Evidence for surface water ice in the lunar polar regions using reflectance measurements from the Lunar Orbiter Laser Altimeter and temperature measurements from the Diviner Lunar Radiometer Experiment

Abstract

We find that the reflectance of the lunar surface within 5 ° of latitude of the South Pole increases rapidly with decreasing temperature, near ~110K, behavior consistent with the presence of surface water iceThe North polar region does not show this behavior, nor do South polar surfaces at latitudes more than 5° from the pole. This South pole reflectance anomaly persists when analysis is limited to surfaces with slopes less than 10° to eliminate false detection due to the brightening effect of mass wasting, and also when the very bright south polar crater Shackleton is excluded from the analysis. We also find that south polar regions of permanent shadow that have been reported to be generally brighter at 1064 nm do not show anomalous reflectance when their annual maximum surface temperatures are too high to preserve water ice. This distinction is not observed at the North Pole. The reflectance excursion on surfaces with maximum temperatures below 110K is superimposed on a general trend of increasing reflectance with decreasing maximum temperature that is present throughout the polar regions in the north and south; we attribute this trend to a temperature or illumination-dependent space weathering effect (e.g. Hemingway et al. 2015). We also find a sudden increase in reflectance with decreasing temperature superimposed on the general trend at 200K and possibly at 300K. This may indicate the presence of other volatiles such as sulfur or organics. We identified and mapped surfaces with reflectances so high as to be unlikely to be part of an ice-free population. In this south we find a similar distribution found by Hayne et al. 2015 based on UV properties. In the north a cluster of pixels near that pole may represent a limited frost exposure.

Figure 1.

Plots illustrating the relationship between Diviner maximum temperature and LOLA 1064 nm reflectance at the lunar poles (defined as regions within 20 latitude of the pole). Top: Two-dimensional histograms of 1064 nm normal albedo vs. maximum temperature for the lunar North (a), and South (b) polar regions. Normal Albedo bin width=0.002, Maximum Temperature bin width=1K; Center: Two-dimensional histograms of 1064 nm normal albedo vs. maximum temperature where data within each temperature bin is normalized to the total area within that temperature bin, for the lunar North (a), and South (b) polar regions. Normal Albedo bin width=0.002, Maximum Temperature bin width=3K. Histograms showing maximum temperature distribution for the North (e) and South (f) poles

Figure 2.

Average reflectance plotted as functions of maximum temperature for the (a) North and (b) South poles. Maximum temperature bin width=1K. Average reflectance curves are plotted with 2σ standard error of the mean.

Figure 3.

Two-dimensional histogram that displays mode-normalized LOLA reflectance distribution as a function of surface slope for the North and South pole. Reflectance bin width=0.005 and slope bin width=1 degree . The strong influence of surface slope is a potential source of temperature-independent bias on the reflectance-temperature relationships observed in the polar regions. [1 column]

Figure 4.

Average 1064 nm reflectance as a function of maximum temperature for low slope (slope < 10°) and high slope (slope > 20°) areas for the North (a) and South (b) polar regions. Maximum Temperature bin width=2K. Reflectance average plotted with 2σ errors (standard error of the mean). [1.5 column]

Figure 5.

Average 1064 nm reflectance as a function of maximum temperature (maximum temperature bin width=3K) for low slope (slope < 10°) areas free of mass wasting influence for areas within 5 degrees of each pole, and between 5 and 20 degrees from each pole., illustrating spatial heterogeneity in polar reflectance-temperature relationships. a) North polar data; b) South polar data. Curves plotted with 2σ standard error of the mean. Volatility thresholds for elemental sulfur and water ice are included for comparison. Thresholds were derived by Zhang and Paige [2010] a sublimation rate of 1mm/Gyr.

Figure 6.

Plot comparing LAMP water band ratio value (positively correlated with increasing water abundance) as a function of maximum Diviner temperature [Hayne et al., 2015], with average South polar LOLA reflectance as a function maximum Diviner temperature derived in this study. Maximum temperature bin width=2K. LOLA reflectance data shown include all data within 20 degrees of the South Pole. The volatility threshold for water ice, derived by Zhang & Paige [2010] for 1mm/Gyr is included for comparison. [1 column]

Figure 7.

a) Scatter plot of South polar reflectance vs. maximum temperature, with points in red derived from a small subset chosen to isolate South-polar Shackleton crater. Region included as red points are shown in (b). c) Curves comparing average 1064 nm reflectance as a function of maximum temperature for the south polar dataset within 20 degrees of the pole (black), and the same data with the ~25×25 km Shackleton subset shown in (b) removed (red). Maximum temperature bin width=2K. Curves are plotted with 2σ standard error of the mean. Exclusion of Shackleton shifts, but does not eliminate, the sharp uptick in reflectance below 110K. [2 column]

Figure 8.

a,b) Maximum temperature distribution of all PSR regions (defined as having average incident solar flux of 0%), compared to the maximum temperature distribution of low slope (slope < 10°) PSR regions, for the North(a) and South(b) poles (maximum temperature bin width=2K). Many portions of these PSRs exhibit maximum temperatures far above the stability threshold for water ice. c,d) LOLA normal albedo distribution of low temperature PSR regions capable of sustaining surface water frost (maximum temperature < 110K, blue), compared to reflectance distribution of high temperature PSR regions incapable of sustaining surface frost (maximum temperature > 125K magenta) at the North(c) and South(d) poles (reflectance bin width=0.005). Reflectance distribution of PSR regions with no temperature constraints included for comparison purposes (black) (reflectance bin width=0.005]) e,f) same as c, d but constrained to low slopes (< 10 degrees) to minimize mass wasting influence. [1.5 column]

Figure 9.

Results showing the effect of varying our three ice model input parameters: ‘ice reflectance’ (a), ‘sub-pixel ice cover’ (b), and ‘percent ice bearing pixels’ (ice-heterogeneity)(c). This model applies ice-brightened material to an ice free distribution via linear mixing.The ice-free starting distribution was we chose the reflectance distribution of low slope (slope <10°) South-polar PSRs (average incident flux=0%), with maximum temperatures high enough (>125K) to preclude the preservation of surface ice. Reflectance bin width of all histograms equals 0.005. [1.5 column]

Figure 10.

Results of some combinations of ice-modeling parameters consistent with plausible modes of ice preservation, compared to the real reflectance distributions of low slope (slope <10°) South-polar PSRs (average incident flux=0%), with maximum temperatures low enough (<110K) to preserve surface ice (filled blue), and high enough (>125K) to preclude the preservation of surface ice (filled orange). Reflectance bin width=0.005. a) results of adding sub-pixel scale, heterogeneously distributed icy material to the reflectance distribution of ice-free south polar PSRs (black solid line). b) results of adding sub-pixel scale, homogeneously distributed icy material (black dashed line), and supra-pixel scale heterogeneously distributed icy material (blue dashed line) to the reflectance distribution of ice-free south polar PSRs. [1.5 column]

Figure 11.

Detection maps of anomalously bright pixels in the north (top) and south (bottom). Cyan pixels are show the locations of surfaces that are brighter than 2-sigma than the mean of the 125–175K population of pixels that have maximum temperatures between 125 and 175K, and have maximum temperatures less than 110K and slopes less than 10˚. At both poles there are concentrations of pixels that indicate a common brightening process. Being brighter than 2-sigma, these clusters are unlikely to be part of the background ice-free population.