JOURNAL OF GEOPHYSICAL RESEARCH: PLANETS, VOL. 118, 1397–1413, doi:10.1002/jgre.20101, 2013 Band-limited Bouguer gravity identifies new basins on the Moon W. E. Featherstone,1 C. Hirt,1 and M. Kuhn 1 Received 7 November 2012; revised 30 May 2013; accepted 2 June 2013; published 28 June 2013. [1] Spectral domain forward modeling is used to generate topography-implied gravity for the Moon using data from the Lunar Orbiter Laser Altimeter instrument operated on board the Lunar Reconnaissance Orbiter mission. This is subtracted from Selenological and Engineering Explorer (SELENE)-derived gravity to generate band-limited Bouguer gravity maps of the Moon so as to enhance the gravitational signatures of anomalous mass densities nearer the surface. This procedure adds evidence that two previously postulated basins on the lunar farside, Fitzgerald-Jackson (25°N, 191°E) and to the east of Debye (50°N, 180°E), are indeed real. When applied over the entire lunar surface, band-limited Bouguer gravity reveals the locations of 280 candidate basins that have not been identified when using full-spectrum gravity or topography alone, showing the approach to be of utility. Of the 280 basins, 66 are classified as distinct from their band-limited Bouguer gravity and topographic signatures, making them worthy of further investigation. Citation: Featherstone, W. E., C. Hirt, and M. Kuhn (2013), Band-limited Bouguer gravity identifies new basins on the Moon, J. Geophys. Res. Planets, 118, 1397–1413, doi:10.1002/jgre.20101. 1. Introduction [4] In this article, we instead compute band-limited Bouguer gravity (BGG) from the newer SGM100i gravity [2] Understanding the structure and evolution of the Moon fi model [Goossens et al., 2011] and LOLA (Lunar Orbiter has bene tted greatly from satellite-based topographic map- Laser Altimeter) topography [Smith et al., 2010]. This ping and satellite-based gravimetry. For instance, there are reveals basins that were previously masked when using full- (i) asymmetry between the nearside and the farside hemi- fi spectrum Bouguer gravity from SELENE alone [cf. spheres with a ~1.9 km offset between the center of gure Matsumoto et al., 2010] or topography alone [cf. Frey, (from topography) and the center of mass (from gravimetry) “ ” 2011]. We present three case studies to exemplify: (i) the [Smith et al., 1997], (ii) the early detection of mascons masking when using full-spectrum Bouguer gravity on its [Muller and Sjogren, 1968], and (iii) information on the own, (ii) validation of the band-limited approach using an long-wavelength isostatic compensation state of the lunar already known lunar basin, and (iii) adding more evidence crust [e.g., Wieczorek, 2007]. Overviews of the various lunar for the presence of two farside basins [Fitzgerald-Jackson satellite missions are given in, e.g., Floberghagen [2002], (25°N, 191°E) and to the east of Debye (50°N, 180°E)] using Matsumoto et al. [2010], Sinha et al. [2010], Vondrak et al. band-limited Bouguer gravity. Sensitivity analyses are [2010], and Zuber et al. [2013]. conducted by varying the topographic mass density used in [3] Early satellite-based studies of the farside had previ- the forward modeling and the degrees of band limitation, in- ously been hampered by the inability to track lunar-orbiting dicating these identifications to be robust. satellites from the Earth due to the synchronous rotation [5] The band-limited Bouguer gravity technique is then ex- and revolution of the Moon. This was redressed by the tended to the entire lunar surface, corroborating the presence Selenological and Engineering Explorer (SELENE) lunar of small-scale (~100 km to ~300 km in diameter) basins: gravity mission [e.g., Namiki et al., 2009], which has already Some are already known, some are probable, and some are revealed several new features on the farside. Most of these are uncertain but remain candidates for future investigations. large-scale basins because smaller structures could not be dis- criminated from noise in the high-degree spherical harmonic fi coef cients of the SGM100h gravity model [e.g., Matsumo 2. Methods and Data et al., 2010]. In addition, small-scale structures can be ob- scured by the long-wavelength gravitational signatures. [6] Namiki et al. [2009], Matsumoto et al. [2010], and sev- eral others generally use full-spectrum Bouguer gravity, 1Western Australian Centre for Geodesy and Institute for Geoscience where the spherical harmonic summations of gravity and Research, Curtin University of Technology, Perth, Western Australia, topography begin at degree n = 2. The method used in this Australia. paper starts the summations in equations (1) and (4) at > Corresponding author: W. E. Featherstone, Western Australian Centre for arbitrarily higher degrees (n1 2), thus enhancing the me- Geodesy and Institute for Geoscience Research, Curtin University of dium- and shorter-wavelength signals that are generated by Technology, GPO Box U1987, Perth, WA 6845, Australia. (w.featherstone@ near-surface mass anomalies (section 2.3). However, both curtin.edu.au) methods are inevitably restricted by the maximum reliable ©2013. American Geophysical Union. All Rights Reserved. degree of the lunar gravitational model and noise in the 2169-9097/13/10.1002/jgre.20101 high-degree coefficients. 1397 FEATHERSTONE ET AL.: BAND-LIMITED LUNAR GRAVITY [7] The use of band-limited data in the planetary sciences is difference between model gravity and reference gravity eval- not novel, however. For instance, Frey et al. [1996] corre- uated at exactly the same point, and the anomaly is the differ- lated band-limited free-air gravity with topography-implied ence evaluated at different points. However, the use of gravity in spherical harmonic bands on Mars. Han [2008] anomalies or disturbances has little consequence here, as used high-pass-filtered free-air gravity in the context of local- we are seeking to locate the basins, rather than apply the sub- ized spherical harmonic functions on the Moon. Zuber et al. tleties of geodetic approaches and terminology to the plane- [2013] emphasized shorter-scale Bouguer gravity signatures tary sciences. Hereafter, we simply use the term “gravity” through high-pass filtering, also on the Moon. However, to denote the radial derivatives of the gravitational potential band-limited Bouguer gravity is not used routinely in prac- at r = R from equation (1). tice. Finally, we note that the band-limited approach is con- ceptually quite similar to regional-residual separation that 2.2. Band-Limited Gravity From Lunar Topography has been applied to Bouguer gravity on Earth [e.g., Griffin, [11] The lunar topography model used herein comes from 1949; Nettleton, 1954]. the 2010 release produced by the LOLA instrument on board the Lunar Reconnaissance Orbiter (LRO) mission [Vondrak 2.1. Band-Limited Gravity From Lunar et al., 2010]. The LRO configuration delivers an ~18 m Gravitational Models along-track and an ~1.8 km across-track spacing at the equator ðÞ ðÞ fi [8] Spherical harmonic syntheses of SGM100i are used to [Smith et al., 2010]. The HC p nm and HS p nm coef cients generate band-limited lunar gravity disturbances, which are were taken from http://pds-geosciences.wustl.edu/lro/lro-l- synonymously termed radial derivatives of the gravitational lola-3-rdr-v1/lrolol_1xxx/data/lola_shadr/ (version lro_ltm02 potential in the planetary science literature [e.g., Wieczorek, _sha.tab), which are based on around 1 year of LOLA observa- 2007, p. 6]; this is tions (start date: 13 July 2009; end date: 20 August 2010) (G. Neumann, personal communication, 2013). While the maxi- ÂÃ n2 n n mum spherical harmonic degree available is 719, it has been δ n2 ¼ GM ∑ R ðÞþ ∑ gn1 SGM 2 n 1 truncated to degree 70 so as to be consistent with the maxi- r n¼n1 r m¼0 ÀÁ mum reliable degree of SGM100i. λ þ λ ðÞθ ðÞ Cnm cosm Snm sinm Pnm cos 1 [12] Band-limited spectral domain forward modeling was used to generate gravity from the LOLA lunar topography. À We have validated this spectral approach with brute force nu- where GM = 4902.80080 × 109 m3 s 2 [Goossens et al., 2011] merical integration of band-limited topography, with both ap- is the product of the universal gravitational constant G and proaches giving comparable results. The fully normalized the lunar mass M for SGM100i, r is the selenocentric radius spherical harmonic coefficients of the gravitational potential to the computation point, R = 1,738,000 m is the SGM100i of the topography are obtained from [e.g., Rummel et al., model's reference radius, n1andn2 denote the lower and upper 1988; Wieczorek and Phillips,1998] degrees of the band-limited syntheses, Cnm and Snm are the fi p fully normalized coef cients of SGM100i (from http://www ∏ ðÞn þ 4 À q TIG 2 p no .miz.nao.ac.jp/rise-pub/en), P ðÞcosθ are the fully normalized C 4π R max q¼1 HCðÞ p nm nm ¼ ρ ∑ nm (2) TIG þ !ðÞþ HSðÞ p associated Legendre functions, and θ and λ are, respectively, Snm 2n 1 M p¼1 p n 3 nm the colatitude and longitude of the computation point. The (n + 1) term in equation (1) delivers gravity disturbances, where pmax is the order of the series expansion, M is the lunar 22 whereas (n À 1) delivers gravity anomalies, which will be mass (7.347 × 10 kg, derived from GM/G for SGM100i), ρ discussed further below. SGM100i is provided to degree is the assumed mass density of the lunar topography, and HC ðÞp and HSðÞ p are derived from spherical harmonic anal- nmax = 100, but Goossens et al. [2011] recommend that it nm nm only be used to degree 70 because of noise in the higher-de- ysis of the surface function fi gree coef cients.