AtmoFacts logo

Tutorials & sandbox datasets

Learn by doing with open tools, free FluxMapper data, and step-by-step guides.

fluxmapPy: analyze FluxMaps in Python

fluxmapPy is our open-source Python package for working with FluxMaps in your own analysis. It reads FluxMap GeoTIFFs together with their quality-flag layers, masks out estimates that fail a quality threshold, and aggregates the result over your areas of interest into statistics and diurnal cycles. It is MIT-licensed and led by AtmoFacts Research Scientist Allen Kaplan.

v0.6 · Open source · MIT license

Four functions cover the everyday workflow

  1. Open FluxMap layers.
  2. Apply quality flags.
  3. Clip to your area of interest: a polygon, or a radius around the tower.
  4. Compute summary statistics, such as half-hourly and daily time series and diurnal cycles.

Start with the tutorial. A step-by-step Jupyter notebook walks through a real research question on public NEON FluxMaps. Open it in Google Colab to run it in your browser with nothing to install.

Install it locally from GitHub (Python 3.10 to 3.13):

python -m pip install "fluxmappy @ git+https://github.com/AtmoFacts/fluxmapPy"

fluxmapPy relies on GDAL-based libraries (rasterio and fiona). If pip has trouble with them, installing those through conda-forge usually solves it. The documentation covers setup and troubleshooting.

v0.6 is an early release: function names may still change on the way to v1.0. A Python Package Index (PyPI) release, conda-forge packaging, and a citable DOI for each release are next.

Sandbox datasets

Explore FluxMapper outputs with public, free-to-download datasets. These curated datasets from completed research projects let you experiment with real flux maps without processing raw data.

BenchFlux: NEON tower network

A collection of FluxMapper outputs across multiple NEON tower sites. Download pre-processed flux maps for CO2, sensible heat, and latent energy from 2019 and 2020 data.

Available sites: HARV (Harvard Forest), STEI (Steigerwald), STER (Sterling), TREE (Treehaven), JORN (Jornada)
Years: 2019, 2020
Format: GeoTIFF layers

What you can do: download spatial flux maps, analyze heterogeneous landscapes, compare sites and years, validate your own analyses against reference datasets, or use them as ground truth for downstream models.

Example file path:

https://storage.cloud.google.com/data-benchflux/TREE/fluxCo2/tree_fluxCo2_20200701.tif