# Solar System Surface Technosignatures: Searching for Signs of Artificial Design on Extraterrestrial Terrain

**Authors:** Ana Henrique, Sofia Sheikh, Benjamin Fields  
**arXiv:** [2610.28687](https://arxiv.org/abs/2610.28687)  
**Date:** 2026-10-20  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2610.28687v1/

## Key points

- The paper proposes searching for long-lasting, non-random surface changes—such as construction, mining, habitation, and waste disposal—as possible technosignatures beyond radio emissions.
- Its goal is to establish quantitative limits on possible artifacts, not to claim that extraterrestrial artifacts have been found.
- Searches could combine optical imagery, LiDAR, ground-penetrating radar, and reflectance spectroscopy to detect unusual shapes, buried structures, or materials.
- The authors recommend pairing unsupervised or self-supervised anomaly detection with supervised classification trained on known human-made structures.
- Human-based training data may help identify familiar structures but could miss unfamiliar technologies or encourage false interpretations.
- Large existing datasets make searches feasible now: LRO’s Narrow Angle Camera has collected over two million lunar images, and Mars Reconnaissance Orbiter’s HiRISE catalog has over 92,000 images at best-case resolution of 30 centimeters per pixel.
- An initial self-supervised method flagged the Apollo 17 lander and unusual natural features, demonstrating anomaly detection—not evidence of extraterrestrial technology.
- On geologically quiet bodies, surface traces might persist for up to hundreds of millions of years; most Solar System surfaces have not been surveyed at sufficient resolution to rule out artifacts.
- Potential search targets include asteroids, rocky moons, Mars, and Kuiper Belt objects; geological activity may have erased traces on Earth or Venus.
- Surface technosignature searches could be a relatively low-cost addition to research and may also reveal scientifically valuable natural formations.
- No candidate extraterrestrial surface technosignature or confirmed artifact has been reported, and no quantitative upper limits have been published.
- The methods remain vulnerable to false positives, false negatives, and assumptions about artificial structures; the authors recommend independent data and, where possible, in situ follow-up.

## Summary

The paper asks whether technosignature searches should look beyond radio emissions for long-lasting changes that technological societies may make to planetary surfaces. Its central idea is that intentional modifications—such as construction, mining, habitation, and waste disposal—can create non-random patterns visible from orbit. Because human societies have left traces on Earth that persist for centuries or millennia, the authors argue that similar features could be sought on other rocky bodies in the Solar System. Their aim is to establish quantitative limits on the presence of such artifacts, not to claim that any have already been found. The proposed approach combines planetary imaging and other remote-sensing data with machine-learning methods. Optical imagery could be searched for unusual shapes or patterns. LiDAR and ground-penetrating radar could reveal buried or subsurface structures, while reflectance spectroscopy could identify materials with unusual chemical or mineralogical properties. The authors recommend developing two types of machine-learning searches in parallel. Unsupervised or self-supervised anomaly detection can identify features that differ from their surroundings without relying on examples of extraterrestrial artifacts. Supervised classification, by contrast, can use known human-made structures as training data. The paper cites terrestrial archaeological research showing that supervised methods can identify structures in satellite imagery. However, the authors caution that training on human examples could cause searches to miss unfamiliar technologies or encourage false interpretations. The opportunity is supported by the scale and resolution of existing planetary datasets. The Lunar Reconnaissance Orbiter’s Narrow Angle Camera has collected more than two million lunar images, amounting to petabytes of high-resolution imagery. The Mars Reconnaissance Orbiter’s HiRISE camera can achieve a best-case resolution of 30 centimeters per pixel, and its public catalog contains more than 92,000 images. The authors also describe an initial self-supervised anomaly-detection method developed for lunar imagery. It identified the Apollo 17 lander, a known human artifact, as well as scientifically interesting natural features, including recent impacts and volcanic pits. This result demonstrates the method’s ability to flag unusual objects, but it is not evidence of extraterrestrial technology. The rationale for these searches partly rests on the potential longevity of surface traces. On bodies with little tectonic or biological activity, artifacts could remain visible far longer than comparable traces on Earth; the paper cites estimates of up to hundreds of millions of years. Most Solar System surfaces, the authors note, have not been explored at a resolution sufficient to rule out artifacts. Potential search locations include asteroids, rocky moons, Mars, and Kuiper Belt objects. The paper also considers the speculative possibility that an earlier technological species existed in the Solar System, while acknowledging that geological processes would likely have erased traces on Earth or Venus. The authors conclude that searches for surface technosignatures are a promising, relatively low-cost addition to technosignature research. Existing archives make it possible to begin searching without waiting for new missions, and the work could also benefit planetary science by revealing unusual natural formations. However, the case remains theoretical: no candidate extraterrestrial surface technosignature or confirmed artifact has been reported, and no quantitative upper limits have been published. The authors do not estimate the probability that such artifacts exist, and the methods are vulnerable to false positives, false negatives, and assumptions about what artificial structures might look like. They therefore recommend combining supervised and unsupervised searches and, where possible, following up with independent data and in situ investigation.