Credits & gratitude

The European Union's Earth observation programme, operated by the European Space Agency. Copernicus runs the Sentinel satellite family — six complementary mission lines covering SAR (Sentinel-1), high-resolution optical (Sentinel-2), ocean + thermal (Sentinel-3), geostationary atmosphere (Sentinel-4 + -5), and altimetry (Sentinel-6). Every byte is published under a free-and-open license that genuinely is free and open. mud2dust depends on Sentinel-1 (via OPERA RTC) and Sentinel-2 (via HLS) plus the Copernicus GLO-30 DEM.

Without the Copernicus open-data policy, mud2dust would not exist as an open project. The decision to publish a continent-scale Earth-observation programme as a public good — free of fees, registration, or commercial gatekeeping — is the load-bearing principle this whole platform sits on.

C-band synthetic aperture radar (SAR) constellation. Sentinel-1A (2014–), Sentinel-1C (2024–), and Sentinel-1D (2025–) provide multi-day repeat coverage at 10–40 m resolution — see-through-cloud day/night observations of every land surface on Earth. NASA OPERA RTC ingests acquisitions from all operational satellites in the constellation; mud2dust uses the resulting σ⁰ as its primary soil-moisture signal.

Sentinel-1 is the load-bearing input to mud2dust's calibrated VWC layer. ESA's commitment to the constellation — three operational satellites in 2026 — is what makes a 30 m moisture map updated every few days realistic.

High-resolution multispectral optical mission. Sentinel-2A (2015–), 2B (2017–), and 2C (2024–) image every land surface on Earth every ~5 days at 10–60 m across 13 spectral bands (visible, near-IR, red-edge, SWIR). mud2dust uses Sentinel-2 reflectance via the NASA HLS (Harmonized Landsat-Sentinel) product for its NDVI / NDWI / NDRE vegetation indices.

Sentinel-2's red-edge bands are what make sub-canopy moisture inference possible — Landsat alone doesn't have them at this resolution. Free, open, every-five-days, 10 m. Remarkable infrastructure to live downstream of.

Ocean and land mission for sea-surface temperature, ocean colour, sea-surface height, vegetation indices at 300 m, and land-surface temperature. Two satellites operational (3A in 2016, 3B in 2018). mud2dust does not currently ingest Sentinel-3 directly — at our 30 m calibration scale, OLCI / SLSTR's 300–1000 m resolution is too coarse — but it's a Phase 6+ candidate for global LST / SST gap-fill where higher-resolution sources aren't available.

Listed for completeness and because the Copernicus open-data licence covers it identically — when CONUS scaling forces us to fill thermal gaps where HLS Landsat / GOES don't reach, Sentinel-3 SLSTR is the best free fallback we know of.

Geostationary atmospheric monitoring instrument hosted on EUMETSAT's MTG-S satellite. Hourly UV-VIS-NIR observations of NO₂, SO₂, ozone, aerosols, and formaldehyde over Europe. Not used by mud2dust today (atmospheric chemistry is outside our soil-moisture remit and the geostationary footprint doesn't cover CONUS) but listed because we depend on the same Copernicus open-data policy that publishes it.

Atmospheric chemistry isn't in mud2dust's scope but it should be in your Earth-observation literacy — the same data policy that gives us Sentinel-1 + -2 gives the atmospheric science community Sentinel-4 + -5P, free.

Validated Level-2 Radiometric Terrain Corrected SAR Backscatter from Sentinel-1, gamma-nought (γ⁰), 30 m resolution, CEOS-ARD compliant. Replaces ad-hoc per-project RTC processing with a single authoritative product.

The OPERA team is paying the heavy compute cost of running RTC processing globally so every downstream project doesn't have to. mud2dust was originally designed assuming Indigo Ag's now-retired RTC bucket; OPERA picked up the mantle and made the project viable again.

NASA Earthdata Distributed Active Archive Center (DAAC) for SAR products: Sentinel-1, OPERA RTC, NISAR (forthcoming), historical SAR archives. Hosts data on AWS S3 in us-west-2.

ASF runs the storage, access, and authentication infrastructure that gates SAR data to the world. They could have built a paywall; they built EDL + free S3 reads instead.

NASA Earthdata DAAC for land-surface remote sensing: HLS (Harmonized Landsat-Sentinel), MODIS, ECOSTRESS, ASTER. Hosted on AWS Earthdata Cloud.

LP-DAAC's HLS product (Phase 2 input for mud2dust) does the cross-sensor harmonization that would otherwise be weeks of work per region. They publish it free.

Federated archive of NASA's Earth-observing data. The Earthdata Login (EDL) authentication service is the single sign-on for OPERA, HLS, SMAP, ECOSTRESS, GEDI, MODIS, GPM, and dozens more datasets.

Earthdata's commitment to free-and-open data is the load-bearing principle that lets small teams build serious science on top. The EDL flow is one signup that unlocks the entire NASA Earth-observation archive.

Global, 250 m, machine-learning-based soil property maps for 6 standardized depth slices (0-5, 5-15, 15-30, 30-60, 60-100, 100-200 cm). Properties: bulk density, cation exchange, clay/sand/silt, coarse fragments, organic carbon, pH, total nitrogen, soil class.

ISRIC publishes SoilGrids globally for free under CC-BY-SA. The original mud2dust plan called for POLARIS (UC Davis) for CONUS detail, but POLARIS is currently offline; SoilGrids 2.0 covers Eastern WA at 250 m and is good enough for the calibration prior. Full credit + thanks to ISRIC for keeping this open and reproducible — the methodology is published, the source code is on GitHub, and the data is a click away.

European Space Agency's global digital elevation model at 30 m resolution, derived from TanDEM-X / TerraSAR-X interferometry plus other sources. Vertical accuracy ~4 m. Hosted on AWS Open Data.

ESA publishes the GLO-30 DEM under a free-and-open license that's actually free and open. AWS hosts it for anyone to read at zero egress cost in-region. mud2dust uses it to derive topographic wetness index and slope/aspect priors for the calibration model.

AWS pays the storage and same-region egress costs for petabytes of public scientific data — Sentinel-1, Landsat, NOAA, NASA EOSDIS, OPERA, NEXRAD, GOES, HRRR, SMAP, and more. Without same-region S3 reads, every download would route through paid egress.

This program is the single biggest reason mud2dust is buildable for the price of a coffee subscription. AWS is paying real money to host public data so people can build on it; they get nothing direct back. Genuine thanks to the team behind that policy and to the contributors at NASA/ESA/USGS/NOAA who push their data into the program every day.