Threading Inputs to Outputs (TITO)

TITO is a framework designed to run the EF5 hydrologic model operationally, integrating satellite data and machine learning techniques to support real-time forecasting and hydrologic analysis.

1. Clone the repository

git clone github.com

2. Navigate to the repository folder

cd TITO/

3. Run the set up code This step might take few minutes.

./setup_tito.sh

4. Add the new conda env to your routines Open pipeline.sh and change the path to the conda environment, should be something like this:

source /Users/$username$/miniconda3/etc/profile.d/conda.sh

After installation, ensure that your TITO folder contains the following subdirectories and files.

This repository is desiged to run EF5 operational over West Africa. Users must populate the required folders with topographic and parameter grids for their region of interest, and modify the EF5 control file (templates/ef5_control_template.txt) accordingly. A guide step-by-step to create the needed layers can be found in: EF5-builder-toolkit.

• westafrica1km_config.py – Configuration file to set up your operational run.
• orchestrator.py – Main Python script that manages the entire workflow.
• pipeline.sh – Bash script that activates the tito_env Conda environment and executes orchestrator.py using settings from westafrica1km_config.py.

• basic/ – Contains DEM, FAC, and FDIR files.
• pet/ – Contains monthly PET (Potential Evapotranspiration) grids.
• parameters/ – Contains distributed parameters for the KW and CREST models.
• states/ – Stores model state files generated during operational runs.
• outputs/ – Output folder where simulation results are saved.
• precip/ – IMERG QPE files are downloaded here; QPF files generated by the nowcasting system are also stored here.
• precipEF5/ – QPE and QPF files are reformatted and copied here to be ingested by EF5.
• templates/ – Stores EF5 control file templates, which are dynamically updated during each run.
• qpf_store/ – Stores QPF files for fallback use in case QPE retrieval fails.
• ML/ – Contains machine learning routines used to generate QPF forecasts.
• tito_utils/ – Collection of utility modules and helper scripts used internally by TITO.

1. Edit the config file: After completing the installation of the required environment and populating the corresponding EF5 folders, open westafrica_1km_config.py file. There are few lines users need to change in this config file to run TITO successfully:

• ef5Path: Update this path to the corresponding ef5's binary path in your system.
• HindCastMode: If you are running an event happened in the PAST, set HindCastMode = True and write the date of interest in HindCastDate, use the format "YYYY-MM-DD HH:MM". If you want to run it in Nowcast Mode (meaning TITO will start running in the present time) set HindCastMode = False
• run_LR: To include QPF in the simulation (options are GFS or WRF), set run_L = True.
  • If the simulation is for a past event (HindCastMode = True), you must provide:
    • QPF start date (StartLRtime)
    • QPF end date (EndLRtime)
    • QPF time step (LR_timestep) in minutes, e.g., 30u
    • Path to your QPF archive (QPF_archive_path)
  • If you are activating this option for real-time operations, TITO uses a predefined QPF time. You can check orchestrator.py to customize it for your convenience.
• email_gpm: This version of TITO uses IMERG Early V07 as QPE, you will need to create and account in GPM server to download precipitation files, please visit NASA GPM registration web p… and follow the instruction provided in the webpage. Important: use your registration email as password so TITO can use it in the routines.

What if I want to use TITO in other regions?

If you plan to run TITO outside the default West Africa domain, there are a few important considerations. The machine learning routines were designed and trained using IMERG V07 data (0.1° resolution) over the West Africa region (xmin = −21.4, xmax = 30.4; ymin = −2.9, ymax = 33.1), corresponding to a grid size of **518 × 360 pixels**.

If you intend to apply TITO to a different region, we recommend selecting an area with the same spatial dimensions (518 × 360 pixels) to ensure compatibility with the input structure. 

2. Run TITO: Run the following line in your terminal:

./pipeline.sh

Please contact Vanessa Robledo at vanessa-robledodelgado@uiowa.edu or the AHWA Laboratory Development team at engr-ahwa-lab@uiowa.edu.

For more updates about this package follow our Github Repository