How data science is helping astronauts prospect for lunar ice
The moon will once again play a crucial role in the space race, but this time as the gatekeeper to the solar system, acting as a link in the planetary supply chain.
Ice deposits found in shadowed craters near the poles can act as resource wells for visiting astronauts, providing not only drinking water for humans and crops, but the key component needed to create rocket fuel.
By manufacturing propellant on our closest neighbour, traversing space could become like a long-haul flight – a stopover at the moon to refuel and rest, before continuing on to your final destination of Mars (cruises to Venus available at the end of this century).
For this to be possible, astronauts must not only know where to find these deposits, but how far under the surface they’re located, the equipment needed to extract them, and their chemical composition.
By applying predictive analytics and data science, researchers from CSIRO and CSIRO’s Data61 are generating detailed interactive maps to identify the most likely places to find ice reserves on the moon.
“By layering this data together, we’ve created a virtual map containing 2D and 3D interactive models showing where ice deposits on the moon can be found,” says project lead Dr Craig Lindley, a computational modelling research scientist at Data61.
“Insights from the data have shown strong evidence that ice is present and possibly exposed on the lunar surface in many permanently shadowed regions such as deep craters in the north and south poles.”
“Our maps will quantify the probability of ice deposits in these areas, calculating how easy they are to reach, how thick they are, their relation to subsurface ice, the deposit’s texture and regolith (layer of material covering solid rock, which can come in the form of dust, soil or broken rock), and the variation of concentration in relation to depth.”
The project can be a key part of the NASA Moon to Mars initiative, with the team using data gathered by NASA’s Lunar Reconnaissance Orbiter and other spacecraft to create a model of potential lunar ice in a prototype space resources data platform that includes a framework for generating 3D material models of asteroids and planetary surfaces.
This suite of tools can also infer the internal composition of a structure and measure potential changes to deposit shape, construction and behaviour induced by mass loss due to resource extraction.
While earth’s nearest celestial neighbour is the focus for now, the system has also been applied to asteroids and can be used for other planets.
“This is all part of developing a more general open source Celestial Object Resource Atlas (CORA) that can include data and analyses,” says Dr Lindley.
Basic processing can be done in existing GIS (Geographic Information System) tools, both commercial (ArcGIS) and open source (QGIS), explains Dr Lindley, with the team currently focussing on building additional analytics tools to determine where resources are most likely to be found, and their 3D structure.
“We’re currently creating a Python library to provide these analyses, which can be added to the merging or ‘burning’ of multiple raster layers into a single visualisation layer (a task currently performed manually in GIS tools).”
“The system has its own web interfaces for interactive visualisation and can also be interfaced with GIS tools that have Python APIs to provide a bridge between custom and off-the-shelf tools.”
A web version of the tool is planned to be released publicly within the next month.
With recent news that there may be life in Venus’s harsh atmosphere, the need to ensure space travellers are as well-prepared as possible to venture into the most challenging conditions known to humankind are crucial next steps as we prepare to explore beyond the moon.
CSIRO’s Data61 is working in collaboration with CSIRO’s Space Technology and Future Science Platform, CSIRO’s Mineral Resources, the Australian Centre for Space Engineering Research (ACSER) at UNSW, the Space Science and Technology Centreand School of Earth and Planetary Sciences, both at Curtin University, the Centre for Astrophysics at the University of Southern Queensland, HEO Robotics P/L, and NEO Resource Atlas P/L (NEORA).
A special thanks to Scott Dorrington who developed the scientific analysis tools, the VoxelNET team led by Charlotte Sennersten, computer scientist, and with Ben Evans, software engineer, working on the interactive visualisations, and CSIRO Mineral Resources Hard Rock Mining Director Ewan Sellers.