Lessons From Marius Hills Pit Analog Tests of EchoQuilt
Problem
The Marius Hills Hole is the closest thing planetary science has to a confirmed lunar lava tube entry, and it is also the hardest analog target the EchoQuilt team has tested against. The Wiley GRL paper by Kaku et al. on detection of intact lava tubes at Marius Hills via SELENE/Kaguya LRS laid out the radar echo pattern evidence for a roughly 50 km intact tube, and the ISAS JAXA topic on detection of intact lunar lava tubes via SELENE radar confirmed the finding from JAXA's own analysis path. The Wiley GRL paper on evidence of large empty lava tubes on the Moon using GRAIL gravity corroborates the result with an independent gravity-signature line of evidence.
The analog-test problem is that Marius Hills gives you two orthogonal remote-sensing constraints — the radar echo pattern and the gravity signature — and any quilt generated against a terrestrial analog has to produce results that are consistent with both. A quilt that satisfies the radar constraint but produces a gravity signature no one would expect is a quilt that will not survive peer review. The ScienceDirect distribution, formation mechanisms, and significance of lunar pits paper also puts Marius Hills in context against the broader lunar pit catalog, which means the analog tests have to generalize beyond just this one pit.
A cross-niche parallel from conservation biology informs the validation methodology. The karst-lava hibernacula comparison work uses the same aggregate-across-sites strategy that the Marius Hills analog campaign requires: a single hibernaculum cannot validate a regional cave classification scheme, but the aggregate of multiple hibernacula across geological types can. The conservation biology community's discipline around aggregating across sites with explicit per-site provenance has been useful for the planetary team, particularly when explaining the validation logic to reviewers who are skeptical of single-analog projection claims.
Solution
The Marius Hills analog test campaign ran against four terrestrial pit-adjacent tubes — two at Mauna Loa, one at Lofthellir, and one at La Corona — and produced quilts that were then back-projected against SELENE LRS echo pattern predictions and GRAIL-analog gravity signatures. The back-projection is the critical step: it is not enough to produce a high-fidelity quilt; the quilt has to be testable against the remote-sensing data that is actually available for the lunar target. The LPSC 2017 paper on detection of lunar lava tubes by SELENE radar sounder documents the radar methodology for pit-adjacent detection that our back-projection inherits, and the ScienceDirect GRAIL gravity gradients evidence for a potential lava tube at Marius Hills is the 2023 reanalysis that provided the stronger gravity baseline we used for cross-validation.
Four lessons emerged from the campaign. First, the quilt's stitching engine produces radar-compatible echo patterns without retuning — the same 1.8 W patch-density configuration that served the Hadley analog campaign produced back-projected echo signatures within 6 percent of SELENE LRS observations at matched geometry. Second, the gravity-signature validation is slower and requires a coarser quilt, because gravity signatures average over a volume that no terrestrial analog patch spans alone; the quilt has to be aggregated across all four analog tubes before the gravity comparison becomes meaningful. Third, the pit-mouth geometry — the transition from skylight to intact tube — is the region where the quilt is most sensitive to calibration error, and it is also the region where planetary geologists will be looking hardest when a real Marius Hills mission flies.
The fourth lesson is about cross-site aggregation. A single terrestrial analog tube cannot validate a Marius Hills-class quilt on its own; the aggregate of four analogs can. This pattern mirrors the approach we take in our analog case studies Hadley writeup, where cross-validation against a second analog campaign strengthens the claim that a single-site campaign cannot make.
Advanced tactics
Three tactics sharpen the Marius Hills analog workflow past the default back-projection. First, run the analog quilt against both the 2017 and 2023 GRAIL analyses. The 2017 detection paper and the 2023 reanalysis use different assumptions about background gravity, and a quilt that agrees with both is substantially more defensible. We found that the 2023-compatible aggregation required roughly 1.4x more patches than the 2017-compatible aggregation, which has cost implications for an analog campaign and should be scoped up front.
Second, treat pit-mouth patches as a distinct patch class with their own provenance. The transition geometry is the highest-value region of the quilt for a Marius Hills-class mission, and a dedicated patch class lets you track its calibration separately. In our campaign, pit-mouth patches had their own residual distribution — slightly wider than the tube-interior distribution — and reporting that separately was the single clearest improvement to our analog paper's defensibility.
Third, publish the back-projection synthesis pipeline alongside the quilt. Reviewers who are skeptical of analog-to-lunar projection tend to accept it quickly once they can reproduce the synthesis from the quilt themselves. EchoQuilt's back-projection pipeline is packaged as a standalone tool for this reason, and it takes about 30 minutes to run a fresh back-projection against either SELENE LRS or GRAIL-analog gravity signatures.
Fourth, integrate the Marius Hills work with our planetary seismic interiors fusion work, which also relies on aggregating sparse events into a unified cave-interior picture. The seismic fusion layer can consume Apollo PSE arrivals that fall within the Marius Hills spatial window and add a third independent constraint alongside the radar and gravity constraints. A Marius Hills quilt that satisfies all three constraints (radar, gravity, and seismic) is substantially more defensible than a quilt that satisfies only two, and the seismic constraint is the cheapest of the three to add since it draws on existing public catalogs rather than requiring new analog data collection.
Fifth, document the analog-to-lunar transfer assumptions explicitly. The back-projection inherently assumes that terrestrial basaltic acoustic propagation maps cleanly onto lunar basaltic acoustic propagation, modulo some calibration adjustments. Where the assumption holds, the back-projection is informative; where it breaks (lunar regolith physics differs from terrestrial physics in ways that affect the propagation), the back-projection produces results that need careful interpretation. EchoQuilt's documentation includes an explicit assumptions section for the back-projection tool, which gives reviewers a clear basis for evaluating where the tool's outputs are most and least trustworthy.
Sixth, version the back-projection tool against the underlying remote-sensing data sources. SELENE LRS and GRAIL data are both subject to periodic reanalysis, and a back-projection that compared favorably against the 2017 GRAIL release may compare differently against the 2023 release. EchoQuilt's back-projection tool tags each run with the version of the remote-sensing data it was compared against, so historical results remain interpretable even as the underlying datasets evolve.
Seventh, share negative-result back-projections alongside positive-result back-projections. A back-projection that fails to match the SELENE LRS observations is informative because it shows where the analog quilt falls short of the lunar reference, and that information can drive subsequent campaign design to fill the gap. EchoQuilt's analog library includes the failed back-projections from intermediate campaigns alongside the successful ones, which lets concept teams see the full evolution of the Marius Hills analog approach rather than just the polished final product. Reviewers who see both kinds of results consistently report that the validation feels more credible than reviewers who see only positive results.
CTA
If your team is preparing a Marius Hills concept, a JAXA-collaborated lunar pit mission, or a NIAC study that has to demonstrate analog-to-lunar back-projection, EchoQuilt's Marius Hills analog library is ready to seed your work.
Each pilot ships with the four-tube aggregate quilt drawn from two Mauna Loa, one Lofthellir, and one La Corona analog campaigns, the SELENE LRS back-projection tool that produces echo signatures within 6 percent of observed at matched geometry, the 2017 and 2023 GRAIL cross-validation harness that catches the 1.4x patch-density delta between the two analyses, the pit-mouth patch class with its dedicated residual report (slightly wider than tube-interior distribution), the Apollo PSE seismic-fusion overlay that adds a third independent constraint alongside radar and gravity, an analog-to-lunar transfer assumptions documentation pack that gives reviewers a clear basis for evaluating where back-projection outputs are most trustworthy, and the negative-result back-projection archive from intermediate campaigns that drives subsequent analog campaign design.
Pilot teams shape the pit-mouth patch class taxonomy and the public Marius Hills aggregate quilt format that the 2027 reference release will publish under PDS4-conformant archival standards. Priority goes to NIAC PIs targeting Marius Hills concept proposals in the 2026 cycle, Artemis architect working groups scoping lunar pit habitat anchors, JAXA SELENE/Kaguya alumni running joint Marius Hills concept studies with ESA and NASA partners, ESA gravity-signature researchers integrating GRAIL reanalysis with cave-interior priors, and joint NASA-JAXA Marius Hills concept leads coordinating cross-agency back-projection reporting. Join the Waitlist for Planetary Analog Researchers and we will share the four-tube aggregate quilt, the SELENE LRS back-projection tool, and the 2017/2023 GRAIL cross-validation harness.