We investigated the ability of a three-band lidar instrument (with wavelengths of 532, 1064, and 1560 nm) to detect and quantify different types of regolith ± water ice mixtures for both mare and highlands regions. A variety of samples were spectrally characterized, focusing on variables such as type of lunar regolith (simulant, lunar meteorite, highland vs mare), grain size (<45 μm, 45–1000 μm, <1 mm), water ice surface coverage and porosity of regolith (discontinuous, continuous, packed), and water ice percentages in intimate mixtures vs areal mixtures. Spectral metrics included absolute reflectance (532, 1064, 1560 nm) and reflectance ratios (1560/532, 1064/532, 1560/1064 nm). The ability to detect water ice in the presence of different types of regolith is critically dependent on the optical properties of end members, the selected wavelength, and the physical nature of the samples, such as grain size and areal vs intimate mixtures. Given the differences in the spectroscopic properties of the end members, both absolute reflectance and reflectance ratios have different predictive powers and both are required for robust analysis. The 1560 nm region is the most sensitive of the three wavelength regions to water ice content, as it falls well within a major water ice absorption band. Spectral contrast affects the ability to detect water ice. It is most sensitive for characterizing water ice + highland materials because the latter are brighter than water ice in this wavelength region. Powdered lunar materials are red-sloped across the 532 to 1064 to 1560 nm regions. By contrast, water ice shows a large downturn in reflectance between 1064 and 1560 nm. Detection limits for water ice in many situations can be as low as 2% and are lower for areal vs intimate mixtures. We found that porosity and angle of incidence with the surface have only minor effects on spectral reflectance properties of regolith. Further confirmation of water ice could be realized by comparing lidar data acquired by transects across permanently-shadowed regions and seeing how absolute reflectance and reflectance ratios vary between sunlit and adjacent shadowed regions. If it is known or assumed that regolith properties do not change as one enters a PSR, with the exception of the presence of surficial water ice, changes in reflectivity and reflectance ratios could provide compelling evidence for the presence of water ice.