environment · agriculture · ocean · family: it worked in the lab
the eye thatdoesn't exist yet
Mapping Earth's ecosystems and minerals from space requires an imaging spectrometer nobody has built
Problem statement
Understanding how terrestrial and aquatic ecosystems respond to climate change, where mineral resources and hazards are located, and how soil and water quality vary across landscapes requires measuring reflected sunlight in hundreds of contiguous spectral bands from 380 to 2500 nm (visible through shortwave infrared) at 30-60 m spatial resolution globally. The Earth Science decadal survey designated Surface Biology and Geology (SBG) as a priority measurement, but no satellite has combined the spectral coverage, spectral resolution (~10 nm), spatial resolution (30 m), signal-to-noise ratio, and global revisit time needed. The engineering challenge is building a spaceborne imaging spectrometer that maintains radiometric calibration to <2% across 200+ spectral channels while scanning the full ~185 km Landsat swath width, producing 5+ terabytes/day of data that must be atmospherically corrected, spectrally unmixed, and delivered to users within hours.
Why this matters
Imaging spectroscopy uniquely identifies the chemical composition of surfaces: chlorophyll absorption features diagnose plant photosynthetic function, mineral absorption features map geological composition, and water absorption features reveal soil moisture and water quality. Multispectral sensors (Landsat, Sentinel-2) sample only 6-12 broad bands and cannot distinguish between spectrally similar materials. Wildfires, droughts, invasive species, harmful algal blooms, and mineral exploration all benefit from the diagnostic specificity of full imaging spectroscopy. NASA's EMIT instrument on the ISS (launched 2022) demonstrated that spaceborne imaging spectroscopy at 60 m resolution is feasible and has already discovered that mineral dust in the atmosphere is more absorptive than models assumed — a finding with direct implications for climate radiative forcing estimates. But EMIT covers only a narrow range of latitudes from the ISS orbit and lacks the global, systematic coverage needed for Earth system science.
What’s been tried and why it hasn’t worked
Airborne imaging spectrometers (AVIRIS, AVIRIS-NG) have demonstrated the science at spatial resolutions of 1-20 m since the 1980s, but airborne campaigns cover tiny areas (~thousands of km² per campaign) and cost ~$50-100/km² — global coverage is economically impossible. The EO-1/Hyperion satellite (2000-2017) carried a 220-band imaging spectrometer but had low signal-to-noise ratio, a narrow 7.7 km swath, and no atmospheric correction pipeline, limiting its scientific utility. The Italian Space Agency's PRISMA (2019) and Germany's EnMAP (2022) demonstrate spaceborne imaging spectroscopy but with 30 km swaths, making global systematic coverage impractical. The core engineering challenge is maintaining spectral calibration stability across thermal cycling in orbit, stray light rejection across a wide field of view, and the detector technology needed for high-SNR measurements in the SWIR (1000-2500 nm) where InGaAs or HgCdTe detector arrays are required. The data processing challenge is equally formidable: atmospheric correction of hyperspectral data requires pixel-by-pixel estimation of water vapor, aerosols, and surface adjacency effects using the spectral data itself — a computationally demanding inverse problem.
What would unlock progress
A Dyson or Offner imaging spectrometer design that achieves low spectral smile and keystone distortion across a wide (~185 km) swath while maintaining SNR >300:1 in the VSWIR. Advances in large-format HgCdTe detector arrays for the SWIR regime, with improved uniformity and lower dark current for space operation. Onboard or near-real-time atmospheric correction processing using physics-informed ML models trained on radiative transfer simulations. Data fusion approaches that combine hyperspectral (high spectral resolution, moderate spatial resolution) with multispectral (moderate spectral, high temporal) data to achieve both spectral specificity and daily-to-weekly revisit.
Entry points for student teams
A student team could process publicly available EMIT or EnMAP hyperspectral data to map a specific surface property (vegetation biochemistry, mineral type, or water quality) in their region, developing and validating atmospheric correction and spectral unmixing workflows using ground truth measurements. Alternatively, a team could design and build a low-cost hyperspectral camera (pushbroom spectrometer with a linear detector array) for drone or ground-based measurements, calibrating it against known spectral standards. Relevant disciplines: optical engineering, remote sensing, signal processing, ecology, geology.
Genes
Structural cousins: stuck the same way, in other fields
Sources
"Thriving on Our Changing Planet: A Decadal Strategy for Earth Observation from Space," National Academies of Sciences, Engineering, and Medicine, 2018. accessed 2026-02-16. Designated Observable TO-18 (Surface Biology and Geology); also NASA SBG Mission Architecture Study, 2021; EMIT mission documentation. go to source ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
- SBG was designated by the Earth Science decadal survey as one of the highest-priority observables. NASA's SBG mission is in formulation as of 2024-2025, with architecture studies evaluating instrument trades.
- The `failure:lab-to-field-gap` tag reflects the 40-year history of airborne imaging spectroscopy demonstrating the science while satellite implementation faces engineering challenges not present in aircraft operations (thermal stability, calibration over years, global data volumes).
- The `failure:not-attempted` tag reflects that no satellite has attempted the full SBG measurement specification (global, 30m, 10nm, high-SNR, VSWIR) — existing missions are either narrower in swath, fewer spectral bands, or lower SNR.
- EMIT's early science results (2022-present) have validated the measurement concept and motivated acceleration of SBG development.
- Cross-domain connection: shares the data-volume challenge with digital-astronomical-transient-alert-processing (instruments generating data faster than pipelines can process) and the airborne-works-but-satellite-doesn't structure with environment-snow-water-equivalent-measurement.
- Note reconciled 2026-08-20: a note above argues for `failure:not-attempted`; the genome now carries `failure:lab-to-field-gap` after a taxonomy revision. The original note is kept verbatim as the tagging rationale of record.
Reconciliation 2026-09-29: Fact repair after a site review read this write-up against its first source. The decadal survey was fetched in full on 2026-09-29 from the National Academies online reader (https://www.nationalacademies.org/read/24938, all chapters and appendixes, pages i to 694), and its citation record from https://api.crossref.org/works/10.17226/24938. (1) The Source line called Surface Biology and Geology "Designated Observable S-4." In the survey, Surface Biology and Geology corresponds to TO-18 in the Targeted Observables table (page 135 and Appendix C). S-4 is a different thing in the survey: a Solid Earth panel question about rates of landscape change. The Source line now says TO-18 in place of S-4. (2) The text said "ESA's PRISMA" and "the Italian-German EnMAP." The survey describes PRISMA as an Italian Space Agency program and EnMAP as a German Aerospace Center mission (page 377). The sentence now says that. Confirmed against the survey and left as written: the title, publisher and year in the Source line; hyperspectral imagery at 30-60 m as the priority for this observable (page 135); a spectral range of 380 to 2500 nm at 10 nm sampling (Appendix B, page 593; page 364 gives 400 to 2500 nm); about 10 nm spectral sampling, 30 m pixels and a 185 km swath at Landsat resolution (pages 364 and 401); 30 km swaths for PRISMA and EnMAP (page 377); Hyperion launched in 2000 with too little signal-to-noise for many uses (page 359); and that there is no global hyperspectral mission (page 376). Confirmed by a fetch of NASA's EMIT pages at earth.jpl.nasa.gov/emit on 2026-09-29, read in part: EMIT launched in 2022, sits on the International Space Station, and covers 380 to 2500 nm. NOT checked: the write-up also cites a 2021 NASA architecture study and EMIT documents without giving an address, and those could not be found and read as named. The figures that are not in the survey may rest on them and are left as written. These are calibration better than 2% across 200 or more channels, 5 or more terabytes per day, signal-to-noise above 300:1 (the survey asks for 400:1 in the visible and near infrared and 250:1 in the shortwave infrared), EMIT's 60 m resolution and its dust finding, the 2019 and 2022 launch years for PRISMA and EnMAP (the 2018 survey gives planned launches of 2018 for PRISMA and 2018 or 2019 for EnMAP, pages 377, 576 and 577), Hyperion's 220 bands, 7.7 km swath and 2017 end date, the airborne cost and coverage figures, the detector and optical design details, and the mission's 2024-2025 status. The write-up says Landsat and Sentinel-2 sample 6-12 broad bands; the survey counts 13 bands of all kinds for Sentinel-2 (page 549), and this was left as written because the write-up counts only broad bands. Genes untouched.