes Optical and SAR data using GEE to obtain surface areas and trends in surface area changes. This data is fed into the TMS-OS algorithm to obtain surface area time-series.
5. Obtains the water level time-series from the Jason-3 altimeter using the Okeowo et al. (2017) methodology.
6. Using the surface area and water level time-series, estimates the ∆S and Outflow.
7. Calulates the Evaporation over the reservoir using the Penman Combination method.
8. Using the modeled Inflow, ∆S and Evaporation, calculates the Outflow for the reservoirs.

The repository is structured as follows:

rat_v2/
├── backend/
│   ├── params/
│   │   ├── metsim/
│   │   ├── routing/
│   │   ├── vic/
│   │   ├── j3_meta.txt
│   │   └── rat_mekong.yml
│   └── scripts/
│       ├── core/
│       ├── data_processing/
│       ├── utils/
│       ├── __init__.py
│       ├── run_rat.py
│       └── run_rat.sh
├── .gitignore
├── LICENSE
├── env.yml
└── README.md

• The backend/ directory contains the scripts and parameter files required by RAT 2.0.
• The backend/params/ directory contains the parameters that are used by RAT 2.0. The paramter file of most importance to users and developers is the backend/params/rat_mekong.yml file. This file contains information about the start date, end date, project location, and similar information that are required to be specified for RAT 2.0 to run.
• The directories in backend/params/ contain sample parameter files that can be used to specify any model specific parameter. Please note that most of the options that can be specified in the files contained in these directories can also be specified in the rat_mekong.yml file (recommended).
• j3_meta.txt file is used by the Altimetry component of RAT 2.0 internally and shouldn't be changed. To intialize, the first line of the file should be set to 0.
• The backend/scripts/core/ contains some of the core functionality, that deal with processing data into model inputs, running the models, running the GEE code and calculation of outflow, etc.
• The backend/scripts/data_processing contain additional code to process data (data downloading, transformation of data, etc.). backend/scripts/utils contain utility functions that are used internally by RAT 2.0.
• The backend/scripts/run_rat.py is the starting point for running the RAT 2.0 model for a single day/multiple days at once. When invoked, it uses the rat_mekong.yml parameter file to run the hydrological modeling component (MetSim, VIC and routing), the remote sensing backend and performs the post-processing for the specified options.
• The backend/scripts/run_rat.sh is a shell script that should be set up as a cron job to run everyday. This script updates the paramter file so that the RAT 2.0 model runs for the latest day. After making suitable changes to the rat_mekong.yml file, it invokes the run_rat.py file.

LICENSE

RAT 2.0 is distributed under the GPL v3 license. You may copy, distribute and modify the software as long as you track changes/dates in sourcefiles. Any modifications to or software including GPL-licensed code must also be made available under the GPL along with build & install instructions. For more information, please see LICENSE.