Quick Start Guide¶

This guide will help you set up and run your first MDT verification workflow.

Installation¶

MDT requires several components from the MONET ecosystem. For the best experience, we recommend using a Conda or Mamba environment.

# Create and activate environment
mamba create -n mdt python=3.9
mamba activate mdt

# Install MONET ecosystem from GitHub
pip install git+https://github.com/noaa-oar-arl/monetio.git
pip install git+https://github.com/noaa-oar-arl/monet.git
pip install git+https://github.com/noaa-oar-arl/monet-stats.git

# Install monet-plots without dependencies to avoid conflicts
pip install git+https://github.com/noaa-oar-arl/monet-plots.git --no-deps

# Install MDT and its direct dependencies
git clone https://github.com/noaa-oar-arl/mvs.git
cd mvs
pip install -e .

Note: MDT also requires pyarrow for Dask/Pandas compatibility.

Running MDT¶

MDT is configured via a YAML file. Once you have a configuration file (e.g., config.yaml), you can run the tool using the command-line interface:

mdt run --config config.yaml

A Minimal Working Example¶

Create a file named simple_eval.yaml with the following content:

data:
  my_model:
    type: "cmaq"
    kwargs:
      fname: "path/to/cmaq_output.nc"
  my_obs:
    type: "aeronet"
    kwargs:
      fname: "path/to/aeronet_data.nc"

pairing:
  eval_pair:
    source: "my_model"
    target: "my_obs"
    method: "interpolate"

statistics:
  basic_stats:
    input: "eval_pair"
    metrics: ["rmse", "bias", "corr"]
    kwargs:
      obs_var: "AOD_500"
      mod_var: "AOD_500"

plots:
  spatial_eval:
    input: "eval_pair"
    type: "spatial"
    kwargs:
      savename: "cmaq_vs_aeronet.png"

Then run it:

mdt run --config simple_eval.yaml

MDT will automatically load the data, pair the model output to the observation locations, compute the requested statistics, and save a spatial plot.