Assessing Observation Coverage and Localization Accuracy of Lunar Impact Flashes with ESA’s LUMIO Mission

Introduction: The Lunar Meteoroid Impact Observer (LUMIO) is a 12U CubeSat mission designed to observe, quantify, and characterize lunar meteoroid impacts [1]. After commissioning and transfer phases, it will operate nominally for one year in a quasi-halo orbit around the Earth-Moon L2 point [2]. By detecting Lunar Impact Flashes (LIFs)— brief bursts of light produced when meteoroids strike the Moon’s surface— from the far side of the Moon, LUMIO will extend the coverage of impact monitoring beyond Earth-based telescopes, which are limited to the nearside and affected by weather conditions [3].

The primary scientific goal of LUMIO is to estimate the meteoroid flux at the Moon and address the lack of data in current kinetic energy spectrum [1,3]. Besides the total number of LIFs detected, it will be necessary to estimate the location of each impact on the lunar surface. This will enable mapping the spatial distribution of impacts and potentially classifying each event as either sporadic or part of a known meteoroid stream.

Problem: Due to LUMIO’s orbit, the camera will record the far side of the Moon’s surface at a variable distance from the Moon while varying the sub-satellite point. Additionally, due to solar illumination, the lunar phases will affect the time period during which LUMIO can actually record LIFs. Moreover, the camera will record LIFs only if the lunar disk is shadowed beyond a set threshold (≈50%, TBD), to avoid excessive frame saturation and to allow the Navigation & Engineering cycle sufficient time for their analysis. The combination of these effects will result in a non-uniform coverage of the Moon’s far-side surface and thus non-uniform detection areas for LIFs. To properly estimate the meteoroid flux at the Moon, these non-uniformity effects must be taken into account. Simulations of the effective observable surface have been conducted to predict the coverage and the losses LUMIO will experience during its one year of operations.

Due to limited downlink capacity, full-frame images cannot be transmitted for each LIF. Instead, cropped pixel-area sequences around the impact are stored and sent. As the classic Moon feature-fitting method [3,4] is not feasible, a kernel-based approach is considered. SPICE data for LUMIO, the Moon, and the Sun, along with camera calibration, allow localization of impacts using detection time and triggered pixels. This can be further refined using additional methods such as limb fitting, feature recognition, or star occultation. To evaluate the suitability of this approach and possible improvements, different methodologies have been implemented to estimate related uncertainties.

Methods and Results: The mission’s primary instrument, LUMIO-Cam, is an optical sensor capable of detecting LIFs in both the visible and near-infrared spectral range (450-950 nm).

Given the LUMIO, Moon, and Sun kernels, tools can be created to compute and study the LUMIO-Cam footprint on the Moon, eventually considering limitations due to solar illumination. A set of analyses can assess the LUMIO’s nominal Moon surface coverage during the operational phase of the mission, to better understand what LUMIO-Cam will be able to observe and with what accuracy. For example, it is possible to estimate which areas will be covered and during which periods, to later compare with expected meteoroid stream events (Fig.1).

Errors in the camera’s calibration (i.e., optical center and focal lengths), LUMIO’s attitude and position, and LIF image processing have been considered to perform perturbation analyses and thus obtain an estimate of kernel-based LIF localization errors through Monte Carlo simulations (Fig.2). These highlight that the primary source of error for LIFs kernel-based localization is the uncertainty in the camera’s optical center position. Poor camera calibration can yield a localization uncertainty of 20 km × 20 km despite accurate navigation, whereas with  perfect calibration,  worst-case navigation uncertainty  limits the area to 3 km × 3 km. In addition to numerical simulations, formal analytical formulas have been derived and validated to describe the area on the lunar surface covered by a specific pixel and thus the best achievable uncertainty for each pixel at a given LUMIO-Moon distance (Fig.3).

After assessing a kernel-based LIF localization method, additional strategies are being developed to decrease localization uncertainties using information from the limited number of pixels downloadable for each event.

Conclusions: A set of tools and analyses has been developed to study what LUMIO-Cam will be able to observe and with what accuracy, and to implement an effective LIF localization strategy. After assessing the reliability of a kernel-based localization method, additional improvements are being developed to refine its accuracy.

Acknowledgment: LUMIO is developed under the European Space Agency’s (ESA), and it is led by Politecnico di Milano (PoliMi). It is supported by the Italian Space Agency (ASI) within the LUMIO project (ASI-PoliMi agreement n. 2024-6-HH.0), the Norwegian Space Agency (NOSA), United Kingdom Space Agency (UKSA), and Swedish National Space Agency (SNSA).

References: [1] Topputo F. et al. (2023) Icarus, 389, 115213. [2] Cipriano A. et al. (2018) Front. Astron. Space Sci., 5, 29. [3] Liakos A. et al. (2024) Astron. Astrophys., 687, A14. [4] Madiedo J.M., Ortiz J.L., Morales N., Cabrera-Cano J.(2015), PSS, 111,105.

Figure 1 2D Equidistant-Azimuthal maps (centered at Lat. 0°, Lon. 180°). Left: Example of LUMIO (red) and Sun (yellow) footprints at a specific time. Right: Colormap of LUMIO’s effective lunar coverage over one year, showing normalized coverage frequency. Superimposed: LUMIO sub-satellite points during Science (red) and Navigation & Engineering (white) cycles.

Figure 2 Example of LUMIO LIF localization and associated uncertainty (blue dots; average in red) for a fixed time and pixel location. Left: Moon far-side view centered at Lat. 0°, Lon. 180°. Right: 2D zoom-in (latitude vs. longitude).

Figure 3 Area covered (δArea) on the lunar surface [km²] by each LUMIO-Cam pixel along the central row (x-axis), evaluated at LUMIO’s minimum (blue) and maximum (orange) distances from the Moon. Dashed lines show twice the minimum area at each distance. The red dash-dot line indicates a threshold used to count pixels (on both sides of the frame) exceeding this value (shown in the top-left corner).