Improving the Apollo 12 landing site mappingwith Chandrayaan M 3 data

Yann Chemin1,2, Ian Crawford1, Roberto Bugiolacchi1, Huma Irfan1 and Louise Alexander1

1Birkbeck College, University of London, U.K. 2IWMI, Pelawatte, Sri Lanka

The geology of the Apollo 12 landing site has been the subject of many studies, including recently by [5] and [9]. This research attempts to bringadditional understanding from a remote sensing perspective using the Moon Mineralogy Mapper (M 3) sensor data, onboard the Chandrayaanlunar orbiter. This has a higher spatial-spectral resolution sensor than the Clementine UV-Vis sensor and provides the opportunity to study thelunar surface with detailed spectral signatures.

Mapping of FeO (wt%) and TiO 2 (wt%) is done using the methods of [7] and [11]. A FeO & TiO 2 processing module (i.feotio2) is madespecifically for this research within the Free & Open Source Software GRASS GIS [8]. Attempts will be made to estimate the lava flow thicknessusing the method of [2] and individual lava layers thicknesses from [10]. Integration of this new information will be put in perspective andintegrated with previous work. Analysis from the combined higher spatial and spectral resolutions will improve the accuracy of the geologicalmapping at the Apollo 12 landing site.

Abstract

[5] reviewed the lunar research done until recently, with a special interest in the Apollo 12 landing site and its vicinity.Among the information newly deducted, are a geological context block diagram 1 (Figure below) along with an Apollo12 landing site interpretive cross-section. [4] completed the digital renovation of the 1:5,000,000 lunar geological mapsseries, enhanced with LOLA-LRO information.

Figure 1: a) Geological block of Apollo 12 landing site from [5].b) Geological maps series with moon nomenclature from [4].

Figure 2: Interpretive cross-section of Apollo 12 landing site from [9].

Previous interpretations

[1] A.M. Baldridge, S.J. Hook, C.I. Grove, and G. Rivera. The {ASTER} spectral library version 2.0. Remote Sensing of Environment, 113(4):711 – 715, 2009.

[2] Roberto Bugiolacchi, Paul D Spudis, and John E Guest. Stratigraphy and composition of lava flows in mare nubium and mare cognitum. Meteoritics & Planetary Science, 41(2):285–304,2006.

[3] Roger Nelson Clark, Gregg A Swayze, R Wise, E Livo, T Hoefen, R Kokaly, and Stephen J Sutley. USGS digital spectral library splib06a. US Geological Survey Denver, CO, 2007.

[4] CM Fortezzo and TM Hare. Completed digital renovation of the 1: 5,000,000 lunar geologic map series. LPI Contributions, 1719:2114, 2013.

[5] Randy L Korotev, Bradley L Jolliff, Ryan A Zeigler, Stephen M Seddio, and Larry A Haskin. Apollo 12 revisited. Geochimica et Cosmochimica Acta, 75(6):1540–1573, 2011.

[6] DJ Lawrence, WC Feldman, RC Elphic, RC Little, TH Prettyman, S Maurice, PG Lucey, and AB Binder. Iron abundances on the lunar surface as measured by the lunar prospectorgamma-ray and neutron spectrometers. Journal of Geophysical Research: Planets (1991–2012), 107(E12):13–1, 2002.

[7] Paul G Lucey, David T Blewett, and Bradley L Jolliff. Lunar iron and titanium abundance algorithms based on final processing of clementine ultraviolet-visible images. Journal of GeophysicalResearch: Planets (1991–2012), 105(E8):20297–20305, 2000.

[8] M. Neteler, M.H. Bowman, M. Landa, and M. Metz. GRASS GIS: a multi-purpose Open Source GIS. Environmental Modelling & Software, 31:124–130, 2012.

[9] Joshua F Snape, Louise Alexander, Ian A Crawford, and Katherine H Joy. Basaltic regolith sample 12003,314: A new member of the apollo 12 feldspathic basalt suite? In Lunar and PlanetaryInstitute Science Conference Abstracts, volume 44, page 1044, 2013.

[10] Shoshana Z Weider, Ian A Crawford, and Katherine H Joy. Individual lava flow thicknesses in oceanus procellarum and mare serenitatis determined from clementine multispectral data.Icarus, 209(2):323–336, 2010.

[11] BB Wilcox, PG Lucey, and JJ Gillis. Mapping iron in the lunar mare: An improved approach. Journal of Geophysical Research: Planets (1991–2012), 110(E11), 2005.

[12] W Zhang and NE Bowles. Mapping lunar tio2 and feo with chandrayaan-1 m3 data. In Lunar and Planetary Institute Science Conference Abstracts, volume 44, page 1212, 2013.

www.bbk.ac.uk grass.osgeo.org

References

1 #!/bin/env python

2 from math import *

3

4 def feolucey2000(uvvis2, uvvis4):

5 return(17.427*(-atan(((uvvis4/uvvis2)-1.19)/(uvvis2-0.08)))-7.565)

6

7 def feolawrence2002(uvvis2, uvvis4):

8 return(5.7*((-0.147+0.372*(-(uvvis4/uvvis2-1.22)/(uvvis2-0.04))+\

9 (-0.036)*pow((-(uvvis4/uvvis2-1.22)/(uvvis2-0.04)),2)))+2.15)

10

11 def feowilcox2005(uvvis2, uvvis4):

12 return(-137.97*((uvvis2*0.9834)+(uvvis4/uvvis2*0.1813))+57.46)

13

14 def feozhang2013(uvvis2,uvvis4):

15 return(17.83*(-atan(((uvvis4/uvvis2)-1.26)/0.01)) - 6.87)

16

17 print "@Surveyor"

18 #M3 at Surveyor crater

19

20 uvvis2=0.05028829

21 uvvis4=0.0579434

22

23 print "@Bench"

24 #M3 at Bench (north outside)

25

26 uvvis2=0.05045871

27 uvvis4=0.05715561

The wt%FeO equations have a large range of resultswhen used with M3 data:

Authors Bench Surveyor[7] -26.64 -23.32[6] 4.72 6.38[11] 22.28 21.82[12] 19.74 19.49[12] (corrected) 15.65 14.76

It is known that the local range of FeO is below10wt% from UVVIS data (see Figure on the lowerleft), only the [6] equation is giving an appropriaterange. [7] equation fails because of the offset of uv-vis4/uvvis2(=1.15 vs offset = 1.19) and the one foruvvis2(=0.05 vs offset = 0.08), as both of the parts arenegative thus giving a positive division output, their in-terplay should bring a negative number for the arctan-gent negative sign to compensate. This clearly does nothappen with M3 data in the Bench and Surveyor cra-ters, the results from the arctangent being positive, andthe sign change brings the main part of the equation toa negative, further reduced by the negative final offset.

Upon using the equation from [6], it turned out that though the wt%FeO range was satisfactory for few crater points,the rate of change was actually reverse to the expected one, reducing to null towards the South of the landing site andincreasing towards the North.

The equation from [12] is giving an appropriate rate of change, however, the local values are too large compared to theClementine derived FeO values. Comparing the zonal contrast between Clementine FeO map from [7] the M3 from[12] leads to an interesting prospect as even the shapes are not consistent (see Figures & ). This indicates that theinner work of the angle equation is not sensitive to the same type of distance between R950 and R750 as they are inClementine and in M3. In an ideal condition, the calibration of both sensor bands would be identical, and the sensorresponse for each band would also be the same. This is not a reality, from the design and manufacture inheritance ofmulti-spectral sensors and hyperspectral sensors, to the data collection flight and its environment. Finally, some accidentor issues can damage the operation.

After Contacting W. Zhang of [12] about the issue in its algorithm, he promptly answered that the article had a typo inthe equation. After correction, the output were in a better range (about 15 for the two craters in the table above).

Figure 4: Subset of Clementine FeO using [7].

Figure 5: Subset of M3 FeO using [12].

Figure 6: Region of M3 FeO using [12] corrected.

M3 derived FeO from standard equations

Isolating the 95% top and 5% bottom of the destriped FeO image brought the possibility to suggest the Fresh CraterEjecta (FCE) values. The FCE was found near Lansberg P in the surroundings of (-22.8215621403E,-2.09875487116N)with 11 samples averaging to 17.87 FeO wt%.

Figure 7 (left): Locating FCE above Lansberg P, with the upper 95% FeO wt%Figure 8 (right): Locating lowest E Westward Copernicus, with the lower 5% FeO wt%.

The H value is the lower boundary of the FeO wt%in the region, typically found in the nearest highlandregion. There are two solid constraints to it value.H should be strictly less than E and strictly morethan zero (0 < H < E). The E value was found inCopernicus ejecta ring and actually taken from theFeO wt% map. The minimum value found in theFeO M3G20090111T013904.destripe map is 3.88 wt%,belonging to the ejecta ring of Copernicus. Thus itcan be inferred that 0 < H < 3.88, strictly, in thisCopernican work. Looking at over M3 FeO tiles, thereis no lower value to the crater ejecta of Copernicus.Some Alpes Formation and Dark Mantling Material inthe vicinity of Gay Lussac do not go lower than this.This is obviously shows the limit of this statisticalsearch method. Another attempt was undertakenusing the original tile FeO M320090610T113334 avisual search for H in the vicinity of Montes Carpatus(Northwestern of Copernicus with -23.6070634574E,14.571534275N) was done and from 14 samplesfound H = 5.88. Another visual search in the ejectablanket of Copernicus for a FCE value, 22 sampleswere collected, giving an average FCE value of 7.16.Thus, it is found that 0 < H < FCE, which is alogical outcome. It is thus proposed to carry forwardthis parametrisation for mapping the depth of theCopernican lava flow with the following equation 1.

Mh =D∗8× E −H

FCE −H[m] (1)

Which realizes as the following equation 2.

Mh =117000

8× E − 5.88

17.87− 5.88[m] (2)

Figure 9: Attempt at Maria Depth mapping

Lava Flow Depth

Some attempts at using M3 data to assess Apollo 12 landing site properties has proven to be constrained at every stepsof the study. The cartographic, radiometric and data noise issues were of the most difficult to address. Combined, theseissues reduced the mosaicking potential, thus the area covered.

The Lava Flow Depth estimation, to this date is still very doubtful and has to be research from scratch. Differences inestimation of FeO (wt%) from M3 and UVVis are not negligible. This is particularly troublesome.

Conclusions

Clementine data was downloaded frompdsimage.wr.usgs.gov, UV-Vis 5bands covering tiles ui03s333.img andui03s339.img, both 3 degrees Southand surrounding the 336E meridian.Ingestion of UV-Vis 5 bands in Isis(isis.astrogeology.usgs.gov) wasdone by using the pds2isis conversionmodule, and import in GRASS GISdone by r.in.gdal. This can be done inISIS-GRASS bridge mode.

An object-based classification is run usingthe i.segment module on the UVVisdataset. The output is vectorized withr.to.vect and FeO statistics are extractedper vector polygons (also called zonalstatistics) using v.rast.stats. The meanzonal FeO statistics indicate a very clearthreshold increase South of Apollo 12landing site.

Figure 10 a): UVVIS RGB composite

Figure 10 b): Object-oriented classification(i.segment)

Figure 10 c): Zonal statistics (v.rast.stats)

Clementine data

Chandrayaan M3 data was downloaded from PDShttp://pds-imaging.jpl.nasa.gov/. The tileM3G20090111T013904 V01 RFL.IMG was im-ported in GRASS GIS, resulting in 85 bands. Thespectrum for four craters where extracted in theFigure . Bench and Surveyor are both together in thelower reflectance curves. On the higher side, Sharpand Middle Crescent are consistently above the twoothers. This higher broadband Albedo is increasingtowards the West, where one of the ray is clearlydrawn (Oblique white strip in the middle of TheClementine upper map. Compared to the spectraldata from Figure the reflectance is overall lower andthe obvious shape differences are there. Additionalspectrum signature data from [3] is in the process ofbeing extracted for analysis.

Figure 11:Spectral signature of M3 from four craters.

Figure 12:Spectral signature from Apollo 12 (Speclib v2.0).

Figure 13:Spectral signature from Apollo 12 from [1].

M3 Signal at Apollo 12 landing site