remote sensing

Review Satellite Remote Sensing of the Greenland Ice Sheet Ablation Zone: A Review

Matthew G. Cooper 1,* and Laurence C. Smith 1,2,3

1 Department of Geography, University of California, Los Angeles, Los Angeles, CA 90095, USA; [email protected] 2 Department of Earth, Environmental & Planetary Sciences, Brown University, Providence, RI 02912, USA 3 Institute at Brown for Environment and Society, Brown University, Providence, RI 02912, USA * Correspondence: [email protected]



Received: 15 August 2019; Accepted: 9 October 2019; Published: 16 October 2019


Abstract: The Greenland Ice Sheet is now the largest land ice contributor to global sea level rise, largely driven by increased surface meltwater runoff from the ablation zone, i.e., areas of the ice sheet where annual mass losses exceed gains. This small but critically important area of the ice sheet has expanded in size by ~50% since the early 1960s, and satellite remote sensing is a powerful tool for monitoring the physical processes that influence its surface mass balance. This review synthesizes key remote sensing methods and scientific findings from satellite remote sensing of the Greenland Ice Sheet ablation zone, covering progress in (1) radar altimetry, (2) laser (lidar) altimetry, (3) gravimetry, (4) multispectral optical imagery, and (5) microwave and thermal imagery. Physical characteristics and quantities examined include surface elevation change, gravimetric mass balance, reflectance, albedo, and mapping of surface melt extent and glaciological facies and zones. The review concludes that future progress will benefit most from methods that combine multi-sensor, multi-wavelength, and cross-platform datasets designed to discriminate the widely varying surface processes in the ablation zone. Specific examples include fusing laser altimetry, radar altimetry, and optical stereophotogrammetry to enhance spatial measurement density, cross-validate surface elevation change, and diagnose radar elevation bias; employing dual-frequency radar, microwave scatterometry, or combining radar and laser altimetry to map seasonal snow depth; fusing optical imagery, radar imagery, and microwave scatterometry to discriminate between snow, liquid water, refrozen meltwater, and bare ice near the equilibrium line altitude; combining optical reflectance with laser altimetry to map supraglacial lake, stream, and crevasse bathymetry; and monitoring the inland migration of snowlines, surface melt extent, and supraglacial hydrologic features.

Keywords: ablation zone; Greenland; ice sheet; surface mass balance; mass balance; altimetry; albedo; scatterometry; lidar; sea level rise

1. Introduction The Greenland Ice Sheet (GrIS) is the second-largest ice mass on earth. If the entire ice sheet were to melt, global sea level would rise by about seven meters [1]. At present, the GrIS is losing mass, consistent with its expected response to anthropogenic climate warming [2–9]. Its mass loss averaged 1 1 –171 Gt yr− between 1991 and 2015, equivalent to 0.47 0.23 mm yr− global eustatic sea level rise ± 1 (SLR) [1]. The rate of SLR contribution accelerated to >1 mm yr− during the latter half of this period with a maximum 1.2 mm contribution in the record melting year 2012 [1,10–13]. Since 2013, satellite gravimetry data suggest mass loss acceleration has stalled but remains negative with an average 0.69 1 mm yr− SLR contribution during the 2012–2016 period [10,14,15]. Consequently, there is an urgent

Remote Sens. 2019, 11, 2405; doi:10.3390/rs11202405 www.mdpi.com/journal/remotesensing Remote Sens. 2019, 11, 2405 2 of 50 need to measure and diagnose changes in the GrIS mass balance, and satellite remote sensing is a powerful tool for this purpose [16]. The observed GrIS mass loss is being driven by both an increase in solid ice discharge (the physical movement of solid ice into surrounding oceans), and by decreasing surface mass balance (the surplus of surface meltwater runoff and evaporation over snowfall) [1,12–14,17–21]. About 60% of total mass loss since 1991 is explained by increased surface meltwater runoff from the ablation zone and lower elevation areas of the accumulation zone [1,22,23]. This increase is attributed in part to a prevailing pattern of warm, dry, clear-sky conditions during summer in (primarily) west Greenland associated with a persistently negative North Atlantic Oscillation (NAO) atmospheric circulation pattern [24,25]. Since about 2000, the NAO has exacerbated background regional warming of the surface air temperature, leading to enhanced surface melt, reduced summer snowfall, and larger areas of exposed bare ice that further enhance surface melt via melt-albedo feedbacks [14,23–30]. In the coming decades, regional warming is expected to drive a continued acceleration of surface meltwater runoff [2–9], whereas solid ice discharge is expected to decline relative to meltwater runoff as the marine-terminating ice sheet margins and outlet glaciers thin and retreat [14,18,31]. Consequently, surface processes in the ablation zone, particularly those that control the exposure and melting of bare ice, will play an enhanced role in determining the long-term GrIS mass balance [1,14,26,30,32]. Already, some 78–85% of the total liquid runoff produced from surface melting is generated in this zone, despite it covering ~22% of the ice sheet’s surface area, up from ~15% in the early 1960s [20,30,33–35]. Remote sensing plays a critical role in validating ice sheet mass balance and the associated SLR contribution [16]. For over three decades, satellite sensors operating in the visible, infrared, and microwave wavelengths of electromagnetic radiation have observed the GrIS [36]. The first spaceborne measurements of GrIS surface elevation were collected in the late 1970s by the National Aeronautics and Space Administration (NASA) Geodetic and Earth Orbiting Satellite-3 (GEOS-3), the NASA Seasat, and the US Navy Geosat oceanographic radar altimeters [37]. More recent radar altimetry missions useful for mapping ice sheet elevation change include the Geosat Follow-On (GFO), and the European Space Agency (ESA) European Remote-sensing Satellites 1 and 2 (ERS-1 and ERS-2), the Environmental Satellite (EnviSat), and the Cryosphere Satellite-2 (CryoSat-2). Radar altimetry now provides the longest record of GrIS volume change, owing to this steady progression of international radar altimeter missions extending to the present day [38]. Historically, ice sheet surface elevation retrieval accuracy from radar altimetry was limited by slope errors in the topographically-complex ablation zone, and uncertain signal penetration depth into snow and firn [39]. The laser altimeters onboard the NASA Ice, Cloud, and land Elevation Satellites 1 and 2 (ICESat-1 and ICESat-2) and the interferometric mode of the Ku-band (~13.9 GHz) radar altimeter onboard CryoSat-2 help redress these issues, and now provide sub-decimeter elevation change accuracy in the ablation zone [40,41]. In addition to new satellite laser and radar altimeters, gravimetry emerged in the early 2000s as a wholly novel remote sensing method for measuring the GrIS mass balance, with the first gravimetric measurements of the GrIS mass balance provided at ~400 km spatial scale by the NASA/German Aerospace Center twin Gravity Recovery and Climate Experiment (GRACE) in 2002 [36]. The GRACE-Follow On (GRACE-FO) satellite launched in May 2018 will continue the GRACE measurement time series for a planned ten years, and adds a new laser ranging interferometer that is expected to improve the satellite-to-satellite distance measurement relative to the GRACE microwave ranging system [42,43]. Optical satellite remote sensing provides the longest continuous spaceborne record of the changing GrIS surface. The first spaceborne photographic images of the GrIS were captured in the early 1960s by United States strategic reconnaissance satellites operated under the classified Corona program, followed by the unclassified Nimbus meteorological satellites launched by NASA in the mid-1960s [44–46]. The first multispectral images of GrIS surface reflectance were collected by the Multispectral Scanner (MSS) instrument on the Earth Resources Technology Satellite-1 (later renamed Landsat 1), launched by NASA in 1972 [47]. Multispectral imaging spectrometers now provide over three decades of ice Remote Sens. 2019, 11, 2405 3 of 50 and snow surface reflectance from which decadal time series of surface albedo are computed [48,49]. Together with thermal, synthetic aperture radar, and microwave scatterometer imagery, the entire GrIS ablation zone ice surface is imaged daily, and regions of unique ice composition (facies or zones) are mapped and monitored for change [35,50]. Crevasse fields, superimposed ice, seasonal snowlines, and “dark zones” of anomalously low albedo are all visible in satellite imagery [26,51–54]. These descriptive zones provide a framework for understanding the spatial organization and physical processes operating on the ablation zone ice surface and their temporal evolution, including surface meltwater presence, outcropping of entrained dust, and the inland migration of the seasonal snowline and supraglacial meltwater lakes [26,29,52,55,56]. Increasingly, ultra-high-resolution commercial satellite and drone imagery is used to recover unprecedented spatial detail, resolving surface features such as cryoconite holes and supraglacial rivers less than one meter in width [57–59]. Previous reviews of satellite remote sensing of glaciers and ice sheets exist [e.g., 46], including recent reviews focused on regional climate modeling of the GrIS surface mass balance [21,60]. Other reviews have focused on satellite altimetry and gravimetry of the combined Greenland and Antarctic ice sheet mass balance [36,61,62], the global land ice contribution to SLR during the satellite era [10], polar science applications of spaceborne wind scatterometers [63], satellite remote sensing of polar climate change [64], principles and theory of remote sensing methods for glaciology [65], remote sensing of Andean mountain glacier mass balance [66], optical remote sensing of Himalayan glaciers [67], glaciological applications of unmanned aerial vehicles [59], and many more specialized reviews of snow and ice optical or hydrologic properties with relevance to glaciological research [68–81]. To the authors’ knowledge, no review has focused specifically on satellite remote sensing the GrIS ablation zone, a small but critically important area of the ice sheet with unique physical processes and strong potential to expand in the coming years. This review summarizes platforms, methods, and data products used for remote sensing of the GrIS ablation zone and some process-level discoveries that these data have produced. Because the ablation zone is defined as those areas where mass losses exceed gains, the review focuses exclusively on mass loss processes. The review is organized as follows: Section2 defines key concepts relevant to the GrIS mass balance, surface mass balance, and energy balance. Section3 reviews radar and laser (lidar) altimetry of GrIS surface elevation change. Section4 reviews satellite measurements of the GrIS mass balance. Section5 reviews optical remote sensing of the GrIS ablation zone surface reflectance and albedo. Section6 reviews active and passive microwave, thermal, and multi-angular remote sensing of glaciological zones, surface melt extent, and surface roughness. Throughout the review, we provide suggestions for future research directions and opportunities in the context of present and future missions. Glossaries of satellite remote sensing sensors, platforms, and managing agencies discussed in this article are provided in AppendixA. The review concludes with a short synthesis and recommendations for future research.

2. Ice Sheet Mass Balance, Surface Mass Balance, and Energy Balance Before reviewing satellite remote sensing of the GrIS ablation zone, we briefly define key concepts and terminology that are discussed throughout the review. Ice sheets and glaciers gain mass through snowfall and lose mass through sublimation, meltwater runoff, and solid ice discharge. Following Lenaerts et al. [21] and van den Broeke et al. [60], the mass balance MB of an ice sheet is usually written as: MB = SMB D (1) − where SMB is surface mass balance and D is solid ice discharged to surrounding oceans. The components 1 1 of MB are commonly expressed in units of mass per time (e.g., kg yr− or Gt yr− ), mass per unit area 2 1 per time (e.g., kg m− yr− ), and mass per unit area per time normalized by density of liquid water 1 (e.g., m water equivalent, m w.e. yr− )[21]. Remote Sens. 2019, 11, 2405 4 of 50

The SMB is the sum of mass inputs and outputs at the ice surface:

SMB = P E ER R (2) − − ds − where precipitation P is the sum of liquid precipitation Pliquid (e.g., rain) and solid precipitation Psolid (e.g., snow), E is the sum of evaporation and sublimation, ERds is net snow erosion, and runoff R is the sum of the liquid water balance:

R = M + P + CO RF RT (3) liquid − − where M is surface meltwater production, CO is condensation, RF is refreezing, and RT is liquid water retention, e.g., water stored in lakes, aquifers, or under capillary tension in snow and ice. The GrIS ablation zone SMB is dominated by Psolid and M with the remaining terms representing smaller but uncertain components [1,82–85]. The accumulation and ablation zones of the ice sheet are defined as those areas where the local SMB > 0 and the local SMB < 0, respectively. The equilibrium line altitude (ELA) separates the two zones where local SMB = 0 [60]. The total change in ice sheet surface elevation over time (the quantity observed by spaceborne altimeters) is usually defined as:

dH SMB BMB D0 = + + + vc + vbr (4) dt ρs ρb ρi

1 2 1 where dH/dt is the total change in surface elevation (m yr− ), SMB (kg m− yr− ) is surface mass 3 2 1 balance, ρs (kg m− ) is surface snow, firn, and/or ice density, BMB (kg m− yr− ) is the change in basal 3 2 1 mass balance, ρb (kg m− ) is basal ice density, D0 (kg m− yr− ) is the local change in ice mass due 1 to ice dynamic motion, ρi is solid ice density, vc (m yr− ) is vertical velocity due to snow and/or firn 1 compaction, and vbr (m yr− ) is vertical bedrock velocity (e.g., glacial isostatic adjustment) [86,87]. The vertical velocity terms contribute to dH/dt, but are not associated with change in MB. Equation (4) is used to estimate volumetric changes in the polar ice sheets, and also to estimate MB. This requires correction for the vertical velocity terms and an estimate of ρs, which for the ablation 3 3 zone is typically taken as a constant value between 830 kg m− and 917 kg m− [88], whereas for snow and firn above the ELA, measured or modeled values of ρs are necessary to convert dH/dt to SMB [86]. In some cases, all but the first term on the right-hand side of Equation (4) can be neglected and the local SMB can be inferred from remotely sensed dH/dt [89]. The energy available for melting snow or ice is determined by the sum of positive and negative 2 heat fluxes at the surface, referred to as the surface energy balance SEB (W m− ):

4 SEB = SWNET + LWNET + SHF + LHF + G = SW (1 α) + LW σT + SHF + LHF + G (5) S in − in − s f c S where SW and LW are shortwave (solar) and longwave (terrestrial) radiation fluxes, SHF and LHF are the turbulent surface fluxes of sensible and latent heat (which are proportional to the aerodynamic 1 surface roughness length, ζ (m− )), GS is subsurface (conductive) heat flux, α is the broadband surface albedo, or the ratio of upwelling (reflected) solar radiation to downwelling (incoming) solar radiation 8 2 2 (SWup/SWin), σ is the Stefan Boltzmann constant (5.67 10− W m− K− ), and Tsfc is the ice surface × 2 kinetic temperature. In Equation (4), all heat fluxes have units W m− and are defined as positive toward the surface [60]. 2 If Tsfc reaches the melting temperature of ice, positive SEB provides melt energy ME (W m− ) and liquid meltwater M is produced. The dominant source of ME for the GrIS is SWNET [90]. Typical values of α are 0.9 for freshly fallen snow, 0.6 for melting snow, and 0.4 for bare glacier ice [73]. This wide variability in α demonstrates the critical role of albedo in modulating the SEB and consequently M in the GrIS ablation zone [26,29,30]. Remote Sens. 2019, 11, 2405 5 of 50

3. Ice Surface Elevation Change The magnitude of GrIS dH/dt and its spatio-temporal variability has been quantified for nearly four decades using spaceborne radar (Radio Detection and Ranging) altimetry [37,38,91–95], and since 2001, using spaceborne laser altimetry [96–101]. Recent advances in laser and radar altimeter precision and spatial resolution allow discrimination of dH/dt in slower-flowing, land-terminating sectors of the GrIS ablation zone, principally caused by a change in SMB rather than D [102,103]. For the accumulation zone, regional climate models and firn compaction models are used to partition the component of dH/dt caused by a change in SMB from those caused by a change in firn density and D [104,105]. The following subsections summarize spaceborne radar altimetry of the GrIS, providing a historical perspective that emphasizes major challenges and advances in the field, and identifies future opportunities for radar altimetry observations of the GrIS ablation zone.

3.1. Radar Altimetry Conventional radar altimeters are nadir pointing single-beam radars that transmit and receive tens to thousands of microwave pulses per second, yielding ground measurement footprint diameters on the order of 1–10 km posted at 0.3–0.6 km along-track spacing [39,106,107]. An estimate of dH/dt is obtained by comparing measurements at crossover points, where an earlier satellite ground track intersects a later one [92]. Two critical factors affect the accuracy of radar altimeter elevation retrievals over ice sheets: (1) variation in topographic slope, which controls the average distance between the altimeter and the measured ground footprint surface [39,108,109], and (2) changes in the electrical permittivity of the ice sheet surface, which controls the reflection, transmission, and absorption of the microwave signal [110–113]. For typical radar altimeter frequencies (Ku-band ~13 GHz), solid ice and liquid water are highly reflective, whereas dry snow and firn are semi-transparent, with Ku-band penetration depths on the order of several meters [114,115]. Spatial and temporal variations in signal penetration depth caused by seasonal bare ice exposure, snow and firn densification, surface meltwater presence, and refreezing of meltwater introduce spurious height change signals that are poorly quantified for the GrIS but are estimated to exceed >0.50 m in some cases [116–118]. Variations in surface slope dominate radar altimeter accuracy in the topographically-complex bare-ice ablation zone, with slope-induced errors that exceed >20 m for slopes >1◦ [39,109,113]. The recent CryoSat-2 radar altimeter mission employs synthetic aperture radar (SAR) and interferometric synthetic aperture radar (InSAR) technologies, which decrease the effective along-track footprint diameter to the order 0.1–0.3 km and provide the cross-track slope from InSAR processing [93]. Together with denser orbital-track spacing and advances in waveform retracking, topographic mapping of ice sheet ablation zone areas is accurate to within a few meters on average, and dH/dt to within sub-meter uncertainty relative to laser altimetry [93,119,120].

3.1.1. Radar Altimetry Sensors and Datasets The first satellite altimeter measurements of surface topography for the Greenland Ice Sheet were made by the GEOS-3, Geosat, and Seasat oceanographic Ku-band altimeters (Table1)[ 37,91,106,121,122]. Their geographical coverage was limited to 72 , but the data they collected showed it was possible to ± ◦ calculate volumetric changes in the polar ice sheets from space, providing proof of the concept for future polar-orbiting altimeters [92]. A first demonstration using GEOS-3, Geosat, and Seasat data found that for areas south of <72 N, the GrIS thickened by 23 6 cm yr 1 between 1978–1986, with enhanced ◦ ± − snowfall in a warmer climate hypothesized to explain the observed thickening [123]. These early missions supported important developments in waveform retracking, slope correction, and physical and empirical models for subsurface volume scattering [106,108,111,124]. These corrections are critical for accurate elevation retrieval and reliable change detection [125,126]. For example, Davis et al. [127] used an improved geodetic model and an improved retracking model to reinterpret the earlier findings Remote Sens.Remote 2019 ,Sens. 11, x 2019FOR, PEER11, x FOR REVIEW PEER REVIEW 5 of 52 5 of 52 the accumulationthe accumulation zone, regional zone, regional climate modelsclimate andmodels firn andcompaction firn compaction models aremodels used are to partitionused to partition the the componentcomponent of 푑퐻/푑푡 of caused푑퐻/푑푡 bycaused change by in change 푆푀퐵 infrom 푆푀퐵 those from caused those by caused change by in change firn density in firn anddensity 퐷 and 퐷 [104,105][104,105]. The following. The following subsections subsections summarize summarize spacebor spaceborne radarne altimetry radar altimetry of the GrIS, of the providing GrIS, providing an an historicalhistorical perspective perspective that emphasizes that emphasizes major challenges major challenges and advances and advances in the field, in the and field, identifies and identifies future opportunitiesfuture opportunities for radar for altimetry radar altimetry observations observations of the GrIS of the ablation GrIS ablationzone. zone.

3.1. Radar3.1. Altimetry Radar Altimetry ConventionaConventional radar altimetersl radar altimeters are nadir are pointing nadir pointing single-beam single radars-beam that radars transmit that transmit and receive and receive tens to tens thousands to thousands of microwave of microwave pulses per pulses second, per second, yielding yielding ground ground measurement measurement footprint footprint diametersdiameters on the order on the 1– order10 km 1 posted–10 km at posted 0.3–0.6 at km 0.3 along–0.6 km-track along spacing-track [39,106,107]spacing [39,106,107]. An estimate. An estimate of of 푑퐻/푑푡 is푑퐻 obtained/푑푡 is obtained by comparing by comparing measurements measurements at crossover at crossover points, where points, an where earlier an sat earlierellite groundsatellite ground track intersectstrack intersects a later one a later [92] .one Two [92] critical. Two factors critical affect factors the affect accuracy the a ccuracyof radar ofaltimeter radar altimeter elevation elevation retrievalsretrievals over ice over sheets: ice 1) sheets: variation 1) variation in topographic in topographic slope, which slope, controls which controls the average the average distance distance betweenbetween the altimeter the altimeter and the measuredand the measured ground footprintground footprint surface [39,108,109]surface [39,108,109], and 2) changes, and 2) changesin the in the electricalelectrical permittivity permittivity of the ice of sheetthe ice surface, sheet surface, which con whichtrols con thetrols reflection, the reflection, transmission, transmission, and and absorptionabsorption of the microwave of the microwave signal [110 signal–113] [110. –113]. For typicalFor radar typical altimeter radar altimeter frequencies frequencies (Ku-band (Ku ~13-band GHz), ~13 solid GHz), ice solidand liquid ice and water liquid are water highly are highly reflectivereflective whereas whereas dry snow dry and snow firn andare semifirn are-transparent, semi-transparent, with Ku -withband Ku penetration-band penetration depths on depths the on the order of orderseveral of meters several [114,115] meters [114,115]. Spatial .and Spatial temporal and temporal variations variations in signal in penetration signal penetration depth caused depth caused by seasonalby seasonal bare ice bare exposure, ice exposure, snow and snow firn and densification, firn densification, surface surfacemeltwater meltwater presence, presence, and and refreezingrefree of meltwaterzing of meltwater introduce introduce spurious spurious height change height signalschange thatsignals are thatpoorly are quantified poorly quantified for the for the GrIS butGrIS are but estimated are estimated to exceed to >0.50 exceed m >0.50 in some m in cases some [116 cases–118] [116. Variations–118]. Variations in surface in surfaceslope slope dominatedominate radar altimeter radar altimeter accuracy accuracy in the topographically in the topographically-complex-complex bare-ice bareablation-ice ablationzone, with zone, slope with- slope- induced inducederrors that errors exceed that >20exceed m for >20 slopes m for >1 slopes° [39,109,113] >1° [39,109,113]. The recent. The CryoSat recent -CryoSat2 radar -altimeter2 radar altimeter mission employsmission employs synthetic synthetic aperture aperture radar (SAR) radar and (SAR) interferometric and interferometric synthetic synthetic aperture aperture radar (InSAR) radar (InSAR) technologiestechnologies which decrease which decrease the effective the effective along-track along footprint-track footprint diameter diameter to the order to the 0.1 order–0.3 km 0.1 –and0.3 km and provide providecross-track cross slope-track from slope InSAR from proc InSARessing proc [93]essing. Together [93]. Together with denser with orbital denser-track orbital spacing-track andspacing and advancesadvances in waveform in waveform retracking, retracking, topographic topographic mapping mapping of ice sheet of ice ablation sheet ablationzone areas zone is accurateareas is accurate to withinto a within few meters a few on meters average, on average, and 푑퐻 / and푑푡 to푑퐻 within/푑푡 to sub within-meter sub uncertainty-meter uncertainty relative relative to laser to laser altimetryaltimetry [93,119,120] [93,119,120]. .

3.1.1. Radar3.1.1. Altimetry Radar Altimetry Sensors andSensors Datasets and Datasets The firstThe satellite first satellite altimeter altimeter measurements measurements of surface of topography surface topography for the Greenland for the Greenland Ice Sheet Ice Sheet were madewere by made the GEOS by the-3, GEOS Geosat,-3, andGeosat, Seasat and oceanographic Seasat oceanographic Ku-band Ku altimeters-band altimeters (Error! Reference (Error! Reference source notsource found. not) [37,91,106,121,122]found.) [37,91,106,121,122]. Their geographical. Their geographical coverage coverage was limited was tolimited ±72o but to ±72theo databut the data they collectedthey collected showed showedit was possible it was possibleto calculate to calculate volumetric volumetric changes changesin the polar in the ice polar sheets ice from sheets from space, providingspace, providing proof of proof concept of concept for future for polar future-orbiting polar- orbiting altimeters altimeters [92]. A first[92] . demo A firstnstration demonstration using GEOSusing-3, GEOS Geosat,-3, andGeosat, Seasat and data Seasat found data that found for areasthat for south areas of south<72oN, of the <72 GrISoN, the thickened GrIS thickened by 23 by 23 ± 6 cm yr±- 16 between cm yr-1 between 1978–1986, 1978 with–1986, enhanced with enhanced snowfall snowfall in a warmer in a warmerclimate hypothesizedclimate hypothesized to explain to explain the observedthe observed thickening thickening [123]. These [123] early. These missions early missions supported supported important important developments developments in waveform in waveform retracking,retracking, slope corre slopection, corre andction, physical and physical and empirical and empirical models formodels subsurface for subsurface volume volume scattering scattering [106,108,111,124]. These corrections are critical for accurate elevation retrieval and reliable change [106,108,111,124]Remote Sens. 2019, 11. These, 2405 corrections are critical for accurate elevation retrieval and reliable change6 of 50 detectiondetection [125,126] [125,126]. For example,. For example, Davis et Davis al. [127] et al. used [127] an used improved an improved geodetic geodetic model and model an and an improvedimproved retracking retracking model to model reinterpret to reinterpret the earlier the findings earlier findingsof Zwally of etZwally al. [92] et, findingal. [92], afinding smaller a smaller 1 of Zwally1.5 ± et0.5-1 al. cm [92 yr],-1 finding thickening a smaller for the 1.5 period0.5 1978 cm yr–1986thickening that highlighted for the the period importance 1978–1986 of slope that - and 1.5 ± 0.5 cm yr thickening for the period 1978±–1986 that −highlighted the importance of slope- and sighighlightednal penetrationsignal thepenetration importance-correction.-correction. of slope- and signal penetration-correction.

TableTable 1. 1.TSummaryableSummary 1. Summary of of radar radar ofand and radar laser laser and altimeter altimeter laser altimeter remote remote sensing remote sensing platforms, sensing platforms, platforms, instruments, instruments, instruments, temporal temporal temporal coverage,coverage,coverage, observed observed observedwavelength, wavelength, wavelength, and and managing managing and managing agencies. agencies. agencies. Temporal Observed TemporalŦ Temporal ObservedObserved §,Ɣ PlatformPlatform * *Platform InstrumentInstrument * Instrument Ŧ Agency §,ƔAgency CoverageCoverage Coverage WavelengthWavelengthWavelength Agency Radar Altimeters Radar altimetersRadar Altimeters

GEOS-3 ALT 1975–1978 13.9 GHz (Ku) NASA Seasat ALT 1978 (110 days) 13.6 GHz (Ku) NASA Geosat GRA 1985–1990 13.5 GHz (Ku) DoD/NASA ERS-1 RA 1991–2006 13.8 GHz (Ku) ESA 13.6 GHz (Ku), 3.2 ERS-2 RA-2 1995–2011 ESA GHz (S) Remote Sens. 2019, 11, x FOR PEER REVIEW 6 of 52 Remote Sens.GFO 2019, 11, x FOR PEER REVIEW GFO-RA 1998–2008 13.5 GHz (Ku)6 of 52 DoD/NASA 13.6 GHz (Ku), 3.2 EnviSatGEOS-3GEOS -3 ALT RA-2 ALT 1975–19781975 –1978 2002–2012 13.9 GHz13.9 (Ku) GHz (Ku) NASA NASA ESA Seasat Seasat ALT ALT 1978 (1101978 days) (110 days) 13.6 GHz13.6 (Ku) GHz (Ku) NASAGHz NASA (S) CryoSat-2Geosat Geosat GRA SIRAL GRA 1985–19901985 –1990 2010–present 13.5 GHz13.5 (Ku) GHz (Ku) 13.9DoD/NASA GHzDoD/NASA (Ku) ESA ERS-1 ERS-1 RA RA 1991–20061991 –2006 13.8 GHz13.8 (Ku) GHz (Ku) ESA ESA SARALERS-2 ALtiKaRA-2 1995–2011 2013–present 13.6 GHz (Ku), 3.2 GHz 36 (S)GHz ESA (Ka) ISRO/CNES ERS-2 RA-2 1995–2011 13.6 GHz (Ku), 3.2 GHz (S)13.6 GHzESA (Ku), 5.4 Sentinel-3AGFO/B GFO GFO SRAL-RAGFO -RA 1998–20081998 –2008 2016–present 13.5 GHz13.5 (Ku) GHz (Ku) DoD/NASADoD/NASA ESA EnviSatEnviSat RA-2 RA-2 2002–20122002 –201213. 6 GHz13. (Ku),6 GHz 3.2 (Ku), GHz 3.2 (S) GHz (S)GHz ESA (C)ESA CryoSatCryoSat-2 -2 SIRAL SIRAL 2010–present2010–present 13.9 GHz13.9 (Ku) GHz (Ku) ESA ESA Laser altimetersSARALSARAL ALtiKaALtiKa 2013–present2013–present 36 GHz36 (Ka) GHz (Ka) ISRO/CNESISRO/CNES Sentinel-3A/B SRAL 2016–present 13.6 GHz (Ku), 5.4 GHz (C) ESA ICESatSentinel-3A/B SRAL GLAS 2016–present 13.6 2003–2009 GHz (Ku), 5.4 GHz (C) 1.064 µESAm, 0.532 µm NASA Laser altimetersLaser altimeters ICESat-2ICESat ICESat GLAS ATLAS GLAS 2003–20092003 –2009 2018–present 1.064 μm,1.064 0.532 μm, μm 0.532 1.064 μm NASAµm,0.532 NASA µm NASA AircraftICESat-ICESat2 -2 ATLAS ATMATLAS 2018 –present2018–present 1977–present 1.064 μm,1.064 0.532 μm, μm 0.532 μm 1.064NASA NASAµm NASA AircraftAircraft ATM ATM 1977–present1977–present 1.064 μm1.064 μm NASA NASAµ AircraftAircraft LVISLVIS 1998–present 1998–present 1.064 μm 1.064 NASAm NASA Aircraft LVIS 1998–present 1.064 μm NASAµ µ AircraftAircraftAircraft MABEL MABELMABEL 2012–20142012 –2014 2012–2014 1.064 μm,1.064 0.532 μm, μm 0.532 1.064 μm NASAm,0.532 NASA m NASA *,Ŧ,§ See Abbreviations at end of article for expanded acronyms. Ɣ Managing agencies are identified *,*,Ŧ,§ See Abbreviations Abbreviations at atend end of article of article for expanded for expanded acronyms. acronyms. Ɣ ManagingManaging agencies agencies are identified are identified by the WMO OSCARby the byWMO databasethe WMOOSCAR (TableOSCAR database A1 database) (Table and may (Table A1) notand A1) reflectmay and not may joint reflect not collaborations. reflectjoint collaborations. joint collaborations. The ESAThe EuropeanESA European Remote Remote-sensing-sensing Satellite Satellite (ERS -1) (ERS carried-1) carried the first the polar first - polarorbiting-orbiting rada r radar o altimeter, extending coverage of the polar ice sheets to o±81.5 [95]. Together with the follow-on ERS- altimeter,The ESAextending European coverage Remote-Sensing of the polar ice sheets Satellite to ±81.5 (ERS-1) [95]. Together carried with the the first follow polar-orbiting-on ERS- radar altimeter, 2 and2 EnviSat and EnviSat missions, missions, these these satellite satellites provideds provided near complete near complete GrIS coverageGrIS coverage and 21 and years 21 years of of extending coverage of the polar ice sheets to 81.5 [95]. Together with the follow-on ERS-2 and continuouscontinuous data acquisitiondata acquisition [38]. The[38] .first The digital first digital elevation elevation± model◦ model (DEM) (DEM) spanning spanning the entire the entire GrIS GrIS EnviSatsurfacesurface was missions, producedwas produced these from satellitesfromcombined combined ERS provided- 1ERS and-1 Geosat and near-complete Geosat altimetry altimetry data, GrIS data,supplemented coverage supplemented andby stereo 21 by yearsstereo- - of continuous dataphotogrammetric acquisitionphotogrammetric and [38 cartographic ].and The cartographic first data digital sets, data with sets, elevation pan with-GrIS pan model -averageGrIS average (DEM)accuracy accuracy spanning-0.33 ±- 0.336.97 ±m 6.97 the [128,129] m entire [128,129]. GrIS. surface was ERS-1/2ERS altimetry-1/2 altimetry data were data usedwere toused discriminate to discriminate thinning thinning rates inrates the inablation the ablation zone (definedzone (defined as <1500 as <1500 produced from combined ERS-1-1 and Geosat altimetry data, supplemented by stereo-photogrammetric-1 m a.s.l.)m a.s.l.)of -2.0 of ± -0.92.0 cm± 0.9 yr cm-1 from yr accumulationfrom accumulation zone (>1500zone (>1500 m a.s.l.) m a.s.l.)thickening thickening of 6.4 of± 0.26.4 cm± 0.2 yr cm-1 yr and cartographic data sets, with pan-GrIS average-1 accuracy of 0.33 6.97 m [128,129]. ERS-1/2 with anwith overall an overall net thickening net thickening of 5.4 of ± 5.40.2 ±cm 0.2 yr cm-1 for yr the for period the period 1992– 19922003– [130]2003− . [130] To± extend. To extend the the altimetrytemporaltemporal coverage data coverage were and usedincrease and increase to spatial discriminate spatial measurement measurement thinning density, density, rates Khvorostovsky inKhvorostovsky the ablation and Jo andhannessen zone Johannessen (defined [131] [131] as <1500 m a.s.l.) developed an algorithm1 to merge ERS-1/2 and EnviSat datasets and detect and eliminate inter- 1 ofdeveloped –2.0 0.9 an cmalgorithm yr− from to merge accumulation ERS-1/2 and EnviSat zone( >datasets1500 m and a.s.l.) detect thickening and eliminate of inter 6.4- 0.2 cm yr− with satellitesatellite± biases. biases. For the For period the period 1992– 19922008,– 2008,their datasettheir1 dataset suggests suggests net thinning net thinning of the of GrIS the initiatedGrIS initiated± an overall net thickening of 5.4 0.2 cm yr− for the period 1992–2003 [130-1 ]. To extend the temporal aroundaround 2000, 2000,and accelerated and accelerated between between± 2006– 20062008,– 2008,with awith thickening a thickening of 4.0 of± 0.24.0 cm± 0.2 yr cm-1 above yr above 1500 m1500 m coverage and increase spatial measurement-1 density, Khvorostovsky and Johannessen [131] developed a.s.l. anda.s.l. thinning and thinning of -7.0 of ± -1.07.0 cm± 1.0 yr cm-1 below yr below 1500 m1500 a.s.l. m [94] a.s.l.. [94]. an algorithmConventionalConventional to radar merge radaraltimeters ERS-1 altimeters /such2 and assuch the EnviSat as instruments the instruments datasets included andincluded on detect ERS on- 1/2ERS and and-1/2 eliminate EnviSat and EnviSat were inter-satellite were biases. Foroptimized theoptimized period for detecting for 1992–2008, detecting elevation elevation their over dataset openover openoceans suggests oceans and low and net-gradient low thinning-gradient polar of polarice the sheet ice GrIS sheetinteriors initiated interiors with with around 2000, and o mean accuracies typically <0.2 m for ice sheet surface slopeso <0.2 [39,109]. Elevation accuracy and acceleratedmean accuracies between typically 2006–2008, <0.2 m for ice with sheet a thickeningsurface slopes of <0.2 4.0 [39,109]0.2 cm. Elevation yr 1 above accuracy 1500 and m a.s.l. and thinning 푑퐻/푑푡 uncertainty is much higher for the topographically complex ablation− zone due to their large 푑퐻/푑푡 uncertainty is much1 higher for the topographically complex± ablation zone due to their large of –7.0 1.0 cm yr below 1500 m a.s.l. [94]. o spatialspatial measurement± measurement −footprint footprint and wide and orbitalwide orbital track spacing,track spacing, with slopewith slopeerrors errors >20 m >20 for mslopes for slopes >1o >1 and Conventional systematicand systematic biases biases radar over roughover altimeters, rough surfaces surfaces such [39,95,109,132] as[39,95,109,132] the instruments. Careful. Careful processing included processing has onbeen has ERS-1 been used used/ to2 and to EnviSat, were optimizeddiscriminatediscriminate forablation detecting ablation zone thinningzone elevation thinning rates usingrates over usingERS open-1/2 ERS oceansand-1/2 EnviSat and and EnviSat altimetry low-gradient altimetry data, includingdata, polar including repea ice trepea sheet- t- interiors with track trackanalysis, analysis, but the but accuracy the accuracy of these of thesedata anddata their and processtheir process-based-based interpretation interpretation is ambiguous is ambiguous mean accuracies typically of <0.2 m for ice sheet surface slopes <0.2◦ [39,109]. Elevation accuracy owingowing to the to aforementioned the aforementioned slope slopeerrors errors and sparse and sparse spatial spatial measurement measurement density density in the in ablation the ablation andzonedH [118,133zone/dt [118,133uncertainty–135].– 135] . are much higher for the topographically complex ablation zone due to their largeThe spatial CryoSatThe CryoSat measurement-2 satellite-2 satellite (launched (launched footprint April April2010) and 2010)provides wide provides orbitalcontinuity continuity track with ERS with spacing,-1/2 ERS and-1/2 withEnviSat and EnviSat slope and and errors >20 m for slopesemploysemploys> 1 SARo and SARdelay systematic - delayDoppler-Doppler along biases - alongtrack over- track processing rough processing and surfaces interferometric and interferometric [39,95,109 cross,132 - crosstrack]. Careful- track processing processing processing has been [120,136][120,136]. CryoSat. CryoSat-2 is the-2 isfirst the ESA first radar ESA radaraltimeter altimeter dedicated dedicated to polar to polarstudies, studies, with orbitalwith orbital coverage coverage used to discriminateo ablation zone thinningo rates using ERS-1/2 and EnviSat altimetry data, including to ±88too and±88 1.6 and km 1.6 cross km -crosstrack- trackspacing spacing at 70 oat. CryoSat 70 . CryoSat-2 carries-2 carries two Ku two-band Ku- bandSAR/Interferometric SAR/Interferometric repeat-trackRadarRadar Altimeter Altimeter analysis, (SIRAL) (SIRAL) butsystems. the systems.accuracy SIRAL SIRAL operates of operates these in three data in threedistinct and distinct their modes process-basedmodes depending depending on location: interpretation on location: 1) 1) is ambiguous low resolutionlow resolution mode mode (LRM (LRM) over) open over oceanopen ocean and ice and sheet ice sheetinteriors, interiors, 2) synthetic 2) synthetic aperture aperture radar radarmode mode (SAR)(SAR) over sea over ice sea and ice coastal and coastal regions, regions, and 3) and SAR 3) interferometric SAR interferometric mode mode (SARIn) (SARIn) over ice over sheet ice sheet margins.margins. The SARIn The SARIn mode mode employs employs two SAR two SARinstruments instruments oriented oriented across across the satellite the satellite track. track. InterferometricInterferometric phase phase processing processing of the of dual the waveforms dual waveforms provides provides cross -crosstrack- slopetrack slopeat ~0.3 at km ~0.3 along km -along- track resolutiontrack resolution [137,138] [137,138]. . o Over mildOver slopesmild slopes (<1o), phase(<1 ), phase unwrapping unwrapping [136] of[136] CryoSat of CryoSat-2 interferometric-2 interferometric data provides data provides 5 km 5 km wide-wideswath-swath elevation elevation retrievals retrievals with twowith orders two orders of magnitude of magnitude more moreindividual individual measurements measurements per per surfacesurface area than area anythan prior any priorradar radaraltimeter altimeter [119,120,137] [119,120,137]. The first. The spatially first spatially continuous continuous swath swath DEMs DEMs

Remote Sens. 2019, 11, 2405 7 of 50 owing to the aforementioned slope errors and sparse spatial measurement density in the ablation zone [118,133–135]. The CryoSat-2 satellite (launched April 2010) provides continuity with ERS-1/2 and EnviSat, and employs SAR delay-Doppler along-track processing and interferometric cross-track processing [120,136]. CryoSat-2 is the first ESA radar altimeter dedicated to polar studies, with orbital coverage to 88 and ± ◦ 1.6 km cross-track spacing at 70◦. CryoSat-2 carries two Ku-band SAR/Interferometric Radar Altimeter (SIRAL) systems. SIRAL operates in three distinct modes depending on location: (1) low resolution mode (LRM) over open ocean and ice sheet interiors, (2) synthetic aperture radar mode (SAR) over sea ice and coastal regions, and (3) SAR interferometric mode (SARIn) over ice sheet margins. The SARIn mode employs two SAR instruments oriented across the satellite track. Interferometric phase processing of the dual waveforms provides a cross-track slope at ~0.3 km along-track resolution [137,138]. Over mild slopes (<1◦), phase unwrapping [136] of CryoSat-2 interferometric data provides 5 km wide-swath elevation retrievals with two orders of magnitude more individual measurements per surface area than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs were generated from CryoSat-2 data for areas of the Devon Ice Cap and western GrIS with 3 m ± precision [119,137]. Swath processing provides information about within-swath topography that is not provided by traditional radar processing techniques. For example, Gray et al. [119] demonstrated a novel method for supraglacial lake water surface elevation retrievals in the ablation zone from CryoSat-2 swath retrievals and provided a detailed technical description of swath processing. 1 The SIRAL system nominally resolves dH/dt to 3.3 cm yr− near the ice sheet margins and 0.7 cm 1 2 2 yr− in the ice sheet interior, at a 104 km and 106 km spatial scale, respectively [138]. Using three years (2011–2014) of CryoSat-2 standard retrievals and an updated threshold retracking algorithm, a new pan-GrIS DEM was generated with a 3 15 m elevation accuracy relative to ICESat over 80% ± of the GrIS and 5 65 m accuracy over 90% of the GrIS (Figure1)[ 93]. These data suggest reliable ± CryoSat-2 elevation retrievals remain limited to areas of the ablation zone with a surface slope <1.5o (Figure1). Elevation change from these data suggests a 2.5 factor increase in pan-GrIS volume loss (–375 24 km3 yr 1) compared with the ICESat (2003–2009) period, with large losses concentrated in ± − the west and southeast marginal ablation zones [93]. A separate analysis of the same data combined with a firn density model found an equivalent mass loss of 269 51 Gt yr 1 [105]. ± − Currently operating beyond its design lifetime, CryoSat-2 will be succeeded by the SAR Radar Altimeter (SRAL) onboard Sentinel-3A/B, and by the AltiKa Ka-band radar altimeter onboard the Satellite with ARgos and ALtiKa (SARAL) (Table1). SARAL is a joint French Space Agency (CNES) and Indian Space Research Organisation (ISRO) oceanographic mission launched in 2013 with the secondary goal of monitoring polar ice sheet surface elevations [139]. SARAL has the same 35-day repeat orbit as EnviSat and ERS-1/2, but the AltiKa instrument operates at Ka-band (~36 GHz), thereby offering unique opportunities for cross-platform validation of prior Ku-band radar altimeters [140]. In particular, the Ka-band should improve diagnosis of Ku-band signal penetration bias, owing to its higher (~36 GHz) signal frequency with ~10 lower theoretical penetration depth into dry × snow relative to Ku-band, higher spatial resolution (~8 km footprint and 175 m along-track spacing), and higher pulse-repetition frequency [135,141–143]. SARAL also provides an independent record of ice sheet dH/dt. Using the 1 km ICESat DEM for the period 2003–2005 [144] as a baseline, data from SARAL-AltiKa for the period 2014–2016 was used to estimate a pan-GrIS volume loss rate of 3 1 247 km yr− , with the largest basin-scale volumetric decreases found in the north and northwest [145]. Remote Sens. 2019, 11, 2405 8 of 50

Remote Sens. 2019, 11, x FOR PEER REVIEW 8 of 52

Figure 1. (a) Digital elevation model (DEM) of the Greenland Ice Sheet surface created from the Figure 1. (a) Digital elevation model (DEM) of the Greenland Ice Sheet surface created from the European Space Agency (ESA) level 1B CryoSat-2 radar altimetry waveform product [93], with slope Europeanindicated Space Agency by shaded (ESA) relief; level (b) 1Bmedian CryoSat-2 error (difference radar altimetry between waveform CryoSat-2 DEM product and ICESat [93], with slope indicated byelevations) shaded vs. relief; surface (b slope,) median with errorelevation (di ffindicatederence betweenby colormap CryoSat-2 in (a); (c) s DEMtandard and deviation ICESat ofelevations) vs. surfaceerror slope, vs. withsurface elevation slope, with indicated elevation by indicated colormap by colormapin (a); (c ) in standard (a). The mediandeviation and of standard error vs. surface o slope, withdeviation elevation of error indicated are calculated by colormap from the CryoSat in (a).-2 TheDEM median error grid and binned standard by a slope deviation with a 0.01 of error are bin size, following Helm et al. [93] (Figure 9). The median and standard deviation of erroro increase calculatedwith from slope thes with CryoSat-2 a step shift DEM toward error higher griderror at binned slopes >1.5 byo.a Elevation slope withand slope a 0.01 valuesbin calculated size, following Helm et al.from [93 the] (Figure DEM indicate 9). The that median 16% of the and ice sheet standard area has deviation a surface slope of error >1.5o, increase and nearly with all (95%) slopes with a step shiftsuch toward areas higher have elevation error at <2000 slopes m a.s.l.,>1.5 o which. Elevation is generally and representative slope values of calculated the ablation from zone the DEM indicate that(calculat 16%ions of performed the ice by sheet the first area author). has aThe surface error grid slope was produced>1.5o, by and first nearly calculating all elevation (95%) such areas differences between the CryoSat-2 DEM and individual ICESat elevations (campaign 3F, 3G, and 3H), have elevation <2000 m a.s.l., which is generally representative of the ablation zone (calculations corrected for elevation change between the individual ICESat observation and the DEM reference performedtime by (1 the Jul first 2012) author). and then Thecalculating error a grid weighted was producederror as a function by first of calculatingsurface roughness, elevation surface di fferences between theslope, CryoSat-2 and number DEM of and cross individual-validation data ICESat points elevations [93]. The DEM (campaign and associated 3F, 3G, error and 3H), grid are corrected for elevation changepublicly available between at thehttps:/ individual/doi.pangaea.de/10.1594/PANGAEA.831394 ICESat observation and the [146] DEM. reference time (1 July 2012) and then calculating a weighted error as a function of surface roughness, surface slope, and number 3.1.2. Current Challenges and Future Opportunities of cross-validation data points [93]. The DEM and associated error grid are publicly available at Progress in waveform retracking and SAR/InSAR technology has improved radar altimetry https://doi.pangaea.de/10.1594/PANGAEA.831394[146]. performance over complex terrain, but retracking errors over rough surfaces and spatiotemporal 3.1.2. Currentvariations Challenges in signal penetration and Future depth Opportunities remain important sources of uncertainty. Reliable performance in the ablation zone appears limited to areas with surface slope <1.5° (Figure 1), and signal Progresspenetration in waveform biases may retrackingexceed 0.5 m in and the lower SAR /accumulationInSAR technology zone [116,118,119,147] has improved. Nevertheless, radar altimetry radar altimetry provides the longest record of GrIS 푑퐻/푑푡 , and two challenges stand out for performance over complex terrain, but retracking errors over rough surfaces and spatiotemporal leveraging this unique dataset: (1) homogenization of the inter-mission and cross-platform methods variationsand in signaldatasets penetration used to detect depth 푑퐻/푑푡 remain[133,148] important, and (2) better sources understanding of uncertainty. of spatial Reliableand temporal performance in the ablationchanges zone in appearssnow, firn, limited and ice to permittivity areas with caused surface by slope surface< 1.5 meltwater◦ (Figure presence,1), and meltwater signal penetration biases maypercolation, exceed 0.5and mice inlens the formation lower [116,147] accumulation. These challenge zone [s116 are, 118 closely,119 related.,147]. For Nevertheless, example, radar methods such as waveform smoothing, waveform model fitting, and optimal thresholding increase altimetry provides the longest record of GrIS dH/dt, and two challenges stand out for leveraging this absolute accuracy, but at the expense of the homogenization required for statistical change detection unique dataset:[125,126,149] (1) homogenization. In other cases, waveform of the retracking inter-mission may mask andreal 푑퐻 cross-platform/푑푡 signals [150] methods. Similarly, real and datasets used to detectchangesdH in/ dtsnow[133 and,148 firn], permittivity and (2) better may introduce understanding false 푑퐻/ of푑푡 spatialsignals, andsuch as temporal the apparent changes 56 ± in snow, firn, and ice26 cm permittivity elevation increase caused attributed by surface to ice lens meltwater formation presence, in the percolation meltwater zone following percolation, the 2012 and ice lens melt event [116]. Here, the SARAL-AltiKa Ka-band altimeter may provide new opportunities for formation [116,147]. These challenges are closely related. For example, methods such as waveform diagnosing radar penetration bias and waveform interpretation. In addition, its replication of the smoothing, waveform model fitting, and optimal thresholding increase absolute accuracy, but at the expense of the homogenization required for statistical change detection [125,126,149]. In other cases, waveform retracking may mask real dH/dt signals [150]. Similarly, real changes in snow and firn permittivity may introduce false dH/dt signals, such as the apparent 56 26 cm elevation increase ± attributed to ice lens formation in the percolation zone following the 2012 melt event [116]. Here, the SARAL-AltiKa Ka-band altimeter may provide new opportunities for diagnosing radar penetration bias and waveform interpretation. In addition, its replication of the ERS-1/2 and EnviSat repeat orbit will improve the statistical reliability of along-track change detection [140]. Remote Sens. 2019, 11, 2405 9 of 50

In contrast to the percolation zone, the ablation zone is typically conceptualized as homogenous solid ice with uniform electromagnetic properties. Ku-band backscatter over the bare ice ablation zone is dominated by surface scattering from the rough (and seasonally saturated) ice surface, but volume scattering has received little direct study and may be important over superimposed ice, seasonally-snow covered surfaces, or areas with multi-modal surface roughness distributions [151]. Penetration depths at C- and L-band exhibit an order of magnitude range over bare ice surfaces on the GrIS, possibly reflecting variations in ice thermal structure and surface roughness [152]. Similar comparisons at the Ku-band do not exist to our knowledge but could be facilitated by dual-frequency radar altimeters such as SRAL or RA-2, or by comparison with SARAL-AltiKa Ka-band altimetry (Table1). Factors that affect bare ice microwave permittivity include its grain size, temperature, porosity, water content, crystal structure, and chemical and physical impurity content [153–156]. Although slope errors dominate elevation retrieval uncertainty in the ablation zone, seasonal and spatial variations in ablation zone surface properties and their effect on radar backscatter may be an overlooked source of uncertainty [117,135]. Conversely, this variation and its effect on waveform shape could provide new opportunities for understanding the ablation zone, such as detecting water surface elevation change in supraglacial lakes, inferring changes in near-surface ice structure related to physical weathering of solid ice, or estimating seasonal snow thickness using a dual-frequency radar, as was recently demonstrated for Arctic sea ice using combined CryoSat-2 and SARAL-AltiKa retrievals [119,156,157]. At the ice-sheet scale, knowledge of snow, firn, and ice density is the main source of uncertainty for conversion between ice sheet volume change and mass change [36]. Density changes are caused by snow and firn compaction and by meltwater refreezing, which both affect waveform interpretation by changing the scattering properties of the snow and firn [105,116]. Field investigations suggest ice lenses are widespread in the percolation zone and have thickened in recent decades [20,158]. The vertical and horizontal distribution of meltwater percolation and refreezing is difficult to model and may not be accurately represented by regional climate models [158,159]. Although these features present challenges for dH/dt detection, they also provide unique opportunities for characterizing snow and firn processes, including detection of extreme melt events, ice lens formation, and snow accumulation rates following ice lens formation [115].

3.2. Laser Altimetry Laser altimeters use lidar (light detection and ranging) technology to measure the two-way travel time of narrow-beam monochromatic laser pulses transmitted between the altimeter and the earth’s surface. In contrast to radar altimetry, signal penetration in ice and snow is minimal at common lidar wavelengths (visible and near-infrared), and the narrow laser beam illuminates a much smaller ground area, which reduces slope errors over complex terrain [39,160]. The small laser footprint increases absolute accuracy but may also introduce uncertainty because interpolation between narrow footprints is needed to obtain spatially continuous elevations [99]. Laser altimeters are sensitive to atmospheric scattering and lack the all-weather capability of radar altimeters but are uniquely adept at measuring local surface elevation in the topographically complex ablation zone, allowing resolution of thinning rates at the scale of individual outlet glaciers and providing benchmark datasets for altimeter accuracy [97,161]. The following subsections summarize spaceborne laser altimetry of the GrIS, emphasizing the operational performance of ICESat, and future opportunities for laser altimetry observations of the GrIS ablation zone from the recently launched ICESat-2.

3.2.1. Laser Altimetry Sensors, Methods, and Datasets The NASA Lidar In-Space Technology Experiment (LITE), flown on the Discovery Space Shuttle in September 1994, was the first spaceborne lidar [162]. The LITE was focused on the vertical structure of clouds and aerosols and provided operational proof of the concept. The NASA Geoscience Laser Altimeter System (GLAS) onboard the ICESat satellite (2003–2009) (Table1) was the first spaceborne lidar designed for polar science. The GLAS transmitted short pulses of near infrared (1064 nm) Remote Sens. 2019, 11, 2405 10 of 50 light for surface altimetry and visible green light (532 nm) for vertical distribution of clouds and aerosols [101]. The ICESat mission was marked by several important innovations in polar altimetry, including unprecedented spatial resolution (70 m footprint, 172 m along-track spacing, and 3 cm vertical resolution), geographic coverage to 86 , and the use of a narrow-beam 1064 nm laser which ± ◦ reduced sensitivity to both slope errors and signal penetration into snow and firn [163]. An important design objective of ICESat was to measure ice sheet surface elevation change over regions of high surface slope and complex topography, with primary mission objectives to measure spatially-averaged (104 km2) ice sheet surface elevation to <15 cm absolute accuracy and elevation 1 change to <1.5 cm yr− accuracy [36,101]. Field validation suggests operational absolute accuracy achieved 2 3 cm for optimal conditions [164]. At-a-point measurement precision is 3 cm over the ± ± ice sheets and the repeat crossover measurement uncertainty is typically 10–15 cm [163,165], and up to 59 cm in regions of high surface slope [39]. Performance is affected by forward scattering of the return signal by intervening clouds, detector saturation, uncertain laser pointing angle, and laser transmit power decline [39,164,166]. The ICESat payload included three separate laser systems. Technical issues with the lasers prevented planned continuous operation, and instead, data collection took place during 18 separate campaigns [166,167]. Campaigns were designated by the laser system used and the campaign-specific orbital repeat frequency. For example, the first laser system operated for 38 days in an 8-day calibration and validation mode repeat frequency before it failed [167]. The second and third laser systems operated in campaign mode with a 91-day repeat frequency constituting individual campaigns ~33 days in length separated by 4–6 months. Inter-campaign range biases up to ~20 cm have been found [165,168]. The biases owe in part to a coding error in the ICESat signal processing algorithm that has since been corrected [168]. On-orbit and post-processing bias corrections were developed to compensate for orbital drift and systematic laser orientation (pointing) errors, but residual inter-campaign biases are an important measurement uncertainty in the ICESat data [86,165,169]. As with all spaceborne altimeters, the ICESat orbit followed ascending and descending orbital reference tracks with crossover reference points at the ascending and descending orbit intersections (Figure2). Several factors a ffect the geolocation of ICESat ground footprints at repeat-track and crossover locations. These include orbital vibrations, orbital drift, random errors in ICESat’s laser orientation determination system, and declines in laser transmit power through time that modify the illuminated footprint diameter [164,167,170]. These factors produce both real and apparent deviations between illuminated ground tracks and reference tracks up to a few hundred meters [163,171]. Consequently, exact co-located repeat track and crossover point measurements are extremely rare, and spatial interpolation is required to recover dH/dt at-a-point, which has become a central challenge for ICESat data interpretation [171]. To estimate pan-GrIS dH/dt, an additional spatial interpolation of the point dH/dt estimates is typically performed, which is complicated by temporal differences among the point dH/dt estimates e.g., [86]. As detailed by Felikson et al. [99], four methods have been developed to recover dH/dt from ICESat data: (1) repeat tracks (RTs), (2) crossovers (XOs), (3) overlapping footprints (OFPs), and (4) surface fitting methods, such as triangulated irregular networks (TINs). The Surface Elevation Reconstruction and Change (SERAC) method [171] combines elements of RT and XO [99]. Alberti and Biscaro [172] summarize the ICESat orbital parameters in detail and describe a flexible algorithm for determining RT and XO points from them. The XO method calculates dH/dt at the ascending and descending orbital crossover points. The method was originally developed for radar altimetry and was expected to be the primary method for interpreting ICESat data [101]. The method was used to estimate a relative elevation accuracy for ICESat of 0.14–0.59 m depending on the surface slope [39]. Using ICESat XO points as reference values, absolute accuracies for ERS-1/2 and EnviSat radar retrievals are ~0.10–0.56 m for surface slopes <0.1◦, but are 2.27 m on average and up to 30 m for surface slopes exceeding 0.7◦ [39]. These differences demonstrate the utility of laser altimetry for validating radar altimetry but also suggest caution when Remote Sens. 2019, 11, 2405 11 of 50 combining retrievals at discrete locations with high surface slopes and/or topographic variability c.f.Remote [109 Sens.]. 2019, 11, x FOR PEER REVIEW 11 of 52

FigureFigure 2. 2.Example Example ofof GLASGLAS/ICESat/ICESat orbitalorbital reference reference tracks, tracks, crossover crossover point, point, and and 27,766 27,766 unique unique surface surface elevationelevation measurements measurements collected collected on on 20 20 February February 2003 2003 obtained obtained from from the GLASthe GLAS/ICESat/ICESat L2 Antarctic L2 Antarctic and Greenlandand Greenland Ice Sheet Ice altimetry Sheet altimetry Data (GLAH12), Data (GLAH12), Version Version 34 (https: 34// nsidc.org (https://nsidc.org/data/GLAH12/)./data/GLAH12/). ICESat dataICESat are data produced are produced by the GLAS by the Science GLAS TeamScience at Team the ICESat at the Science ICESat Investigator-ledScience Investigator Processing-led Processing System (I-SIPS)System at(I- NASASIPS) at/GSFC NASA/GSFC and are archived and are archived by the National by the National Snow and Snow Ice Data and CenterIce Data (NSIDC) Center (NSIDC) DAAC. DAAC. The RT method calculates dH/dt from repeat measurements at reference points along the reference tracks,As which detailed increases by Felikson spatial measurementet al. [99], four density methods and have coverage been relative developed to XO to points recover (Figure 푑퐻/3푑푡)[ 173from]. However,ICESat data: RT points(1) repeat are tracks typically (RTs), further (2) crossovers apart than (XOs), XO points, (3) overlapping which may increasefootprints slope (OFPs), errors and [99 (4)]. Owingsurface to fitting higher methods measurement, such density, as triangulated the RT method irregular has seennetworks greater (TINs). operational The Surface use than Elevation the XO methodReconstruction and is more and appropriate Change (SERAC) for local method or regional [171] scale combines analysis elements [99]. The of RT RT method and XO was [99] combined. Alberti withand AdvancedBiscaro [172] Spaceborne summarize Thermal the ICESat Emission orbital and parameters Reflection in Radiometerdetail and describe (ASTER) a flexible digital elevationalgorithm modelfor determining differencing, RT which and XO together points suggest from them. an average volume loss in southeast Greenland of ~108 km3 1 yr− forThe the XO period method 2003–2005 calculates [161 푑퐻]./ Concentrated푑푡 at the ascending thinning and of descending narrow outlet orbital glaciers crossover contributed points. lessThe volumemethod loss was than originally dispersed, developed but smaller for radar thinning altimetry over largerand was interior expected areas, to highlightingbe the primary the method enhanced for capabilityinterpreting of laserICESat altimetry data [101] for observing. The method local-scale was used processes to estimate in the a ablation relative zoneelevation [161]. accuracy for ICESatThe of OFP 0.14 method–0.59 m uses depending both XO on and theRT surface points slope but requires[39]. Using that ICESat the laser XO footprints points as of reference repeat measurementsvalues, absolute overlap. accuracies ICESat for ERS footprints-1/2 and are EnviSat ellipses radar with retrievals a 50–90 are m major ~0.10–0.56 axis m length for surface [174]. Theslopes overlapping <0.1°, but are criterion 2.27 m is on determined average and by up a predefinedto 30 m for semi-majorsurface slopes axis exceeding distance 0.7 between° [39]. These laser footprintsdifferences [100 demonstrate]. In principle, the theutility method of laser does altimetry not account for validating for the slope radar between altimetry footprints, but also but suggest slope correctionscaution when have combining been applied, retrievals for example,at discrete to locations validate with ICESat high with surface NASA slopes Airborne and/or Topographic topographic Mappervariability (ATM) c.f. [109] laser. altimeter data [175]. Using a similar method, an average 0.07 m offset was found betweenThe NASA RT method airborne calculates LVIS lidar 푑퐻 and/푑푡 ICESatfrom repeat elevation measurements retrievals over at reference the GrIS points [176]. Thealong OFP the methodreference has tracks, greater which spatial increases coverage spatial owing measurement to the combined density use ofand RT coverage and XO points.relative Consequently, to XO points (Figure 3) [173]. However, RT points are typically further apart than XO points, which may increase slope errors [99]. Owing to higher measurement density, the RT method has seen greater operational use than the XO method and is more appropriate for local or regional scale analysis [99]. The RT method was combined with Advanced Spaceborne Thermal Emission and Reflection Radiometer

Remote Sens. 2019, 11, 2405 12 of 50

Remote Sens. 2019, 11, x FOR PEER REVIEW 12 of 52 dH/dt uncertainty appears to be the lowest among methods [99]. The method was used to estimate (ASTER) digital elevation model differencing, which together suggest an average volume loss in an average thickening of 0.02 m yr 1 (equivalent to 21 km3 yr 1 volumetric change) for areas above southeast Greenland of ~108 km3 yr−−1 for the period 2003–2005 [161]− . Concentrated thinning of narrow 2000 m a.s.l., and an average thinning of 0.24 m yr 1 (168 km3 yr 1) below 2000 m a.s.l. for the period outlet glaciers contributed less volume loss− than dispersed− , but smaller− thinning over larger interior 2003–2007 [100]. Thinning rates near the margin were found to be slightly larger than estimates from areas, highlighting the enhanced capability of laser altimetry for observing local-scale processes in other methods due to greater sample size and coverage in the ablation zone. the ablation zone [161].

FigureFigure 3. IllustrationIllustration of of altimeter altimeter (e.g., (e.g., ICESat) ICESat) reference reference track track and and repeat repeat tracks. tracks. Repeat Repeat tracks tracks are areparallel parallel to reference to reference tracks tracks but separated but separated in the in cross the cross-track-track direction. direction. An area An of area interest of interest in the along in the- along-tracktrack direction direction is used is usedto collect to collect ground ground footprint footprint measurements measurements that that are are spatially spatially interpolated interpolated to recoverrecover thethe area-averagearea-average surfacesurface elevationelevation andand subsequentsubsequent changechange inin area-averagearea-average surfacesurface elevationelevation dH푑퐻//dt푑푡.. InIn practice,practice, repeatrepeat trackstracks areare oftenoften notnot parallelparallel owingowing toto orbitalorbital variations,variations, andand inin thethe casecase ofof ICESatICESat laserlaser altimetry,altimetry, footprintfootprint diameterdiameter maymay changechange owingowing toto laserlaser transmit transmit power power variation. variation.

TheThe TINOFP methodmethod [103uses] isboth similar XO and to RT RT as points it aggregates but requir all availablees that the along-track laser footprints measurements of repeat withinmeasurements a predefined overlap. distance ICESat around footprints each RTare reference ellipses point.with a Unlike50–90 them major standard axis RTlength method, [174] a. TINThe surfaceoverlapping is fitted criterion to a subset is determined of points at each by a RT predefined reference point semi- collectedmajor axis within distance a reference between two-year laser dH dt timefootprints period. [100]./ In principlis estimatede, the from method all combinations does not account of surface for the elevation slope between measurements footprints outside, but slope the referencecorrections period have relativebeen applied, to the TINfor example surface.,The to validate TIN surface ICESat implicitly with NASA accounts Airborne for cross-track Topographic and along-trackMapper (ATM) slope laser and altimeter sacrifices data temporal [175]. Using resolution a similar for both method, spatial an resolutionaverage 0.07 and m referenceoffset was period found consistency. Using the TIN method, ICESat dH/dt uncertainty was estimated to be 0.07 m yr 1 at 1σ between NASA airborne LVIS lidar and ICESat elevation retrievals over the GrIS± [176]. The− OFP levelmethod [103 has]. However, greater spatial the method coverage may owing systematically to the combined under-sample use of someRT and areas XO ofpoints. the ice Consequently, sheet margin with푑퐻/푑푡 insu uncertaintyfficient points appears to fit to the be TIN, the leadinglowest among to an underestimated methods [99]. The total method ice sheet was volume used changeto estimate [99]. Qualitatively,an average thickening each method of 0.02 shows m yr−1 a similar(equivalent pattern to 21 of km increasing3 yr−1 volumetric ice thinning change) along for the areas ice above sheet margins2000 m a.s.l. with, and the highestan average thinning thinning rates of along −0.24 the m southernyr−1 (168 coasts,km3 yr−1 especially) below 2000 in the m marine-terminatinga.s.l. for the period Jakobshavn2003–2007 [100] Isbrae. Thinning region and rates in near southeast the margin Greenland. were found Across to methods, be slightly pan-GrIS larger than volume estimates loss ranges from 3 1 fromother ~200–275methods kmdue toyr −greaterduring sample the ICESat size and era coverage [99]. in the ablation zone. The TIN method [103] is similar to RT as it aggregates all available along-track measurements 3.2.2. ICESat-2 and Future Opportunities within a predefined distance around each RT reference point. Unlike the standard RT method, a TIN surfaceThe is Advancedfitted to a subset Topographic of points Laser at each Altimeter RT reference System point (ATLAS) collected was within launched a reference in September two-year 2018time period. onboard 푑퐻 the/푑푡 Ice, is estimated Cloud, and from land all Elevationcombinations Satellite-2 of surface (ICESat-2) elevation (Table measurements1). ATLAS outside is a dual-beamthe reference single-photon period relative counting to the laser TIN altimeter surface. markedThe TIN by surface several implicitly improvements accounts to the for ICESat cross/GLAS-track measurementand along-track strategy slope and [40,177 sacrifices]. These temporal include itsresolution dual-beam for laser,both s organizedpatial resolution into three and pairs reference with aperiod 90 m consistency. beam separation Using andthe TIN 3.3 method, km pair ICESat separation. 푑퐻/푑푡Each uncertainty beam has was a estimated nominal to 17 be m ±0.07 spatial m footprintyr−1 at 1σ diameterlevel [103] and. However, 0.7 m along-track the method measurement may systematically spacing (Figure under4-).sample This configuration some areas of permits the ice sheet margin with insufficient points to fit the TIN, leading to an underestimated total ice sheet volume change [99]. Qualitatively, each method shows a similar pattern of increasing ice thinning

Remote Sens. 2019, 11, 2405 13 of 50 the determination of cross-track slope, increases spatial measurement density, and increases local measurement accuracy [177]. ICESat-2 mission requirements include the determination of ice sheet 1 1 2 dH/dt to 0.4 cm yr− accuracy (0.25 cm yr− for outlet glaciers) when averaged over 100 km areas [177]. Data collection began on Oct 14, 2018, with initial ICESat-2 data products released on 7 June 2019, including an initial release of the ATLAS/ICESat-2 L3A Land Ice Height (ATL06), Version 1 [178], which provides mean ice sheet surface elevation averaged along 40 m segments of ground track posted at 20 m along-track spacing (Figure5)(https: //nsidc.org/data/ATL06/). Remote Sens. 2019, 11, x FOR PEER REVIEW 14 of 52

FigureFigure 4. 4.( (aa)) ExampleExample ofof ATLASATLAS/ICESat/ICESat-2-2 Level-3ALevel-3A datadata productproduct LandLand IceIce Height,Height, VersionVersion 11 (ATL06),(ATL06), representingrepresenting 304,550304,550 uniqueunique surface elevation elevation measurements measurements collected collected on on 18 18 October October 2018 2018 [178] [178; (b];) (bexample) example crossover crossover location location (red (red box box inset inset in (a in)) (showinga)) showing the theICESat ICESat-2-2 multi multi-beam-beam configuration configuration (two (twobeam beam-pairs-pairs shown, shown, the the third third beam beam-pair-pair was was not not included included in in ATL06 ATL06 Version Version 1 1 at this time time and and location).location). Single-beamSingle-beam altimetersaltimeters suchsuch asas ICESatICESat givegive oneone crossovercrossover measurementmeasurement atat eacheach crossover crossover location,location, andand the the cross-track cross-track slope slope cannot cannot be be determined determined in in the the along-track along-track direction. direction. The The ICESat-2 ICESat-2 multi-beammulti-beam configuration configuration allows allows determination determination of cross-track of cross- slopetrack and slope gives and multiple gives unique multiple crossover unique measurementscrossover measurements at each crossover at each location crossover (up location to nine (up possible, to nine four possible, shown four in this shown example). in this example).

InIn addition addition to to an an improvedimproved multi-beammulti-beam configuration,configuration, thethe ATLASATLAS instrument instrument will will transmit transmit laserlaser pulsespulses at at 532 532 nm nm wavelength,wavelength, which which willwill enhanceenhance bothboth photon-returnphoton-return density density and andoptical opticalpenetration penetration capabilitycapability relativerelative to to the theGLAS GLASinstrument instrument onon ICESat.ICESat. Water,Water,both bothin inliquid liquidand and solidsolid form,form, isis nearlynearly transparenttransparent to to 532 532 nm nm light. light. ForFor example,example, thethe pathpath lengthlength inin purepure waterwater requiredrequired to to attenuate attenuate 532 532 nm nm lightlight to to 50% 50% of of its its incident incident intensity intensity is ~16 is ~16 m, whereasm, whereas this paththis path length lengt is ~0.01h is m~0.01 for 1064m for nm 1064 light, nm owing light, toowing an ~1400x to an respective ~1400x respective increase in increase the imaginary in the index imaginary of refraction index of of water refraction [179, 180 of ]. water Consequently, [179,180]. ICESat-2Consequently, laser pulsesICESat at-2 laser 532 nmpulses will at e 532ffectively nm will “see effectively through” “see liquid through” water liquid and water thereby and facilitate thereby spacebornefacilitate spaceborne measurement meas ofurement supraglacial of supraglacial lake, stream, lake, stream, and water-filled and water crevasse-filled crevasse bathymetry bathymetry [181]. Conversely,[181]. Conversely, the reflectivity the reflectivity of glacier of ice glacier and snow ice and is near snow maximum is near maximum at 532 nm at owing 532 nm to the owing extremely to the lowextremely absorption low absorption coefficient coefficient of ice and of extremely ice and extremely high scattering high scattering efficiency efficien from icecy grainsfrom ice and grains air bubblesand air [182 bubbles]. This [182] enhanced. This reflectivity enhanced should reflectivity produce should higher produce photon-return higher photon density-return relative density to the 1064relative nm to laser the but1064 will nm also laser include but will a also non-zero include volume a non- scatteringzero volume component scattering that component could introduce that could a rangeintroduce bias similara range tobias radar similar penetration to radarinto penetration snow and into firn snow [183 ,and184]. firn [183,184].

Remote Sens. 2019, 11, 2405 14 of 50 Remote Sens. 2019, 11, x FOR PEER REVIEW 15 of 52

Figure 5. 5. (a)( aExample) Example of ATLAS/ICESat of ATLAS/ICESat-2-2 Level Level-3A-3A data product data product Land Ice Land Height, Ice Height,Version 1 Version (ATL06) 1 (ATL06)[178], collected [178], collected on 24 October on 24 October2018 in 2018the Isunnguata in the Isunnguata Sermia Sermiaand Kangerlussuaq and Kangerlussuaq catchments catchments in the southin thewestern southwestern Greenl Greenlandand Ice Sheet Ice ablation Sheet ablation zone. Background zone. Background image imageis Landsat is Landsat 8 Operational 8 Operational Land ImagerLand Imager 30 m resolution 30 m resolution image image (RGB: (RGB: band 4 band (red), 4 (red),3 (green), 3 (green), and 2 and(blue)) 2 (blue)) collected collected on 24 onAugust 24 August 2018. Inset2018. is Inset Greenland is Greenland Ice Mapping Ice Mapping Project Project (GIMP) (GIMP) ice mask ice [185] mask with [185 ]red with box red sho boxwing showing extent extentof image of area;image (b area;) elevation (b) elevation profile profilefor ground for ground track profile track profile 3, right 3, beam right beam(‘gt3r’), (‘gt3r’), for the for north the northto south to southtrack showntrack shown in (a), in showing (a), showing the the rough, rough, crevassed crevassed ice ice surface surface in in the the marginal marginal ablation ablation zone. zone. ATL06 ATL06 elevations represent meanmean surfacesurface elevationelevation averaged averaged along along 40 40 m m segments segments of of ground ground track, track, posted posted at at20 20 m m along-track along-track spacing. spacing.

ICESatICESat-2-2 research to date has focused on pre-missionpre-mission proof of concept, airborne and ground validation, and sensor calibration [186 [186–188]188].. In In addition addition to to its its unique abil abilityity to map surface water bathymetry, other novel applications of laser altimetry to the ablation zone that will benefit benefit from ICESatICESat-2-2 continuitycontinuity includeinclude assimilation assimilation of ofdH 푑퐻/dt/푑푡observations observations into into transient transient ice flow ice flow simulations simulations [189], [189]fusion, fusion of multi-sensor of multi-sensor (e.g., stereophotogrammetry (e.g., stereophotogrammetry or radar or altimetry)radar altimetry) datasets datasets to increase to increase the spatial the spatialdensity density of surface of surface elevation elevation measurements measurements [174], [174] and, application and application of lidar-based of lidar-based snow snow grain grain size sizeand and surface surface roughness roughness retrievals retrievals independent independent of of or or in in combination combination with with opticaloptical sensors [190,191] [190,191].. Comparison between ice sheet surface elevation change estimates by the laser altimetry and regional climate model (RCM) SMB is another area that that is is underexploredunderexplored [[89]89].. Such comparisons require accurate knowledge of the surface density and change in surface surface elevation elevation due due to to ice ice dynamic dynamic motion, motion, andand,, therefore,therefore, are are most most appropriate appropriate for bare for icebare areas ice during areas duringsummer summer when surface when melting surface dominates melting dominatesthe elevation the change elevation signal change [89]. signal Finally, [89] it. Finally, was recently it was reportedrecently reported that the CryoSat-2that the CryoSat science-2 teamscience is teamconsidering is considering adjusting adjusting the satellite’s the satellite’s orbit to overlap orbit to its overlap ground its tracks ground with tracks those with of ICESat-2 those of once ICESat every-2 ~2once days every [21]. ~2 If so, days these [21] overlapping. If so, these data overlapping would provide data cross-platform would providHde/ dt crossvalidation-platform and 푑퐻 could/푑푡 validationbe used to and infer could changes be used in the to CryoSat-2 infer changes Ku-band in the penetration CryoSat-2 Ku depth-band into penetration snow and depth firn. into snow and firn.

Remote Sens. 2019, 11, 2405 15 of 50

4. Remote Sensing of Mass Balance Three methods are used to estimate MB from satellite remote sensing: (1) changes in ice sheet surface elevation from altimetry, (2) the mass budget or input–output method, and (3) time-variable gravimetry [192]. Satellite and airborne altimeters measure dH/dt, which is converted to MB using modeled firn density and modeled SMB to partition D. The input–output method uses remotely-sensed ice sheet surface velocity and ice sheet thickness to calculate D for outlet glacier drainage basins and uses modeled SMB to calculate MB. Satellite gravimetry measures changes in the earth’s gravity field, to directly estimate MB. The methods are complementary and largely independent, and together provide a comprehensive view of spatial and temporal patterns in the GrIS MB, since 2002 for gravimetry, 1992 for altimetry, and 1972 for IOM [192]. We review each method below and compare estimates of MB from each.

4.1. Converting Ice Surface Elevation Change to Mass Change Satellite altimeters observe the total change in ice sheet surface elevation, which includes both real and apparent changes in ice sheet thickness dH. Glaciers and ice sheets thicken from snow accumulation and thin by melting, sublimation, horizontal ice flow, snow redistribution, and increases in ice, snow, and firn density. Apparent changes in ice sheet thickness are caused by the vertical motion of underlying bedrock. Converting altimetric measurements of dH/dt to change in ice volume and mass balance requires an estimate of each individual contribution to dH/dt, and a correction for each term in Equation (4) that is not associated with a change in ice volume and/or change in mass balance. The typical procedure for estimating each term on the right-hand side of Equation involves obtaining SMB from a climate model and defining the accumulation and ablation zones as those areas where SMB is positive and negative, respectively, separated by the ELA where modeled SMB = 0 [86]. Above the ELA, positive dH/dt is caused by net accumulation and ρs is estimated with a firn density model, whereas negative dH/dt is caused by ice dynamics and ρs is solid ice density (typically taken 3 as 917 kg m− ). Below the ELA, dH/dt is caused by both negative SMB and ice dynamics, and solid ice density is used to convert both to mass c.f. [86,97,105,193]. The vc term causes an apparent change in mass that is estimated with a firn compaction model and removed from dH/dt [194]. BMB is negligible 1 in the accumulation zone and very small (~0.015–0.020 m yr− ) in the ablation zone, and vbr contributes 1 about 1 GT yr− in apparent mass change that can be corrected or ignored [86,195]. When corrected for the terms in Equation, the ICESat altimetry data suggest mass loss rates of –191 23 Gt yr 1 to –240 28 Gt yr 1 for the period 2003–2008 [86]. The spread arises from the choice ± − ± − of dH/dt interpolation method (Section 3.2.1) and, more substantially, the choice of firn density model. 1 Removal of vc reduces mass loss rates by 33–36 Gt yr− , almost all of which is due to compaction in the zone between the ELA and 2000 m a.s.l. [86]. Despite the uncertainties associated with firn densification, a good agreement is found between altimetry and mass balance estimates from GRACE, which suggest a mass loss of –230 33 Gt yr 1 during the period 2002–2009 [196], –179 25 Gt ± − ± yr 1 for 2002–2007 [197], and –237 20 Gt yr 1 for 2003–2008 [17]. Applying the same methods as − ± − Sørensen et al. [86], Greenland’s peripheral ice caps lost mass at –28 11 Gt yr 1 (0.08 0.03 mm yr 1 ± − ± − SLR equivalent), representing 20% of total GrIS mass loss during the ICESat era [96]. A combined analysis of ICESat data and airborne laser altimeter data from the NASA Airborne Thematic Mapper (ATM) suggests a mass loss rate of –243 18 Gt yr 1 (0.68 0.05 mm yr 1 SLR ± − ± − equivalent) for the 2003–2009 period [97]. The SERAC method was used to combine ICESat and ATM XO and RT points, and an SMB model was used to diagnose spatial differences in thinning, likely representing the most comprehensive analysis of spatiotemporal variability in dH/dt and its attribution to SMB vs. D for the ICESat era. Negative SMB accounted for 52% of total mass loss and accelerated during the study period. Persistent SMB decreases were concentrated in the southwest and southeast marginal ablation zones, but an acceleration in thinning due to SMB was also observed in the northwest. On average, dynamic thinning of outlet glaciers decelerated during the study period, but large spatiotemporal variability was found with some outlets accelerating, owing to local scale Remote Sens. 2019, 11, 2405 16 of 50 controls on ice dynamics (e.g., bed geometry). Regionally, the largest contribution to D was from southeast outlet glaciers, with the largest thinning rates observed on Helheim and Kangerdlugssuaq glaciers and on Sermeq Kujalleq (Jakobshavn Isbræ) in the west. As with prior studies, e.g., [198], a slight thickening of the interior ice sheet above the ELA was found.

4.2. The Input–Output Method Satellite altimetry is unique in that it provides spatially resolved estimates of mass changes, but attribution of dH/dt to SMB vs. D requires estimating SMB with a regional climate model and treating D as a residual. Alternatively, the input–output method (IOM) is used to calculate D directly from remotely sensed ice surface velocity and outlet glacier geometry [12,17,18,192,199–201]. Ice surface velocity is obtained from offset-feature tracking of optical [202–206] or SAR imagery [207–212], and outlet glacier (flux gate) geometry is obtained from airborne radar soundings of ice sheet thickness [213–215], or a mass conservation approach that combines the two, recently with fjord bathymetry and for the entire ice sheet [216–219]. As with altimetric surveys of dH/dt, the IOM reveals a complex spatial pattern of ice sheet mass loss concentrated in narrow outlet glaciers along the coastal margins, with longer term (e.g., 1990–present) mass losses concentrated in the southeast and the Jakobshavn basin in the west, and recent (post-2005) increases concentrated in the northwest [18,192,200,220,221]. In general, a small number of isolated glaciers drive the majority of D with just four (Sermeq Kujalleq, Kangerdlugssuaq, Køge Bugt, and Ikertivaq South) accounting for ~50% of total D acceleration during the period 2000–2012 [18,192,222]. The IOM provides the longest record of D and, therefore, uniquely places recent MB trends in a long term context. For example, although negative SMB accounts for approximately 60% of MB for the period 1990–present, the proportion is reversed for the period 1972–2018 [192,220,221]. The recent availability of digital bed elevation models, ice surface velocity mosaics, and digital surface elevation models for the entire ice sheet provide new insight into the spatial and temporal drivers of GrIS MB at the scale of individual outlet glaciers and their upstream drainage basins [185,206,212,218]. Mouginot et al. [192] reconstruct 46 years of MB for 260 individual outlet glaciers for the period 1972–2018, using surface velocity constructed from all available SAR (ALOS/PALSAR, ENVISAT/ASAR, RADARSAT-1/2, and Sentinel-1b) and Landsat-8 optical imagery [212], ice thickness from BedMachine v3 [218], and surface topography from the Greenland Ice Mapping Project DEM [185], WorldView DEMs, and historical orthophotographs [223]. Their results show that MB switched from +47 21 Gt yr 1 in ± − 1972–1980 to –187 17 Gt yr 1 in 2000–2008, and –286 20 Gt yr 1 in 2010–2018, within the bounds of ± − ± − most estimates from altimetry and GRACE for the latter two periods. As in Enderlin et al. [18], they find four glaciers control half the mass loss during 2000–2012, but for the 1972–2018 study period, Ikertivaq South gained mass, while Steenstrup-Dietrichson in northwest, Humboldt in north, and Midgårdgletscher in southeast join the three others as dominant drivers of mass loss. The largest single historical contributor is Sermeq Kujalleq (Jakobshavn Isbræ), which has experienced long term retreat and episodic rapid acceleration [224], including the disintegration of its floating ice tongue between 2000–2003 and rapid retreat of its calving front (Figure6)[ 224–229]. Whereas D has historically been dominated by the southeast, northwest, and western sectors, they conclude that the north and northeast sectors are of greatest importance to future sea level rise owing to their present-day slow velocities and potential for large increases in D. Remote Sens. 2019, 11, 2405 17 of 50 Remote Sens. 2019, 11, x FOR PEER REVIEW 18 of 52

FigureFigure 6.6. The largest single contributor of ice discharge D퐷 fromfrom thethe GreenlandGreenland IceIce SheetSheet toto thethe globalglobal oceanocean duringduring recentrecent decadesdecades isis SermeqSermeq KujalleqKujalleq (Jakobshavn(Jakobshavn Isbræ),Isbræ ), shown herehere in a 2727 JulyJuly 20172017 LandsatLandsat 88 OperationalOperational LandLand ImagerImager 3030 m resolution image (RGB: band 4 (red), 3 (green), and 2 (blue)). TheThe annual annual average average ice ice surface surface velocity velocity for the for period the period 2011–2016 2011 is– shown2016 is for shown upstream for upstreamareas, calculated areas, fromcalculated Landsat-8 from (optical), Landsat Sentinel-1,-8 (optical), and Sen RADARSAT-2tinel-1, and (interferometricRADARSAT-2 (interferometric SAR) image processing SAR) image [212]. 1 −1 Theprocessing maximum [212]. surface The maximum velocity value surface is ~12 velocity km yr −value. Calving is ~12 fronts, km yr which. Calving mark fronts, the terminus which mark position the ofterminus the outlet position glacier of the where outlet ice glacier is discharged where ice to isthe discharged ocean, are to the mapped ocean, fromare mapped Landsat-7, from Landat-8, Landsat- and7, Landat Sentinel-2B-8, and opticalSentinel imagery-2B optical [229 imagery]. Sermeq [229]. Kujalleq Sermeq has Kujalleq lost 137 has km lost2 of 137 surface km2 of area surface between area 1998–2018between 1998 (compare–2018 thick(compare red line thick to thick redblack line line), to thick resulting black in near-completeline), resulting disintegration in near-complete of its floatingdisintegration ice tongue of its [224 floating,227]. ice Its flowtongue speed [224,227]. has nearly Its flow doubled speed since has thenearly early doubled 1990s, withsince upstream the early 1 −1 thinning1990s, with rates upstream exceeding thinning 15 m yrrates− [exceeding226]. Inset 15 is m the yr Greenland [226]. Inset Ice is Mapping the Greenland Project Ice (GIMP) Mapping ice maskProject [185 (GIMP)] with ice the mask red box [185] showing with the the red extent box showing of the image the extent area. of the image area.

4.3.4.3. Time VariableVariable GravimetryGravimetry andand thethe Twin-GRACETwin-GRACE MissionMission TheThe NASANASA/German/German AerospaceAerospace CenterCenter twintwin Gravity Recovery and Climate Experiment (GRACE) satellitessatellites (2002–2017)(2002–2017) areare conceptuallyconceptually unique from all otherother remoteremote sensingsensing platformsplatforms because they dodo notnot measuremeasure thethe interactioninteraction betweenbetween thethe earthearth andand electromagneticelectromagnetic radiationradiation [[230]230].. AsAs thethe twintwin GRACEGRACE satellitessatellites orbit,orbit, theirtheir proximalproximal distancedistance variesvaries withwith earth’searth’s gravitationalgravitational pull.pull. TheseThese smallsmall changeschanges inin theirtheir accelerationacceleration areare usedused toto measuremeasure variationsvariations inin thethe densitydensity ofof earthearth atat ~400–500~400–500 kmkm spatialspatial resolution. Consequently, GRACE provides the only only direct direct measurement measurement of of ice ice sheet sheet 푀퐵MB andand independentindependent validationvalidation of ofMB 푀퐵estimates estimates made made from from satellite satellite altimetry altimetry and and RCMs RCMs (Figure (Figure7)[ 231 7) –[231233].– 233].

Remote Sens. 2019, 11, 2405 18 of 50 Remote Sens. 2019, 11, x FOR PEER REVIEW 19 of 52

Figure 7. Comparison of ice surface elevation change 푑퐻/푑푡 [m yr1−1] from CryoSat-2 radar altimetry Figure 7. Comparison of ice surface elevation change dH/dt [m yr− ] from CryoSat-2 radar altimetry with mass change 푑푀/푑푡 [(m w.e. yr1−1) from GRACE. Both missions observe similar spatial patterns, with mass change dM/dt [(m w.e. yr− ) from GRACE. Both missions observe similar spatial patterns, withwith mass mass losses losses concentrated concentratedin in the the southeast southeast and and western western sectors, sectors, andand aa slight slight thickening thickening ofofthe the interior.interior. AltimetryAltimetry resolvesresolves rapidrapid thinningthinning concentratedconcentrated inin narrownarrow outletoutlet glaciersglaciers alongalong thethe coastal coastal margins.margins. TheThe CryoSat-2 CryoSat-2 Surface Surface Elevation Elevation Change Change (SEC) (SEC) product product version version 2.2 2.2 is is based based on on the the ESA ESA BaselineBaseline C C CryoSat-2 CryoSat- product2 product and and is provided is provided at 1 kmat 1 gridkm spacinggrid spacing as a five-year as a five average-year average for the periodfor the 2011–2015period 2011 [234–2015]. The [234] GRACE. The mass GRACE balance mass product balance is producedproduct is by produced the Danish by Institutethe Danish of Space Institute (DTU of Space)Space and(DTU is providedSpace) and as is a five-yearprovided average as a five for-year the average period 2012–2016 for the period with a2012 500– km2016 nominal with a spatial500 km resolutionnominal spatial [235]. Bothresolution datasets [235] are. Both Essential datasets Climate are Essential Variables Climate provided Variables by the ESA provided Climate by Changethe ESA InitiativeClimate Change and are availableInitiative onlineand are at available http://products.esa-icesheets-cci.org online at http://products.esa. -icesheets-cci.org.

AsAs an an independent independent dataset, dataset, many many studies studies have have combined combined or compared or compared GRACE GRACEMB with 푀퐵 satellite with altimetrysatellite altimetryor other traditional or other traditionalremote sensing remote techniques sensing [ techniques236]. GRACE [236]MB. GRACEhas been 푀퐵 compared has been to ERScompared radar altimetry to ERS radar [237], altimetry to surface [237] melt, extentto surface from melt MODIS extent thermal from imageryMODIS thermal [238], and imagery to MB from[238], combinedand to 푀퐵 RCM fromSMB combinedand InSAR RCM derived 푆푀퐵 Dand[239 InSAR]. These derived comparisons 퐷 [239] confirmed. These comparisons earlier findings confirmed from bothearlier satellite findings altimetry from both and RCMssatellite that altimetry the GrIS andMB RCMsdeclined that at the an acceleratingGrIS 푀퐵 declined rate during at an the accelerating post-2002 periodrate during [196,237 the,240 post].- GRACE2002 periodMB estimates[196,237,240] range. GRACE from 191푀퐵 estimates21 Gt yr range1 between from 2002–2009 −191 ± 21 [Gt98] yr to−1 − ± − between244 6 Gt2002 yr–12009between [98] to 2002–2016, −244 ± 6 Gt with yr−1 an between acceleration 2002– of2016,28 with9 Gtan yracceleration2 during thisof − period28 ± 9 Gt [241 yr].−2 − ± − − ± − Anduring updated this period analysis [241] of GRACE. An updated data suggestsanalysis of the GRACE negative dataMB suggestsacceleration the negative halted in푀퐵 2013 acceleration owing to thehalted suppression in 2013 owing of surface to the runo suppressionff production of surface in west runoff Greenland production caused byin west a shift Greenland in the North caused Atlantic by a Oscillationshift in the to North its cool Atlantic phase Oscillation [14,93]. to its cool phase [14,93]. TheThe accuracy accuracy of of GRACE GRACE data data is is inherently inherently limited limited by by its its spatial spatial resolution, resolution, with with accuracy accuracy decreasingdecreasing as as spatial spatial resolution resolution increases increases [ 197[197,232,,232,242242].]. Sophisticated Sophisticated signal signal processing processing methods methods are are requiredrequired to to determine determine the the optimal optimal spatial spatial resolution resolution and and toto quantify quantify the the spatial spatial dependence dependence ofof thethe signal processing error [240,243,244]. For example, north–south oriented spatially correlated errors (“striping”) are a persistent issue [245]. Most early methods used the Mascon approach to extract the

Remote Sens. 2019, 11, 2405 19 of 50 signal processing error [240,243,244]. For example, north–south oriented spatially correlated errors (“striping”) are a persistent issue [245]. Most early methods used the Mascon approach to extract the GRACE signal onto a geometric grid [233,246]. Recent work has shown spherical Slepian functions theoretically maximize the spatial resolution of the GRACE signal [241]. The Slepian solutions were used to extract MB trends at a previously unresolvable scale, showing nearby Baffin Island and 1 Ellesmere Island MB trends of 22 2 and 38 2 Gt yr− , respectively [241]. In addition to signal − ± − ± 1 processing errors, the glacial isostatic adjustment effect on the GRACE signal is ~5–20 Gt yr− for the GrIS, but is uncertain [235,247]. The GRACE Follow-On (GRACE-FO) mission (launched 22 May 2018) is the successor to the GRACE mission [248]. GRACE-FO uses the same orbital acceleration measurement technology as the original GRACE mission and will continue the GRACE measurement time series for a planned ten years. In addition to providing gravity time series continuity, GRACE-FO will test a laser ranging interferometer that is expected to improve the satellite-to-satellite distance measurement relative to the GRACE microwave ranging system [42,43]. Following a period of in-orbit checks, GRACE-FO entered the science phase of its mission in January 2019 [249]. The first GRACE-FO Level-1 data products were released on 24 May 2019, and Level-2 gravity field products were released on 10 June 2019, with planned monthly release updates thereafter [250].

5. Remote Sensing of Ice Surface Reflectance and Albedo Surface albedo modulates the absorption of shortwave radiation by ice and snow surfaces [75]. Shortwave radiation is the dominant contributor to melt energy in the GrIS ablation zone [90]. Consequently, albedo is an important control on the production of surface meltwater and, by extension, the SMB [93]. Satellite remote sensing instruments measure multispectral reflectance, which is used to estimate albedo [251]. Remotely sensed albedo is used to understand spatial patterns in surface melting, as input to land surface models and regional climate models that simulate the SMB, and as a diagnostic marker of the changing ice surface [26,30,252–254]. The following subsections summarize key terminology relevant to spaceborne measurements of surface reflectance and albedo, the sensors and datasets used to retrieve the GrIS surface albedo, observed changes in GrIS albedo and their diagnosis, and current challenges and future opportunities for understanding the GrIS ablation zone albedo.

5.1. Definition of Reflectance, BRDF, and Albedo Standardized nomenclature for reflectance quantities were defined in terms of incident and reflected beam geometry by Nicodemus et al. [255] and later adapted for common optical remote sensing measurement configurations [256–259]. Here we briefly review key definitions of reflectance quantities, following Schaepman-Strub et al. [259]. Spectral radiance, L, is the radiant flux in a beam 2 1 1 per unit wavelength and per unit area and solid angle of that beam, with SI units [W m− sr− nm− ]. 2 Reflectance is the ratio of exitent radiant flux density (radiant exitance, M) [W m− ] to incident radiant 2 flux density (irradiance, E) [W m− ], where both the incident and exitent radiance are integrated across the beam geometry [sr] and radiant spectrum [nm] [259]. For passive optical remote sensing measurements, the incident beam geometry is hemispherical and is composed of both direct-beam and diffuse solar radiation (Figure8). The exitent (reflected) beam geometry is conical, defined by the sensor instantaneous field of view (IFOV), and is composed of both direct-beam and diffuse reflected solar radiation, corresponding to the hemispherical-conical reflectance function (HCRF) described by Schaepman-Strub et al. [259]. 1 The bidirectional reflectance distribution function BRDF [sr− ] describes the scattering of an infinitesimal beam of incident light from one direction in a hemisphere into another direction, and is Remote Sens. 2019, 11, 2405 20 of 50

2 1 defined (omitting spectral dependence) as the ratio of directional reflected radiance Lr [W m− sr− ] to 2 incident irradiance Ei [W m− ]: dLr(θi, φi, θr, φr) BRDF = (6) dEi(θi, φi) whereRemote θSens.i and 2019θ, r11are, x FOR the PEER incident REVIEW and reflected viewing angle, respectively, and φi and φr are21 theof 52 incident and reflected azimuth angles, respectively. As a ratio of infinitesimal quantities, the BRDF is a theoreticala theoretical construct construct that cannot that cannot be directly be directly measured measured but describes but describes the intrinsic the reflectance intrinsic properties reflectance ofproperties a surface from of a which surface measurable from which quantities measurable can be quantities derived [259 can]. be For derived example, [259 integration]. For example, of the BRDFintegration across ofθr the(0 BRDFπ/2) acrossand φ r휃(0r(0 →2ππ)/2yields) and bihemispherical 휙r(0 → 2휋) yields reflectance, bihemispherical or what reflectance, is commonly or → → calledwhat albedo.is commonly called albedo.

FigureFigure 8. 8.Diagram Diagram ofofthe theincident incidentand andreflected reflected viewing viewingangles anglesfor for a a directional directional light light source source with with directdirect and and di diffuseffuse (hemispherical) (hemispherical) incoming incoming radiance radiance and and conical conical reflected reflected radiance, radiance, corresponding corresponding toto the the hemispherical-conicalhemispherical-conical reflectance reflectance factor factor (HCRF) (HCRF) described described by by Schaepman Schaepman-Strub-Strub et al. et [ al.259 []259. The]. The incident viewing angle (θi), reflected viewing angle (θr), and the azimuth angle (φr) together define incident viewing angle (휃i), reflected viewing angle (휃r), and the azimuth angle (휙) together define the angular coordinates of both the directional illumination and the light reflected toward the sensor, the angular coordinates of both the directional illumination and the light reflected toward the sensor, with respect to the surface normal, z. with respect to the surface normal, z. Viewing geometry is important because every natural earth surface is an anisotropic reflector. Viewing geometry is important because every natural earth surface is an anisotropic reflector. For example, snow and ice are strongly forward scattering at the particle scale, especially at visible For example, snow and ice are strongly forward scattering at the particle scale, especially at visible wavelengths, but ice surfaces can be strongly backward scattering due to surface roughness [260]. wavelengths, but ice surfaces can be strongly backward scattering due to surface roughness [260]. Consequently, an estimate of the BRDF is required to convert multi-angular reflectance measured by Consequently, an estimate of the BRDF is required to convert multi-angular reflectance measured by satellite remote sensing instruments to albedo. To achieve this, the BRDF for each satellite ground satellite remote sensing instruments to albedo. To achieve this, the BRDF for each satellite ground footprint is estimated (not measured) from repeat multi-angular reflectance measurements collected footprint is estimated (not measured) from repeat multi-angular reflectance measurements collected over time and, therefore, under varying illumination conditions [251] or collected instantaneously by over time and, therefore, under varying illumination conditions [251] or collected instantaneously by spaceborne multi-angular instruments [261]. A complete technical discussion of the BRDF is available spaceborne multi-angular instruments [261]. A complete technical discussion of the BRDF is available from Schaepman-Strub et al. [259]. from Schaepman-Strub et al. [259].

5.2. Optical Reflectance and Albedo Sensors and Datasets Spatially- and temporally-consistent repeat measurements of surface reflectance, and estimates of BRDF and albedo for the Greenland Ice Sheet are provided by satellite-borne imaging spectrometers, including the NASA Moderate Resolution Imaging Spectroradiometer (MODIS), the NASA Multi-angle Imaging SpectroRadiometer (MISR), and the National Oceanic and Atmospheric Administration’s Advanced Very High Resolution Radiometers (AVHRR) (Table 2) [261–264]. Other notable spaceborne sensors providing surface reflectance that have been used to study the GrIS

Remote Sens. 2019, 11, x FOR PEER REVIEWRemote Sens. 2019, 11, x FOR PEER REVIEW 5 of 52 5 of 52 the accumulation zone, regional climatethe accumulation models and zone, firn regionalcompaction climate models models are used and firnto partition compaction the models are used to partition the component of 푑퐻/푑푡 caused by componentchange in 푆푀퐵 of 푑퐻 from/푑푡 thosecaused caused by change by change in 푆푀퐵 in firnfrom density those causedand 퐷 by change in firn density and 퐷 [104,105]. The following subsections[104,105] summarize. The following spacebor nesubsections radar altimetry summarize of the spacebor GrIS, providingne radar an altimetry of the GrIS, providing an historical perspective that emphasizeshistorical major perspective challenges that and emphasizes advances major in the challenges field, and identifiesand advances in the field, and identifies future opportunities for radar altimetryfuture opportunities observations offor the radar GrIS altimetry ablation observations zone. of the GrIS ablation zone.

3.1. Radar Altimetry 3.1. Radar Altimetry Conventional radar altimeters areConventiona nadir pointingl radar single altimeters-beam radars are nadir that pointing transmit single and receive-beam radars that transmit and receive tens to thousands of microwavetens pulses to thousands per second, of microwave yielding ground pulses measurement per second, yielding footprint ground measurement footprint diameters on the order 1–10 km posteddiameters at 0.3 on–0.6 the km order along 1–10-track km spacingposted at [39,106,107] 0.3–0.6 km. alongAn estimate-track spacing of [39,106,107]. An estimate of 푑퐻/푑푡 is obtained by comparing 푑퐻measurements/푑푡 is obtained at crossover by comparing points, measurements where an earlier at crossover satellite ground points, where an earlier satellite ground track intersects a later one [92]. Twotrack criticalintersects factors a later affect one the [92] accuracy. Two critical of radar factors altimeter affect elevationthe accuracy of radar altimeter elevation retrievals over ice sheets: 1) variationretrievals in topographic over ice sheets: slope, 1) variationwhich controls in topographic the average slope, distance which controls the average distance between the altimeter and the measuredbetween ground the altimeter footprint and surface the measured [39,108,109] ground, and footprint 2) changes surface in the [39,108,109] , and 2) changes in the electrical permittivity of the iceelectrical sheet surface, permittivity which of con thetrols ice the sheet reflection, surface, transmission, which controls and the reflection, transmission, and absorption of the microwave signalabsorption [110–113] of. the microwave signal [110–113]. For typical radar altimeter frequenciesFor typical (Ku-band radar ~13 altimeter GHz), solidfrequencies ice and (Ku liquid-band water ~13 areGHz), highly solid ice and liquid water are highly reflective whereas dry snow andreflective firn are semi whereas-transparent, dry snow with and Ku firn-band are semipenetration-transparent, depths with on theKu -band penetration depths on the order of several meters [114,115].order Spatial of andseveral temporal meters variations [114,115] .in Spatial signal and penetration temporal depth variations caused in signal penetration depth caused by seasonal bare ice exposure,by snow seasonal and firn bare densification, ice exposure, surface snow and meltwater firn densification, presence, and surface meltwater presence, and refreezing of meltwater introducerefree spuriouszing of height meltwater change introduce signals that spurious are poorly height quantified change signals for the that are poorly quantified for the GrIS but are estimated to exceedGrIS >0.50 but m are in estimated some cases to exceed[116–118] >0.50. Variations m in some in surface cases [116 slope–118] . Variations in surface slope dominate radar altimeter accuracydominate in the topographically radar altimeter-complex accuracy bare in the-ice topographically ablation zone, with-complex slope bare- -ice ablation zone, with slope- induced errors that exceed >20 minduced for slopes errors >1 °that [39,109,113] exceed >20. The m recentfor slopes CryoSat >1° -[39,109,113]2 radar altimeter. The recent CryoSat-2 radar altimeter mission employs synthetic aperturemission radar employs (SAR) and synthetic interferometric aperture synthetic radar (SAR) aperture and interferometric radar (InSAR) synthetic aperture radar (InSAR) technologies which decrease the technologieseffective along which-track decrease footprint the diameter effective to along the order-track 0.1 footprint–0.3 km diameterand to the order 0.1–0.3 km and provide cross-track slope from InSARprovide proc crossessing-track [93] slope. Together from withInSAR denser proc essingorbital [93]-track. Together spacing withand denser orbital-track spacing and advancesRemote Sens.in waveform2019, 11, retracking, 2405 advances topographic in waveform mapping retracking, of ice sheet topographic ablation zone mapping areas is of accurate ice sheet ablation zone areas21 is of accurate 50 to within a few meters on average,to within and 푑퐻 a few/푑푡 metersto within on sub average,-meter and uncertainty 푑퐻/푑푡 to relative within to sub laser-meter uncertainty relative to laser altimetry [93,119,120]. altimetry [93,119,120]. 5.2. Optical Reflectance and Albedo Sensors and Datasets 3.1.1. Radar Altimetry Sensors and3.1.1. Datasets Radar Altimetry Sensors and Datasets Spatially- and temporally-consistent repeat measurements of surface reflectance, and estimates of The first satellite altimeter measurementsThe first satellite of surface altimeter topography measurements for the Greenland of surface Ice topography Sheet for the Greenland Ice Sheet wereBRDF made and by the albedo GEOS for-3, Geosat, thewere Greenland and made Seasat by oceanographic Icethe GEOS Sheet-3, are Geosat, Ku provided-band and altimeters Seasat by satellite-borneoceanographic (Error! Reference Ku imaging-band altimeters spectrometers, (Error! Reference sourceincluding not found. the) NASA [37,91,106,121,122] Moderatesource. Their Resolutionnot found. geographical) [37,91,106,121,122] Imaging coverage Spectroradiometer was. Their limited geographical to ±72o(MODIS), but coveragethe data the was NASA limited Multi-angleto ±72o but the data theyImaging collected SpectroRadiometer showed it was possiblethey collected (MISR),to calculate showed and volumetric the it was National possiblechanges Oceanictoin calculatethe polar and volumetricice Atmosphericsheets from changes in Administration’s the polar ice sheets from space, providing proof of concept for future polar-orbiting altimeters [92]. A first demonstration space,Advanced providing Very proof High of concept Resolution for future Radiometers polar-orbiting (AVHRR) altimeters (Table [92]. A2 )[ first261 demo–264nstration]. Other notable spaceborne using GEOS-3, Geosat, and Seasatusing data GEOSfound- 3,that Geosat, for areas and south Seasat of data <72 ofoundN, the that GrIS for thickened areas south by 23 of <72oN, the GrIS thickened by 23 ± 6sensors cm yr-1 between providing 1978–1986, surface with± 6 reflectancecmenhanced yr-1 between snowfall that 1978 in have– a1986, warmer been with climate enhanced used hypothesized to snowfall study thein to a warmerexplain GrIS ablation climate hypothesized zone include to explain thethe observed Enhanced thickening Thematic [123]. These Mapperthe observedearly missions Plus thickening (ETM supported+ [123]) onboard .important These early Landsat developments missions 7, thesupported in Operational waveform important Land developments Imager in (OLI) waveform retracking,onboard slope Landsat correction, 8, the and ASTERretracking, physical onboard and slope empirical corre Terra,ction, models the and High for physical subsurface Resolution and empirical volume imagers scattering models onboard for subsurface the Satellite volume Pour scattering [106,108,111,124]. These corrections are critical for accurate elevation retrieval and reliable change [106,108,111,124]l’Observation. These de lacorrections Terre (SPOT are critical 1-7), for the accu Globalrate elevation Land retrieval Imager and (GLI) reliable onboard change the Advanced Earth detection [125,126]. For example,detection Davis et[125,126] al. [127]. For used example, an improved Davis et geodetic al. [127] model used and an improved an geodetic model and an improvedObserving retracking Satellite model II to (ADEOS-2), reinterpretimproved the retracking the earlier Multispectral findings model to of reinterpret Zwally Imager et althe (MSI). [92] earlier, finding onboard findings a smaller of Sentinel-2, Zwally et al and. [92], the finding Ocean a smaller 1.5Colour ± 0.5 cm and yr-1 thickening Land Instrument for the1.5 period ± 0.5 (OLCI) cm 1978 yr–-1 onboard1986 thickening that highlighted Sentinel-3A for the period the/ Bimportance 1978 (Table–19862 )[ thatof265 slope highlighted–271- and]. the importance of slope- and signal penetration-correction. signal penetration-correction. Table 2. Summary of optical and near-infrared remote sensing platforms, instruments, temporal Table 1. Summary of radar and laserTable altimeter 1. Summary remote of sensing radar and platforms, laser altimeter instruments, remote temporal sensing platforms, instruments, temporal coverage, observed wavelengths, and managing agencies. coverage, observed wavelength, andcoverage, managing observed agencies. wavelength, and managing agencies. Temporal Observed Temporal Temporal ObservedŦ Observed §,Ɣ PlatformPlatform * * InstrumentInstrument Ŧ Platform * Instrument Agency §,Ɣ Agency Coverage Coverage Wavelength CoverageWavelength Wavelength Agency Radar Altimeters Radar Altimeters Nimbus 1-4 AVCS/IDCS 1964–1980 (1 band) 0.45–0.65 µm NASA Landsat 1-5 MSS 1972–2013 (4 bands): 0.55–0.955 µm NASA/USGS GOES 1-15 VISSR 1975–present (5 bands): 0.62–12 µm NASA/NOAA Seasat VIRR 1978 (110 days) (2 bands): 0.49-0.94, 10.5–12.5 µm NASA TIROS-N, AVHRR 1978–2001 (4 bands): 0.63–11 µm NOAA/NASA NOAA-6,8,10 NOAA-7,9,11-14 AVHRR/2 1981–present (5 bands): 0.63–12 µm NOAA Landsat 4-5 TM 1982–2013 (7 bands): 0.485–11.45 µm NASA Spot 1-3 HRV 1985–2009 (3 bands): 0.55–0.83 µm CNES/Spot Image ERS-1 ATSR 1991–2000 (4 bands): 1.6–12.0 µm ESA JERS-1 OPS 1992–1998 (8 bands): 0.52–2.40 µm JAXA ERS-2 ATSR-2 1995–2011 (7 bands): 0.55–12.0 µm ESA ADEOS POLDER 1996–1997 (9 bands): 0.44–0.91 µm CNES ADEOS AVNIR 1996–1997 (4 bands): 0.42–0.89 µm JAXA Spot 4 HRVIR 1998–2013 (4 bands): 0.55–1.63 µm CNES/Spot Image NOAA-15-19, AVHRR/3 1998–present (6 bands): 0.63–12.0 µm NOAA/EMSO MetOp A-C Landsat 7 ETM+ 1999–present (7 bands): 0.49–11.45 µm NASA/USGS Ikonos-2 OSA 1999–present (4 bands): 0.45–0.86 µm DigitalGlobe EO-1 Hyperion 2000–2017 (242 channels): 0.35–2.5 µm NASA/USGS EO-1 ALI 2000–2017 (9 bands): 0.44–2.2 µm NASA/USGS Aqua, Terra MODIS 2000–present (36 bands): 0.412–14.2 µm NASA Terra MISR 2000–present (4 bands) ¥: 0.446–0.867 µm NASA Terra ASTER 2000–present (14 bands): 0.56–11.3 µm NASA QuickBird-2 BGIS2000 2001–2015 (4 bands): 0.45–0.9 µm DigitalGlobe ADEOS-2 POLDER 2002–2003 (9 bands): 0.44–0.91 µm CNES Remote Sens. 2019, 11, x FOR PEER REVIEW 6 of 52 µ ADEOS-2Remote Sens. 2019, 11, x FOR GLI PEER REVIEW 2002–2003 (36 bands): 0.38–11.95 m 6 of 52 JAXA EnviSat MERIS 2002–2012 (15 bands): 0.39–1.04 µm ESA EnviSatGEOS-3 AATSRALT 1975 2002–2012–1978 13.9 GHz (Ku) (7 bands): 0.55–12.0NASA µm ESA GEOS-3 ALT 1975–1978 13.9 GHz (Ku) NASA Spot 5Seasat HRG/HRSALT 1978 (110 2002–2015 days) 13.6 GHz (Ku) (4 bands): 0.55–1.63NASA µm CNES/Spot Image Seasat ALT 1978 (110 days) 13.6 GHz (Ku) NASA Geosat GRA 1985–1990 13.5 GHz (Ku) DoD/NASAµ ALOSGeosat AVNIR-2 GRA 1985 2006–2011–1990 13.5 GHz (Ku) (4 bands): 0.46–0.82DoD/NASAm JAXA ERS-1 RA 1991–2006 13.8 GHz (Ku) ESA µ WorldView-1ERS-1WV60 RA 2007–present1991–2006 13.8 GHz (Ku) (1 band): 0.45–0.90ESA m DigitalGlobe ERS-2 RA-2 1995–2011 13.6 GHz (Ku), 3.2 GHz (S) ESA GeoEye-1ERS-2 GISRA-2 2008–present1995–2011 13.6 GHz (Ku), 3.2 (4 GHz bands): (S) 0.45–9.2ESA µm DigitalGlobe GFO GFO-RA 1998–2008 13.5 GHz (Ku) DoD/NASA WorldView-2GFOWV110 GFO-RA 2009–present1998–2008 13.5 GHz (Ku) (8 bands): 0.40–1.04DoD/NASAµm DigitalGlobe Suomi NPPEnviSatEnviSat VIIRS RA-2 RA-2 2002 2011–present2002–2012–2012 13.613. GHz6 GHz (Ku), (Ku), 3.2 3.2 (22GHz GHz bands): (S) (S) 0.41–12.01ESAESA µm NASA Spot 6CryoSatCryoSat-2 NAOMI-2 SIRALSIRAL 2010 2012–present2010–present–present 13.913.9 GHz GHz (Ku) (Ku) (4 bands): 0.48–0.82ESAESA µm CNES/Spot Image Landsat 8SARALSARAL OLIALtiKaALtiKa 2013 2013–present2013–present–present 36 GHz36 GHz (Ka) (Ka) (9 bands):ISRO/CNES 0.44–2.19ISRO/CNESµm USGS/NASA Landsat 8SentinelSentinel-3A/B -3A/B TIRS SRALSRAL 2016 2013–present2016–present–present 13.613.6 GHz GHz (Ku), (Ku), 5.4 5.4 GHz (2 GHz bands): (C) (C) 10.9,ESAESA 12 µm USGS/NASA Spot 7Laser altimetersLaser altimeters NAOMI 2014–present (4 bands): 0.48–0.82 µm CNES/Spot Image WorldView-3ICESatICESat WV-3 ImagerGLASGLAS 2003 2014–present2003–2009–2009 1.0641.064 μm, μm, 0.532 0.532 (29 μm bands):μm 0.40–2.365NASANASA µm DigitalGlobe Sentinel-2A/BICESatICESat-2 -2 MSI ATLASATLAS 2018 2015–present2018–present–present 1.0641.064 μm, μm, 0.532 0.532 (13 μm bands):μm 0.444–2.202NASANASA µm ESA GeoEye-2 AircraftAircraftSpaceView TMATM110ATM 19772016–present1977–present–present 1.0641.064 μm μm (4 bands): 0.45–9.2NASANASA µ m DigitalGlobe Sentinel-3A/BAircraftAircraft OLCI LVIS LVIS 1998 2016–present1998–present–present 1.0641.064 μm μm (21 bands): 0.40–1.02NASANASA µm ESA GCOM-C1AircraftAircraft SGLI MABELMABEL 2012 2017–present2012–2014–2014 1.0641.064 μm, μm, 0.532 0.532 (19 μm μm bands): 0.38–12.0NASANASA µm JAXA *,Ŧ,§ See Abbreviations at end of article for expanded acronyms. Ɣ Managing agencies are identified *,*,Ŧ,§ See Abbreviations at at end end of of article article for for expanded expanded acronyms. acronyms. Ɣ ManagingManaging agencies agencies are areidentified identified by the WMO by the WMO OSCAR database (Table A1) and may not reflect joint collaborations.¥ OSCARby the WMO database OSCAR (Table database A1) and (Table may A1) not and reflect may jointnot reflect collaborations. joint collaborations.Observed fore and aft of nadir @ 26◦, 45◦, 60◦, 70.5◦. The ESAThe ESAEuropean European Remote Remote-sensing-sensing Satellite Satellite (ERS (ERS-1)- 1) carried carried the the first first polar polar-orbiting-orbiting rada radar altimeter, extending coverage of the polar ice sheets to ±81.5o [95]. Together with the follow-on ERS- altimeter, extending coverage of the polar ice sheets to ±81.5o [95]. Together with the follow-on ERS- 2 and EnviSat missions, these satellites provided near complete GrIS coverage and 21 years of 2 and EnviSat missions, these satellites provided near complete GrIS coverage and 21 years of continuous data acquisition [38]. The first digital elevation model (DEM) spanning the entire GrIS continuous data acquisition [38]. The first digital elevation model (DEM) spanning the entire GrIS surface was produced from combined ERS-1 and Geosat altimetry data, supplemented by stereo- surface was produced from combined ERS-1 and Geosat altimetry data, supplemented by stereo- photogrammetric and cartographic data sets, with pan-GrIS average accuracy -0.33 ± 6.97 m [128,129]. photogrammetricERS-1/2 altimetry and cartographic data were used data to discriminatesets, with pan thinning-GrIS average rates in theaccuracy ablation -0.33 zone ± 6.97(defined m [128,129] as <1500. ERS-1/2m altimetrya.s.l.) of -2.0 data ± 0.9were cm used yr-1 fromto discriminate accumulation thinning zone (>1500 rates in m the a.s.l.) ablation thickening zone of(defined 6.4 ± 0.2 as cm <1500 yr-1 m a.s.l.)with of an-2.0 overall ± 0.9 cm net yr thickening-1 from accumulation of 5.4 ± 0.2 cmzone yr -1(>1500 for the m period a.s.l.) thickening 1992–2003 [130]of 6.4. To ± 0.2 extend cm yr the-1 with antemporal overall coverage net thickening and increase of 5.4 spatial ± 0.2 measurement cm yr-1 for density, the period Khvorostovsky 1992–2003 and[130] Jo. hannessen To extend [131] the temporaldeveloped coverage an and algorithm increase to spatial merge measurement ERS-1/2 and EnviSat density, datasets Khvorostovsky and detect and and Johannessen eliminate inter[131]- developedsatellite an biases. algorithm For the to period merge 1992 ERS–-1/22008, and their EnviSat dataset datasets suggests andnet thinning detect and of the elim GrISinate initiated inter- satellitearound biases. 2000, For and the accelerated period 1992 between–2008, 2006 their–2008, dataset with suggests a thickening net ofthinning 4.0 ± 0.2 of cm the yr -1GrIS above initiated 1500 m arounda.s.l. 2000, and and thinning accelerated of -7.0 between± 1.0 cm yr 2006-1 below–2008, 1500 with m a.s.l.a thickening [94]. of 4.0 ± 0.2 cm yr-1 above 1500 m a.s.l. and thinningConventional of -7.0 radar ± 1.0 altimeters cm yr-1 below such 1500as the m instruments a.s.l. [94]. included on ERS-1/2 and EnviSat were Conventionaloptimized for radardetecting altimeters elevation such over as open the instrumentsoceans and low included-gradient on polar ERS- ice1/2 sheetand EnviSatinteriors werewith o optimizedmean for accuracies detecting typically elevation <0.2 over m for open ice sheetoceans surface and lowslopes-gradient <0.2 [39,109] polar .ice Elevation sheet interiors accuracy with and 푑퐻/푑푡 mean accuracies uncertainty typically is much<0.2 m higher for ice for sheet the topographically surface slopes complex<0.2o [39,109] ablation. Elevation zone due accuracy to their large and spatial measurement footprint and wide orbital track spacing, with slope errors >20 m for slopes >1o 푑퐻/푑푡 uncertainty is much higher for the topographically complex ablation zone due to their large and systematic biases over rough surfaces [39,95,109,132]. Careful processing has been used to spatial measurement footprint and wide orbital track spacing, with slope errors >20 m for slopes >1o discriminate ablation zone thinning rates using ERS-1/2 and EnviSat altimetry data, including repeat- and systematic biases over rough surfaces [39,95,109,132]. Careful processing has been used to track analysis, but the accuracy of these data and their process-based interpretation is ambiguous discriminate ablation zone thinning rates using ERS-1/2 and EnviSat altimetry data, including repeat- owing to the aforementioned slope errors and sparse spatial measurement density in the ablation track analysis,zone [118,133 but –the135] accuracy. of these data and their process-based interpretation is ambiguous owing to theThe aforementioned CryoSat-2 satellite slope (launched errors April and 2010)sparse provides spatial continuitymeasurement with ERSdensity-1/2 andin the EnviSat ablation and zone [118,133employs– 135] SAR. delay-Doppler along-track processing and interferometric cross-track processing The[120,136] CryoSat. CryoSat-2 satellite-2 is the(launched first ESA April radar 2010) altimeter provides dedicated continuity to polar with studies, ERS- with1/2 and orbital EnviSat coverage and employsto ±88 SARo and delay 1.6 -kmDoppler cross- track along spacing-track at processing 70o. CryoSat and-2 carries interferometric two Ku-band cross SAR/Interferometric-track processing [120,136]Radar. CryoSat Altimeter-2 is (SIRAL) the first systems. ESA radar SIRAL altimeter operates dedicated in three distinctto polar modes studies, depending with orbital on location: coverage 1) to ±88lowo and resolution 1.6 km modecross- (LRMtrack )spacing over open at ocean70o. CryoSat and ice sheet-2 carries interiors, two 2)Ku synthetic-band SAR/Interferometric aperture radar mode Radar(SAR) Altimeter over (SIRAL) sea ice andsystems. coastal SIRAL regions, operates and 3) in SAR three interferometric distinct modes mode depending (SARIn) on over location: ice sheet 1) low resolutionmargins. mode The SARIn (LRM ) modeover open employs ocean two and SAR ice sheet instruments interiors, oriented 2) synthetic across aperture the satellite radar mode track. (SAR)Interferometric over sea ice and phase coastal processing regions, of the and dual 3) waveforms SAR interferometric provides cross mode-track (SARIn) slope at over~0.3 km ice along sheet- margins.track The resolution SARIn [137,138] mode . employs two SAR instruments oriented across the satellite track. o InterferometricOver mildphase slopes processing (<1 ), phase of the unwrapping dual waveforms [136] of provides CryoSat -cross2 interferometric-track slope data at ~0.3 provides km along 5 km- wide-swath elevation retrievals with two orders of magnitude more individual measurements per track resolution [137,138]. surface area than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs Over mild slopes (<1o), phase unwrapping [136] of CryoSat-2 interferometric data provides 5 km wide-swath elevation retrievals with two orders of magnitude more individual measurements per surface area than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs

Remote Sens. 2019, 11, 2405 22 of 50

The NOAA Extended AVHRR Polar Pathfinder (App-X) product provides surface broadband all-sky albedo and a suite of surface radiative flux and cloud property variables on a 25 km equal area grid twice-daily for the period 1982 to present for the Arctic and Antarctic regions [272]. The AVHRR App-X product provides a continuous record of GrIS surface albedo and an important climatological dataset [253]. An early demonstration of AVHRR surface albedo retrieval revealed a region of anomalously dark ice in the southwestern GrIS ablation zone later termed the “Dark Zone” (Section 5.3)[273]. Although rigorous quality control procedures are applied during dataset production [49], its utility for investigations of GrIS ablation zone albedo may be limited by geolocation errors, four-channel spectral resolution, and inter-mission biases [274,275]. Spatial and spectral resolution are both improved with MODIS, which measures surface reflectance in 30 narrow spectral bands at 250–500 m spatial resolution. Higher spectral and/or spatial resolution reflectance products are provided by Landsat, MISR, SPOT, Sentinel, and other spaceborne spectrometers (Table2). Various albedo estimations exist for most sensors [276]. Many rely on the MODIS BRDF or empirical conversions from narrowband reflectance to broadband albedo, and for others (e.g., MISR), a unique BRDF and physically-based albedo estimation exists [261,277]. Most studies of the GrIS albedo have used MODIS albedo owing to the development of operational, publicly available data products spanning nearly two decades. Two MODIS albedo products are currently available: (1) the 500-m Terra/Aqua MODIS 8-day BRDF/Albedo product (MCD43) and, (2) the 500-m Terra daily snow/ice cover and BRDF/Albedo product (MOD10A) [251,263,278]. The MCD43 algorithm uses cloud-free, multi-date, atmospherically corrected input data from combined Terra (MOD09) and Aqua (MYD09) multi-angular surface reflectance retrievals to estimate a unique BRDF for each MODIS pixel in the seven MODIS ‘Land’ bands (1–7) every eight days. Albedo is calculated using this BRDF estimation for each of bands 1–7, and also for three broadbands (0.4–0.7, 0.7–3.0, and 0.4–3.0 um) [251,278]. The MOD10A algorithm is a daily product that uses the discrete-ordinate radiative transfer (DISORT) model to convert MOD09 daily surface reflectance to spectral albedo [279]. The daily frequency of the MOD10A product makes it particularly useful for studies of the GrIS ablation zone, where day-to-day albedo variability is substantial [26,280]. Both the MCD43 and MOD10A retrieval algorithms require cloud-free images, as identified by the MOD35 cloud mask. For the MCD43 product, data gaps within the 16-day period window may result in under-sampling of the BRDF, requiring the use of a backup BRDF algorithm. Albedo values estimated with the backup algorithm are generally considered less reliable, and quality flags are provided to distinguish between them [281]. The accuracy of MODIS albedo over the GrIS has been evaluated by comparison with surface-based measurements of albedo at Greenland Climate Network (GC-Net) automatic weather stations (AWS) [282]. The MCD43 product has an average root mean squared difference (RMSD) of 0.07 ± and mean bias +0.022 relative to AWS albedo, which has 0.035 RMSD uncertainty relative to ± precision pyranometer measurements [278,283]. The MCD43 RMSD is reduced to 0.04 when only ± the highest-quality flagged values are used [283]. MODIS albedo accuracy is reduced at a high solar zenith angle (SZA) [263,283]. For example, Wang and Zender [284] found a large MCD43 albedo bias relative to GC-Net AWS albedo in the dry-snow zone for SZA > 55◦. However, these biases are mostly eliminated if only the highest-quality flagged values are used as recommended [281]. In general, the MODIS albedo products are not recommended for comparison with surface measurements for SZA > 70◦ [281,283]. Relative to three AWS stations in the southwestern GrIS ablation zone, the MOD10A albedo is lower by up to 0.10 and by 0.06 on average during July for the period 2009–2016 [280]. The low bias is attributed to the inadequate sampling of surface heterogeneity by the AWS, which are preferentially located on flat areas of bare ice rather than areas of lower albedo, such as meltwater channels or crevasses that influence the satellite albedo [280]. Using one day of high quality in situ spectral albedo measurements (N = 232) collected within two MODIS satellite pixel ground footprints in the southwest GrIS ablation zone, Moustafa et al. [285] showed that the MODIS Collection V006 daily blue-sky albedo Remote Sens. 2019, 11, 2405 23 of 50 was accurate to within 0.04 to +0.07 for the homogeneous surface of one pixel and was biased low by − 0.04 to 0.07 for the heterogeneous surface of the second pixel. As in Ryan et al. [280], the low bias − − was attributed to under sampling of dark meltwater channels and shadowed crevasses. In addition to spatial heterogeneity, comparisons between satellite albedo and surface measurements are complicated by factors including sensor spectral sensitivity and whether hemispherical or angular reflectance is measured [281,283]. Owing to these varied processes, ground validation of satellite albedo products for the GrIS ablation zone is challenging, and satellite retrieval uncertainty is enhanced relative to the interior snow-covered accumulation-zone surface, especially at sub-pixel scale [280,285,286].

5.3. Dark Ice in the Ablation Zone: Albedo Trends and Drivers Satellite albedo retrievals provide an important marker of the changing ice surface. The MOD43 data suggest the GrIS area-averaged albedo for June, July, and August (JJA) decreased by 0.044 − between 2000–2012 [278]. The JJA albedo anomaly during the record 2012 pan-GrIS melt event [287] was more than two standard deviations below the period mean. The most negative albedo trends are observed for the southern and western ice sheet ablation area where July albedo trends are 0.12 to − 0.24 per decade, attributed to both reduced seasonal snow cover duration and increased bare ice − exposure during this period [29,30,278]. The MOD10A product indicates a similar JJA albedo trend of 0.083 averaged over the GrIS ablation zone for the period 2000–2010 [30] and 0.078 for the period − − 2000–2013 [288]. Although MOD10A reflectance data suggest a modest brightening of the ablation zone since 2013 that likely reflects anomalously cold and snowier spring and summer conditions during this period [26,51], there is strong potential for melt-albedo feedbacks to accelerate negative SMB in the coming years [30,32]. For the ablation zone, spatial and temporal albedo variability is driven foremost by the seasonal snowline position and its control on bare ice exposure (Figure9)[ 26]. Where snow is present, snow grain growth is the strongest modifier of surface albedo [252,253]. Bare ice albedo is scale dependent but is principally controlled by air bubble size and shape distribution, shadowing by cracks and crevasses [285,289], and by ice grain metamorphism, surface meltwater presence, and exposure and deposition of mineral and biological light-absorbing impurities (LAI) [51,52,58,253,290–292]. The darkening effect of LAI is particularly apparent in satellite images of the GrIS ablation zone that reveal foliated bands of “dark ice” with anomalously low albedo relative to surrounding ice (Figure 10)[52,273,293,294]. The wavy appearance of these foliated bands and their static position indicate their provenance as outcropping ice layers rich in dust deposited during past millennia [52,294–296]. The spatial extent of bare ice and albedo variability within the bare ice zone exerts a primary control on ice sheet albedo and surface meltwater production in the ablation zone [26,30]. Shimada et al. [297] used MODIS/Terra surface reflectance to quantify the regional distribution of bare ice and dark ice extent during July for the period 2000–2014. The spatial extent of bare ice ranged from 5–16% 1 of the GrIS surface area, dark ice ranged from 4–10%, and both exhibited a positive trend (4.4% yr− 1 for bare ice and 7.6% yr− for dark ice) with the greatest expansion in the southwest ablation zone. Bare ice extent was strongly correlated (r = 0.66) with air temperature, whereas dark ice extent was weakly correlated with air temperature and was negatively correlated with solar radiation, suggesting that bare ice weathering by solar radiation may reduce dark ice extent and increase surface albedo in the GrIS ablation zone (Figure 11). Recent work highlights the darkening effect of biological LAI on bare ice albedo, including assemblages of biologically-active dust termed “cryoconite” that melt quasi-cylindrical holes into the ice (Figure 11)[298] and distributed communities of algae and cyanobacteria that inhabit the ice surface [299]. For example, Tedstone et al. [51] applied the bare ice and dark ice detection algorithm from Shimada et al. [297] to MODIS surface reflectance imagery of the southwest ablation zone during June–July–August for the period 2000–2016. In contrast to Shimada et al. [297], they conclude that distributed algae blooms, rather than bare ice weathering and cryoconite hole growth, likely explains Remote Sens. 2019, 11, 2405 24 of 50 dark ice dynamics in their study region, owing to the synchronous and abrupt timing of dark ice exposure across the study area and the progressive rather than episodic increase in dark ice extent from June to August. In general, the deepening of cryoconite holes into the ice surface appears to limit their regional effect on albedo, especially at a low sun angle (Figure 11d), whereas distributed LAI that accumulate on the ice surface are considered stronger agents of bare ice darkening in the GrIS ablation zone [58,300]. Understanding seasonal and interannual relationships between bare ice structure, mineral and biological LAI, and bare ice albedo driven by cryoconite hole deepening and removal, weathering crust growth and decay, rotting and removal of superimposed ice, and variations in ice grain and air bubble geometry remain important areas of research that currently elude confident detection by satellite remote sensing [51,85,297,301,302]. Remote Sens. 2019, 11, x FOR PEER REVIEW 25 of 52

Figure 9. The average end-of-summer (maximum) snowline position for the southwest sector of the Figure 9. The average end-of-summer (maximum) snowline position for the southwest sector of the Greenland Ice Sheet during the period 2001–2017 as determined from MODIS MOD09GA surface Greenland Ice Sheet during the period 2001–2017 as determined from MODIS MOD09GA surface reflectance, reprinted with permission from Ryan et al. [26] (courtesy Johnathan Ryan, Brown University). reflectance, reprinted with permission from Ryan et al. [26] (courtesy Johnathan Ryan, Brown Daily reflectance maps for June, July, and August were classified into bare ice, snow-covered, and University). Daily reflectance maps for June, July, and August were classified into bare ice, snow- water-covered pixels using supervised random forest classification. The bare ice presence index is an covered, and water-covered pixels using supervised random forest classification. The bare ice exposure frequency representing the fraction of total days classified as bare ice for each pixel. The presence index is an exposure frequency representing the fraction of total days classified as bare ice average end-of-summer snowline elevation is 1520 113 m in this sector, with interannual variation of for each pixel. The average end-of-summer snowline± elevation is 1520 ± 113 m in this sector, with 385 m. Interannual snowline variability explains 53% of MOD10A albedo variability [26]. interannual± variation of ±385 m. Interannual snowline variability explains 53% of MOD10A albedo variability [26].

The spatial extent of bare ice and albedo variability within the bare ice zone exerts a primary control on ice sheet albedo and surface meltwater production in the ablation zone [26,30]. Shimada et al. [297] used MODIS/Terra surface reflectance to quantify the regional distribution of bare ice and dark ice extent during July for the period 2000–2014. The spatial extent of bare ice ranged from 5– 16% of the GrIS surface area, dark ice ranged from 4–10%, and both exhibited a positive trend (4.4% yr−1 for bare ice and 7.6% yr−1 for dark ice) with the greatest expansion in the southwest ablation zone. Bare ice extent was strongly correlated (푟 = 0.66) with air temperature, whereas dark ice extent was weakly correlated with air temperature and was negatively correlated with solar radiation, suggesting that bare ice weathering by solar radiation may reduce dark ice extent and increase surface albedo in the GrIS ablation zone (Figure 11).

Remote Sens. 2019, 11, 2405 25 of 50

Remote Sens. 2019, 11, x FOR PEER REVIEW 26 of 52

Figure 10.FigureSatellite 10. Satellite images images of the of ablationthe ablation zone zone proximalproximal to toInglefield Inglefield LandLand in northwest in northwest Greenland Greenland (78.64°N, 65.89°W), showing bands of outcropping dust in bare ablating ice, supraglacial lakes and (78.64◦ N,rivers 65.89 indicating◦ W), showing melting bandsice, and ofwhat outcropping appears to be dust snow, in firn, bare or ablatingotherwise non ice,-melting supraglacial ice that lakes and rivers indicatingmay indicate melting the approximateice, and what location appears of the to summer be snow, snowline. firn, or (a otherwise) Landsat-8 non-meltingOperational Land ice that may indicate theImager approximate 30 m resolution location image (RGB: of the bands summer 4 (red), snowline. 3 (green), and (a )2 Landsat-8(blue)) acquired Operational on 16 July 2016. Land Imager 30 m resolutionElevations image in this (RGB: image bands range from 4 (red), 600 m 3 (green),a.s.l. at the and ice sheet 2 (blue)) edge to acquired ~1500 m a.s.l.; on 16 (b July) same 2016. as (a),Elevations but for detail box; (c) MODIS/Terra MOD09A1 500 m resolution 8-day composite image (RGB: bands in this image range from 600 m a.s.l. at the ice sheet edge to ~1500 m a.s.l.; (b) same as (a), but for detail 1 (red), 4 (green), and 3 (blue)) acquired on 27 July–03 August 2016; (d) Sentinel-2 Multispectral box; (c) MODISImager/Terra 10 m resolution MOD09A1 image 500 (RGB: m resolution bands 4 (red), 8-day 3 (green), composite and 2 (blue)) image acquired (RGB: on bands 23 July 1 (red),2016. 4 (green), and 3 (blue))(e) WorldViewacquired- on2 image 27 July–03 © 2019 August DigitalGlobe, 2016; Inc (d. ) (RGB: Sentinel-2 bands Multispectral5 (red), 3 (green), Imager and 2 10(blue)) m resolution image (RGB:resampled bands to 4 1.8 (red), m resolution 3 (green), (native and resolution: 2 (blue)) 0.5 acquired m) acquired on 03 23 September July 2016. 2019. (e Images) WorldView-2 in (b)– image (e) demonstrate the range of spatial resolutions typically used for studying the ablation zone. © 2019 DigitalGlobe, Inc. (RGB: bands 5 (red), 3 (green), and 2 (blue)) resampled to 1.8 m resolution (native resolution:Recent work 0.5 m) highlight acquireds the 03 darkening September effect 2019. of Images biological in (LAIb–e ) on demonstrate bare ice albedo, the range including of spatial resolutionsassemblages typically of biologically used for studying-active dust the termed ablation “cryoconite” zone. that melt quasi-cylindrical holes into the Remoteice (Figure Sens. 2019 11, )11 [,298 x FOR] and PEER distributed REVIEW communities of algae and cyanobacteria that inhabit 27 the of ice52 surface [299]. For example, Tedstone et al. [51] applied the bare ice and dark ice detection algorithm from Shimada et al. [297] to MODIS surface reflectance imagery of the southwest ablation zone during June–July–August for the period 2000–2016. In contrast to Shimada et al. [297], they conclude that distributed algae blooms, rather than bare ice weathering and cryoconite hole growth, likely explains dark ice dynamics in their study region, owing to the synchronous and abrupt timing of dark ice exposure across the study area and the progressive rather than episodic increase in dark ice extent from June to August. In general, the deepening of cryoconite holes into the ice surface appears to limit their regional effect on albedo, especially at a low sun angle (Figure 11d), whereas distributed LAI that accumulate on the ice surface are considered stronger agents of bare ice darkening in the GrIS ablation zone [58,300]. Understanding seasonal and interannual relationships between bare ice structure, mineral and biological LAI, and bare ice albedo driven by cryoconite hole deepening and removal, weathering crust growth and decay, rotting and removal of superimposed ice, and variations in ice grain and air bubble geometry remain important areas of research that currently elude confident detection by satellite remote sensing [51,85,297,301,302].

Figure 11. Images of cryoconite hole-studded ice surface in the western Greenland Ice Sheet ablation Figure 11.zone,Images collected of cryoconite at (a) ~850 m hole-studded a.s.l.; (b) ~950 m ice a.s.l. surface; and (c) in~1200 the m western a.s.l. along Greenland an elevation Ice transect Sheet ablation zone, collectedoutside at Kangerlussuaq (a) ~850 m a.s.l.;(Søndre (b Strømfjord).) ~950 m a.s.l.;At the and low (andc) ~1200 mid-elevation m a.s.l. sites, along the icean surface elevation is transect outside Kangerlussuaqglazed and smooth, (Søndre and the water Strømfjord). table within At the the cryoconite low and holes mid-elevation is nearly coincident sites, with the the iceice surface is glazed andsurf smooth,ace. At the and high the-elevation water site table (c), the within surface the is rougher, cryoconite and evidence holes is of nearly nocturnal coincident refreezing withis the ice visible. (d) Same location as (c), showing the rough, weathered ice surface at low sun angle, reprinted c surface. Atfrom the Cooper high-elevation et al. [85]. Quadrat site ( shown), the in surface (b) is 3 m is wide. rougher, Cryoconite and evidenceholes in (c) ofare nocturnalon the orderrefreezing of is visible. (d1)– Same5 cm wide. location Seasonal as (c weathering), showing of the the rough, ice surface weathered, including ice cryoconite surface hole at low deepening sun angle, and reprinted from Cooperremoval, et al. exerts [85]. a Quadratprimary control shown on inice (roughnessb) is 3 m and wide. grain Cryoconite morphology holesbut its ineffect (c) on are bare on ice the order of 1–5 cm wide.albedo Seasonal has received weathering little direct of study the [29 ice7 surface,]. including cryoconite hole deepening and removal,

exerts5.4. a primary Current Challenges control on and ice Future roughness Opportunities and grain morphology but its effect on bare ice albedo has received little direct study [297]. Contemporary research on GrIS ablation zone surface reflectance and albedo is focused on understanding and diagnosing agents of albedo change [253]. Important questions remaining unresolved include whether observed reductions in GrIS snow albedo are caused by enhanced snow grain metamorphism due to a warming climate [252], or by deposition of LAI such as dust and black carbon from industrial emissions and/or forest fire [292,303]. These questions cannot currently be answered owing to the spectral and radiometric limitations of existing spaceborne optical sensors and inadequate surface validation measurements [280,304]. For example, it is unlikely that the albedo-reducing effect of industrial black carbon emissions on relatively clean accumulation zone snow can be detected from space [305,306]. Similarly, it is unknown if bare ice albedo reduction due to deposition or emergence of inorganic mineral LAI can be distinguished using satellite imagery from albedo reduction due to biological LAI [292,301]. To date, efforts to remotely sense changing bare ice albedo have focused on quasi-physical proxies such as “darkening” in broad spectral bands [51,58]. Hyperspectral reflectance is used to map the albedo-reducing effect of LAI, including snow algae and dust in mountain environments and valley glaciers elsewhere [307,308]. However, ice algae are taxonomically distinct from snow algae [291], and optical methods developed by the snow albedo community need to be adapted to detect ice algae or separate its effect on ice albedo from inorganic impurities or variations in ice grain size and structure [292,301]. At present, no spaceborne hyperspectral imager operates over the polar regions, although several hyperspectral missions are planned for the coming years [309]. Recently, multispectral reflectance imagery from the OLCI onboard Sentinel-3A was used to infer the spatiotemporal distribution of ice algal blooms in the southwest GrIS ablation zone, providing insight into the capability of enhanced spectral resolution for diagnosing albedo change in the polar regions

Remote Sens. 2019, 11, 2405 26 of 50

5.4. Current Challenges and Future Opportunities Contemporary research on GrIS ablation zone surface reflectance and albedo is focused on understanding and diagnosing agents of albedo change [253]. Important questions remaining unresolved include whether observed reductions in GrIS snow albedo are caused by enhanced snow grain metamorphism due to a warming climate [252], or by deposition of LAI such as dust and black carbon from industrial emissions and/or forest fire [292,303]. These questions cannot currently be answered owing to the spectral and radiometric limitations of existing spaceborne optical sensors and inadequate surface validation measurements [280,304]. For example, it is unlikely that the albedo-reducing effect of industrial black carbon emissions on relatively clean accumulation zone snow can be detected from space [305,306]. Similarly, it is unknown if bare ice albedo reduction due to deposition or emergence of inorganic mineral LAI can be distinguished using satellite imagery from albedo reduction due to biological LAI [292,301]. To date, efforts to remotely sense changing bare ice albedo have focused on quasi-physical proxies such as “darkening” in broad spectral bands [51,58]. Hyperspectral reflectance is used to map the albedo-reducing effect of LAI, including snow algae and dust in mountain environments and valley glaciers elsewhere [307,308]. However, ice algae are taxonomically distinct from snow algae [291], and optical methods developed by the snow albedo community need to be adapted to detect ice algae or separate its effect on ice albedo from inorganic impurities or variations in ice grain size and structure [292,301]. At present, no spaceborne hyperspectral imager operates over the polar regions, although several hyperspectral missions are planned for the coming years [309]. Recently, multispectral reflectance imagery from the OLCI onboard Sentinel-3A was used to infer the spatiotemporal distribution of ice algal blooms in the southwest GrIS ablation zone, providing insight into the capability of enhanced spectral resolution for diagnosing albedo change in the polar regions [269]. While this represents an important first step toward bioalbedo detection from space, additional ground validation of cellular optical properties specific to the microbial communities on the GrIS surface is needed to discriminate the individual drivers of bare ice albedo reduction using satellite remote sensing [301,310].

6. Mapping Surface Melt and Glaciological Zones Mapping of surface zones or facies, such as dry snow, wet snow, and bare ice, was pioneered by Benson [35], and later mapped across the GrIS using both ERS SAR imagery [50] and Seasat-A scatterometer backscatter [311]. In contrast to the climatological facies of Benson [35], the concept of radar glacier zones was introduced to distinguish dynamic zones driven by the seasonal cycle of surface melt onset, surface roughening, and snowline migration, thereby revealing the seasonal extent of the bare ice ablation zone [312–314]. Multi-angular optical reflectance, together with derived surface roughness, provides an independent method to map glacier zones on the GrIS with particular relevance to ablation zone studies owing to the unique ability of angular reflectance to detect crevasse fields and the lower limit of superimposed ice [53]. These classification approaches are complemented by an extensive legacy of surface melt detection studies using active microwave backscatter, and thermal and passive microwave brightness temperature, which together provide a comprehensive view of the ablation zone surface and changes in ice sheet melt extent with time [55,315–319].

6.1. Active Microwave Detection of Surface Melt and Glacier Zones SAR imagers measure radar backscatter (image brightness) at (typically) C-band (5 GHz) and Ku-band (14 GHz) frequencies (Table3). Backscatter strength, or the normalized radar cross-section (σ◦), is principally controlled by the complex electrical permittivity of the ice surface and by geometric properties, including ice surface roughness, grain size, incidence angle, and ice thermal and physical structure [50,319]. Variations in σ◦ associated with these controls are used to map zones of unique snow and ice composition on glacier and ice sheet surfaces [320]. An early demonstration of the method used Remote Sens. 2019, 11, x FOR PEER REVIEW 5 of 52

the accumulation zone, regional climate models and firn compaction models are used to partition the component of 푑퐻/푑푡 caused by change in 푆푀퐵 from those caused by change in firn density and 퐷 [104,105]. The following subsections summarize spaceborne radar altimetry of the GrIS, providing an historical perspective that emphasizes major challenges and advances in the field, and identifies future opportunities for radar altimetry observations of the GrIS ablation zone.

3.1. Radar Altimetry Conventional radar altimeters are nadir pointing single-beam radars that transmit and receive tens to thousands of microwave pulses per second, yielding ground measurement footprint diameters on the order 1–10 km posted at 0.3–0.6 km along-track spacing [39,106,107]. An estimate of 푑퐻/푑푡 is obtained by comparing measurements at crossover points, where an earlier satellite ground track intersects a later one [92]. Two critical factors affect the accuracy of radar altimeter elevation retrievals over ice sheets: 1) variation in topographic slope, which controls the average distance between the altimeter and the measured ground footprint surface [39,108,109], and 2) changes in the electrical permittivity of the ice sheet surface, which controls the reflection, transmission, and absorption of the microwave signal [110–113]. For typical radar altimeter frequencies (Ku-band ~13 GHz), solid ice and liquid water are highly reflective whereas dry snow and firn are semi-transparent, with Ku-band penetration depths on the order of several meters [114,115]. Spatial and temporal variations in signal penetration depth caused by seasonal bare ice exposure, snow and firn densification, surface meltwater presence, and refreezing of meltwater introduce spurious height change signals that are poorly quantified for the GrIS but are estimated to exceed >0.50 m in some cases [116–118]. Variations in surface slope dominate radar altimeter accuracy in the topographically-complex bare-ice ablation zone, with slope- induced errors that exceed >20 m for slopes >1° [39,109,113]. The recent CryoSat-2 radar altimeter mission employs synthetic aperture radar (SAR) and interferometric synthetic aperture radar (InSAR) technologies which decrease the effective along-track footprint diameter to the order 0.1–0.3 km and provide cross-track slope from InSAR processing [93]. Together with denser orbital-track spacing and advances in waveform retracking, topographic mapping of ice sheet ablation zone areas is accurate to within a few meters on average, and 푑퐻/푑푡 to within sub-meter uncertainty relative to laser altimetry [93,119,120].

3.1.1. Radar Altimetry Sensors and Datasets The first satellite altimeter measurements of surface topography for the Greenland Ice Sheet were made by the GEOS-3, Geosat, and Seasat oceanographic Ku-band altimeters (Error! Reference source not found.) [37,91,106,121,122]. Their geographical coverage was limited to ±72o but the data Remote Sens. 2019, 11, 2405 Remote Sens. 2019, 1127, xof FOR 50 PEER REVIEW 6 of 52 they collected showed it was possible to calculate volumetric changes in the polar ice sheets from

space, providing proof of concept for future polar-orbiting altimeters [92]. A first demonstration GEOS-3 ALT 1975–1978 13.9 GHz (Ku) NASA using GEOS-3, Geosat, and Seasat data found that for areas south of <72oN, the GrIS thickened by 23 Seasat ALT 1978 (110 days) 13.6 GHz (Ku) NASA variations± 6 cm inyr-1 C-band between SAR1978–1986, image with brightness enhanced snowfall to map in zones a warmer of dry climate snow, hypothesized percolation, to explain wet snow, andGeosat bare GRA 1985–1990 13.5 GHz (Ku) DoD/NASA ice, extendingthe observed the thickening early facies[123]. These work early of missions Benson supported [35] into important the satellite developments era [50 ].in waveform Cold, dry, fine-grainedERS-1 RA 1991–2006 13.8 GHz (Ku) ESA ERS-2 RA-2 1995–2011 13.6 GHz (Ku), 3.2 GHz (S) ESA snowretracking, in the accumulation slope correction, zone and appears physical dark and empirical in C-band models imagery, for subsurface whereas volume the percolation scattering zone andGFO wet GFO-RA 1998–2008 13.5 GHz (Ku) DoD/NASA snow[106,108,111,124] zone appear.bright, These corrections owing to are reflections critical for fromaccurate refrozen elevation ice retrieval lenses and in reliable the snow change/firn columnEnviSat [316 ]. RA-2 2002–2012 13.6 GHz (Ku), 3.2 GHz (S) ESA detection [125,126]. For example, Davis et al. [127] used an improved geodetic model and an CryoSat-2 SIRAL 2010–present 13.9 GHz (Ku) ESA The bareimproved ice zone retracking is distinguished model to reinterpret from thethe earlier wet snow findings zone of Zwally using et di alfferential. [92], finding brightness a smallerbetween SARAL winter ALtiKa 2013–present 36 GHz (Ka) ISRO/CNES Sentinel-3A/B SRAL 2016–present 13.6 GHz (Ku), 5.4 GHz (C) ESA and summer1.5 ± 0.5 cm images yr-1 thickening and the for bright the period reflections 1978–1986 of that rough, highlighted crevassed the importance surfaces of [50 slope,321- ].and Laser altimeters signal penetration-correction. ICESat GLAS 2003–2009 1.064 μm, 0.532 μm NASA Table 3. Summary of active microwave (synthetic aperture radar and microwave scatterometer) remoteICESat-2 ATLAS 2018–present 1.064 μm, 0.532 μm NASA Table 1. Summary of radar and laser altimeter remote sensing platforms, instruments, temporal Aircraft ATM 1977–present 1.064 μm NASA sensing platforms, instruments, temporal coverage, observed frequencies, and managing agencies. coverage, observed wavelength, and managing agencies. Aircraft LVIS 1998–present 1.064 μm NASA Aircraft MABEL 2012–2014 1.064 μm, 0.532 μm NASA TemporalTemporal Observed Observed PlatformPlatform * * InstrumentInstrument Ŧ Agency §,Ɣ Agency*,Ŧ,§ See Abbreviations at end of article for expanded acronyms. Ɣ Managing agencies are identified CoverageCoverage Wavelength Frequency by the WMO OSCAR database (Table A1) and may not reflect joint collaborations. SyntheticRadar Aperture Altimeters Radar The ESA European Remote-sensing Satellite (ERS-1) carried the first polar-orbiting radar o Seasat SAR 1978 (110 days) 1.275 GHz (L)altimeter, NASA extending coverage of the polar ice sheets to ±81.5 [95]. Together with the follow-on ERS- ERS-1 AMI-SAR 1991–2006 5.3 GHz (C)2 and ESA EnviSat missions, these satellites provided near complete GrIS coverage and 21 years of JERS-1 SAR 1992–1998 1.2 GHz (L)continuous JAXA data acquisition [38]. The first digital elevation model (DEM) spanning the entire GrIS ERS-2 AMI-SAR 1995–2011 5.3 GHz (C)surface ESA was produced from combined ERS-1 and Geosat altimetry data, supplemented by stereo- RADARSAT-1 SAR 1995–2013 5.3 GHz (C) CSA photogrammetric and cartographic data sets, with pan-GrIS average accuracy -0.33 ± 6.97 m [128,129]. Remote Sens.EnviSat 2019, 11, x FOR PEER REVIEW ASAR 2002–2012 3.2 (S), 5.3 (C), 13.6 GHz (Ku) 6 ofESA 52 ALOS PALSAR 2006–2011 1.2 GHz (L)ERS JAXA-1/2 altimetry data were used to discriminate thinning rates in the ablation zone (defined as <1500 -1 -1 RADARSAT-2GEOS-3 SARALT 1975 2007–present–1978 13.9 GHz (Ku) 5.3 GHz (C)NASA m CSAa.s.l.) of -2.0 ± 0.9 cm yr from accumulation zone (>1500 m a.s.l.) thickening of 6.4 ± 0.2 cm yr TerraSAR-XSeasat SAR-XALT 1978 (110 2007–present days) 13.6 GHz (Ku) 9.6 GHz (X)NASA DLRwith/EADS an overall net thickening of 5.4 ± 0.2 cm yr-1 for the period 1992–2003 [130]. To extend the TanDEM-XGeosat SAR-XGRA 1985 2010–present–1990 13.5 GHz (Ku) 9.6 GHz (X)DoD/NASA DLRtemporal/EADS coverage and increase spatial measurement density, Khvorostovsky and Johannessen [131] CryoSat-2 SIRAL 2010–present 13.9 GHz (Ku) ESA ERS-1 RA 1991–2006 13.8 GHz (Ku) ESA developed an algorithm to merge ERS-1/2 and EnviSat datasets and detect and eliminate inter- Sentinel-1AERS-2 SAR-CRA-2 1995 2014–present–2011 13.6 GHz (Ku), 3.2 5.4GHz GHz (S) (C)ESA ESA ALOS-2GFO PALSAR-2GFO-RA 1998 2014–present–2008 13.5 GHz (Ku) 1.2 GHz (L)DoD/NASA satellite JAXA biases. For the period 1992–2008, their dataset suggests net thinning of the GrIS initiated Sentinel-1BEnviSat SAR-CRA-2 2002 2016–present–2012 13.6 GHz (Ku), 3.2 5.4GHz GHz (S) (C)ESA around ESA 2000, and accelerated between 2006–2008, with a thickening of 4.0 ± 0.2 cm yr-1 above 1500 m -1 ScatterometersCryoSat-2 SIRAL 2010–present 13.9 GHz (Ku) ESA a.s.l. and thinning of -7.0 ± 1.0 cm yr below 1500 m a.s.l. [94]. SARAL ALtiKa 2013–present 36 GHz (Ka) ISRO/CNES Conventional radar altimeters such as the instruments included on ERS-1/2 and EnviSat were SeasatSentinel-3A/B SASSSRAL 2016 1978–present (110 days)13.6 GHz (Ku), 5.4 14.6 GHz GHz (C) (Ku)ESA NASA optimized for detecting elevation over open oceans and low-gradient polar ice sheet interiors with ERS-1Laser altimeters AMI-SCAT 1991–2006 5.3 GHz (C) ESA o ERS-2ICESat AMI-SCATGLAS 2003– 1995–20112009 1.064 μm, 0.532 5.3 μm GHz (C)NASA mean ESA accuracies typically <0.2 m for ice sheet surface slopes <0.2 [39,109]. Elevation accuracy and ADEOSICESat-2 NSCATATLAS 2018–present 1996–1997 1.064 μm, 0.532 14.0 μm GHz (Ku)NASA 푑퐻 JAXA/푑푡 uncertainty is much higher for the topographically complex ablation zone due to their large QuikSCATAircraft SeaWindsATM 1977–present 1999–2009 1.064 μm 13.4 GHz (Ku)NASA spatial NASA measurement footprint and wide orbital track spacing, with slope errors >20 m for slopes >1o ADEOS-2Aircraft SeaWindsLVIS 1998–present 2002–2003 1.064 μm 13.4 GHz (Ku)NASA JAXA/NASA and systematic biases over rough surfaces [39,95,109,132]. Careful processing has been used to MetOp A-CAircraft ASCATMABEL 2012 2006–present–2014 1.064 μm, 0.532 5.3 μm GHz (C)NASA EOMS/ESA discriminate ablation zone thinning rates using ERS-1/2 and EnviSat altimetry data, including repeat- *,Ŧ,§ See Abbreviations at end of article for expanded acronyms. Ɣ Managing agencies are identified *, , See Abbreviations at end of article for expanded acronyms. Managing agencies are identifiedtrack by analysis, the WMO but the accuracy of these data and their process-based interpretation is ambiguous by the WMO OSCAR database (Table A1) and may not reflect joint collaborations. OSCAR database (Table A1) and may not reflect joint collaborations. owing to the aforementioned slope errors and sparse spatial measurement density in the ablation The ESA European Remote-sensing Satellite (ERS-1) carried the first polar-orbiting radazoner [118,133–135]. altimeter, extending coverage of the polar ice sheets to ±81.5o [95]. Together with the follow-on ERS-The CryoSat-2 satellite (launched April 2010) provides continuity with ERS-1/2 and EnviSat and C- and Ku-band wind scatterometers measure σ◦ at both vertical (VV) and horizontal (HH) 2 and EnviSat missions, these satellites provided near complete GrIS coverage and 21 yearsemploys of SAR delay-Doppler along-track processing and interferometric cross-track processing polarization and are used to map the timing of melt onset and the seasonal progression of surface melt continuous data acquisition [38]. The first digital elevation model (DEM) spanning the entire [120,136]GrIS . CryoSat-2 is the first ESA radar altimeter dedicated to polar studies, with orbital coverage extentsurface [313,322 was]. produced As with from SAR, combined the method ERS-1 exploitsand Geosat the altimetry strong data, reduction supplemented in σ◦ caused by stereoto ±88 by- o theand presence1.6 km cross-track spacing at 70o. CryoSat-2 carries two Ku-band SAR/Interferometric photogrammetric and cartographic data sets, with pan-GrIS average accuracy -0.33 ± 6.97 m [128,129]. of liquid water at or near the ice surface. Relative to SAR imagers, scatterometers provideRadar Altimeterσ◦ at coarse (SIRAL) systems. SIRAL operates in three distinct modes depending on location: 1) ERS-1/2 altimetry data were used to discriminate thinning rates in the ablation zone (defined as <1500 spatial resolution but high temporal resolution, for example, a 25 km grid size twice-dailylow resolution formode the (LRM) over open ocean and ice sheet interiors, 2) synthetic aperture radar mode m a.s.l.) of -2.0 ± 0.9 cm yr-1 from accumulation zone (>1500 m a.s.l.) thickening of 6.4 ± 0.2 cm(SAR) yr-1 over sea ice and coastal regions, and 3) SAR interferometric mode (SARIn) over ice sheet SeaWinds on NASA’s Quick Scatterometer (QuikSCAT)-1 [322]. Resolution enhancement techniques with an overall net thickening of 5.4 ± 0.2 cm yr for the period 1992–2003 [130]. To extendmargins. the The SARIn mode employs two SAR instruments oriented across the satellite track. increasetemporal the e coverageffective and resolution increase spatial to ~8–10 measurement km [323 density,]. Khvorostovsky and Johannessen Interferometric[131] phase processing of the dual waveforms provides cross-track slope at ~0.3 km along- Surfacedeveloped melt an algorithm on the to GrIS merge and ERS-1/2 peripheral and EnviSatice datasets caps and has detect been and detected eliminate intertrack by- resolution the Seasat-A [137,138]. satellite biases. For the period 1992–2008, their dataset suggests net thinning of the GrIS initiated o scatterometer (SASS), the Advanced Microwave Instrument-Scatterometer (AMI-SCAT)Over on mild ERS-1 slopes/2, (<1 ), phase unwrapping [136] of CryoSat-2 interferometric data provides 5 km around 2000, and accelerated between 2006–2008, with a thickening of 4.0 ± 0.2 cm yr-1 above 1500wide m- swath elevation retrievals with two orders of magnitude more individual measurements per and thea.s.l. SeaWinds and thinning scatterometer of -7.0 ± 1.0 cm yr on-1 below both 1500 NASA’s m a.s.l. Quick[94]. Scatterometer (QuikSCAT)surface and area ADEOS-2 than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs (Table3)[311Conventional,320,322 ,radar324– 326altimeters]. Methods such as the to instruments detect surface included melt on includeERS-1/2 and single-channel EnviSat were absolute σ◦ optimized for detecting elevation over open oceans and low-gradient polar ice sheet interiors with thresholds [324], diurnal σ◦ variability (relative thresholds) [322], physical and statistical model-based mean accuracies typically <0.2 m for ice sheet surface slopes <0.2o [39,109]. Elevation accuracy and methods [326], and dual-frequency/polarization thresholds that also use the diurnal difference between 푑퐻/푑푡 uncertainty is much higher for the topographically complex ablation zone due to their large ascendingspatial andmeasurement descending footprint orbital and wide passes orbital [327 track]. Thespacing, diurnal with slope method errors exploits>20 m for slopes contrasts >1o in σ◦ caused by diurnaland systematic melt-freeze biases overcycles. rough The surfaces use [39,95,109,132] of relative. thresholds Careful processing reduces has errorsbeen used from to sensor drift, cross-missiondiscriminate biases, ablation or zone step thinning changes rates using in surface ERS-1/2 and properties, EnviSat altimetry such asdata, ice including lens formation repeat- that affect track analysis, but the accuracy of these data and their process-based interpretation is ambiguous absolute thresholds [115]. In addition to discrete melt onset, time-integrated σ reduction is used to owing to the aforementioned slope errors and sparse spatial measurement density in the ◦ablation infer seasonalzone [118,133 melt–135] intensity. [324–326]. In additionThe CryoSat to mapping-2 satellite (launched surface meltApril onset2010) provides and extent, continuity seasonal with ERS changes-1/2 and in EnviSatσ◦ are and used to infer the timingemploys and spatial SAR delay extent-Doppler of ice layer along-track formation processing in snow and andinterferometric firn [115 ,316 cross,328-track]. Understanding processing ice layer [120,136]. CryoSat-2 is the first ESA radar altimeter dedicated to polar studies, with orbital coverage to ±88o and 1.6 km cross-track spacing at 70o. CryoSat-2 carries two Ku-band SAR/Interferometric Radar Altimeter (SIRAL) systems. SIRAL operates in three distinct modes depending on location: 1) low resolution mode (LRM) over open ocean and ice sheet interiors, 2) synthetic aperture radar mode (SAR) over sea ice and coastal regions, and 3) SAR interferometric mode (SARIn) over ice sheet margins. The SARIn mode employs two SAR instruments oriented across the satellite track. Interferometric phase processing of the dual waveforms provides cross-track slope at ~0.3 km along- track resolution [137,138]. Over mild slopes (<1o), phase unwrapping [136] of CryoSat-2 interferometric data provides 5 km wide-swath elevation retrievals with two orders of magnitude more individual measurements per surface area than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs

Remote Sens. 2019, 11, x FOR PEER REVIEW 5 of 52

the accumulation zone, regional climate models and firn compaction models are used to partition the component of 푑퐻/푑푡 caused by change in 푆푀퐵 from those caused by change in firn density and 퐷 [104,105]. The following subsections summarize spaceborne radar altimetry of the GrIS, providing an historical perspective that emphasizes major challenges and advances in the field, and identifies future opportunities for radar altimetry observations of the GrIS ablation zone.

3.1. Radar Altimetry

Remote Sens.Conventiona2019, 11,l 2405radar altimeters are nadir pointing single-beam radars that transmit and receive 28 of 50 tens to thousands of microwave pulses per second, yielding ground measurement footprint diameters on the order 1–10 km posted at 0.3–0.6 km along-track spacing [39,106,107]. An estimate of formation푑퐻/푑푡 is obtained is important by comparing because measurements meltwater at refreezing crossover points, increases where firn an earlier density satellite without ground reducing mass and track intersects a later one [92]. Two critical factors affect the accuracy of radar altimeter elevation raisesretrievals the e overffective ice sheets: backscattering 1) variationsurface in topographic detected slope, by which radar controls altimeters, the average leading distance to errors in radar mass balancebetween estimates the altimeter (Section and the 3.1.2measured). Nghiem ground footprint et al. [115 surface] developed [39,108,109] a, field-validatedand 2) changes in the method that relates thresholdelectrical increases permittivity in of QuickSCAT the ice sheet HH-polarized surface, which σ cono beforetrols the and reflection, after melt transmission, seasons andto ice layer formation inabsorption the GrIS percolationof the microwave zone. signal The [110 method–113]. also provides a basis for estimating snow accumulation by For typical radar altimeter frequencies (Ku-band ~13 GHz), solid ice and liquid water are highly σo σo integratingreflective whereasreduction dry snow following and firn are ice semi layer-transparent, formation. with Ku The-band method penetration assumes depths threshold on the increases in areorder caused of several by enhanced meters [114,115] reflections. Spatial fromand temporal newly variations formed in ice signal layers, penetration whereas depth the caused gradual attenuation of σ◦byfollowing seasonal bareice layer ice exposure, formation snow is causedand firn by densification, new snow surface accumulation. meltwater presence, Wang et and al. [328] applied the methodrefreezing to fiveof meltwater years of introduce enhanced-resolution spurious height change QuickSCAT signals that imagery are poorly [329 quantified] and found for the extensive increases GrIS but are estimated to exceed >0.50 m in some cases [116–118]. Variations in surface slope indominate ice layer radar formation altimeter following accuracy in athe short topographically three-day-complex melt event bare- inice 2002,ablation highlighting zone, with slope the- disproportionate impactinduced of errors extreme that meltexceed events >20 m for on slopes ice layer >1° [39,109,113] formation.. The recent CryoSat-2 radar altimeter mission employs synthetic aperture radar (SAR) and interferometric synthetic aperture radar (InSAR) 6.2.technologies Passive Microwave which decrease and the Thermal effective Radiometry along-track footprint diameter to the order 0.1–0.3 km and provide cross-track slope from InSAR processing [93]. Together with denser orbital-track spacing and advancesAt thermal in waveform and microwaveretracking, topographic wavelengths mapping (beyond of ice sheet about ablation 3 um) zone spectral areas is accurate radiance (converted via theto Planck within a function few meters to on brightness average, and temperature, 푑퐻/푑푡 to withinTb sub) is-meter approximately uncertainty relative linear to with laserT sfc: Tb = ε Tsfc, altimetry [93,119,120]. ∗ where ε is the material emissivity [330]. Whereas σ◦ is dramatically reduced by liquid water in snow,3.1.1.T Radarb is dramatically Altimetry Sensors increased, and Datasets forming the basis for surface melt detection threshold algorithms, typically using a threshold value below freezing to indicate melting [315,331,332]. Passive microwave The first satellite altimeter measurements of surface topography for the Greenland Ice Sheet radiometerswere made by measure the GEOS background-3, Geosat, and microwave Seasat oceanographic emission Ku at-band (primarily) altimeters ( K-bandError! Reference (~19 GHz) and Ka-band (~37source GHz) not frequenciesfound.) [37,91,106,121,122] and various. Their combinations geographical coverage of VV was and limited HH polarization to ±72o but the data (Table 4). In contrast tothey the collected higher showed spatial it resolutionwas possible (~10to calculate km2) volumetric but the lowerchanges temporalin the polar resolutionice sheets from (weeks to months) space, providing proof of concept for future polar-orbiting altimeters [92]. A first demonstration of spaceborne active microwave sensors, passive microwave sensors provide T twice-daily at using GEOS-3, Geosat, and Seasat data found that for areas south of <72oN, the GrIS thickened by 23 b 2 Remote Sens. 2019, 11, x FOR PEER REVIEW 6 of 52 ~25–50± 6 cm kmyr-1 betweenspatial 1978 resolution–1986, with andenhanced complete snowfall ice in a sheet warmer coverage. climate hypothesized Melt detection to explain methods using Tb includethe observed single-channel thickening [123] thresholds. These early [332 missions,333], dualsupported frequency important/polarization developments combinations in waveform (cross polarizedGEOS-3 ALT 1975–1978 13.9 GHz (Ku) NASA Seasat ALT 1978 (110 days) 13.6 GHz (Ku) NASA gradientretracking, ratio slope XPGR) correction, [315 , and334 ], physical and diurnal and empirical amplitude models variations for subsurface (DAV) volume on ascending scattering and descendingGeosat GRA 1985–1990 13.5 GHz (Ku) DoD/NASA [106,108,111,124]. These corrections are critical for accurate elevation retrieval and reliable change ERS-1 RA 1991–2006 13.8 GHz (Ku) ESA passes [335,336]. As with σ◦, microwave Tb is strongly modified by liquid meltwater presenceERS-2 at RA-2 1995–2011 13.6 GHz (Ku), 3.2 GHz (S) ESA detection [125,126]. For example, Davis et al. [127] used an improved geodetic model and an GFO GFO-RA 1998–2008 13.5 GHz (Ku) DoD/NASA theimproved ice sheet retracking surface model but to typically reinterpret doesthe earlier not findings provide of Zwally information et al. [92] about, finding the a smaller internal snowEnviSat or firn RA-2 2002–2012 13.6 GHz (Ku), 3.2 GHz (S) ESA CryoSat-2 SIRAL 2010–present 13.9 GHz (Ku) ESA structure1.5 ± 0.5 [cm324 yr,337-1 thickening]. for the period 1978–1986 that highlighted the importance of slope- and SARAL ALtiKa 2013–present 36 GHz (Ka) ISRO/CNES signal penetration-correction. Sentinel-3A/B SRAL 2016–present 13.6 GHz (Ku), 5.4 GHz (C) ESA Laser altimeters Table 4. Summary of passive microwave remote sensing platforms, instruments, temporal coverage,ICESat GLAS 2003–2009 1.064 μm, 0.532 μm NASA Table 1. Summary of radar and laser altimeter remote sensing platforms, instruments, temporal ICESat-2 ATLAS 2018–present 1.064 μm, 0.532 μm NASA observedRemote Sens. frequencies,2019, 11, x FOR PEER and REVIEW managing agencies. 6 of 52 Aircraft ATM 1977–present 1.064 μm NASA coverage, observed wavelength, and managing agencies. Aircraft LVIS 1998–present 1.064 μm NASA Aircraft MABEL 2012–2014 1.064 μm, 0.532 μm NASA GEOS-3 ALT Temporal1975Temporal– 1978 Observed13.9 GHz (Ku)Observed NASA Ŧ §,Ɣ *,Ŧ,§ See Abbreviations at end of article for expanded acronyms. Ɣ Managing agencies are identified PlatformPlatform *Seasat * InstrumentInstrumentALT 1978 (110 days) 13.6 GHz (Ku) AgencyNASA Agency CoverageCoverage Wavelength Frequency by the WMO OSCAR database (Table A1) and may not reflect joint collaborations. Radar AltimetersGeosat GRA 1985–1990 13.5 GHz (Ku) DoD/NASA Nimbus-5ERS-1 ESMRRA 1991 1972–1983–2006 13.8 GHz (Ku) 19 GHzESA NASAThe ESA European Remote-sensing Satellite (ERS-1) carried the first polar-orbiting radar Nimbus-6ERS-2 ESMRRA-2 1995 1975–1983–2011 13.6 GHz (Ku), 3.2 37 GHz GHz (S) ESA altimeter, NASA extending coverage of the polar ice sheets to ±81.5o [95]. Together with the follow-on ERS- SeasatGFO SMMRGFO-RA 19781998 (110–2008 days) 13.5 GHz 7, (Ku) 10, 18, 21, 37DoD/NASA 2 NASA and EnviSat missions, these satellites provided near complete GrIS coverage and 21 years of Nimbus-7EnviSat SMMRRA-2 2002 1978–1994–2012 13.6 GHz (Ku), 7, 10,3.2 GHz 19, 37 (S) GHz ESA NASAcontinuous/NOAA data acquisition [38]. The first digital elevation model (DEM) spanning the entire GrIS DMSP F08,10-15,18CryoSat-2 SSM/ISIRAL 2010 1987–present–present 13.9 GHz 19, 22,(Ku) 37, 86 GHzESA DoDsurface/NOAA was produced from combined ERS-1 and Geosat altimetry data, supplemented by stereo- DMSP F16-19SARAL SSMISALtiKa 2013 1987–present–present 36 GHz 19, (Ka) 22, 37, 92 GHzISRO/CNES DoDphotogrammetric/NOAA and cartographic data sets, with pan-GrIS average accuracy -0.33 ± 6.97 m [128,129]. ERS-1Sentinel-3A/B ATSRSRAL 2016 1991–2006–present 13.6 GHz (Ku), 5.4 24, GHz 37 GHz (C) ESA ERS ESA-1/2 altimetry data were used to discriminate thinning rates in the ablation zone (defined as <1500 GFOLaser altimeters WVR 1998–2008 22, 37 GHz DoDm /a.s.l.)NASA of -2.0 ± 0.9 cm yr-1 from accumulation zone (>1500 m a.s.l.) thickening of 6.4 ± 0.2 cm yr-1 ADEOS-2ICESat AMSRGLAS 2003 2002–2003–2009 1.064 7, 10, μm, 19, 0.532 24, 37,μm 89A, 89BNASA GHz JAXAwith/NASA an overall net thickening of 5.4 ± 0.2 cm yr-1 for the period 1992–2003 [130]. To extend the AquaICESat-2 AMSR-EATLAS 2018 2002–2011–present 1.064 7,μm, 10, 0.532 19, 24,μm37, 89 GHzNASA temporal NASA coverage and increase spatial measurement density, Khvorostovsky and Johannessen [131] ERS-2, EnviSatAircraft MWRATM 1977 2002–2012–present 1.064 μm 24, 37 GHzNASA developed ESA an algorithm to merge ERS-1/2 and EnviSat datasets and detect and eliminate inter- GCOM-W1Aircraft AMSR-2LVIS 1998 2012–present–present 71.064 (dual), μm 10, 19, 37, 89 GHzNASA NASAsatellite/JAXA biases. For the period 1992–2008, their dataset suggests net thinning of the GrIS initiated Sentinel 3A/BAircraft MWRMABEL 2016–present2012–2014 1.064 μm, 0.532 24, μm 37GHz NASA around ESA 2000, and accelerated between 2006–2008, with a thickening of 4.0 ± 0.2 cm yr-1 above 1500 m a.s.l. and thinning of -7.0 ± 1.0 cm yr-1 below 1500 m a.s.l. [94]. *,Ŧ,§ See Abbreviations at end of article for expanded acronyms. Ɣ Managing agencies are identified *, , See Abbreviations at end of article for expanded acronyms. Managing agencies are identified byConventional the WMO radar altimeters such as the instruments included on ERS-1/2 and EnviSat were by the WMO OSCAR database (Table A1) and may not reflect joint collaborations. OSCAR database (Table A1) and may not reflect joint collaborations. optimized for detecting elevation over open oceans and low-gradient polar ice sheet interiors with o The ESA European Remote-sensing Satellite (ERS-1) carried the first polar-orbiting radameanr accuracies typically <0.2 m for ice sheet surface slopes <0.2 [39,109]. Elevation accuracy and 푑퐻/푑푡 uncertainty is much higher for the topographically complex ablation zone due to their large o altimeter, extending coverage of the polar ice sheets to ±81.5 [95]. Together with the follow-on ERS- o With more than 30 years of continuous data collection, pan-GrIS spatial coverage,spatial and measurement all-weather footprint and wide orbital track spacing, with slope errors >20 m for slopes >1 2 and EnviSat missions, these satellites provided near complete GrIS coverage and 21 years and of systematic biases over rough surfaces [39,95,109,132]. Careful processing has been used to capability,continuous spaceborne data acquisition passive [38] microwave. The first digital radiometers elevation provide model (DEM) a unique spanning insight the entire into theGrISdiscriminate climatic ablation drivers zone thinning rates using ERS-1/2 and EnviSat altimetry data, including repeat- of icesurface sheet surfacewas produced mass from balance combined processes, ERS-1 and including Geosat altimetry changes data, in supplemented the location by and stereo extenttrack- analysis, of surfacebut the accuracy of these data and their process-based interpretation is ambiguous owing to the aforementioned slope errors and sparse spatial measurement density in the ablation meltingphotogrammetric (Figure 12)[ and315 cartographic,317,318,334 data,336 sets,,338 with,339 pan].-GrIS Data average from accuracy the Special -0.33 ± 6.97 Sensor m [128,129] Microwavezone. [118,133/–Imager135]. ERS-1/2 altimetry data were used to discriminate thinning rates in the ablation zone (defined as <1500 (SSM/I) suggests the area of active surface melt over the GrIS has approximately doubled sinceThe CryoSat the-2 early satellite (launched April 2010) provides continuity with ERS-1/2 and EnviSat and m a.s.l.) of -2.0 ± 0.9 cm yr-1 from accumulation zone (>1500 m a.s.l.) thickening of 6.4 ± 0.2 cm yremploys-1 SAR delay-Doppler along-track processing and interferometric cross-track processing 2 1 1990s,with with an a overall 40,000 net km thickeningyr− trend of 5.4 during ± 0.2 cm this yr period-1 for the [ 318 period,336 1992]. Data–2003 from[130]. the To extend Scanning the[120,136] Multichannel. CryoSat-2 is the first ESA radar altimeter dedicated to polar studies, with orbital coverage to ±88o and 1.6 km cross-track spacing at 70o. CryoSat-2 carries two Ku-band SAR/Interferometric temporal coverage and increase spatial measurement density, Khvorostovsky and Johannessen [131]Radar Altimeter (SIRAL) systems. SIRAL operates in three distinct modes depending on location: 1) developed an algorithm to merge ERS-1/2 and EnviSat datasets and detect and eliminate interlow- resolution mode (LRM) over open ocean and ice sheet interiors, 2) synthetic aperture radar mode satellite biases. For the period 1992–2008, their dataset suggests net thinning of the GrIS initiated(SAR) over sea ice and coastal regions, and 3) SAR interferometric mode (SARIn) over ice sheet around 2000, and accelerated between 2006–2008, with a thickening of 4.0 ± 0.2 cm yr-1 above 1500margins. m The SARIn mode employs two SAR instruments oriented across the satellite track. Interferometric phase processing of the dual waveforms provides cross-track slope at ~0.3 km along- -1 a.s.l. and thinning of -7.0 ± 1.0 cm yr below 1500 m a.s.l. [94]. track resolution [137,138]. Conventional radar altimeters such as the instruments included on ERS-1/2 and EnviSat were Over mild slopes (<1o), phase unwrapping [136] of CryoSat-2 interferometric data provides 5 km optimized for detecting elevation over open oceans and low-gradient polar ice sheet interiors withwide -swath elevation retrievals with two orders of magnitude more individual measurements per mean accuracies typically <0.2 m for ice sheet surface slopes <0.2o [39,109]. Elevation accuracy andsurface area than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs 푑퐻/푑푡 uncertainty is much higher for the topographically complex ablation zone due to their large spatial measurement footprint and wide orbital track spacing, with slope errors >20 m for slopes >1o and systematic biases over rough surfaces [39,95,109,132]. Careful processing has been used to discriminate ablation zone thinning rates using ERS-1/2 and EnviSat altimetry data, including repeat- track analysis, but the accuracy of these data and their process-based interpretation is ambiguous owing to the aforementioned slope errors and sparse spatial measurement density in the ablation zone [118,133–135]. The CryoSat-2 satellite (launched April 2010) provides continuity with ERS-1/2 and EnviSat and employs SAR delay-Doppler along-track processing and interferometric cross-track processing [120,136]. CryoSat-2 is the first ESA radar altimeter dedicated to polar studies, with orbital coverage to ±88o and 1.6 km cross-track spacing at 70o. CryoSat-2 carries two Ku-band SAR/Interferometric Radar Altimeter (SIRAL) systems. SIRAL operates in three distinct modes depending on location: 1) low resolution mode (LRM) over open ocean and ice sheet interiors, 2) synthetic aperture radar mode (SAR) over sea ice and coastal regions, and 3) SAR interferometric mode (SARIn) over ice sheet margins. The SARIn mode employs two SAR instruments oriented across the satellite track. Interferometric phase processing of the dual waveforms provides cross-track slope at ~0.3 km along- track resolution [137,138]. Over mild slopes (<1o), phase unwrapping [136] of CryoSat-2 interferometric data provides 5 km wide-swath elevation retrievals with two orders of magnitude more individual measurements per surface area than any prior radar altimeter [119,120,137]. The first spatially continuous swath DEMs

Remote Sens. 2019, 11, 2405 29 of 50

Microwave Radiometer (SMMR, 1978-1987) and SSM/I were used to characterize the effects of the Mt. Pinatubo eruption on GrIS surface melt patterns [340]. Surface melt extent from SSM/I has been correlated with downstream sediment plume concentrations in Greenland fjords driven by ice sheet meltwater discharge [341], to validate surface melt extent calculated from surface energy balance models [342,343], and to quantify extreme events such as the record July 2012 melt event when 98.6% of the GrIS surface was actively melting [287,339]. Remote Sens. 2019, 11, x FOR PEER REVIEW 31 of 52

.

FigureFigure 12. SurfaceSurface melt melt presence presence frequency frequency-of-occurrence-of-occurrence during during the summer the summer melting melting season season (June– (June–August)August) for the for period the period 1972– 1972–20122012 from frompassive passive microwave microwave brightness brightness temperature temperature observed observed by the by theScanning Scanning Multichannel Multichannel Microwave Microwave Radiometer Radiometer (SMMR), (SMMR), the the Special Special Sensor Microwave Microwave/Imager/Imager (SSM(SSM/I),/I), andand thethe SpecialSpecial SensorSensor MicrowaveMicrowave ImagerImager/Sounder/Sounder (SSMIS).(SSMIS). AreasAreas experiencingexperiencing zerozero meltmelt presencepresence frequency frequency are are colored colored white. white. Surface Surface melt presence melt presence provides provides a sensitive a sensitive indicator indicatorof changing of climaticchanging conditions climatic conditions over the Greenlandover the Greenland Ice Sheet, Ice including Sheet, including the July the 2012 July extreme 2012 extreme melt event melt when event surfacewhen surface melt prevailed melt prevailed over the over entire the iceentire sheet ice forsheet the for first the time first in time the in satellite the satellite era [287 era]. [ Surface287]. Surface melt frequencymelt frequency is calculated is calculated by the by authors the authors from the from NASA the NASA MEaSUREs MEaSUREs Greenland Greenland Surface Surface Melt Daily Melt 25 Daily km EASE-Grid25 km EASE 2.0-Grid data 2.0 set data [344 set]. [344].

Thermal radiance (~3–14 µm) is used to map Tsfc and provides an additional method for surface Thermal radiance (~3–14 μm) is used to map 푇sfc and provides an additional method for surface melt detection [345]. Variations in thermal and optical radiance form the basis for mapping “reflectance melt detection [345]. Variations in thermal and optical radiance form the basis for mapping zones”“reflectance to characterize zones” to meltcharacterize presence melt and presence changes inand the changes thermal in structure the thermal on glacier structure surfaces on glacier [346]. Whereas σ◦ and microwave TB are primarily diagnostic of melt presence, both at or near the surface, surfaces [346]. Whereas σ° and microwave 푇B are primarily diagnostic of melt presence, both at or thermalnear the radiance surface, isthermal diagnostic radiance of the is surfacediagnostic “skin” of the temperature, surface “skin” and providestemperature, little and to no provides information little about subsurface temperature, meltwater presence, or snow and firn structure. Consequently, thermal to no information about subsurface temperature, meltwater presence, or snow and firn structure. Tsfc is strictly an indicator of surface melt and, together with the higher spatial and radiometric Consequently, thermal 푇sfc is strictly an indicator of surface melt and, together with the higher resolution of thermal sensors such as MODIS, provides an independent method for validating active spatial and radiometric resolution of thermal sensors such as MODIS, provides an independent andmethod passive for microwave validating meltactive presence and passive products microwave [347,348 ]. melt When presence combined products with microwave[347,348]. When melt presence and remotely sensed albedo, thermal Tsfc may improve the discrimination of surface and combined with microwave melt presence and remotely sensed albedo, thermal 푇sfc may improve the subsurfacediscrimination melt of areas surface and diurnaland subsurface variations melt in melt-freeze areas and diurnal cycles [ 347variations]. As with in melt microwave-freeze surfacecycles [ melt347]. detection,As with microwave thermal radiance surface melt detectiondetection, hasthermal been radiance used to validate melt detection modeled has surface been used melt to [318 validate,342], andmodeled to study surface the melt spatiotemporal [318,342], and variation to study of the surface spatiotemporal melt extent variation and duration of surface on the melt GrIS extent and and its relationduration to on climatic the GrIS variability and its relation [55,238 ,to348 climatic–350]. variability [55,238,348–350].

The primary spaceborne thermal sensors used to obtain 푇sfc are MODIS and AVHRR. The NOAA Extended AVHRR Polar Pathfinder (App-X) product provides 푇sfc on a 25 km polar equal- area grid twice-daily for the period 1982 to present [272]. As with the AVHRR albedo product (Section 5.2), there are spectral, radiometric, and inter-mission homogenization issues that limit its utility [351]. The MODIS land surface temperature product (MOD11A1) uses the split-window technique

developed for AVHRR to calculate 푇sfc from radiance at 10.78 μ m and 11.77 μ m [352]. The MOD11A1 data are provided on a 1 km grid globally for clear-sky conditions as discriminated by the MOD35 cloud mask. The MOD11A1 product is accurate to ± 1 °C on average over snow and ice

Remote Sens. 2019, 11, 2405 30 of 50

The primary spaceborne thermal sensors used to obtain Tsfc are MODIS and AVHRR. The NOAA Extended AVHRR Polar Pathfinder (App-X) product provides Tsfc on a 25 km polar equal-area grid twice-daily for the period 1982 to present [272]. As with the AVHRR albedo product (Section 5.2), there are spectral, radiometric, and inter-mission homogenization issues that limit its utility [351]. The MODIS land surface temperature product (MOD11A1) uses the split-window technique developed for AVHRR to calculate Tsfc from radiance at 10.78 µm and 11.77 µm [352]. The MOD11A1 data are provided on a 1 km grid globally for clear-sky conditions as discriminated by the MOD35 cloud mask. The MOD11A1 product is accurate to 1 C on average over snow and ice surfaces and agrees to within ± ◦ 0.5 C with T calculated from ASTER and Landsat ETM+ thermal radiance over the GrIS [350]. ± ◦ sfc Surface melt extent from MOD11A1 corresponds closely to surface melt inferred from QuickSCAT using the diurnal σo method [347]. The Greenland Ice Surface Temperature (IST) product is an enhanced version of MOD11A1 that provides IST and binary melt absence/presence for the period 2000–2014 on a 1.25 km polar stereographic grid for the GrIS [349,351]. For values of T near 0 C, the IST data are 0.5 C sfc ◦ − ◦ cooler than surface-based Tsfc measurements collected at the Summit Station [353,354]. The bias increases to 5.0 C for values of T near 60 C and increases with SZA, suggesting the bias may be − ◦ sfc − ◦ related to reduced accuracy of the MOD35 cloud mask at high SZA. Under-sampling of Tsfc during warm inversions with dense cloud coverage may also introduce bias but this effect has not been systematically evaluated [353,354]. In general, spaceborne thermal Tsfc retrievals are effective for detecting surface melt during clear-sky conditions but are sensitive to the atmospheric aerosol and water vapor profile [351,355]. To overcome cloud-cover data gaps, Välisuo et al. [348] gap-filled the IST product with modeled values of Tsfc from the European Centre for Medium-Range Weather Forecasts ERA-Interim reanalysis.

6.3. Multi-Angular Reflectance and Surface Roughness The Multi-angle Imaging SpectroRadiometer (MISR) measures coincident-in-time bi-directional radiance in four spectral bands between 450–850 nm at 0◦ (nadir), 26.1◦, 45.6◦, 60.0◦, and 70.5◦ fore and aft of nadir. Reflectance is provided on a 275 m grid and nine-day global coverage to 83 (Table2)[ 260]. ± ◦ The normalized difference angular index (NDAI) was developed to estimate ice surface roughness from MISR angular reflectance [260]. The method uses an empirical relationship between MISR red band reflectance at 60o fore and aft of nadir and ATM lidar-derived surface roughness to develop spatially-continuous maps of surface roughness over ice sheets and sea ice [356]. The MISR NDAI and near-infrared albedo (Section 5.2) were used to map unique signatures of surface glaciological features in the western GrIS ablation zone, including crevasse fields, wet snow, and bare glacier ice [53]. With this approach, MISR angular reflectance appears to provide a unique method for detecting the superimposed ice zone and may improve the identification of changes in crevasse field roughness relative to SAR and optical sensors [53,260]. Surface roughness controls ice–atmosphere interactions via the aerodynamic roughness length and the net vapor flux, an understudied component of the GrIS SMB [60,82,357]. The MISR surface roughness product [356] was used to define the aerodynamic roughness length of the BMF13 vapor flux model to improve the spatial realism of net vapor flux in the ablation zone where roughness is highly variable [357]. The average annual modeled vapor flux was 14.6 3.6 Gt yr 1, or 6 2% of ± − ± annual SMB for the period 2003–2014. The average annual difference between modeled vapor flux with and without MISR roughness was 30 15%. In addition to angular reflectance, surface roughness ± has been quantified from ICESat and ATM laser altimetry waveforms [173,358,359]. The forthcoming ICESat-2 laser altimeter may provide additional capability for measuring surface roughness on the GrIS, for example using the multi-sensor lidar-angular reflectance approach of Nolin and Mar [356] and Nolin et al. [260], which may also be useful for radar altimetry waveform interpretation of the leading edge of the beam footprint [93]. Remote Sens. 2019, 11, 2405 31 of 50

6.4. Future Opportunities for Mapping the Changing GrIS Ablation Zone Surface The GrIS surface is undergoing rapid change, driven by increased surface meltwater production in the ablation zone. Spaceborne SAR imagers, wind scatterometers, and passive microwave, thermal, and angular radiometers are used to map and monitor diagnostic features of change on the ablation zone surface, including surface melt presence and extent, subsurface ice layer formation, ice surface temperature, and ice surface roughness. These characteristic features are used to define and map the dry snow zone, wet snow zone, percolation zone, superimposed ice zone, and the bare ice zone. Mapping of supraglacial hydrologic features, including meltwater lakes, rivers, and moulins stands out as an additional research priority c.f. [68], in particular, their expansion and inland migration toward sensitive (e.g., high elevation) areas of the ice sheet [56,360,361]. Other diagnostic markers of change detectable in satellite imagery include the end of summer snowline position (Figure9)[ 26], the lower limit of superimposed ice [53], and the lower limit of the slush zone [293]. Multi-sensor methods, for example, optical imagery combined with SAR imagery or scatterometry, shows promise for detecting dynamic regions characterized by changing snow, firn, and ice surface types, including supraglacial lakes and slush fields obscured by snow or clouds, and may reduce detection bias caused by cloud cover [362–364].

7. Conclusions For over forty years, earth-observing satellites sensitive to visible, infrared, and microwave electromagnetic radiation, together with gravimetry, have documented the patterns of change on the GrIS ablation zone surface. Satellite remote sensing data show an ablation zone expanded in size, its albedo and surface elevation lower in response to enhanced melting and ice discharge, and an ice sheet transition from steady state to negative mass balance that now represents the largest land ice contributor to global sea level rise. Already, some 78–85% of the total liquid runoff produced from surface melting is generated in the bare ice ablation zone, despite it covering ~22% of the ice sheet’s total surface area, up from ~15% in the early 1960s [20,30,33–35]. Although often conceptualized as a uniform surface of solid ice, the ablation zone is a dynamic region with widely varying electromagnetic properties controlled by diverse physical, biological, and hydrologic processes. Future progress in remote sensing the ablation zone will likely benefit most from methods that directly address this complexity, for example, using multi-sensor, multi-wavelength, and cross-platform datasets. Examples include fusing radar and laser altimetry with optical stereophotogrammetry to discriminate and diagnose causes of surface elevation change [174], or fusing radar and laser backscatter with optical imagery to discriminate snow, ice, liquid water, and refrozen meltwater in sensitive areas near the equilibrium line altitude [362,363]. Other areas of opportunity recommended for future research include spaceborne detection of subsurface refrozen meltwater and its effects on radar backscatter, which requires additional in-situ validation [115,116], the partitioning of ablation zone thinning into ice-dynamic and surface mass balance components [97], cross-validation of ice surface elevation change from altimetry with modeled surface mass balance [89] and modeled ice dynamic motion [189], spaceborne diagnosis of changing bare ice albedo [269] and grain size [191], and monitoring the inland migration of snowlines, surface melt extent, and surface hydrologic features including lakes, streams, and moulins [26,56].

Author Contributions: Conceptualization, M.G.C. and L.C.S.; writing—original draft preparation, M.G.C.; writing—review and editing M.G.C. and L.C.S.; visualization, M.G.C.; supervision, L.C.S.; funding acquisition L.C.S. and M.G.C. Funding: This research was funded by NASA Cryosphere Program grant 80NSSC19K0942, managed by Thorsten Markus and Colene Haffke, and a graduate fellowship from the NASA Earth and Space Sciences Fellowship Program managed by Lin Chambers. Acknowledgments: The authors thank the Polar Geospatial Center, University of Minnesota, for providing WorldView Imagery © 2019 DigitalGlobe, Inc, Jonathan Ryan, Brown University, for providing Figure9, the many public and private agencies, universities, and scientists who provided satellite data, processed datasets, and Remote Sens. 2019, 11, 2405 32 of 50 image processing tools that contributed to this work, and three anonymous reviewers for valuable feedback that substantially improved this work. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations

Abbreviations for select remote sensing satellite platforms. ADEOS ADvanced Earth Observing Satellite ADEOS-2 ADvanced Earth Observing Satellite-2 ALOS Advanced Land Observing Satellite CryoSat Cryosphere Satellite DMSP Defense Meteorological Satellite Program EnviSat Environmental Satellite EO-1 Earth Observation-1 ERS-1 European Remote-sensing Satellite-1 ERS-2 European Remote-sensing Satellite-2 GCOM-W Global Change Observation Mission—Water “Shizuku” GCOM-C Global Change Observation Mission—Climate “Shikisai” GEOS-3 Geodetic and Earth Orbiting Satellite-3 GEOSAT GEOdetic SATellite GFO GEOdetic SATellite Follow On GOES Geostationary Operational Environmental Satellites GRACE Gravity Recovery and Climate Experiment ICESat Ice, Cloud, and land Elevation Satellite ICESat-2 Ice, Cloud, and land Elevation Satellite-2 JERS-1 Japanese Earth Resource Satellite-1 Landsat Land Satellite MetOp Meteorological Operational satellite program NOAA National Oceanic and Atmospheric Administration QuickSCAT Quick Scatterometer RADARSAT Radar Satellite of the Canadian Space Agency SPOT Satellite Pour l’Observation de la Terre SARAL Satellite with Argos and ALtiKa Suomi NPP Suomi National Polar-orbiting Partnership TerraDEM-X TerraSAR-X add on for Digital Elevation Measurements TerraSAR-X Synthetic Aperture Radar X-band TIROS Television Infrared Operational Satellite Abbreviations for select remote sensing satellite sensors. AATSR Advanced Along Track Scanning Radiometer ALI Advanced Land Imager ALT Radar Altimeter ALtiKa Ka-Band Altimeter AMI Advanced Microwave Instrument AMSR Advanced Microwave Scanning Radiometer AMSR-2 Advanced Microwave Scanning Radiometer 2 AMSR-E Advanced Microwave Scanning Radiometer—Earth Observing System ASAR Advanced Synthetic-Aperture Radar ASCAT Advanced Scatterometer ASTER Advanced Spaceborne Thermal Emission and Reflection radiometer ATLAS Advanced Topographic Laser Altimeter System ATM Airborne Topographic Mapper ATSR Along Track Scanning Radiometer ATSR-2 Along Track Scanning Radiometer-2 AVCS Advanced Vidcon Camera System AVHRR Advanced Very High Resolution Radiometer Remote Sens. 2019, 11, 2405 33 of 50

AVNIR Advanced Visible and Near Infrared Radiometer AVNIR-2 Advanced Visible and Near Infrared Radiometer type 2 BGIS2000 Ball Global Imaging System 2000 C-SAR C-band Synthetic Aperture Radar ESMR Electrically Scanning Multichannel Radiometer ETM+ Enhanced Thematic Mapper Plus GIS GeoEye Imaging System GLAS Geoscience Laser Altimeter System GLI Global Land Imager GRA Geosat Radar Altimeter GRACE Gravity Recovery and Climate Experiment HRG High Resolution Geometric HRS High Resolution Stereoscopic HRV High Resolution Visible HRVIR High Resolution Visible and Infrared Hyperion Hyperspectral Imager IDCS Image Dissector Camera System LVIS Land, Vegetation and Ice Sensor MERIS MEdium Resolution Imaging Spectrometer MISR Multi-angle Imaging SpectroRadiometer MODIS Moderate Resolution Imaging Spectroradiometer MSI Multispectral Imager MSS Multispectral Scanner MWR Microwave Radiometer NAOMI New AstroSat Optical Modular Instrument NSCAT NASA Scatterometer OLCI Ocean and Land Colour Instrument OLI Operational Land Imager OSA Optical Sensor Assembly OPS Optical Sensor PALSAR Phased-Array type L-band SAR PALSAR-2 Phased-Array type L-band SAR POLDER Polarization and Directionality of the Earth’s Reflectances RA Radar Altimeter RA-2 Radar Altimeter 2 SAR Synthetic Aperture Radar SASS SEASAT-A Satellite Scatterometer SGLI Second generation Global Imager SIRAL Synthetic Aperture Interferometric Radar Altimeter SMMR Scanning Multichannel Microwave Radiometer SSM/I Special Sensor Microwave/Imager SSMIS Special Sensor Microwave Imager/Sounder TIRS Thermal Infrared Sensor TM Thematic Mapper VIIRS Visible/Infrared Imager Radiometer Suite VIRR Visible and Infrared Radiometer VISSR Visible and Infrared Spin Scan Radiometer WV3-Imager WorldView-3 Imager WV60 WorldView-60 camera WV110 WorldView-110 camera WVR Water Vapor Radiometer Abbreviations for public and private remote sensing sponsoring agencies. CNES French Space Agency CSA Canadian Space Agency DLR German Aerospace Center Remote Sens. 2019, 11, 2405 34 of 50

DoD Department of Defense (United States) EADS European Aeronautic Defense and Space Company EOMS European Organisation for Meteorological Satellites EOSAT Earth Observation Satellite Company ESA European Space Agency ISRO Indian Space Research Organisation JAXA Japanese Aerospace Exploratory Agency JPL Jet Propulsion Laboratory NASA National Aeronautics and Space Administration (US) NASDA National Space Development Agency of Japan NOAA National Oceanic and Atmospheric Administration (US) USGS United States Geological Survey

Appendix A Table A1. List of online repositories of satellite remote sensing platforms, sensors, and managing agency information.

Managing Organization Repository Name URL World Meteorological Observing Systems Capability Analysis and https://www.wmo-sat.info/oscar/spacecapabilities Organization Review (OSCAR) NASA Space Science Data Coordinated Archive NASA https://nssdc.gsfc.nasa.gov/nmc/ (NSSDCA) https://directory.eoportal.org/web/eoportal/ ESA Earth Observation Portal (eoPortal) satellite-missions

References

1. Van den Broeke, M.R.; Enderlin, E.M.; Howat, I.M.; Kuipers Munneke, P.; Noël, B.P.Y.; van de Berg, W.J.; van Meijgaard, E.; Wouters, B. On the recent contribution of the Greenland ice sheet to sea level change. Cryosphere 2016, 10, 1933–1946. [CrossRef] 2. Hanna, E.; Huybrechts, P.; Steffen, K.; Cappelen, J.; Huff, R.; Shuman, C.; Irvine-Fynn, T.; Wise, S.; Griffiths, M. Increased Runoff from Melt from the Greenland Ice Sheet: A Response to Global Warming. J. Clim. 2008, 21, 331–341. [CrossRef] 3. Vizcaíno, M.; Mikolajewicz, U.; Gröger, M.; Maier-Reimer, E.; Schurgers, G.; Winguth, A.M.E. Long-term ice sheet–climate interactions under anthropogenic greenhouse forcing simulated with a complex Earth System Model. Clim. Dyn. 2008, 31, 665–690. [CrossRef] 4. Ettema, J.; van den Broeke, M.R.; van Meijgaard, E.; van de Berg, W.J.; Bamber, J.L.; Box, J.E.; Bales, R.C. Higher surface mass balance of the Greenland ice sheet revealed by high-resolution climate modeling. Geophys. Res. Lett. 2009, 36, L12501. [CrossRef] 5. Vizcaíno, M.; Lipscomb, W.H.; Sacks, W.J.; van Angelen, J.H.; Wouters, B.; van den Broeke, M.R. Greenland Surface Mass Balance as Simulated by the Community Earth System Model. Part I: Model Evaluation and 1850–2005 Results. J. Clim. 2013, 26, 7793–7812. [CrossRef] 6. Vizcaíno, M.; Lipscomb, W.H.; Sacks, W.J.; van den Broeke, M. Greenland Surface Mass Balance as Simulated by the Community Earth System Model. Part II: Twenty-First-Century Changes. J. Clim. 2013, 27, 215–226. [CrossRef] 7. Vizcaíno, M.; Mikolajewicz, U.; Ziemen, F.; Rodehacke, C.B.; Greve, R.; van den Broeke, M.R. Coupled simulations of Greenland ice sheet and climate change up to AD 2300. Geophys. Res. Lett. 2015, 42, 3927–3935. [CrossRef] 8. Trusel, L.D.; Das, S.B.; Osman, M.B.; Evans, M.J.; Smith, B.E.; Fettweis, X.; McConnell, J.R.; Noël, B.P.Y.; Broeke, M.R. van den Nonlinear rise in Greenland runoff in response to post-industrial Arctic warming. Nature 2018, 564, 104–108. [CrossRef] 9. Pattyn, F.; Ritz, C.; Hanna, E.; Asay-Davis, X.; DeConto, R.; Durand, G.; Favier, L.; Fettweis, X.; Goelzer, H.;

Golledge, N.R.; et al. The Greenland and Antarctic ice sheets under 1.5 ◦C global warming. Nat. Clim. Chang. 2018, 8, 1053–1061. [CrossRef] 10. Bamber, J.L.; Westaway, R.M.; Marzeion, B.; Wouters, B. The land ice contribution to sea level during the satellite era. Environ. Res. Lett. 2018, 13, 063008. [CrossRef] Remote Sens. 2019, 11, 2405 35 of 50

11. Hanna, E.; Navarro, F.J.; Pattyn, F.; Domingues, C.M.; Fettweis, X.; Ivins, E.R.; Nicholls, R.J.; Ritz, C.; Smith, B.; Tulaczyk, S.; et al. Ice-sheet mass balance and climate change. Nature 2013, 498, 51–59. [CrossRef][PubMed] 12. Shepherd, A.; Ivins, E.R.; Geruo, A.; Barletta, V.R.; Bentley, M.J.; Bettadpur, S.; Briggs, K.H.; Bromwich, D.H.; Forsberg, R.; Galin, N.; et al. A Reconciled Estimate of Ice-Sheet Mass Balance. Science 2012, 338, 1183–1189. [CrossRef][PubMed] 13. Rignot, E.; Velicogna, I.; van den Broeke, M.R.; Monaghan, A.; Lenaerts, J.T.M. Acceleration of the contribution of the Greenland and Antarctic ice sheets to sea level rise. Geophys. Res. Lett. 2011, 38, L05503. [CrossRef] 14. Bevis, M.; Harig, C.; Khan, S.A.; Brown, A.; Simons, F.J.; Willis, M.; Fettweis, X.; van den Broeke, M.R.; Madsen, F.B.; Kendrick, E.; et al. Accelerating changes in ice mass within Greenland, and the ice sheet’s sensitivity to atmospheric forcing. Proc. Natl. Acad. Sci. USA 2019, 116, 1934–1939. [CrossRef][PubMed] 15. Wouters, B.; Bamber, J.L.; van den Broeke, M.R.; Lenaerts, J.T.M.; Sasgen, I. Limits in detecting acceleration of ice sheet mass loss due to climate variability. Nat. Geosci. 2013, 6, 613–616. [CrossRef] 16. Kjeldsen, K.K.; Korsgaard, N.J.; Bjørk, A.A.; Khan, S.A.; Box, J.E.; Funder, S.; Larsen, N.K.; Bamber, J.L.; Colgan, W.; van den Broeke, M.; et al. Spatial and temporal distribution of mass loss from the Greenland Ice Sheet since AD 1900. Nature 2015, 528, 396–400. [CrossRef][PubMed] 17. van den Broeke, M.; Bamber, J.; Ettema, J.; Rignot, E.; Schrama, E.; van de Berg, W.J.; van Meijgaard, E.; Velicogna, I.; Wouters, B. Partitioning Recent Greenland Mass Loss. Science 2009, 326, 984–986. [CrossRef] 18. Enderlin, E.M.; Howat, I.M.; Jeong, S.; Noh, M.-J.; van Angelen, J.H.; van den Broeke, M.R. An improved mass budget for the Greenland ice sheet. Geophys. Res. Lett. 2014, 41, 866–872. [CrossRef] 19. Andersen, M.L.; Stenseng, L.; Skourup, H.; Colgan, W.; Khan, S.A.; Kristensen, S.S.; Andersen, S.B.; Box, J.E.; Ahlstrøm, A.P.; Fettweis, X.; et al. Basin-scale partitioning of Greenland ice sheet mass balance components (2007–2011). Earth Planet. Sci. Lett. 2015, 409, 89–95. [CrossRef] 20. Machguth, H.; MacFerrin, M.; van As, D.; Box, J.E.; Charalampidis, C.; Colgan, W.; Fausto, R.S.; Meijer, H.A.J.; Mosley-Thompson, E.; van de Wal, R.S.W. Greenland meltwater storage in firn limited by near-surface ice formation. Nat. Clim. Chang. 2016, 6, 390–393. [CrossRef] 21. Lenaerts, J.T.M.; Medley, B.; van den Broeke, M.R.; Wouters, B. Observing and Modeling Ice Sheet Surface Mass Balance. Rev. Geophys. 2019, 57, 376–420. [CrossRef][PubMed] 22. Fettweis, X.; Box, J.E.; Agosta, C.; Amory, C.; Kittel, C.; Lang, C.; van As, D.; Machguth, H.; Gallée, H. Reconstructions of the 1900–2015 Greenland ice sheet surface mass balance using the regional climate MAR model. Cryosphere 2017, 11, 1015–1033. [CrossRef] 23. Van Angelen, J.H.; van den Broeke, M.R.; Wouters, B.; Lenaerts, J.T.M. Contemporary (1960–2012) Evolution of the Climate and Surface Mass Balance of the Greenland Ice Sheet. Surv. Geophys. 2014, 35, 1155–1174. [CrossRef] 24. Hanna, E.; Cropper, T.E.; Hall, R.J.; Cappelen, J. Greenland Blocking Index 1851–2015: A regional climate change signal. Int. J. Climatol. 2016, 36, 4847–4861. [CrossRef] 25. Fettweis, X.; Hanna, E.; Lang, C.; Belleflamme, A.; Erpicum, M.; Gallée, H. Brief communication: Important role of the mid-tropospheric atmospheric circulation in the recent surface melt increase over the Greenland ice sheet. Cryosphere 2013, 7, 241–248. [CrossRef] 26. Ryan, J.C.; Smith, L.C.; van As, D.; Cooley, S.W.; Cooper, M.G.; Pitcher, L.H.; Hubbard, A. Greenland Ice Sheet surface melt amplified by snowline migration and bare ice exposure. Sci. Adv. 2019, 5, eaav3738. [CrossRef] 27. Hofer, S.; Tedstone, A.J.; Fettweis, X.; Bamber, J.L. Decreasing cloud cover drives the recent mass loss on the Greenland Ice Sheet. Sci. Adv. 2017, 3, e1700584. [CrossRef] 28. Lim, Y.-K.; Schubert, S.D.; Nowicki, S.M.J.; Lee, J.N.; Molod, A.M.; Cullather, R.I.; Zhao, B.; Velicogna, I. Atmospheric summer teleconnections and Greenland Ice Sheet surface mass variations: Insights from MERRA-2. Environ. Res. Lett. 2016, 11, 024002. [CrossRef] 29. Tedesco, M.; Fettweis, X.; van den Broeke, M.R.; van de Wal, R.S.W.; Smeets, C.J.P.P.; van de Berg, W.J.; Serreze, M.C.; Box, J.E. The role of albedo and accumulation in the 2010 melting record in Greenland. Environ. Res. Lett. 2011, 6, 014005. [CrossRef] 30. Box, J.E.; Fettweis, X.; Stroeve, J.C.; Tedesco, M.; Hall, D.K.; Steffen, K. Greenland ice sheet albedo feedback: Thermodynamics and atmospheric drivers. Cryosphere 2012, 6, 821–839. [CrossRef] 31. Fürst, J.J.; Goelzer, H.; Huybrechts, P. Ice-dynamic projections of the Greenland ice sheet in response to atmospheric and oceanic warming. Cryosphere 2015, 9, 1039–1062. [CrossRef] Remote Sens. 2019, 11, 2405 36 of 50

32. Franco, B.; Fettweis, X.; Erpicum, M. Future projections of the Greenland ice sheet energy balance driving the surface melt. Cryosphere 2013, 7, 1–18. [CrossRef] 33. Steger, C.R.; Reijmer, C.H.; van den Broeke, M.R. The modelled liquid water balance of the Greenland Ice Sheet. Cryosphere 2017, 11, 2507–2526. [CrossRef] 34. Bader, H. The Greenland Ice Sheet; Cold Regions Research and Engineering Lab: Hanover, NH, USA, 1961; pp. 1–17. 35. Benson, C.S. Stratigraphic Studies in the Snow and Firn of the Greenland Ice Sheet; Cold Regions Research and Engineering Lab: Hanover, NH, USA, 1962; pp. 13–37. 36. Pritchard, H.D.; Luthcke, S.B.; Fleming, A.H. Understanding ice-sheet mass balance: Progress in satellite altimetry and gravimetry. J. Glaciol. 2011, 56, 1151–1161. [CrossRef] 37. Brooks, R.L.; Campbell, W.J.; Ramseier, R.O.; Stanley, H.R.; Zwally, H.J. Ice sheet topography by satellite altimetry. Nature 1978, 274, 539–543. [CrossRef] 38. Sørensen, L.S.; Simonsen, S.B.; Forsberg, R.; Khvorostovsky, K.; Meister, R.; Engdahl, M.E. 25 years of elevation changes of the Greenland Ice Sheet from ERS, Envisat, and CryoSat-2 radar altimetry. Earth Planet. Sci. Lett. 2018, 495, 234–241. [CrossRef] 39. Brenner, A.C.; DiMarzio, J.P.; Zwally, H.J. Precision and Accuracy of Satellite Radar and Laser Altimeter Data Over the Continental Ice Sheets. IEEE Trans. Geosci. Remote Sens. 2007, 45, 321–331. [CrossRef] 40. Abdalati, W.; Zwally, H.J.; Bindschadler, R.; Csatho, B.; Farrell, S.L.; Fricker, H.A.; Harding, D.; Kwok, R.; Lefsky, M.; Markus, T.; et al. The ICESat-2 Laser Altimetry Mission. Proc. IEEE 2010, 98, 735–751. [CrossRef] 41. Drinkwater, M.R.; Francis, R.; Ratier, G.; Wingham, D.J. The European Space Agency’s Earth Explorer Mission CryoSat: Measuring variability in the cryosphere. Ann. Glaciol. 2004, 39, 313–320. [CrossRef] 42. Flechtner, F.; Neumayer, K.-H.; Dahle, C.; Dobslaw, H.; Fagiolini, E.; Raimondo, J.-C.; Güntner, A. What can be Expected from the GRACE-FO Laser Ranging Interferometer for Earth Science Applications? In Remote Sensing and Water Resources; Cazenave, A., Champollion, N., Benveniste, J., Chen, J., Eds.; Space Sciences Series of ISSI; Springer International Publishing: Cham, Switzerland, 2016; pp. 263–280. ISBN 978-3-319-32449-4. 43. Turyshev, S.G.; Sazhin, M.V.; Toth, V.T. General relativistic laser interferometric observables of the GRACE-Follow-On mission. Phys. Rev. D 2014, 89, 105029. [CrossRef] 44. Bjørk, A.A.; Kjær, K.H.; Korsgaard, N.J.; Khan, S.A.; Kjeldsen, K.K.; Andresen, C.S.; Box, J.E.; Larsen, N.K.; Funder, S. An aerial view of 80 years of climate-related glacier fluctuations in southeast Greenland. Nat. Geosci. 2012, 5, 427–432. [CrossRef] 45. Gallaher, D.; Campbell, G.G.; Meier, W.; Moses, J.; Wingo, D. The process of bringing dark data to light: The rescue of the early Nimbus satellite data. GeoResJ 2015, 6, 124–134. [CrossRef] 46. Bindschadler, R. Monitoring ice sheet behavior from space. Rev. Geophys. 1998, 36, 79–104. [CrossRef] 47. Freden, S.C.M. Significant Results Obtained from the Earth Resources Technology Satellite-1. Volume 1: Technical presentations, sections A and B. In Proceedings of the Symposium on Significant Results Obtained from the Earth Resources Technology Satellite-1, New Carrollton, MD, USA, 5–9 March 1973. 48. He, T.; Liang, S.; Yu, Y.; Wang, D.; Gao, F.; Liu, Q. Greenland surface albedo changes in July 1981–2012 from satellite observations. Environ. Res. Lett. 2013, 8, 044043. [CrossRef] 49. Key, J.; Wang, X.; Liu, Y.; Dworak, R.; Letterly, A. The AVHRR Polar Pathfinder Climate Data Records. Remote Sens. 2016, 8, 167. [CrossRef] 50. Fahnestock, M.; Bindschadler, R.; Kwok, R.; Jezek, K. Greenland Ice Sheet Surface Properties and Ice Dynamics from ERS-1 SAR Imagery. Science 1993, 262, 1530–1534. [CrossRef] 51. Tedstone, A.J.; Bamber, J.L.; Cook, J.M.; Williamson, C.J.; Fettweis, X.; Hodson, A.J.; Tranter, M. Dark ice dynamics of the south-west Greenland Ice Sheet. Cryosphere 2017, 11, 2491–2506. [CrossRef] 52. Wientjes, I.G.M.; Oerlemans, J. An explanation for the dark region in the western melt zone of the Greenland ice sheet. Cryosphere 2010, 4, 261–268. [CrossRef] 53. Nolin, A.W.; Payne, M.C. Classification of glacier zones in western Greenland using albedo and surface roughness from the Multi-angle Imaging SpectroRadiometer (MISR). Remote Sens. Environ. 2007, 107, 264–275. [CrossRef] 54. Yang, K.; Smith, L.C. Supraglacial Streams on the Greenland Ice Sheet Delineated From Combined Spectral-Shape Information in High-Resolution Satellite Imagery. IEEE Geosci. Remote Sens. Lett. 2013, 10, 801–805. [CrossRef] Remote Sens. 2019, 11, 2405 37 of 50

55. Hall, D.K.; Williams, R.S.; Casey, K.A.; DiGirolamo, N.E.; Wan, Z. Satellite-derived, melt-season surface temperature of the Greenland Ice Sheet (2000–2005) and its relationship to mass balance. Geophys. Res. Lett. 2006, 33, L11501. [CrossRef] 56. Gledhill, L.A.; Williamson, A.G. Inland advance of supraglacial lakes in north-west Greenland under recent climatic warming. Ann. Glaciol. 2018, 59, 66–82. [CrossRef] 57. Legleiter, C.J.; Tedesco, M.; Smith, L.C.; Behar, A.E.; Overstreet, B.T. Mapping the bathymetry of supraglacial lakes and streams on the Greenland ice sheet using field measurements and high-resolution satellite images. Cryosphere 2014, 8, 215–228. [CrossRef] 58. Ryan, J.C.; Hubbard, A.L.; Stibal, M.; Irvine-Fynn, T.D.; Cook, J.; Smith, L.C.; Cameron, K.; Box, J.E. Dark zone of the Greenland Ice Sheet controlled by distributed biologically-active impurities. Nat. Commun. 2018, 9, 1065. [CrossRef][PubMed] 59. Bhardwaj, A.; Sam, L.; Akanksha; Martín-Torres, F.J.; Kumar, R. UAVs as remote sensing platform in glaciology: Present applications and future prospects. Remote Sens. Environ. 2016, 175, 196–204. [CrossRef] 60. Van den Broeke, M.; Box, J.; Fettweis, X.; Hanna, E.; Noël, B.; Tedesco, M.; van As, D.; van de Berg, W.J.; van Kampenhout, L. Greenland Ice Sheet Surface Mass Loss: Recent Developments in Observation and Modeling. Curr. Clim. Chang. Rep. 2017, 3, 345–356. [CrossRef] 61. Quincey, D.J.; Luckman, A. Progress in satellite remote sensing of ice sheets. Prog. Phys. Geogr. 2009, 33, 547–567. [CrossRef] 62. Rémy, F.; Parouty, S. Antarctic Ice Sheet and Radar Altimetry: A Review. Remote Sens. 2009, 1, 1212–1239. [CrossRef] 63. Long, D.G. Polar Applications of Spaceborne Scatterometers. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2017, 10, 2307–2320. [CrossRef] 64. Hall, D.K. Assessment of polar climate change using satellite technology. Rev. Geophys. 1988, 26, 26–39. [CrossRef] 65. König, M.; Winther, J.-G.; Isaksson, E. Measuring snow and glacier ice properties from satellite. Rev. Geophys. 2001, 39, 1–27. [CrossRef] 66. Bamber, J.L.; Rivera, A. A review of remote sensing methods for glacier mass balance determination. Glob. Planet. Chang. 2007, 59, 138–148. [CrossRef] 67. Racoviteanu, A.E.; Williams, M.W.; Barry, R.G. Optical Remote Sensing of Glacier Characteristics: A Review with Focus on the Himalaya. Sensors 2008, 8, 3355–3383. [CrossRef][PubMed] 68. Pitcher, L.H.; Smith, L.C. Supraglacial Streams and Rivers. Annu. Rev. Earth Planet. Sci. 2019, 47, 421–452. [CrossRef] 69. Chu, V.W. Greenland ice sheet hydrology: A review. Prog. Phys. Geogr. Earth Environ. 2014, 38, 19–54. [CrossRef] 70. Rennermalm, A.K.; Moustafa, S.E.; Mioduszewski, J.; Chu, V.W.; Forster, R.R.; Hagedorn, B.; Harper, J.T.; Mote, T.L.; Robinson, D.A.; Shuman, C.A.; et al. Understanding Greenland ice sheet hydrology using an integrated multi-scale approach. Environ. Res. Lett. 2013, 8, 015017. [CrossRef] 71. Irvine-Fynn, T.D.L.; Hodson, A.J.; Moorman, B.J.; Vatne, G.; Hubbard, A.L. Polythermal Glacier Hydrology: A Review. Rev. Geophys. 2011, 49, RG4002. [CrossRef] 72. Nolin, A.W. Recent advances in remote sensing of seasonal snow. J. Glaciol. 2010, 56, 1141–1150. [CrossRef] 73. Gardner, A.S.; Sharp, M.J. A review of snow and ice albedo and the development of a new physically based broadband albedo parameterization. J. Geophys. Res. 2010, 115, F01009. [CrossRef] 74. Bell, R.E. The role of subglacial water in ice-sheet mass balance. Nat. Geosci. 2008, 1, 297–304. [CrossRef] 75. Warren, S.G. Optical properties of snow. Rev. Geophys. 1982, 20, 67–89. [CrossRef] 76. Carsey, F. Remote sensing of ice and snow: Review and status. Int. J. Remote Sens. 1992, 13, 5–11. [CrossRef] 77. Deems, J.S.; Painter, T.H.; Finnegan, D.C. Lidar measurement of snow depth: A review. J. Glaciol. 2013, 59, 467–479. [CrossRef] 78. Frei, A.; Tedesco, M.; Lee, S.; Foster, J.; Hall, D.K.; Kelly, R.; Robinson, D.A. A review of global satellite-derived snow products. Adv. Space Res. 2012, 50, 1007–1029. [CrossRef] 79. Richter-Menge, J.A.; Colbeck, S.C.; Jezek, K.C. Recent Progress in Snow and Ice Research. Rev. Geophys. 1991, 29, 218–226. [CrossRef] 80. Dietz, A.J.; Kuenzer, C.; Gessner, U.; Dech, S. Remote sensing of snow—A review of available methods. Int. J. Remote Sens. 2012, 33, 4094–4134. [CrossRef] Remote Sens. 2019, 11, 2405 38 of 50

81. Dozier, J.; Painter, T.H. Multispectral and Hyperspectral Remote Sensing of Alpine Snow Properties. Ann. Rev. Earth Planet. Sci. 2004, 32, 465–494. [CrossRef] 82. Box, J.E.; Steffen, K. Sublimation on the Greenland Ice Sheet from automated weather station observations. J. Geophys. Res. Atmos. 2001, 106, 33965–33981. [CrossRef] 83. Forster, R.R.; Box, J.E.; van den Broeke, M.R.; Miège, C.; Burgess, E.W.; van Angelen, J.H.; Lenaerts, J.T.M.; Koenig, L.S.; Paden, J.; Lewis, C.; et al. Extensive liquid meltwater storage in firn within the Greenland ice sheet. Nature Geosci 2014, 7, 95–98. [CrossRef] 84. Rennermalm, A.K.; Smith, L.C.; Chu, V.W.; Box, J.E.; Forster, R.R.; Van den Broeke, M.R.; Van As, D.; Moustafa, S.E. Evidence of meltwater retention within the Greenland ice sheet. Cryosphere 2013, 7, 1433–1445. [CrossRef] 85. Cooper, M.G.; Smith, L.C.; Rennermalm, A.K.; Miège, C.; Pitcher, L.H.; Ryan, J.C.; Yang, K.; Cooley, S.W. Meltwater storage in low-density near-surface bare ice in the Greenland ice sheet ablation zone. Cryosphere 2018, 12, 955–970. [CrossRef] 86. Sørensen, L.S.; Simonsen, S.B.; Nielsen, K.; Lucas-Picher, P.; Spada, G.; Adalgeirsdottir, G.; Forsberg, R.; Hvidberg, C.S. Mass balance of the Greenland ice sheet (2003–2008) from ICESat data—The impact of interpolation, sampling and firn density. Cryosphere 2011, 5, 173–186. [CrossRef] 87. Zwally, H.J.; Li, J.; Brenner, A.C.; Beckley, M.; Cornejo, H.G.; DiMarzio, J.; Giovinetto, M.B.; Neumann, T.A.; Robbins, J.; Saba, J.L.; et al. Greenland ice sheet mass balance: Distribution of increased mass loss with climate warming; 2003–07 versus 1992–2002. J. Glaciol. 2011, 57, 88–102. [CrossRef] 88. Cuffey, K.M.; Paterson, W.S.B. The Physics of Glaciers, 4th ed.; Academic Press: Amsterdam, The Nederlands, 2010. 89. Sutterley, T.C.; Velicogna, I.; Fettweis, X.; Rignot, E.; Noël, B.; van den Broeke, M.R. Evaluation of reconstructions of snow/ice melt in Greenland by regional atmospheric climate models using laser altimetry data. Geophys. Res. Lett. 2018, 45, 8324–8333. [CrossRef] 90. Van den Broeke, M.; Smeets, P.; Ettema, J.; van der Veen, C.; van de Wal, R.; Oerlemans, J. Partitioning of melt energy and meltwater fluxes in the ablation zone of the west Greenland ice sheet. Cryosphere 2008, 2, 179–189. [CrossRef] 91. Zwally, H.J.; Bindschadler, R.A.; Brenner, A.C.; Martin, T.V.; Thomas, R.H. Surface elevation contours of Greenland and Antarctic Ice Sheets. J. Geophys. Res. Oceans 1983, 88, 1589–1596. [CrossRef] 92. Zwally, H.J.; Bindschadler, R.A.; Brenner, A.C.; Major, J.A.; Marsh, J.G. Growth of Greenland Ice Sheet: Measurement. Science 1989, 246, 1587–1589. [CrossRef] 93. Helm, V.; Humbert, A.; Miller, H. Elevation and elevation change of Greenland and Antarctica derived from CryoSat-2. Cryosphere 2014, 8, 1539–1559. [CrossRef] 94. Khvorostovsky, K.S. Merging and Analysis of Elevation Time Series Over Greenland Ice Sheet From Satellite Radar Altimetry. IEEE Trans. Geosci. Remote Sens. 2012, 50, 23–36. [CrossRef] 95. Thomas, R.; Davis, C.; Frederick, E.; Krabill, W.; Li, Y.; Manizade, S.; Martin, C. A comparison of Greenland ice-sheet volume changes derived from altimetry measurements. J. Glaciol. 2008, 54, 203–212. [CrossRef] 96. Bolch, T.; Sørensen, L.S.; Simonsen, S.B.; Mölg, N.; Machguth, H.; Rastner, P.; Paul, F. Mass loss of Greenland’s glaciers and ice caps 2003–2008 revealed from ICESat laser altimetry data. Geophys. Res. Lett. 2013, 40, 875–881. [CrossRef] 97. Csatho, B.M.; Schenk, A.F.; van der Veen, C.J.; Babonis, G.; Duncan, K.; Rezvanbehbahani, S.; van den Broeke, M.R.; Simonsen, S.B.; Nagarajan, S.; van Angelen, J.H. Laser altimetry reveals complex pattern of Greenland Ice Sheet dynamics. Proc. Natl. Acad. Sci. USA 2014, 111, 18478–18483. [CrossRef][PubMed] 98. Ewert, H.; Groh, A.; Dietrich, R. Volume and mass changes of the Greenland ice sheet inferred from ICESat and GRACE. J. Geodyn. 2012, 59–60, 111–123. [CrossRef] 99. Felikson, D.; Urban, T.J.; Gunter, B.C.; Pie, N.; Pritchard, H.D.; Harpold, R.; Schutz, B.E. Comparison of Elevation Change Detection Methods From ICESat Altimetry Over the Greenland Ice Sheet. IEEE Trans. Geosci. Remote Sens. 2017, 55, 5494–5505. [CrossRef] 100. Slobbe, D.C.; Lindenbergh, R.C.; Ditmar, P. Estimation of volume change rates of Greenland’s ice sheet from ICESat data using overlapping footprints. Remote Sens. Environ. 2008, 112, 4204–4213. [CrossRef] 101. Zwally, H.J.; Schutz, B.; Abdalati, W.; Abshire, J.; Bentley, C.; Brenner, A.; Bufton, J.; Dezio, J.; Hancock, D.; Harding, D.; et al. ICESat’s laser measurements of polar ice, atmosphere, ocean, and land. J. Geodyn. 2002, 34, 405–445. [CrossRef] Remote Sens. 2019, 11, 2405 39 of 50

102. Sole, A.; Payne, T.; Bamber, J.; Nienow, P.; Krabill, W. Testing hypotheses of the cause of peripheral thinning of the Greenland Ice Sheet: Is land-terminating ice thinning at anomalously high rates? Cryosphere 2008, 2, 205–218. [CrossRef] 103. Pritchard, H.D.; Arthern, R.J.; Vaughan, D.G.; Edwards, L.A. Extensive dynamic thinning on the margins of the Greenland and Antarctic ice sheets. Nature 2009, 461, 971–975. [CrossRef] 104. Zwally, H.J.; Giovinetto, M.B.; Li, J.; Cornejo, H.G.; Beckley, M.A.; Brenner, A.C.; Saba, J.L.; Yi, D. Mass changes of the Greenland and Antarctic ice sheets and shelves and contributions to sea-level rise: 1992–2002. J. Glaciol. 2005, 51, 509–527. [CrossRef] 105. McMillan, M.; Leeson, A.; Shepherd, A.; Briggs, K.; Armitage, T.W.K.; Hogg, A.; Kuipers Munneke, P.; Broeke, M.; Noël, B.; Berg, W.J.; et al. A high-resolution record of Greenland mass balance. Geophys. Res. Lett. 2016, 43, 7002–7010. [CrossRef] 106. Martin, T.V.; Zwally, H.J.; Brenner, A.C.; Bindschadler, R.A. Analysis and retracking of continental ice sheet radar altimeter waveforms. J. Geophys. Res. Oceans 1983, 88, 1608–1616. [CrossRef] 107. Ekholm, S.; Forsberg, R.; Brozena, J.M. Accuracy of satellite altimeter elevations over the Greenland Ice Sheet. J. Geophys. Res. Oceans 1995, 100, 2687–2696. [CrossRef] 108. Brenner, A.C.; Blndschadler, R.A.; Thomas, R.H.; Zwally, H.J. Slope-induced errors in radar altimetry over continental ice sheets. J. Geophys. Res. Oceans 1983, 88, 1617–1623. [CrossRef] 109. Bamber, J.L.; Ekholm, S.; Krabill, W. The accuracy of satellite radar altimeter data over the Greenland Ice Sheet determined from airborne laser data. Geophys. Res. Lett. 1998, 25, 3177–3180. [CrossRef] 110. Mätzler, C.; Wegmüller, U. Dielectric properties of freshwater ice at microwave frequencies. J. Phys. D Appl. Phys. 1987, 20, 1623–1630. [CrossRef] 111. Ridley, J.K.; Partington, K.C. A model of satellite radar altimeter return from ice sheets. Int. J. Remote Sens. 1988, 9, 601–624. [CrossRef] 112. Davis, C.H. The effect of sub-surface volume scattering on the accuracy of ice-sheet altimeter retracking algorithms. In Proceedings of the IGARSS ’93-IEEE International Geoscience and Remote Sensing Symposium, Tokyo, Japan, 18–21 August 1993; Volume 3, pp. 1053–1057. 113. Davis, C.H. Temporal change in the extinction coefficient of snow on the Greenland ice sheet from an analysis of Seasat and Geosat altimeter data. IEEE Trans. Geosci. Remote Sens. 1996, 34, 1066–1073. [CrossRef] 114. Davis, C.H.; Zwally, H.J. Geographic and seasonal variations in the surface properties of the ice sheets by satellite-radar altimetry. J. Glaciol. 1993, 39, 687–697. [CrossRef] 115. Nghiem, S.V.; Steffen, K.; Neumann, G.; Huff, R. Mapping of ice layer extent and snow accumulation in the percolation zone of the Greenland ice sheet. J. Geophys. Res. Earth Surf. 2005, 110, F02017. [CrossRef] 116. Nilsson, J.; Vallelonga, P.; Simonsen, S.B.; Sørensen, L.S.; Forsberg, R.; Dahl-Jensen, D.; Hirabayashi, M.; Goto-Azuma, K.; Hvidberg, C.S.; Kjær, H.A.; et al. Greenland 2012 melt event effects on CryoSat-2 radar altimetry. Geophys. Res. Lett. 2015, 42, 3919–3926. [CrossRef] 117. Scott, J.B.T.; Nienow, P.; Mair, D.; Parry, V.; Morris, E.; Wingham, D.J. Importance of seasonal and annual layers in controlling backscatter to radar altimeters across the percolation zone of an ice sheet. Geophys. Res. Lett. 2006, 33, L24502. [CrossRef] 118. Sørensen, L.S.; Simonsen, S.B.; Meister, R.; Forsberg, R.; Levinsen, J.F.; Flament, T. Envisat-derived elevation changes of the Greenland ice sheet, and a comparison with ICESat results in the accumulation area. Remote Sens. Environ. 2015, 160, 56–62. [CrossRef] 119. Gray, L.; Burgess, D.; Copland, L.; Dunse, T.; Langley, K.; Moholdt, G. A revised calibration of the interferometric mode of the CryoSat-2 radar altimeter improves ice height and height change measurements in western Greenland. Cryosphere 2017, 11, 1041–1058. [CrossRef] 120. Gourmelen, N.; Escorihuela, M.J.; Shepherd, A.; Foresta, L.; Muir, A.; Garcia-Mondéjar, A.; Roca, M.; Baker, S.G.; Drinkwater, M.R. CryoSat-2 swath interferometric altimetry for mapping ice elevation and elevation change. Adv. Space Res. 2018, 62, 1226–1242. [CrossRef] 121. Partington, K.C.; Ridley, J.K.; Rapley, C.G.; Zwally, H.J. Observations of the Surface Properties of the Ice Sheets by Satellite Radar Altimetry. J. Glaciol. 1989, 35, 267–275. [CrossRef] 122. Thomas, R.H.; Martin, T.V.; Zwally, H.J. Mapping Ice-Sheet Margins from Radar Altimetry Data. Ann. Glaciol. 1983, 4, 283–288. [CrossRef] 123. Zwally, H.J. Growth of Greenland Ice Sheet: Interpretation. Science 1989, 246, 1589–1591. [CrossRef] [PubMed] Remote Sens. 2019, 11, 2405 40 of 50

124. Davis, C.H. A surface and volume scattering retracking algorithm for ice sheet satellite altimetry. IEEE Trans. Geosci. Remote Sens. 1993, 31, 811–818. [CrossRef] 125. Davis, C.H. Growth of the Greenland ice sheet: A performance assessment of altimeter retracking algorithms. IEEE Trans. Geosci. Remote Sens. 1995, 33, 1108–1116. [CrossRef] 126. Davis, C.H. A robust threshold retracking algorithm for measuring ice-sheet surface elevation change from satellite radar altimeters. IEEE Trans. Geosci. Remote Sens. 1997, 35, 974–979. [CrossRef] 127. Davis, C.H.; Kluever, C.A.; Haines, B.J. Elevation Change of the Southern Greenland Ice Sheet. Science 1998, 279, 2086–2088. [CrossRef][PubMed] 128. Ekholm, S. A full coverage, high-resolution, topographic model of Greenland computed from a variety of digital elevation data. J. Geophys. Res. Solid Earth 1996, 101, 21961–21972. [CrossRef] 129. Bamber, J.L.; Ekholm, S.; Krabill, W.B. A new, high-resolution digital elevation model of Greenland fully validated with airborne laser altimeter data. J. Geophys. Res. Solid Earth 2001, 106, 6733–6745. [CrossRef] 130. Johannessen, O.M.; Khvorostovsky, K.; Miles, M.W.; Bobylev, L.P. Recent Ice-Sheet Growth in the Interior of Greenland. Science 2005, 310, 1013–1016. [CrossRef][PubMed] 131. Khvorostovsky, K.S.; Johannessen, O.M. Merging of ERS-1, ERS-2 and Envisat Altimeter Data over the Greenland Ice Sheet; Nansen Environmental and Remote Sensing Center: Bergen, Norway, 2009; pp. 1–34. 132. Arthern, R.J.; Wingham, D.J.; Ridout, A.L. Controls on ERS altimeter measurements over ice sheets: Footprint-scale topography, backscatter fluctuations, and the dependence of microwave penetration depth on satellite orientation. J. Geophys. Res. Atmos. 2001, 106, 33471–33484. [CrossRef] 133. Levinsen, J.F.; Khvorostovsky, K.; Ticconi, F.; Shepherd, A.; Forsberg, R.; Sørensen, L.S.; Muir, A.; Pie, N.; Felikson, D.; Flament, T.; et al. ESA ice sheet CCI: Derivation of the optimal method for surface elevation change detection of the Greenland ice sheet—Round robin results. Int. J. Remote Sens. 2015, 36, 551–573. [CrossRef] 134. Su, X.; Shum, C.K.; Kuo, C.; Yi, Y. Improved Envisat Altimetry Ice Sheet Elevation Change Data Processing Algorithms Using Repeat-Track Analysis. IEEE Geosci. Remote Sens. Lett. 2016, 13, 1099–1103. [CrossRef] 135. Su, X.; Luo, Z.; Zhou, Z. Assessing backscatter change due to backscatter gradient over the Greenland ice sheet using Envisat and SARAL altimetry. J. Geodyn. 2018, 117, 41–48. [CrossRef] 136. Wingham, D.J.; Francis, C.R.; Baker, S.; Bouzinac, C.; Brockley, D.; Cullen, R.; de Chateau-Thierry, P.; Laxon, S.W.; Mallow, U.; Mavrocordatos, C.; et al. CryoSat: A mission to determine the fluctuations in Earth’s land and marine ice fields. Adv. Space Res. 2006, 37, 841–871. [CrossRef] 137. Gray, L.; Burgess, D.; Copland, L.; Cullen, R.; Galin, N.; Hawley, R.; Helm, V. Interferometric swath processing of Cryosat data for glacial ice topography. Cryosphere 2013, 7, 1857–1867. [CrossRef] 138. Parrinello, T.; Shepherd, A.; Bouffard, J.; Badessi, S.; Casal, T.; Davidson, M.; Fornari, M.; Maestroni, E.; Scagliola, M. CryoSat: ESA’s ice mission—Eight years in space. Adv. Space Res. 2018, 62, 1178–1190. [CrossRef] 139. Verron, J.; Sengenes, P.; Lambin, J.; Noubel, J.; Steunou, N.; Guillot, A.; Picot, N.; Coutin-Faye, S.; Sharma, R.; Gairola, R.M.; et al. The SARAL/AltiKa Altimetry Satellite Mission. Mar. Geod. 2015, 38, 2–21. [CrossRef] 140. Rémy, F.; Flament, T.; Michel, A.; Verron, J. Ice sheet survey over Antarctica using satellite altimetry: ERS-2, Envisat, SARAL/AltiKa, the key importance of continuous observations along the same repeat orbit. Int. J. Remote Sens. 2014, 35, 5497–5512. [CrossRef] 141. Rémy, F.; Flament, T.; Michel, A.; Blumstein, D. Envisat and SARAL/AltiKa Observations of the Antarctic Ice Sheet: A Comparison between the Ku-band and Ka-band. Mar. Geod. 2015, 38, 510–521. [CrossRef] 142. Chander, S.; Mishra, S.K.; Chauhan, P.; Ajai. Ice Height and Backscattering Coefficient Variability over Greenland Ice Sheets Using SARAL Radar Altimeter. Mar. Geod. 2015, 38, 466–476. [CrossRef] 143. Rajkumar, K.S.; Maheshwari, M.; Pallipad, J.; Rajak, D.R.; Kumar, R.; Oza, S.R. Concurrent Use of OSCAT and AltiKa to Characterize Antarctic Ice Surface Features. Mar. Geod. 2015, 38, 497–509. [CrossRef] 144. DiMarzio, J.P. GLAS/ICESat 1 km Laser Altimetry Digital Elevation Model of Greenland; Version 1; NSIDC National Snow and Ice Data Center: Boulder, CO, USA, 2007. 145. Suryawanshi, M.R.; Chander, S.; Oza, S.R.; Bahuguna, I.M. Volume loss of the Greenland ice sheet revealed by SARAL/AltiKa repeat passes radar altimetry. J. Earth Syst. Sci. 2019, 128, 187. [CrossRef] 146. Helm, V.; Humbert, A.; Miller, H. Elevation Models of Antarctica and Greenland derived from CryoSat-2 in the period 2011 to 2013. Supplement to: Helm, V. Elevation and elevation change of Greenland and Antarctica derived from CryoSat-2. Cryosphere 2014, 8, 1539–1559. [CrossRef] Remote Sens. 2019, 11, 2405 41 of 50

147. Gray, L.; Burgess, D.; Copland, L.; Demuth, M.N.; Dunse, T.; Langley, K.; Schuler, T.V. CryoSat-2 delivers monthly and inter-annual surface elevation change for Arctic ice caps. Cryosphere 2015, 9, 1895–1913. [CrossRef] 148. Schröder, L.; Horwath, M.; Dietrich, R.; Helm, V.; van den Broeke, M.R.; Ligtenberg, S.R.M. Four decades of Antarctic surface elevation changes from multi-mission satellite altimetry. Cryosphere 2019, 13, 427–449. [CrossRef] 149. Davis, C.H.; Kluever, C.A.; Haines, B.J.; Perez, C.; Yoon, Y.T. Improved elevation-change measurement of the southern Greenland ice sheet from satellite radar altimetry. IEEE Trans. Geosci. Remote Sens. 2000, 38, 1367–1378. [CrossRef] 150. Alley, R.B.; Spencer, M.K.; Anandakrishnan, S. Ice-sheet mass balance: Assessment, attribution and prognosis. Ann. Glaciol. 2007, 46, 1–7. [CrossRef] 151. Rott, H.; Mätzler, C. Possibilities and Limits of Synthetic Aperture Radar for Snow and Glacier Surveying. Ann. Glaciol. 1987, 9, 195–199. [CrossRef] 152. Rignot, E.; Echelmeyer, K.; Krabill, W. Penetration depth of interferometric synthetic-aperture radar signals in snow and ice. Geophys. Res. Lett. 2001, 28, 3501–3504. [CrossRef] 153. Pettersson, R.; Jansson, P.; Blatter, H. Spatial variability in water content at the cold-temperate transition surface of the polythermal Storglaciären, Sweden. J. Geophys. Res. 2004, 109, F02009. [CrossRef] 154. Irvine-Fynn, T.D.L.; Moorman, B.J.; Williams, J.L.M.; Walter, F.S.A. Seasonal changes in ground-penetrating radar signature observed at a polythermal glacier, Bylot Island, Canada. Earth Surf. Proc. Landf. 2006, 31, 892–909. [CrossRef] 155. Warren, S.G.; Brandt, R.E. Optical constants of ice from the ultraviolet to the microwave: A revised compilation. J. Geophys. Res. 2008, 113, D14220. [CrossRef] 156. Kendrick, A.K.; Schroeder, D.M.; Chu, W.; Young, T.J.; Christoffersen, P.; Todd, J.; Doyle, S.H.; Box, J.E.; Hubbard, A.; Hubbard, B.; et al. Surface Meltwater Impounded by Seasonal Englacial Storage in West Greenland. Geophys. Res. Lett. 2018, 45, 10474–10481. [CrossRef] 157. Guerreiro, K.; Fleury, S.; Zakharova, E.; Rémy, F.; Kouraev, A. Potential for estimation of snow depth on Arctic sea ice from CryoSat-2 and SARAL/AltiKa missions. Remote Sens. Environ. 2016, 186, 339–349. [CrossRef] 158. Vandecrux, B.; MacFerrin, M.; Machguth, H.; Colgan, W.T.; van As, D.; Heilig, A.; Stevens, C.M.; Charalampidis, C.; Fausto, R.S.; Morris, E.M.; et al. Firn data compilation reveals widespread decrease of firn air content in western Greenland. Cryosphere 2019, 13, 845–859. [CrossRef] 159. Van As, D.; Box, J.E.; Fausto, R.S. Challenges of Quantifying Meltwater Retention in Snow and Firn: An Expert Elicitation. Front. Earth Sci. 2016, 4, 101. [CrossRef] 160. Brenner, A.C.; Zwally, H.J.; Bentley, C.R.; Csatho, B.M.; Harding, D.J.; Hofton, M.A.; Minster, J.-B.; Roberts, L.; Saba, J.L.; Thomas, R.H.; et al. The Algorithm Theoretical Basis Document for the Derivation of Range and Range Distributions from Laser Pulse Waveform Analysis for Surface Elevations, Roughness, Slope, and Vegetation Heights; Goddard Space Flight Center: Greenbelt, MD, USA, 2012; pp. 1–134. 161. Howat, I.M.; Smith, B.E.; Joughin, I.; Scambos, T.A. Rates of southeast Greenland ice volume loss from combined ICESat and ASTER observations. Geophys. Res. Lett. 2008, 35, L17505. [CrossRef] 162. Winker, D.M.; Couch, R.H.; McCormick, M.P. An overview of LITE: NASA’s Lidar In-space Technology Experiment. Proc. IEEE 1996, 84, 164–180. [CrossRef] 163. Shuman, C.A.; Zwally, H.J.; Schutz, B.E.; Brenner, A.C.; DiMarzio, J.P.; Suchdeo, V.P.; Fricker, H.A. ICESat Antarctic elevation data: Preliminary precision and accuracy assessment. Geophys. Res. Lett. 2006, 33, L07501. [CrossRef] 164. Fricker, H.A.; Borsa, A.; Minster, B.; Carabajal, C.; Quinn, K.; Bills, B. Assessment of ICESat performance at the salar de Uyuni, Bolivia. Geophys. Res. Lett. 2005, 32, L21S06. [CrossRef] 165. Siegfried, M.R.; Hawley, R.L.; Burkhart, J.F. High-Resolution Ground-Based GPS Measurements Show Intercampaign Bias in ICESat Elevation Data Near Summit, Greenland. IEEE Trans. Geosci. Remote Sens. 2011, 49, 3393–3400. [CrossRef] 166. Abshire, J.B.; Sun, X.; Riris, H.; Sirota, J.M.; McGarry, J.F.; Palm, S.; Yi, D.; Liiva, P. Geoscience Laser Altimeter System (GLAS) on the ICESat Mission: On-orbit measurement performance. Geophys. Res. Lett. 2005, 32, L21S02. [CrossRef] 167. Schutz, B.E.; Zwally, H.J.; Shuman, C.A.; Hancock, D.; DiMarzio, J.P. Overview of the ICESat Mission. Geophys. Res. Lett. 2005, 32, L21S01. [CrossRef] Remote Sens. 2019, 11, 2405 42 of 50

168. Borsa, A.A.; Moholdt, G.; Fricker, H.A.; Brunt, K.M. A range correction for ICESat and its potential impact on ice-sheet mass balance studies. Cryosphere 2014, 8, 345–357. [CrossRef] 169. Ewert, H.; Popov, S.V.; Richter, A.; Schwabe, J.; Scheinert, M.; Dietrich, R. Precise analysis of ICESat altimetry data and assessment of the hydrostatic equilibrium for subglacial Lake Vostok, East Antarctica. Geophys. J. Int. 2012, 191, 557–568. [CrossRef] 170. Luthcke, S.B.; Rowlands, D.D.; Williams, T.A.; Sirota, M. Reduction of ICESat systematic geolocation errors and the impact on ice sheet elevation change detection. Geophys. Res. Lett. 2005, 32, L21S05. [CrossRef] 171. Schenk, T.; Csatho, B. A New Methodology for Detecting Ice Sheet Surface Elevation Changes From Laser Altimetry Data. IEEE Trans. Geosci. Remote Sens. 2012, 50, 3302–3316. [CrossRef] 172. Alberti, M.; Biscaro, D. Height variation detection in polar regions from ICESat satellite altimetry. Comput. Geosci. 2010, 36, 1–9. [CrossRef] 173. Yi, D.; Zwally, H.J.; Sun, X. ICESat measurement of Greenland ice sheet surface slope and roughness. Ann. Glaciol. 2005, 42, 83–89. [CrossRef] 174. Schenk, T.; Csatho, B.; van der Veen, C.; McCormick, D. Fusion of multi-sensor surface elevation data for improved characterization of rapidly changing outlet glaciers in Greenland. Remote Sens. Environ. 2014, 149, 239–251. [CrossRef] 175. Thomas, R.; Frederick, E.; Krabill, W.; Manizade, S.; Martin, C.; Mason, A. Elevation changes on the Greenland ice sheet from comparison of aircraft and ICESat laser-altimeter data. Ann. Glaciol. 2005, 42, 77–82. [CrossRef] 176. Hofton, M.A.; Blair, J.B.; Luthcke, S.B.; Rabine, D.L. Assessing the performance of 20–25 m footprint waveform lidar data collected in ICESat data corridors in Greenland. Geophys. Res. Lett. 2008, 35, L24501. [CrossRef] 177. Markus, T.; Neumann, T.; Martino, A.; Abdalati, W.; Brunt, K.; Csatho, B.; Farrell, S.; Fricker, H.; Gardner, A.; Harding, D.; et al. The Ice, Cloud, and land Elevation Satellite-2 (ICESat-2): Science requirements, concept, and implementation. Remote Sens. Environ. 2017, 190, 260–273. [CrossRef] 178. Smith, B.; Fricker, H.A.; Gardner, A.S.; Siegfried, M.R.; Adusumilli, S.; Csathó, B.M.; Holschuh, N.; Nilsson, J.; Paolo, F.S. ICESat-2 Science Team ATLAS/ICESat-2 L3A Land Ice Height, Version 1; NSIDC National Snow and Ice Data Center: Boulder, CO, USA, 2019. 179. Pope, R.M.; Fry, E.S. Absorption spectrum (380–700 nm) of pure water. II. Integrating cavity measurements. Appl. Opt. AO 1997, 36, 8710–8723. [CrossRef] 180. Hale, G.M.; Querry, M.R. Optical Constants of Water in the 200-nm to 200-µm Wavelength Region. Appl. Opt. AO 1973, 12, 555–563. [CrossRef] 181. Brunt, K.M.; Neumann, T.A.; Amundson, J.M.; Kavanaugh, J.L.; Moussavi, M.S.; Walsh, K.M.; Cook, W.B.; Markus, T. MABEL photon-counting laser altimetry data in Alaska for ICESat-2 simulations and development. Cryosphere 2016, 10, 1707–1719. [CrossRef] 182. Warren, S.G.; Brandt, R.E.; Grenfell, T.C. Visible and near-ultraviolet absorption spectrum of ice from transmission of solar radiation into snow. Appl. Opt. AO 2006, 45, 5320–5334. [CrossRef][PubMed] 183. Greeley, A.; Kurtz, N.T.; Neumann, T.; Markus, T. Estimating Surface Elevation Bias Due to Subsurface Scattered Photons from Visible Wavelength Laser Altimeters. In Proceedings of the AGU Fall Meeting Abstracts, New Orleans, LA, USA, 11–15 December 2017; Volume 51. 184. Gardner, A.S.; Smith, B.E.; Brunt, K.M.; Harding, D.J.; Neumann, T.; Walsh, K. ICESat2 subsurface-scattering biases estimated based on the 2015 SIMPL/AVRIS campaign. In Proceedings of the AGU Fall Meeting Abstracts, San Francisco, CA, USA, 14–18 December 2015; Volume 41, p. C41C-0710. 185. Howat, I.M.; Negrete, A.; Smith, B.E. The Greenland Ice Mapping Project (GIMP) land classification and surface elevation data sets. Cryosphere 2014, 8, 1509–1518. [CrossRef] 186. Farrell, S.L.; Markus, T.; Kwok, R.; Connor, L. Laser altimetry sampling strategies over sea ice. Ann. Glaciol. 2011, 52, 69–76. [CrossRef] 187. Brunt, K.M.; Neumann, T.A.; Walsh, K.M.; Markus, T. Determination of Local Slope on the Greenland Ice Sheet Using a Multibeam Photon-Counting Lidar in Preparation for the ICESat-2 Mission. IEEE Geosci. Remote Sens. Lett. 2014, 11, 935–939. [CrossRef] 188. Brunt, K.M.; Hawley, R.L.; Lutz, E.R.; Studinger, M.; Sonntag, J.G.; Hofton, M.A.; Andrews, L.C.; Neumann, T.A. Assessment of NASA airborne laser altimetry data using ground-based GPS data near Summit Station, Greenland. Cryosphere 2017, 11, 681–692. [CrossRef] 189. Larour, E.; Utke, J.; Csatho, B.; Schenk, A.; Seroussi, H.; Morlighem, M.; Rignot, E.; Schlegel, N.; Khazendar, A. Inferred basal friction and surface mass balance of the Northeast Greenland Ice Stream using data assimilation Remote Sens. 2019, 11, 2405 43 of 50

of ICESat (Ice Cloud and land Elevation Satellite) surface altimetry and ISSM (Ice Sheet System Model). Cryosphere 2014, 8, 2335–2351. [CrossRef] 190. Nolin, A.W. Mapping snow grain size on the ice sheets with a laser altimeter. In Proceedings of the IEEE International Geoscience and Remote Sensing (IGARSS ’98) Symposium on Sensing and Managing the Environment, Seattle, WA, USA, 6–10 July 1998; Volume 4, pp. 2264–2266. 191. Yang, Y.; Marshak, A.; Han, M.; Palm, S.P.; Harding, D.J. Snow grain size retrieval over the polar ice sheets with the Ice, Cloud, and land Elevation Satellite (ICESat) observations. J. Quant. Spectrosc. Radiat. Transf. 2017, 188, 159–164. [CrossRef] 192. Mouginot, J.; Rignot, E.; Bjørk, A.A.; van den Broeke, M.; Millan, R.; Morlighem, M.; Noël, B.; Scheuchl, B.; Wood, M. Forty-six years of Greenland Ice Sheet mass balance from 1972 to 2018. Proc. Natl. Acad. Sci. USA 2019, 116, 9239–9244. [CrossRef] 193. Kuipers Munneke, P.; Ligtenberg, S.R.M.; Noël, B.P.Y.; Howat, I.M.; Box, J.E.; Mosley-Thompson, E.; McConnell, J.R.; Steffen, K.; Harper, J.T.; Das, S.B.; et al. Elevation change of the Greenland Ice Sheet due to surface mass balance and firn processes, 1960–2014. Cryosphere 2015, 9, 2009–2025. [CrossRef] 194. Li, J.; Zwally, H.J. Modeling of firn compaction for estimating ice-sheet mass change from observed ice-sheet elevation change. Ann. Glaciol. 2011, 52, 1–7. [CrossRef] 195. Fahnestock, M.; Abdalati, W.; Joughin, I.; Brozena, J.; Gogineni, P. High Geothermal Heat Flow, Basal Melt, and the Origin of Rapid Ice Flow in Central Greenland. Science 2001, 294, 2338–2342. [CrossRef][PubMed] 196. Velicogna, I. Increasing rates of ice mass loss from the Greenland and Antarctic ice sheets revealed by GRACE. Geophys. Res. Lett. 2009, 36, L19503. [CrossRef] 197. Wouters, B.; Chambers, D.; Schrama, E.J.O. GRACE observes small-scale mass loss in Greenland. Geophys. Res. Lett. 2008, 35, L20501. [CrossRef] 198. Thomas, R.; Frederick, E.; Krabill, W.; Manizade, S.; Martin, C. Progressive increase in ice loss from Greenland. Geophys. Res. Lett. 2006, 33, L10503. [CrossRef] 199. Rignot, E.; Thomas, R.H. Mass Balance of Polar Ice Sheets. Science 2002, 297, 1502–1506. [CrossRef] 200. Rignot, E.; Braaten, D.; Gogineni, S.P.; Krabill, W.B.; McConnell, J.R. Rapid ice discharge from southeast Greenland glaciers. Geophys. Res. Lett. 2004, 31, L10401. [CrossRef] 201. Rignot, E.; Box, J.E.; Burgess, E.; Hanna, E. Mass balance of the Greenland ice sheet from 1958 to 2007. Geophys. Res. Lett. 2008, 35, L20502. [CrossRef] 202. Lucchitta, B.K.; Ferguson, H.M. Antarctica: Measuring Glacier Velocity from Satellite Images. Science 1986, 234, 1105–1108. [CrossRef] 203. Bindschadler, R.A.; Scambos, T.A. Satellite-Image-Derived Velocity Field of an Antarctic Ice Stream. Science 1991, 252, 242–246. [CrossRef] 204. Rosenau, R.; Scheinert, M.; Dietrich, R. A processing system to monitor Greenland outlet glacier velocity variations at decadal and seasonal time scales utilizing the Landsat imagery. Remote Sens. Environ. 2015, 169, 1–19. [CrossRef] 205. Fahnestock, M.; Scambos, T.; Moon, T.; Gardner, A.; Haran, T.; Klinger, M. Rapid large-area mapping of ice flow using Landsat 8. Remote Sens. Environ. 2016, 185, 84–94. [CrossRef] 206. Joughin, I.; Smith, B.E.; Howat, I.M. A complete map of Greenland ice velocity derived from satellite data collected over 20 years. J. Glaciol. 2018, 64, 1–11. [CrossRef][PubMed] 207. Rignot, E.; Jezek, K.C.; Sohn, H.G. Ice flow dynamics of the Greenland Ice Sheet from SAR interferometry. Geophys. Res. Lett. 1995, 22, 575–578. [CrossRef] 208. Joughin, I.R.; Winebrenner, D.P.; Fahnestock, M.A. Observations of ice-sheet motion in Greenland using satellite radar interferometry. Geophys. Res. Lett. 1995, 22, 571–574. [CrossRef] 209. Joughin, I.R.; Kwok, R.; Fahnestock, M.A. Interferometric estimation of three-dimensional ice-flow using ascending and descending passes. IEEE Trans. Geosci. Remote Sens. 1998, 36, 25–37. [CrossRef] 210. Joughin, I.R.; Fahnestock, M.A.; Bamber, J.L. Ice flow in the northeast Greenland ice stream. Ann. Glaciol. 2000, 31, 141–146. [CrossRef] 211. Joughin, I. Ice-sheet velocity mapping: A combined interferometric and speckle-tracking approach. Ann. Glaciol. 2002, 34, 195–201. [CrossRef] 212. Mouginot, J.; Rignot, E.; Scheuchl, B.; Millan, R. Comprehensive Annual Ice Sheet Velocity Mapping Using Landsat-8, Sentinel-1, and RADARSAT-2 Data. Remote Sens. 2017, 9, 364. [CrossRef] Remote Sens. 2019, 11, 2405 44 of 50

213. Gogineni, S.; Tammana, D.; Braaten, D.; Leuschen, C.; Akins, T.; Legarsky, J.; Kanagaratnam, P.; Stiles, J.; Allen, C.; Jezek, K. Coherent radar ice thickness measurements over the Greenland ice sheet. J. Geophys. Res. 2001, 106, 33761–33772. [CrossRef] 214. Gogineni, P.; Yan, J.-B. Remote sensing of ice thickness and surface velocity. In Remote Sensing of the Cryosphere; Wiley-Blackwell: Hoboken, NJ, USA, 2014; pp. 187–230. ISBN 978-1-118-36890-9. 215. Bamber, J.L.; Griggs, J.A.; Hurkmans, R.T.W.L.; Dowdeswell, J.A.; Gogineni, S.P.; Howat, I.; Mouginot, J.; Paden, J.; Palmer, S.; Rignot, E.; et al. A new bed elevation dataset for Greenland. Cryosphere 2013, 7, 499–510. [CrossRef] 216. Morlighem, M.; Rignot, E.; Seroussi, H.; Larour, E.; Dhia, H.B.; Aubry, D. A mass conservation approach for mapping glacier ice thickness. Geophys. Res. Lett. 2011, 38, L19503. [CrossRef] 217. Morlighem, M.; Willis, J. Improving Bed Topography Mapping of Greenland Glaciers Using NASA’s Oceans Melting Greenland (OMG) Data. Oceanography 2016, 29, 62–71. [CrossRef] 218. Morlighem, M.; Williams, C.N.; Rignot, E.; An, L.; Arndt, J.E.; Bamber, J.L.; Catania, G.; Chauché, N.; Dowdeswell, J.A.; Dorschel, B.; et al. BedMachine v3: Complete Bed Topography and Ocean Bathymetry Mapping of Greenland From Multibeam Echo Sounding Combined With Mass Conservation. Geophys. Res. Lett. 2017, 44, 11051–11061. [CrossRef][PubMed] 219. Millan, R.; Rignot, E.; Mouginot, J.; Wood, M.; Bjørk, A.A.; Morlighem, M. Vulnerability of Southeast Greenland Glaciers to Warm Atlantic Water From Operation IceBridge and Ocean Melting Greenland Data: Vulnerability of southeast greenland. Geophys. Res. Lett. 2018, 45, 2688–2696. [CrossRef][PubMed] 220. King, M.D.; Howat, I.M.; Jeong, S.; Noh, M.J.; Wouters, B.; Noël, B.; van den Broeke, M.R. Seasonal to decadal variability in ice discharge from the Greenland Ice Sheet. Cryosphere 2018, 12, 3813–3825. [CrossRef] 221. Mankoff, K.D.; Colgan, W.; Solgaard, A.; Karlsson, N.B.; Ahlstrøm, A.P.; Van As, D.; Box, J.E.; Abbas Khan, S.; Kjeldsen, K.K.; Mouginot, J.; et al. Greenland Ice Sheet solid ice discharge from 1986 through 2017. Earth Syst. Sci. Data 2019, 11, 769–786. [CrossRef] 222. Howat, I.M.; Ahn, Y.; Joughin, I.; van den Broeke, M.R.; Lenaerts, J.T.M.; Smith, B. Mass balance of Greenland’s three largest outlet glaciers, 2000–2010. Geophys. Res. Lett. 2011, 38, L12501. [CrossRef] 223. Korsgaard, N.J.; Nuth, C.; Khan, S.A.; Kjeldsen, K.K.; Bjørk, A.A.; Schomacker, A.; Kjær, K.H. Digital elevation model and orthophotographs of Greenland based on aerial photographs from 1978–1987. Sci. Data 2016, 3, 160032. [CrossRef] 224. Joughin, I.; Abdalati, W.; Fahnestock, M. Large fluctuations in speed on Greenland’s Jakobshavn Isbræ glacier. Nature 2004, 432, 608–610. [CrossRef] 225. Luckman, A.; Murray, T. Seasonal variation in velocity before retreat of Jakobshavn Isbræ, Greenland. Geophys. Res. Lett. 2005, 32, L08501. [CrossRef] 226. Alley, R.B. Ice-Sheet and Sea-Level Changes. Science 2005, 310, 456–460. [CrossRef][PubMed] 227. Joughin, I.; Howat, I.M.; Fahnestock, M.; Smith, B.; Krabill, W.; Alley, R.B.; Stern, H.; Truffer, M. Continued evolution of Jakobshavn Isbrae following its rapid speedup. J. Geophys. Res. 2008, 113, F04006. [CrossRef] 228. Moon, T.; Joughin, I. Changes in ice front position on Greenland’s outlet glaciers from 1992 to 2007. J. Geophys. Res. Earth Surf. 2008, 113, F02022. [CrossRef] 229. Andersen, J.K.; Hansen, K.; Box, J.E.; Andersen, S.B.; Ahlstrøm, A.P.; van As, D.; Citterio, M.; Colgan, W.; Karlsson, N.B.; Kjeldsen, K.K.; et al. Update of annual calving front lines for 47 marine terminating outlet glaciers in Greenland (1999–2018). GEUS Bull. 2019, 43, e2019430202. [CrossRef] 230. Tedesco, M. Remote sensing and the cryosphere. In Remote Sensing of the Cryosphere; Wiley-Blackwell: Hoboken, NJ, USA, 2014; pp. 1–16. ISBN 978-1-118-36890-9. 231. Ramillien, G.; Lombard, A.; Cazenave, A.; Ivins, E.R.; Llubes, M.; Remy, F.; Biancale, R. Interannual variations of the mass balance of the Antarctica and Greenland ice sheets from GRACE. Glob. Planet. Chang. 2006, 53, 198–208. [CrossRef] 232. Ramillien, G.; Famiglietti, J.S.; Wahr, J. Detection of Continental Hydrology and Glaciology Signals from GRACE: A Review. Surv. Geophys. 2008, 29, 361–374. [CrossRef] 233. Velicogna, I.; Wahr, J. Acceleration of Greenland ice mass loss in spring 2004. Nature 2006, 443, 329–331. [CrossRef] 234. Simonsen, S.B.; Sørensen, L.S. Implications of changing scattering properties on Greenland ice sheet volume change from Cryosat-2 altimetry. Remote Sens. Environ. 2017, 190, 207–216. [CrossRef] Remote Sens. 2019, 11, 2405 45 of 50

235. Barletta, V.R.; Sabadini, R.; Bordoni, A. Isolating the PGR signal in the GRACE data: Impact on mass balance estimates in Antarctica and Greenland. Geophys. J. Int. 2008, 172, 18–30. [CrossRef] 236. Kjeldsen, K.K.; Khan, S.A.; Wahr, J.; Korsgaard, N.J.; Kjær, K.H.; Bjørk, A.A.; Hurkmans, R.; van den Broeke, M.R.; Bamber, J.L.; van Angelen, J.H. Improved ice loss estimate of the northwestern Greenland ice sheet. J. Geophys. Res. Solid Earth 2013, 118, 698–708. [CrossRef] 237. Su, X.; Shum, C.K.; Guo, J.; Duan, J.; Howat, I.; Yi, Y. High resolution Greenland ice sheet inter-annual mass variations combining GRACE gravimetry and Envisat altimetry. Earth Planet. Sci. Lett. 2015, 422, 11–17. [CrossRef] 238. Hall, D.K.; Williams, R.S.; Luthcke, S.B.; Digirolamo, N.E. Greenland ice sheet surface temperature, melt and mass loss: 2000–2006. J. Glaciol. 2008, 54, 81–93. [CrossRef] 239. Sasgen, I.; van den Broeke, M.; Bamber, J.L.; Rignot, E.; Sørensen, L.S.; Wouters, B.; Martinec, Z.; Velicogna, I.; Simonsen, S.B. Timing and origin of recent regional ice-mass loss in Greenland. Earth Planet. Sci. Lett. 2012, 333–334, 293–303. [CrossRef] 240. Velicogna, I.; Wahr, J. Time-variable gravity observations of ice sheet mass balance: Precision and limitations of the GRACE satellite data. Geophys. Res. Lett. 2013, 40, 3055–3063. [CrossRef] 241. Harig, C.; Simons, F.J. Ice mass loss in Greenland, the Gulf of Alaska, and the Canadian Archipelago: Seasonal cycles and decadal trends. Geophys. Res. Lett. 2016, 43, 3150–3159. [CrossRef] 242. Harig, C.; Simons, F.J. Mapping Greenland’s mass loss in space and time. Proc. Natl. Acad. Sci. USA 2012, 109, 19934–19937. [CrossRef] 243. Wahr, J.; Swenson, S.; Velicogna, I. Accuracy of GRACE mass estimates. Geophys. Res. Lett. 2006, 33, L06401. [CrossRef] 244. Watkins, M.M.; Wiese, D.N.; Yuan, D.-N.; Boening, C.; Landerer, F.W. Improved methods for observing Earth’s time variable mass distribution with GRACE using spherical cap mascons. J. Geophys. Res. Solid Earth 2015, 120, 2648–2671. [CrossRef] 245. Swenson, S.; Wahr, J. Post-processing removal of correlated errors in GRACE data. Geophys. Res. Lett. 2006, 33, L08402. [CrossRef] 246. Velicogna, I.; Wahr, J. Greenland mass balance from GRACE. Geophys. Res. Lett. 2005, 32, L18505. [CrossRef] 247. Sasgen, I.; Klemann, V.; Martinec, Z. Towards the inversion of GRACE gravity fields for present-day ice-mass changes and glacial-isostatic adjustment in North America and Greenland. J. Geodyn. 2012, 59–60, 49–63. [CrossRef] 248. Flechtner, F.; Morton, P.; Watkins, M.; Webb, F. Status of the GRACE Follow-On Mission. In Proceedings of the Gravity, Geoid and Height Systems; Marti, U., Ed.; Springer International Publishing: Cham, Switzerland, 2014; pp. 117–121. 249. Greicius, T. GRACE-FO Resumes Data Collection. Available online: http://www.nasa.gov/feature/jpl/grace- fo-resumes-data-collection (accessed on 15 January 2019). 250. NASA/JPL-Caltech GRACE-FO First Gravity Field Data Now. Available online: https://gracefo.jpl.nasa.gov/ news/145/grace-fo-first-gravity-field-data-now-available (accessed on 2 July 2019). 251. Schaaf, C.B.; Gao, F.; Strahler, A.H.; Lucht, W.; Li, X.; Tsang, T.; Strugnell, N.C.; Zhang, X.; Jin, Y.; Muller, J.-P.; et al. First operational BRDF, albedo nadir reflectance products from MODIS. Remote Sens. Environ. 2002, 83, 135–148. [CrossRef] 252. Nolin, A.W.; Stroeve, J. The changing albedo of the Greenland ice sheet: Implications for climate modeling. Ann. Glaciol. 1997, 25, 51–57. [CrossRef] 253. Tedesco, M.; Doherty, S.; Fettweis, X.; Alexander, P.; Jeyaratnam, J.; Stroeve, J. The darkening of the Greenland ice sheet: Trends, drivers, and projections (1981–2100). Cryosphere 2016, 10, 477–496. [CrossRef] 254. Van Angelen, J.H.; Lenaerts, J.T.M.; Lhermitte, S.; Fettweis, X.; Kuipers Munneke, P.; van den Broeke, M.R.; van Meijgaard, E.; Smeets, C.J.P.P. Sensitivity of Greenland Ice Sheet surface mass balance to surface albedo parameterization: A study with a regional climate model. Cryosphere 2012, 6, 1175–1186. [CrossRef] 255. Nicodemus, F.E.; Richmond, J.C.; Hsia, J.J.; Ginsberg, I.W.; Limperis, T. Geometrical Considerations and Nomenclature for Reflectance; National Bureau of Standards: Washington, DC, USA, 1977; pp. 1–51. 256. Martonchik, J.V.; Bruegge, C.J.; Strahler, A.H. A review of reflectance nomenclature used in remote sensing. Remote Sens. Rev. 2000, 19, 9–20. [CrossRef] 257. Snyder, W.C. Definition and invariance properties of structured surface BRDF. IEEE Trans. Geosci. Remote Sens. 2002, 40, 1032–1037. [CrossRef] Remote Sens. 2019, 11, 2405 46 of 50

258. Di Girolamo, L. Generalizing the definition of the bi-directional reflectance distribution function. Remote Sens. Environ. 2003, 88, 479–482. [CrossRef] 259. Schaepman-Strub, G.; Schaepman, M.E.; Painter, T.H.; Dangel, S.; Martonchik, J.V. Reflectance quantities in optical remote sensing—definitions and case studies. Remote Sens. Environ. 2006, 103, 27–42. [CrossRef] 260. Nolin, A.W.; Fetterer, F.M.; Scambos, T.A. Surface roughness characterizations of sea ice and ice sheets: Case studies with MISR data. IEEE Trans. Geosci. Remote Sens. 2002, 40, 1605–1615. [CrossRef] 261. Stroeve, J.C.; Nolin, A.W. New methods to infer snow albedo from the MISR instrument with applications to the Greenland ice sheet. IEEE Trans. Geosci. Remote Sens. 2002, 40, 1616–1625. [CrossRef] 262. Bindschadler, R.; Dowdeswell, J.; Hall, D.; Winther, J.-G. Glaciological applications with Landsat-7 imagery: Early assessments. Remote Sens. Environ. 2001, 78, 163–179. [CrossRef] 263. Stroeve, J.C.; Box, J.E.; Haran, T. Evaluation of the MODIS (MOD10A1) daily snow albedo product over the Greenland ice sheet. Remote Sens. Environ. 2006, 105, 155–171. [CrossRef] 264. Stroeve, J. Assessment of Greenland albedo variability from the advanced very high resolution radiometer Polar Pathfinder data set. J. Geophys. Res. 2001, 106, 33989–34006. [CrossRef] 265. Arnold, N.S.; Banwell, A.F.; Willis, I.C. High-resolution modelling of the seasonal evolution of surface water storage on the Greenland Ice Sheet. Cryosphere 2014, 8, 1149–1160. [CrossRef] 266. Pope, A.; Scambos, T.A.; Moussavi, M.; Tedesco, M.; Willis, M.; Shean, D.; Grigsby, S. Estimating supraglacial lake depth in West Greenland using Landsat 8 and comparison with other multispectral methods. Cryosphere 2016, 10, 15–27. [CrossRef] 267. Yang, K.; Smith, L.C.; Sole, A.; Livingstone, S.J.; Cheng, X.; Chen, Z.; Li, M. Supraglacial rivers on the northwest Greenland Ice Sheet, Devon Ice Cap, and Barnes Ice Cap mapped using Sentinel-2 imagery. Int. J. Appl. Earth Obs. Geoinf. 2019, 78, 1–13. [CrossRef] 268. Sohn, H.-G.; Jezek, K.C. Mapping ice sheet margins from ERS-1 SAR and SPOT imagery. Int. J. Remote Sens. 1999, 20, 3201–3216. [CrossRef] 269. Wang, S.; Tedesco, M.; Xu, M.; Alexander, P.M. Mapping Ice Algal Blooms in Southwest Greenland From Space. Geophys. Res. Lett. 2018, 45, 11779–11788. [CrossRef] 270. Georgiou, S.; Shepherd, A.; McMillan, M.; Nienow, P. Seasonal evolution of supraglacial lake volume from ASTER imagery. Ann. Glaciol. 2009, 50, 95–100. [CrossRef] 271. Hori, M.; Aoki, T.; Stamnes, K.; Li, W. ADEOS-II/GLI snow/ice products—Part III: Retrieved results. Remote Sens. Environ. 2007, 111, 291–336. [CrossRef] 272. Key, J.; Wang, X.; Liu, Y. NOAA Climate Data Record of AVHRR Polar Pathfinder Extended (APP-X), Version 1; NOAA National Centers for Environmental Information: Asheville, NC, USA, 2014. 273. Knap, W.H.; Oerlemans, J. The surface albedo of the Greenland ice sheet: Satellite-derived and in situ measurements in the Søndre Strømfjord area during the 1991 melt season. J. Glaciol. 1996, 42, 364–374. [CrossRef] 274. Stroeve, J.C.; Box, J.E.; Fowler, C.; Haran, T.; Key, J. Intercomparison between In Situ and AVHRR Polar Pathfinder-Derived Surface Albedo over Greenland. Remote Sens. Environ. 2001, 75, 360–374. [CrossRef] 275. Box, J.E.; Bromwich, D.H.; Veenhuis, B.A.; Bai, L.-S.; Stroeve, J.C.; Rogers, J.C.; Steffen, K.; Haran, T.; Wang, S.-H. Greenland Ice Sheet Surface Mass Balance Variability (1988–2004) from Calibrated Polar MM5 Output. J. Clim. 2006, 19, 2783–2800. [CrossRef] 276. Liang, S. Narrowband to broadband conversions of land surface albedo I: Algorithms. Remote Sens. Environ. 2001, 76, 213–238. [CrossRef] 277. Greuell, W.; Oerlemans, J. Narrowband-to-broadband albedo conversion for glacier ice and snow: Equations based on modeling and ranges of validity of the equations. Remote Sens. Environ. 2004, 89, 95–105. [CrossRef] 278. Stroeve, J.; Box, J.E.; Wang, Z.; Schaaf, C.; Barrett, A. Re-evaluation of MODIS MCD43 Greenland albedo accuracy and trends. Remote Sens. Environ. 2013, 138, 199–214. [CrossRef] 279. Klein, A.G.; Stroeve, J. Development and validation of a snow albedo algorithm for the MODIS instrument. Ann. Glaciol. 2002, 34, 45–52. [CrossRef] 280. Ryan, J.C.; Hubbard, A.L.; Irvine-Fynn, T.D.; Doyle, S.H.; Cook, J.M.; Stibal, M.; Box, J.E. How robust are in situ observations for validating satellite-derived albedo over the dark zone of the Greenland Ice Sheet? Geophys. Res. Lett. 2017, 44, 6218–6225. [CrossRef] Remote Sens. 2019, 11, 2405 47 of 50

281. Schaaf, C.B.; Wang, Z.; Strahler, A.H. Commentary on Wang and Zender—MODIS snow albedo bias at high solar zenith angles relative to theory and to in situ observations in Greenland. Remote Sens. Environ. 2011, 115, 1296–1300. [CrossRef] 282. Steffen, K.; Box, J.E.; Abdalati, W. Greenland Climate Network: GC-Net. In Special Report 96-27 Glaciers, Ice Sheets and Volcanoes: A Tribute to Mark F. Meier; Colbeck, S.C., Ed.; CRREL: Hanover, NH, USA, 1996; pp. 98–103. 283. Stroeve, J.; Box, J.E.; Gao, F.; Liang, S.; Nolin, A.; Schaaf, C. Accuracy assessment of the MODIS 16-day albedo product for snow: Comparisons with Greenland in situ measurements. Remote Sens. Environ. 2005, 94, 46–60. [CrossRef] 284. Wang, X.; Zender, C.S. MODIS snow albedo bias at high solar zenith angles relative to theory and to in situ observations in Greenland. Remote Sens. Environ. 2010, 114, 563–575. [CrossRef] 285. Moustafa, S.E.; Rennermalm, A.K.; Román, M.O.; Wang, Z.; Schaaf, C.B.; Smith, L.C.; Koenig, L.S.; Erb, A. Evaluation of satellite remote sensing albedo retrievals over the ablation area of the southwestern Greenland ice sheet. Remote Sens. Environ. 2017, 198, 115–125. [CrossRef] 286. Moustafa, S.E.; Rennermalm, A.K.; Smith, L.C.; Miller, M.A.; Mioduszewski, J.R.; Koenig, L.S.; Hom, M.G.; Shuman, C.A. Multi-modal albedo distributions in the ablation area of the southwestern Greenland Ice Sheet. Cryosphere 2015, 9, 905–923. [CrossRef] 287. Nghiem, S.V.; Hall, D.K.; Mote, T.L.; Tedesco, M.; Albert, M.R.; Keegan, K.; Shuman, C.A.; DiGirolamo, N.E.; Neumann, G. The extreme melt across the Greenland ice sheet in 2012. Geophys. Res. Lett. 2012, 39, L20502. [CrossRef] 288. Alexander, P.M.; Tedesco, M.; Fettweis, X.; van de Wal, R.S.W.; Smeets, C.J.P.P.; van den Broeke, M.R. Assessing spatio-temporal variability and trends in modelled and measured Greenland Ice Sheet albedo (2000–2013). Cryosphere 2014, 8, 2293–2312. [CrossRef] 289. Dadic, R.; Mullen, P.C.; Schneebeli, M.; Brandt, R.E.; Warren, S.G. Effects of bubbles, cracks, and volcanic tephra on the spectral albedo of bare ice near the Transantarctic Mountains: Implications for sea glaciers on Snowball Earth. J. Geophys. Res. Earth Surf. 2013, 118, 1658–1676. [CrossRef] 290. Benning, L.G.; Anesio, A.M.; Lutz, S.; Tranter, M. Biological impact on Greenland’s albedo. Nat. Geosci. 2014, 7, 691. [CrossRef] 291. Yallop, M.L.; Anesio, A.M.; Perkins, R.G.; Cook, J.; Telling, J.; Fagan, D.; MacFarlane, J.; Stibal, M.; Barker, G.; Bellas, C.; et al. Photophysiology and albedo-changing potential of the ice algal community on the surface of the Greenland ice sheet. ISME J. 2012, 6, 2302–2313. [CrossRef] 292. Goelles, T.; Bøggild, C.E. Albedo reduction of ice caused by dust and black carbon accumulation: A model applied to the K-transect, West Greenland. J. Glaciol. 2017, 63, 1063–1076. [CrossRef] 293. Greuell, W.; Knap, W.H. Remote sensing of the albedo and detection of the slush line on the Greenland ice sheet. J. Geophys. Res. Atmos. 2000, 105, 15567–15576. [CrossRef] 294. Wientjes, I.G.M.; De Van Wal, R.S.W.; Schwikowski, M.; Zapf, A.; Fahrni, S.; Wacker, L. Carbonaceous particles reveal that Late Holocene dust causes the dark region in the western ablation zone of the Greenland ice sheet. J. Glaciol. 2012, 58, 787–794. [CrossRef] 295. Wientjes, I.G.M.; Van de Wal, R.S.W.; Reichart, G.J.; Sluijs, A.; Oerlemans, J. Dust from the dark region in the western ablation zone of the Greenland ice sheet. Cryosphere 2011, 5, 589–601. [CrossRef] 296. Bøggild, C.E.; Oerter, H.; Tukiainen, T. Increased ablation of Wisconsin ice in eastern north Greenland: Observations and modelling. Ann. Glaciol. 1996, 23, 144–148. [CrossRef] 297. Shimada, R.; Takeuchi, N.; Aoki, T. Inter-Annual and Geographical Variations in the Extent of Bare Ice and Dark Ice on the Greenland Ice Sheet Derived from MODIS Satellite Images. Front. Earth Sci. 2016, 4, 43. [CrossRef] 298. Hodson, A.; Bøggild, C.; Hanna, E.; Huybrechts, P.; Langford, H.; Cameron, K.; Houldsworth, A. The cryoconite ecosystem on the Greenland ice sheet. Ann. Glaciol. 2010, 51, 123–129. [CrossRef] 299. Uetake, J.; Naganuma, T.; Hebsgaard, M.B.; Kanda, H.; Kohshima, S. Communities of algae and cyanobacteria on glaciers in west Greenland. Polar Science 2010, 4, 71–80. [CrossRef] 300. Stibal, M.; Box, J.E.; Cameron, K.A.; Langen, P.L.; Yallop, M.L.; Mottram, R.H.; Khan, A.L.; Molotch, N.P.; Chrismas, N.A.M.; Quaglia, F.C.; et al. Algae Drive Enhanced Darkening of Bare Ice on the Greenland Ice Sheet. Geophys. Res. Lett. 2017, 44, 11463–11471. [CrossRef] Remote Sens. 2019, 11, 2405 48 of 50

301. Cook, J.M.; Hodson, A.J.; Gardner, A.S.; Flanner, M.; Tedstone, A.J.; Williamson, C.; Irvine-Fynn, T.D.L.; Nilsson, J.; Bryant, R.; Tranter, M. Quantifying bioalbedo: A new physically based model and discussion of empirical methods for characterising biological influence on ice and snow albedo. Cryosphere 2017, 11, 2611–2632. [CrossRef] 302. Chandler, D.M.; Alcock, J.D.; Wadham, J.L.; Mackie, S.L.; Telling, J. Seasonal changes of ice surface characteristics and productivity in the ablation zone of the Greenland Ice Sheet. Cryosphere 2015, 9, 487–504. [CrossRef] 303. Dumont, M.; Brun, E.; Picard, G.; Michou, M.; Libois, Q.; Petit, J.-R.; Geyer, M.; Morin, S.; Josse, B. Contribution of light-absorbing impurities in snow to Greenland’s darkening since 2009. Nat. Geosci. 2014, 7, 509. [CrossRef] 304. Tedesco, M.; Doherty, S.; Warren, S.G.; Tranter, M.; Stroeve, J.C.; Fettweis, X.; Alexander, P. What Darkens the Greenland Ice Sheet? Eos 2015, 96.[CrossRef] 305. Zege, E.; Katsev, I.; Malinka, A.; Prikhach, A.; Polonsky, I. New algorithm to retrieve the effective snow grain size and pollution amount from satellite data. Ann. Glaciol. 2008, 49, 139–144. [CrossRef] 306. Warren, S.G. Can black carbon in snow be detected by remote sensing? J. Geophys. Res. Atmos. 2013, 118, 779–786. [CrossRef] 307. Painter, T.H.; Duval, B.; Thomas, W.H.; Mendez, M.; Heintzelman, S.; Dozier, J. Detection and Quantification of Snow Algae with an Airborne Imaging Spectrometer. Appl. Environ. Microbiol. 2001, 67, 5267–5272. [CrossRef] 308. Takeuchi, N.; Dial, R.; Kohshima, S.; Segawa, T.; Uetake, J. Spatial distribution and abundance of red snow algae on the Harding Icefield, Alaska derived from a satellite image. Geophys. Res. Lett. 2006, 33, L21502. [CrossRef] 309. Transon, J.; d’Andrimont, R.; Maugnard, A.; Defourny, P. Survey of Hyperspectral Earth Observation Applications from Space in the Sentinel-2 Context. Remote Sens. 2018, 10, 157. [CrossRef] 310. Cook, J.M.; Hodson, A.J.; Taggart, A.J.; Mernild, S.H.; Tranter, M. A predictive model for the spectral “bioalbedo” of snow. J. Geophys. Res. Earth Surf. 2017, 122, 434–454. [CrossRef] 311. Long, D.G.; Drinkwater, M.R. Greenland ice-sheet surface properties observed by the Seasat-A scatterometer at enhanced resolution. J. Glaciol. 1994, 40, 213–230. [CrossRef] 312. Forster, R.R.; Isacks, B.L.; Das, S.B. Shuttle imaging radar (SIR-C/X-SAR) reveals near-surface properties of the South Patagonian Icefield. J. Geophys. Res. Planets 1996, 101, 23169–23180. [CrossRef] 313. Smith, L.C.; Forster, R.R.; Isacks, B.L.; Hall, D.K. Seasonal climatic forcing of alpine glaciers revealed with orbital synthetic aperture radar. J. Glaciol. 1997, 43, 480–488. [CrossRef] 314. Zhou, C.; Zheng, L. Mapping Radar Glacier Zones and Dry Snow Line in the Antarctic Peninsula Using Sentinel-1 Images. Remote Sens. 2017, 9, 1171. [CrossRef] 315. Abdalati, W.; Steffen, K. Passive microwave-derived snow melt regions on the Greenland Ice Sheet. Geophys. Res. Lett. 1995, 22, 787–790. [CrossRef] 316. Jezek, K.C.; Gogineni, P.; Shanableh, M. Radar measurements of melt zones on the Greenland Ice Sheet. Geophys. Res. Lett. 1994, 21, 33–36. [CrossRef] 317. Mote, T.L. Greenland surface melt trends 1973–2007: Evidence of a large increase in 2007. Geophys. Res. Lett. 2007, 34, L22507. [CrossRef] 318. Fettweis, X.; Tedesco, M.; van den Broeke, M.; Ettema, J. Melting trends over the Greenland ice sheet (1958–2009) from spaceborne microwave data and regional climate models. Cryosphere 2011, 5, 359–375. [CrossRef] 319. Shi, J.; Dozier, J. Measurements of snow- and glacier-covered areas with single-polarization SAR. Ann. Glaciol. 1993, 17, 72–76. [CrossRef] 320. Ashcraft, I.S.; Long, D.G. Comparison of methods for melt detection over Greenland using active and passive microwave measurements. Int. J. Remote Sens. 2006, 27, 2469–2488. [CrossRef] 321. Partington, K.C. Discrimination of glacier facies using multi-temporal SAR data. J. Glaciol. 1998, 44, 42–53. [CrossRef] 322. Nghiem, S.V.; Steffen, K.; Kwok, R.; Tsai, W.Y. Detection of snowmelt regions on the Greenland ice sheet using diurnal backscatter change. J. Glaciol. 2001, 47, 539–547. [CrossRef] 323. Early, D.S.; Long, D.G. Image reconstruction and enhanced resolution imaging from irregular samples. IEEE Trans. Geosci. Remote Sens. 2001, 39, 291–302. [CrossRef] Remote Sens. 2019, 11, 2405 49 of 50

324. Wismann, V. Monitoring of seasonal snowmelt on Greenland with ERS scatterometer data. IEEE Trans. Geosci. Remote Sens. 2000, 38, 1821–1826. [CrossRef] 325. Smith, L.C.; Sheng, Y.; Forster, R.R.; Steffen, K.; Frey, K.E.; Alsdorf, D.E. Melting of small Arctic ice caps observed from ERS scatterometer time series. Geophys. Res. Lett. 2003, 30, 2034. [CrossRef] 326. Hicks, B.R.; Long, D.G. Inferring Greenland melt and refreeze severity from SeaWinds scatterometer data. Int. J. Remote Sens. 2011, 32, 8053–8080. [CrossRef] 327. Li, X.; Zhang, Y.; Liang, L. Snowmelt detection on the Greenland ice sheet using microwave scatterometer measurements. Int. J. Remote Sens. 2017, 38, 796–807. [CrossRef] 328. Wang, L.; Sharp, M.; Rivard, B.; Steffen, K. Melt season duration and ice layer formation on the Greenland ice sheet, 2000–2004. J. Geophys. Res. 2007, 112, F04013. [CrossRef] 329. Long, D.G.; Hicks, B.R. SeaWinds on QuikScat Enhanced Resolution Image Products (Version 2); Brigham Young University: Provo, UT, USA, 2005. 330. Wan, Z.; Dozier, J. Land-surface temperature measurement from space: Physical principles and inverse modeling. IEEE Trans. Geosci. Remote Sens. 1989, 27, 268–278. 331. Chang, T.C.; Gloersen, P.; Schmugge, T.; Wilheit, T.T.; Zwally, H.J. Microwave Emission From Snow and Glacier Ice. J. Glaciol. 1976, 16, 23–39. [CrossRef] 332. Mote, T.L.; Anderson, M.R.; Kuivinen, K.C.; Rowe, C.M. Passive microwave-derived spatial and temporal variations of summer melt on the Greenland ice sheet. Ann. Glaciol. 1993, 17, 233–238. [CrossRef] 333. Mote, T.L.; Anderson, M.R. Variations in snowpack melt on the Greenland ice sheet based on passive-microwave measurements. J. Glaciol. 1995, 41, 51–60. [CrossRef] 334. Abdalati, W.; Steffen, K. Greenland Ice Sheet melt extent: 1979–1999. J. Geophys. Res. Atmos. 2001, 106, 33983–33988. [CrossRef] 335. Ramage, J.M.; Isacks, B.L. Determination of melt-onset and refreeze timing on southeast Alaskan icefields using SSM/I diurnal amplitude variations. Ann. Glaciol. 2002, 34, 391–398. [CrossRef] 336. Tedesco, M. Snowmelt detection over the Greenland ice sheet from SSM/I brightness temperature daily variations. Geophys. Res. Lett. 2007, 34, L02504. [CrossRef] 337. Remote Sensing of the Cryosphere; Tedesco, M. (Ed.) The Cyrosphere Science Series; Wiley Blackwell: Hoboken, NJ, USA, 2015; ISBN 978-1-118-36885-5. 338. Bhattacharya, I.; Jezek, K.C.; Wang, L.; Liu, H. Surface melt area variability of the Greenland ice sheet: 1979–2008. Geophys. Res. Lett. 2009, 36, L20502. [CrossRef] 339. Tedesco, M.; Fettweis, X.; Mote, T.; Wahr, J.; Alexander, P.; Box, J.E.; Wouters, B. Evidence and analysis of 2012 Greenland records from spaceborne observations, a regional climate model and reanalysis data. Cryosphere 2013, 7, 615–630. [CrossRef] 340. Abdalati, W.; Steffen, K. The apparent effects of the Mt. Pinatubo Eruption on the Greenland Ice Sheet melt extent. Geophys. Res. Lett. 1997, 24, 1795–1797. [CrossRef] 341. Chu, V.W.; Smith, L.C.; Rennermalm, A.K.; Forster, R.R.; Box, J.E.; Reeh, N. Sediment plume response to surface melting and supraglacial lake drainages on the Greenland ice sheet. J. Glaciol. 2009, 55, 1072–1082. [CrossRef] 342. Mernild, S.H.; Liston, G.E.; Hiemstra, C.A.; Christensen, J.H. Greenland Ice Sheet Surface Mass-Balance Modeling in a 131-Yr Perspective, 1950–2080. J. Hydrometeor. 2010, 11, 3–25. [CrossRef] 343. Mernild, S.H.; Mote, T.L.; Liston, G.E. Greenland ice sheet surface melt extent and trends: 1960–2010. J. Glaciol. 2011, 57, 621–628. [CrossRef] 344. Mote, T.L. MEaSUREs Greenland Surface Melt Daily 25km EASE-Grid 2.0; NASA DAAC at the National Snow and Ice Data Center: Boulder, CO, USA, 2014. 345. Stroeve, J.; Haefliger, M.; Steffen, K. Surface Temperature from ERS-1 ATSR Infrared Thermal Satellite Data in Polar Regions. J. Appl. Meteor. 1996, 35, 1231–1239. [CrossRef] 346. Hall, D.K.; Ormsby, J.P.; Bindschadler, R.A.; Siddalingaiah, H. Characterization of Snow and Ice Reflectance Zones On Glaciers Using Landsat Thematic Mapper Data. Ann. Glaciol. 1987, 9, 104–108. [CrossRef] 347. Hall, D.K.; Nghiem, S.V.; Schaaf, C.B.; DiGirolamo, N.E.; Neumann, G. Evaluation of surface and near-surface melt characteristics on the Greenland ice sheet using MODIS and QuikSCAT data. J. Geophys. Res. Earth Surf. 2009, 114, F04006. [CrossRef] 348. Välisuo, I.; Vihma, T.; Pirazzini, R.; Schäfer, M. Interannual Variability of Atmospheric Conditions and Surface Melt in Greenland in 2000–2014. J. Geophys. Res. Atmos. 2018, 123, 10443–10463. [CrossRef] Remote Sens. 2019, 11, 2405 50 of 50

349. Hall, D.K.; Comiso, J.C.; DiGirolamo, N.E.; Shuman, C.A.; Box, J.E.; Koenig, L.S. Variability in the surface temperature and melt extent of the Greenland ice sheet from MODIS. Geophys. Res. Lett. 2013, 40, 2114–2120. [CrossRef] 350. Hall, D.K.; Box, J.E.; Casey, K.A.; Hook, S.J.; Shuman, C.A.; Steffen, K. Comparison of satellite-derived and in-situ observations of ice and snow surface temperatures over Greenland. Remote Sens. Environ. 2008, 112, 3739–3749. [CrossRef] 351. Hall, D.K.; Comiso, J.C.; DiGirolamo, N.E.; Shuman, C.A.; Key, J.R.; Koenig, L.S. A Satellite-Derived Climate-Quality Data Record of the Clear-Sky Surface Temperature of the Greenland Ice Sheet. J. Clim. 2012, 25, 4785–4798. [CrossRef] 352. Wan, Z.; Li, Z.-L. Radiance-based validation of the V5 MODIS land-surface temperature product. Int. J. Remote Sens. 2008, 29, 5373–5395. [CrossRef] 353. Shuman, C.A.; Hall, D.K.; DiGirolamo, N.E.; Mefford, T.K.; Schnaubelt, M.J. Comparison of Near-Surface Air Temperatures and MODIS Ice-Surface Temperatures at Summit, Greenland (2008–13). J. Appl. Meteor. Climatol. 2014, 53, 2171–2180. [CrossRef] 354. Koenig, L.S.; Hall, D.K. Comparison of satellite, thermochron and air temperatures at Summit, Greenland, during the winter of 2008/09. J. Glaciol. 2010, 56, 735–741. [CrossRef] 355. Stroeve, J.; Steffen, K. Variability of AVHRR-Derived Clear-Sky Surface Temperature over the Greenland Ice Sheet. J. Appl. Meteor. 1998, 37, 23–31. [CrossRef] 356. Nolin, A.W.; Mar, E. Arctic Sea Ice Surface Roughness Estimated from Multi-Angular Reflectance Satellite Imagery. Remote Sens. 2019, 11, 50. [CrossRef] 357. Boisvert, L.N.; Lee, J.N.; Lenaerts, J.T.M.; Noël, B.; Broeke, M.R.; Nolin, A.W. Using remotely sensed data from AIRS to estimate the vapor flux on the Greenland ice sheet: Comparisons with observations and a regional climate model. J. Geophys. Res. Atmos. 2017, 122, 202–229. [CrossRef] 358. Li, X.; Xu, K.; Xu, L. Surface slope and roughness measurement using ICESat/GLAS elevation and laser waveform. Meas. Sci. Technol. 2016, 27, 095202. [CrossRef] 359. Van der Veen, C.J.; Ahn, Y.; Csatho, B.M.; Mosley-Thompson, E.; Krabill, W.B. Surface roughness over the northern half of the Greenland Ice Sheet from airborne laser altimetry. J. Geophys. Res. Earth Surf. 2009, 114, F01001. [CrossRef] 360. Poinar, K.; Joughin, I.; Das, S.B.; Behn, M.D.; Lenaerts, J.T.M.; van den Broeke, M.R. Limits to future expansion of surface-melt-enhanced ice flow into the interior of western Greenland. Geophys. Res. Lett. 2015, 42, 1800–1807. [CrossRef] 361. Hoffman, M.J.; Perego, M.; Andrews, L.C.; Price, S.F.; Neumann, T.A.; Johnson, J.V.; Catania, G.; Lüthi, M.P. Widespread Moulin Formation During Supraglacial Lake Drainages in Greenland. Geophys. Res. Lett. 2018, 45, 778–788. [CrossRef] 362. Koenig, L.S.; Lampkin, D.J.; Montgomery, L.N.; Hamilton, S.L.; Turrin, J.B.; Joseph, C.A.; Moutsafa, S.E.; Panzer, B.; Casey, K.A.; Paden, J.D.; et al. Wintertime storage of water in buried supraglacial lakes across the Greenland Ice Sheet. Cryosphere 2015, 9, 1333–1342. [CrossRef] 363. Miles, K.E.; Willis, I.C.; Benedek, C.L.; Williamson, A.G.; Tedesco, M. Toward Monitoring Surface and Subsurface Lakes on the Greenland Ice Sheet Using Sentinel-1 SAR and Landsat-8 OLI Imagery. Front. Earth Sci. 2017, 5, 58. [CrossRef] 364. Cooley, S.W.; Christoffersen, P. Observation Bias Correction Reveals More Rapidly Draining Lakes on the Greenland Ice Sheet. J. Geophys. Res. Earth Surf. 2017, 122, 1867–1881. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).