TCThe CryosphereTCThe Cryosphere1994-0424Copernicus PublicationsGöttingen, Germany10.5194/tc-11-1781-2017Impact of MODIS sensor calibration updates on Greenland Ice Sheet surface
reflectance and albedo trendsCaseyKimberly A.kimberly.a.casey@nasa.govhttps://orcid.org/0000-0002-6115-7525PolashenskiChris M.ChenJustinTedescoMarcoThayer School of Engineering, Dartmouth College, Hanover, NH
03755, USACryospheric Sciences Lab, NASA Goddard Space Flight Center,
Greenbelt, MD 20771, USACold Regions Research and Engineering Laboratory, Alaska Projects
Office, US Army Corps of Engineers, Fairbanks, AK 99709, USADepartment of Computer Science, Stanford University, Stanford, CA 94305, USALamont–Doherty Earth Observatory, Columbia University, NY 10964,
USANASA Goddard Institute for Space Studies, New York, NY 10025, USAnow at: Land Remote Sensing Program, US
Geological Survey, Reston, VA 20192, USAKimberly A. Casey (kimberly.a.casey@nasa.gov)1August20171141781179515March201727March201716June201719June2017This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/This article is available from https://tc.copernicus.org/articles/11/1781/2017/tc-11-1781-2017.htmlThe full text article is available as a PDF file from https://tc.copernicus.org/articles/11/1781/2017/tc-11-1781-2017.pdf
We evaluate Greenland Ice Sheet (GrIS) surface reflectance and
albedo trends using the newly released Collection 6 (C6) MODIS (Moderate
Resolution Imaging Spectroradiometer) products over
the period 2001–2016. We find that the correction of MODIS sensor degradation
provided in the new C6 data products reduces the magnitude of the surface
reflectance and albedo decline trends obtained from previous MODIS data
(i.e.,
Collection 5, C5). Collection 5 and 6 data product analysis over GrIS is
characterized by surface (i.e., wet vs. dry) and elevation (i.e., 500–2000 m,
2000 m and greater) conditions over the summer season from 1 June to 31 August. Notably, the visible-wavelength declining reflectance trends identified
in several bands of MODIS C5 data from previous studies are only slightly
detected at reduced magnitude in the C6 versions over the dry snow area.
Declining albedo in the wet snow and ice area remains over the MODIS record
in the C6 product, albeit at a lower magnitude than obtained using C5 data.
Further analyses of C6 spectral reflectance trends show both reflectance
increases and decreases in select bands and regions, suggesting that several
competing processes are contributing to Greenland Ice Sheet albedo change.
Investigators using MODIS data for other ocean, atmosphere and/or land
analyses are urged to consider similar re-examinations of trends previously
established using C5 data.
Introduction
The Greenland Ice Sheet (GrIS) has experienced substantial mass loss during
the past three decades, resulting in sizeable contribution to sea level rise
(Krabill et al., 2000; Fettweis, 2007; van den Broeke et al., 2009; Rignot et
al., 2011; Velicogna and Wahr, 2013; Enderlin et al., 2014). Partitioned
estimates of GrIS mass loss have shown that surface melt contributes
substantially to annual ice loss (Box, 2013; van den Broeke et al., 2016).
Snow and ice surfaces become darker and have reduced visible-to-near-infrared (VNIR)
reflectance and albedo due to the deposition of particulates, re-emergence of
engrained particulates, biological activity, snow grain metamorphosis and the
presence of melt (LaChapelle, 1969; Warren and Wiscombe, 1980; Kohshima et
al., 1993; Painter et al., 2001; Takeuchi, 2009; Hodson et al., 2017). As
surface melt is both a factor and result of surface darkening, the potential
exists for a variety of melt–albedo feedbacks to further enhance shortwave
absorption and accelerate melt, increasing mass loss and sea level rise
contributions (Wiscombe and Warren, 1980; Tedesco et al., 2015). Moderate
Resolution Imaging Spectroradiometer (MODIS) observations are among the best
tools available to evaluate these albedo feedbacks. Several studies have used
MODIS data to study the magnitude of ongoing albedo trends and assess their
role in enhancing the impact of climate warming on the ice sheet (e.g.,
Tedesco et al., 2011, 2016; Box et al., 2012; Stroeve et al., 2013). A challenge in using long-term satellite records to detect change is
maintaining consistent instrument performance. Recent literature has
discussed MODIS Terra sensor calibration degradation and its impacts on
apparent data trends (e.g., Franz et al., 2008; Wang et al., 2012; Lyapustin
et al., 2014; Polashenski et al., 2015; Sayer et al., 2015). Specifically
related to GrIS trends, a recent work by Polashenski et al. (2015) indicated
that uncorrected sensor degradation in MODIS Collection 5 (C5) data,
particularly on the Terra platform, was contributing significantly to an
apparent albedo declining trend over the GrIS and that the albedo trend in
large areas of the ice sheet which do not experience melt (the dry snow area)
may disappear once Collection 6 (C6) calibrations are applied. This study
compares GrIS surface reflectance, snow and land albedo (bands 1–7) trends
obtained from MODIS C5 with those obtained from the newly released MODIS C6
data to identify and differentiate the impact of sensor calibration drift
from actual surface changes. We present these comparisons by spectral band
and spatial filters and discuss the implications that C6 calibration
enhancements have on our understanding of GrIS albedo decline and the
mechanisms driving this decline. The identified band 1–7, 459–2155 nm, and
broadband albedo spatial and temporal patterns and trends provide insight
toward the physical mechanisms likely dominating GrIS albedo decline, which
is key to understanding the surface energy balance of the Greenland Ice
Sheet.
Background
MODIS instruments operate onboard both the NASA Earth Observing System (EOS)
Terra and Aqua satellites and collect Earth observations in 36 spectral
bands ranging from 0.4–14.4 µm at spatial resolutions of
250–1000 m (Barnes et al., 1998). The MODIS data record is now in excess of
17 years for Terra and 15 years for Aqua; both sensors are operating well
past their 6-year design life. MODIS calibrations are updated periodically
to reflect new understanding of instrument changes, with the entire data
record reprocessed as a new “Collection”. Significant revisions in the
calibration approach were initiated in C6 (Toller et al., 2013), resulting
in a relatively large adjustment to end products. Before launch, both MODIS
sensors were calibrated via laboratory light sources. Because the optics of
the sensor (including reflecting mirrors and electronics) were expected to
degrade over time, the MODIS sensor design included methods for post-launch
onboard calibration. However, the onboard calibration did not sufficiently
account for degradation of the calibrators due to the large degradation of
the solar diffuser (detailed in Lyapustin et al., 2014). This led to a
long-term drift in calibration, most pronounced on the Terra sensor, with the largest
in the blue band (B3), and decreasing with increasing wavelength (Xiong and
Barnes, 2006).
The discovery of systemic nonphysical trends in MODIS Terra products by
science users (e.g., Kwiatkowska et al., 2008; Levy et al., 2010; Wang et
al., 2012) motivated research into independent trend characterization for C6
calibration. Analysis of observations of remote desert targets that were
assumed to be nearly invariant over the long term were used to constrain the
long-term drift of the observations. These so-called pseudo-invariant
targets such as deserts, deep convective clouds and high-elevation ice sheet
investigations led to a new vicarious approach for calibration, which relies
on collection of pseudo-invariant Earth site data continually to provide a
reference over time. The approach permits calibration at multiple angles of
incidence (AOIs) on the challenging-to-characterize scan mirror, rather than
two angles of incidence available through the onboard solar diffuser and
moon observations through the space view port (Sun et al., 2012; Lyapustin
et al., 2014). The resulting calibration is of lower precision than a
well-characterized mirror calibrated on a lunar standard but provides
significant improvement to the C5 data (Lyapustin et al., 2014). Long-term
degradation of the solar diffuser stability monitor and detectors will
continue to be evaluated and addressed as the MODIS record proceeds (Toller
et al., 2013).
The impact of C5 to C6 updates on higher-level MODIS products cannot be
represented by a simple offset because the magnitude of calibration revision
is dependent on mirror side, AOI and time. The correction reaches several
percent (in absolute reflectance values) in the worst cases (i.e., B3, near
the end of the C5 record, roughly 2013 to 2015). Lyapustin et al. (2014)
report C5 to C6 adjustment of B3 top-of-atmosphere (TOA) reflectance reaches
approximately 0.02 (from 0.225 to 0.245) at the Libya 4 pseudo-invariant
site by Terra mission day 5000. Residual error after C6 calibrations (when
compared to pseudo-invariant desert sites) is found to be on the order of
several tenths of 1 % in TOA reflectance (Lyapustin et al., 2014). The
residual error is now within the stated accuracy of the products but may
still be large enough to impact end user's scientific results, particularly
in products derived from band ratios such as aerosol or vegetation indices.
Further possible improvements to sensor calibration have been discussed
(Meister and Franz, 2014; Xiong et al., 2015), including those that address
polarization correction on the Terra sensor (Lyapustin et al., 2014). A
thorough description of the C5 calibration degradation can be found in
Lyapustin et al. (2014), and further details on C6 sensor characterization
can be found in Toller et al. (2013) and at the MODIS Characterization
Support Team website at http://mcst.gsfc.nasa.gov/calibration/information. MODIS C6 updates
included not only sensor calibration algorithms applied to Level 1 data, but
also updates to algorithms used to derive higher-level data products
(including products used in this study) as well as the aerosol retrieval
and correction algorithms, the cloud and cloud shadow detection algorithms,
and quality assurance bands (Levy et al., 2013; Platnick et al., 2015).
Detailed documentation of modifications to the products used in this paper
can be found in the MOD09 user guide at
http://modis-sr.ltdri.org/guide/MOD09_UserGuide_v1.4.pdf, the MOD10 user guide at
http://modis-snow-ice.gsfc.nasa.gov/uploads/C6_MODIS_Snow_User_Guide.pdf and
the MCD43 documentation at
https://www.umb.edu/spectralmass/terra_aqua_modis/v006. Our analysis investigates Terra and Aqua records from three
product types, namely MOD09/MYD09, MOD10/MYD10, and MCD43 for the entire
summer data record, though it does not directly compare MODIS C5 and C6 data
with ground data. Several MODIS cryospheric calibration and validation
investigations, some of which include in situ data, have been performed,
such as Stroeve et al. (2005, 2013), Moody et al. (2007), Hall et al. (2008),
Alexander et al. (2014), Wright et al. (2014),
Polashenski et al. (2015), and Zhan and Davies (2016). Recently, Box et al. (2017)
found Terra MOD10A1 albedo substantially improves in relative accuracy from
C5 to C6, agreeing with the Greenland Climate Network and Programme for Monitoring of the
Greenland Ice Sheet station data from
mid-May through August for the majority of GrIS south of 80∘ N
within 0.04 (unitless albedo from 0–1, see Box et al., 2017, Fig. 5b).
Accuracy of in situ ice sheet automated weather station measurements remains
challenging due to limitations of unattended stations and interference of
factors including ice riming, high wind speeds, low temperatures, ablation,
tilt and ice flow (van den Broeke et al., 2004; van As and Fausto, 2011). Considerable
biases have been reported in GrIS automated weather station data (Stroeve et
al., 2006, 2013; Wang et al., 2016; Ryan et al., 2017).
MethodsProcessing MODIS data
We analyzed the MODIS Terra and Aqua 8-day surface reflectance products
(MOD09A1/MYD09A1; Vermote, 2007, Vermote, 2015a for C6 Terra, Vermote, 2015b for C6 Aqua), Terra and Aqua daily snow
cover and broadband albedo products (MOD10A1/MYD10A1; Hall et al., 2006a for C5 Terra; Hall et al., 2006b for C5 Aqua; Hall and Riggs, 2016a for C6 Terra; Hall and Riggs, 2016b for C6 Aqua), and combined Terra and Aqua platform daily land
surface albedo products (MCD43A3; Schaaf and Wang, 2015) for both C5 and C6
collections over Greenland. The MOD09A1 and MYD09A1 land surface reflectance
products from the Terra and Aqua platforms, respectively, provide surface
reflectance in bands 1–7, corresponding to a wavelength range of 459–2155 nm (Table 1).
Data are provided as an 8-day product, which contains the
best available Level 2 gridded observation during the 8-day period, based on
observation coverage, view angle, clouds and aerosol loading. The
MOD10A1/MYD10A1 Terra/Aqua daily snow cover products include daily broadband
snow albedo and quality assurance observations, which are the focus of our
processing, in addition to a snow cover index calculated from Level 1
radiance data. The MCD43A3 land surface albedo product provides both direct
hemispherical reflectance (black-sky albedo, BSA) and bihemispherical
reflectance (white-sky albedo, WSA) for bands 1–7 from a bidirectional
reflectance distribution function (BRDF) inversion of all available
observations during the 16-day moving window centered on the date of
interest. Data are provided every 8 days in C5 and daily in C6.
Because the high latitude and low solar zenith angles over the GrIS are at
the extreme of MODIS capabilities, we filtered the data to ensure use of
only the highest quality retrievals for our analysis. MOD09A1/MYD09A1
(M*D09A1) data were filtered by using the band quality assurance layer where
all four band quality flags were set to 0 for each band, and solar zenith
angle observations were below 70∘. MOD10A1/MYD10A1 (M*D10A1) daily
snow cover data were filtered using methods similar to those in Box et al. (2012), where values outside the M*D10A1 scientific dataset albedo values are
excluded and a filter removes pixels whose albedo is more than 2 standard
deviations from the 11-day running median or which differ from the 11-day
running mean by more than 0.04 (albedo is a dimensionless number on a scale
of 0.0–1.0). Processing MCD43A3 datasets was carried out following methods
described in Stroeve et al. (2013) Sect. 2.1. We use only data with
the highest quality (inversion flag set to 0 in the MCD43A2 product), which
represents data derived from time periods where sufficient cloud-free, high-quality observations are available for full BRDF inversion. We did not use
data derived with the backup algorithm, even though evidence suggests it
performs almost as well (Stroeve et al., 2013). During the summer season
chosen for the processing timeframe (1 June–31 August), the majority of the data
are acquired near solar noon, ensuring many high-quality inversions in the
MCD43A3 data.
MODIS sensor reflective band 1–7 characteristics.
MODIS bandBandwidth(nm)Band 1620–670Band 2841–876Band 3459–479Band 4545–565Band 51230–1250Band 61628–1652Band 72105–2155
Location and topographic map of the Greenland Ice Sheet with 500 m
(green), 1000 m (black), 2000 m (blue) and 3000 m (black) contour lines
overlaid. Greenland surface elevation from Howat et al. (2014) is displayed,
where darker greys indicate lower elevation (minimum at sea level, 0 m) and
brighter greys and white indicate higher elevation (maximum at 3500 m).
After quality filtering, data are processed to produce an annual mean albedo
for dry and wet snow and ice areas of the ice sheet with methods of
Polashenski et al. (2015). Wet snow and ice are GrIS areas that have
experienced melt at any point during the current year prior to the date of
interest. Dry snow is snow which has not at any point experienced melt. We
do not return wet snow and ice to the dry snow category after it experiences
melt until the following year when we are certain the melt surface has been
buried, due to the residual impacts on albedo caused by melt occurrence. An
elevation mask is applied using a GrIS digital elevation model (Howat
et al., 2014) to group ice sheet areas at two different elevation bands (dry
snow is snow above 2000 m, wet snow and ice is snow and ice within the
elevations of 500 to 2000 m; Fig. 1). To these two elevation areas, high
elevation and low elevation, we apply a melt mask generated from the
regional climate model Modèle Atmosphérique Régional (MAR; Tedesco, 2014) to exclude pixels that did not match the predominant melt
condition. MAR-indicated dry and wet snow and ice pixels were defined as
those with no simulated melt occurring at any time during the summer (dry)
and those showing one or more melt events (wet). The use of a dual filter,
based on elevation and melt state, ensures that the areas discussed
represent iconic surface types without contamination (e.g., for the dry snow
area, the elevation cut off ensures non-melting bare ice at the margins of
the ice sheet does not contribute). From the mosaicked, filtered and masked
data, the average of the daily mean of all remaining pixels from 1 June to 31 August is calculated for each year of the record. Though the time
interval differs from the 15 May–15 July time interval chosen by
Polashenski et al. (2015) to match the time when solar elevation angle is
highest and albedo is most important to the GrIS, it is better aligned with
prior studies and still captures the key behavior of the high-insolation
period. We also conducted the same analysis for 15 May–15 July (not
shown) and found very similar trend revisions and overall behavior of ice
sheet albedo. Linear trends are calculated for display in map form using a
least squares linear regression to the data at each pixel location.
Terra MODIS GrIS C5 (dashed) and C6 (solid) average annual 1 June–31 August MOD09A1 surface reflectance
(B1–B7) and MOD10A1 broadband snow
albedo, denoted “albedo”, for dry (a) and wet snow and ice (b). (Note
the y-axis scale is different for dry vs. wet snow and ice for B1–4, albedo
and B5).
All MODIS data product tiles were downloaded from the NASA US Geological
Survey Land Processes Distributed Active Archive Center (LP DAAC) and National
Snow and Ice Data Center (NSIDC). Data tiles were mosaicked and resampled via nearest-neighbor
method to polar stereographic projection using the MODIS Reprojection Tool.
The MODIS Reprojection Tool utilized to mosaic and resample MODIS tile data
can be found at https://lpdaac.usgs.gov/tools/modis_reprojection_tool and the user guide at
https://lpdaac.usgs.gov/sites/default/files/public/mrt41_usermanual_032811.pdf.
Aqua MODIS GrIS C5 (dashed) and C6 (solid) average annual 1 June–31 August MYD09A1 surface reflectance (B1–B7) and MYD10A1 (broadband snow
albedo, denoted “albedo”) for dry (a) and wet snow and ice (b). (Note
the y-axis scale is different for dry vs. wet snow and ice for B1–4, albedo
and B5).
ResultsAnnual average summer M*D09A1 surface reflectance and M*D10A1 broadband
albedo
Annual average summer (1 June–31 August) GrIS surface reflectance for dry
and wet snow and ice areas from both Terra and Aqua data is presented
in Figs. 2 and 3, respectively, as derived from the M*D09A1 and M*D10A1 C5
and C6 products. The discrepancy between the two data collections is
indicated by the difference between dashed (C5) and solid (C6) lines of the
same color. The adjustment from C5 to C6 is significantly greater for Terra
than Aqua and greatest over the GrIS in the shortest wavelength bands,
consistent with the sensor calibration degradation reported by Lyapustin et al. (2014). C6 reduces the discrepancy between Terra and Aqua data
appreciably. Trends of the plotted data are quantified by linear regression
in Tables 2 and 3 along with their statistical significance. Significant
declining trends found in C5 Terra dry and wet snow and ice data with
magnitude exceeding 0.01 decade-1 are no longer present in C6 data. Very small
dry snow trends remain in C6 data, though not of strong statistical
significance. Thus, these likely do not represent real changes on the
surface. The C6 trend magnitude over the dry snow area is near the
calibration drift of several tenths of a percent, and the trends show an
incoherent pattern of albedo change. Specifically, B3, 459–479 nm, increases
slightly, while B1, 620–670 nm, decreases. This could not be produced by
expected physical mechanisms, for example, absorbing impurity concentration
or grain size changes, which would both be expected to shift B1 and B3 in
the same direction (see Warren and Wiscombe, 1980; Wiscombe and Warren,
1980, respectively). In wet snow and ice, significant trends seen in C5
Terra albedo nearly disappear in C6. Marginally, non-significant trends in
wet snow and ice albedo remain across C6 visible bands, at magnitudes
approximately one-third to one-half those of C5 data. We note that higher
interannual variability (noise) in the wet snow and bare ice area limits
trend significance at the p≤ 0.05 level, even though the absolute
magnitude of trends (signal) is larger than in the dry snow area. The
coherence across visible bands of marginally non-significant trends (nearly all bands decline for both sensors) and magnitude exceeding
sensor accuracy suggest a physically real trend is likely, if not proven
statistically by the satellite data products.
Trends (decade-1) and statistical significance of trends
for M*D09A1 surface reflectance bands 1–7 and M*D10A1 broadband albedo.
Statistically significant (where p≤ 0.05) trends are in bold,
and marginally significant trends (where p= 0.05 to 0.1) are in
italics.
Terra C5 dry snow Terra C6 dry snow Aqua C5 dry snow Aqua C6 dry snow BandTrendSignificanceTrendSignificanceTrendSignificanceTrendSignificanceB1-0.0270.003-0.0070.166-0.0060.377-0.0020.761B2-0.0270.017-0.0040.641-0.0100.289-0.0020.815B3-0.0510.0000.0030.493-0.0010.9020.0040.509B4-0.0180.013-0.0090.073-0.0020.718-0.0010.797B5-0.0090.326-0.0050.600-0.0120.221-0.0120.232B6-0.0080.166-0.0060.166-0.0070.211-0.0070.227B7-0.0040.301-0.0040.162-0.0040.263-0.0040.292Broadband albedo-0.0320.003-0.0020.808-0.0110.231-0.0080.419Terra C5 wet snow and ice Terra C6 wet snow and ice Aqua C5 wet snow and ice Aqua C6 wet snow and ice BandTrendSignificanceTrendSignificanceTrendSignificanceTrendSignificanceB1-0.0370.029-0.0230.102-0.0230.198-0.0200.222B2-0.0420.024-0.0200.194-0.0260.166-0.0220.223B3-0.0540.003-0.0170.227-0.0190.259-0.0180.281B4-0.0290.069-0.0250.073-0.0200.237-0.0200.202B5-0.0270.043-0.0140.257-0.0230.144-0.0240.115B6-0.0140.001-0.0060.067-0.0080.102-0.0090.087B7-0.0080.001-0.0030.075-0.0040.123-0.0050.109Broadband albedo-0.0400.001-0.0130.166-0.0090.364-0.0070.495Annual average summer MCD43A3 albedo
Annual average summer (1 June–31 August) GrIS C5 and C6 MCD43A3 land surface
direct hemispherical reflectances are presented in Fig. 4 for dry and wet
snow and ice, showing results expected from a combination of Terra and Aqua
data. Dashed lines represent C5 data, while solid lines represent C6.
Revisions across the duration of the MODIS record are apparent between the
MCD43A3 C5 and C6 data in both dry and wet snow and ice areas. Similar to
M*D09A1 and M*D10A1 products, the MCD43A3 revisions result in a considerable
decrease in trend magnitude. Statistically significant dry snow direct
hemispherical reflectance declines, which had been apparent in C5 data, are
reduced to magnitudes of or under 0.01 decade-1 in C6. MCD43A3 direct
hemispherical and bihemispherical reflectance trends (Table 3) remain
significant or near-significant for B1 and B4, with incoherent patterns for
other visible bands (B2, B3). As discussed in Sect. 4.1, the spectral
pattern, with conflicting trends in B2 and B3, indicates that these changes are
likely not linked to physical processes. Wet snow and ice areas still
exhibit coherent declining albedo trends after C6 revisions across B1
through B7, albeit of slightly lower magnitude than in C5. Like C5,
individual band wet snow and ice trends have low (B2, B3) to marginal (B1)
statistical significance due to large interannual variability in the wet
snow and ice albedo (Fig. 4). Coherence across bands (as presented in
Table 3 “BSA C6 wet snow and ice” and “WSA C6 wet snow and ice” trend columns), however,
increases confidence in the trends. Impacts of C5 to C6 revision are very
similar on WSA and BSA.
Greenland Ice Sheet average summer (1 June–31 August) MCD43A3 BSA C5
and C6 band 1–7 albedo for dry (a) and wet snow and ice (b). (Note
the y-axis scale is different for dry vs. wet snow and ice for B1–4 and B5).
Interestingly, there is a separation in C5 to C6 B1, B2, B4, B5 and B6 albedo
values throughout the MODIS record.
Spatial pattern of albedo trend
Maps of the decadal M*D10A1 broadband snow albedo trend over the entire
MODIS record are shown in Fig. 5 for C5 (a, d) and C6 (b, c, e, f) data from
both Terra and Aqua sensors. Both sensors and both collections show similar
spatial patterns, with the greatest albedo declines at low elevations of the ice
sheet (Fig. 1), particularly on the western and southeastern margins. A
statistically significant, ice-sheet-wide declining albedo trend in C5 Terra
data is largely absent in C5 Aqua sensor data. The discrepancy between C5
Terra and Aqua data is not spatially dependent. C6 revisions change Terra
trends upward by approximately 0.03 decade-1 across the ice sheet. Aqua
revisions are smaller but also result in trends that are less negative in
C6 than C5. Using revised C6 data, we mask the areas that have negligible
trend (-0.01 to +0.01 decade-1) in Fig. 5c, f. The magnitude of a
“negligible” trend was determined by considering the errors of several
tenths of a percent trend per decade remaining in C6 data collected to that
obtained over pseudo-invariant desert sites (Lyapustin et al., 2014). Trends
below this value must be considered below detection limit. The region of
negligible trends covers nearly all of the dry snow and large portions of
the upper reaches of the wet snow and ice, indicating that, even in some
areas where it is well known that melt frequency is increasing (e.g.,
Fettweis et al., 2011; Box et al., 2012; Fausto et al., 2016), albedo trends
are small enough to be challenging to confirm over the duration of the MODIS
record. Trends with magnitude greater than sensor calibration limits of
∼ 0.01 decade-1 are almost all negative. Significant areas of
the southern ice sheet exhibit trends near -0.01 decade-1, and a narrow band of
substantial albedo decline, reaching -0.04 decade-1, exists in areas around
the periphery of the ice sheet. Based on MAR analysis, the positive albedo
trends over northeastern Greenland are likely associated with a shift from
no trend in accumulation to a trend of increasing accumulation (by 35 Gt yr-1), starting in 2013. Direct hemispherical reflectance trends for select
visible and near-infrared bands of MCD43A3 C6 data are presented in Fig. 6. Both positive and negative trends of statistical significance exist in
all bands around the periphery of the ice sheet. These are discussed in
detail below. Band 1 and band 3 data show spatially uniform trends that are
generally under the sensor calibration accuracy of ∼ 0.01 decade-1
across the interior of the ice sheet. Band 1 red visible light trends are
positive and band 3 blue visible light trends are negative, a behavior
inconsistent with albedo change caused by either light-absorbing impurity
deposition or changes in surface grain properties, strongly suggesting
these changes are not physically real. NIR bands 2 and 5 show trends in the
upper elevations of the ice sheet that regionally exceed sensor calibration
accuracy and significance, with notable spatial patterns. Positive albedo
trends in these NIR bands dominate northeast Greenland, while negative
trends cover most of the remainder of the ice sheet. Band 5, in the near-infrared, is highly sensitive to grain size impacts. Figure 6d shows the
opposing trends of albedo increasing primarily in the northeast and
decreasing in the west and southern periphery. Figure 4 B5 shows modest
interannual variability in the dry snow and more variability in the wet snow
and ice areas, suggesting wet snow and ice albedo is worth investigating on
seasonal and/or annual scales.
Maps (a, b) depict Terra MOD10A1 C5 and C6 2002–2016 decadal trend
in broadband albedo. Maps (d, e) depict Aqua MYD10A1 C5 and C6 2003–2016
decadal trend in broadband albedo. Maps (c, f) show the same C6 decadal
Terra and Aqua trends, respectively, with the sensor trends of ±0.01 decade-1 masked out as white to visualize trends in albedo that are
larger than expected sensor calibration uncertainty, including albedo
decline (southeast and periphery GrIS) and albedo increase (northeast GrIS).
DiscussionImpact of C6 revision on scientific investigation of GrIS
The MODIS C5 to C6 revisions, and new surface reflectance and albedo trends
over GrIS, have substantial ramifications for research seeking to evaluate
the cause of enhanced surface melt, and hence mass loss, from the GrIS. Over
the dry snow area, C5 Terra data and combined sensor data indicated a
decadal trend of declining reflectance at a rate of up to several percent
per decade, with high statistical significance (see Tables 2, 3; Fig. 5a).
The albedo decline was greatest in short wavelength visible bands – a
spectral signature consistent with enhanced dust deposition on the ice sheet
(Dumont et al., 2014). The data, therefore, suggested that a snow albedo
feedback initiated by dry snow processes resulted in increased GrIS melt.
C6 revisions, particularly to short wavelength bands of Terra data,
appreciably reduce these trends. C6 exhibits no statistically significant
trends in visible nor NIR wavelength band surface reflectance from either
sensor over the aggregated dry snow area exceeding the sensor calibration
accuracy of ∼ 0.01 decade-1. Statistical significance of trends
in MCD43 B4 C6 data is misleading. These trends, of a few tenths of a
percent per decade, are within the range of a residual calibration error and
their statistical significance may in fact reflect the ease with which a
calibration error can be detected as a significant trend when superimposed
on the relatively constant albedo of dry snow. Care in examination of the
spatial and spectral variability and coherency between these trends (Figs. 5, 6) is recommended for future regional, in situ and/or process studies.
Trends in C6 dry snow visible albedo (B1 and B3) are suspiciously consistent
across the ice sheet, and opposite trends between these bands are
inconsistent with expected mechanisms of albedo change. We conclude that dry
snow visible band albedo is stable within MODIS' capabilities. This stable
visible-wavelength albedo of the GrIS dry snow area supports conclusions in
Polashenski et al. (2015), who found no in situ evidence of continual
enhanced black carbon or dust deposition to support C5 trends. To note,
forest fire events in North America and Asia have resulted in black carbon
deposition to GrIS (e.g., Zennaro et al., 2014; Thomas et al., 2017), and such
events have been predicted to increase during periods of drought in a
warming climate (Soja et al., 2007; Flannigan et al., 2013). Deposition of
black carbon and other absorbing impurities in the dry snow area is often
buried by new snowfall; however, these impurities often have a stronger
influence in reducing reflectance and albedo in wet snow and ice areas. In
contrast to B1 and B3 decadal trends which are nearly stable in aggregate,
NIR wavelength albedos (B2 and B5) show significant regional trends.
Positive trends in NE Greenland offset negative trends across much of the
remainder of the ice sheet in the aggregate dry snow data. The regional
trends are statistically significant over large areas and exceed the
expected magnitude of remaining calibration errors. The trends appear to
indicate changing snow grain metamorphism. We speculate, based on the
spectral change, that enhanced snowfall in the dry areas of NE Greenland is
causing more rapid surface burial and a lower age of surface grains, while
increased temperatures and occasional melt are increasing grain size on
southern and western areas of GrIS, where deposition was already relatively
frequent. We find that the pattern does not appear to coincide well with the
area of enhanced melt, as detected by passive microwave products (Tedesco et
al., 2014) over the MODIS era (2000–present), indicating that this indeed appears to be
more related to dry snow grain metamorphism and snowfall frequency than melt
(see Fig. 7).
Trends (decade-1) and statistical significance of trends
for MCD43A3 directional hemispherical reflectance (or black-sky albedo, BSA)
and bihemispherical reflectance (or white-sky albedo, WSA). Statistically
significant (where p≤ 0.05) trends are in bold, and
marginally significant trends (where p= 0.05 to 0.1) are in
italics.
MCD43A3 C6 band 3 (459–479 nm), band 1 (620–670 nm), band 2
(841–876 nm) and band 5 (1230–1250 nm) 2003–2016 albedo trends.
Band-specific maps show considerable spatial complexity in albedo trends.
In wet snow and ice areas, the annual average broadband albedo and
visible-wavelength reflectance shown for all products exhibits a downward trend for
C5 and C6 data. The magnitude of these trends is reduced by approximately
one half from C5 to C6 for MOD09A1 and by a small amount in MYD09A1 and
MCD43A3. Statistical significance is harder to establish in the wet snow and
ice areas due to much higher interannual variability (wet snow and ice,
Figs. 2, 3, and 4). The C5 to C6 revisions to wet snow and ice albedo
trends impact our understanding of GrIS surface energy and mass budgets and
suggest a smaller role for albedo feedback in driving Greenland mass loss
than previously indicated. The C5 to C6 revisions do not, however, demand a
change in conclusions about reflectance and albedo declining trend existence
in the wet snow and ice areas. Wet snow and ice areas on GrIS, in aggregate,
still show coherent albedo decline across the nearly all visible and NIR
wavelength bands from both sensors in C6 data (Tables 2, 3). Reduction in
reflectance and albedo of this magnitude (several tenths) has been shown to
result in radiative forcing of several tens of W m-2 (e.g., 0.4
broadband albedo decline from dust on snow resulting in 80 W m-2 radiative forcing, Painter et al., 2007; 0.3 reduction in broadband
visible albedo from black carbon and impurities on snow resulting in 70 W m-2 radiative forcing, Casey et al., 2017). The MODIS C6 reflectance
and albedo data product results provide strong supporting evidence that
enhanced melt processes (including melt-induced snow microstructure changes
and melt-induced light-absorbing impurity accumulation changes) are creating
an albedo feedback on the GrIS periphery.
Average number of melt days experienced from 1 June to 31 August as
determined from spaceborne passive microwave data (Tedesco, 2014). Note that
the NE GrIS high-elevation areas show only a few melt days, almost entirely
from 2012. The pattern of melt is inconsistent with the pattern of NIR
albedo change. Both low and high elevations in central and NE Greenland show
NIR albedo increases while only low elevations experience melt (and the
duration is in fact increasing, not shown). Additionally, the boundary between NIR
albedo positive and negative trends falls in the middle of the area with
little melt but closely tracks the summit ridge of the ice sheet,
suggesting it is accumulation related rather than melt related.
The cause of these changes is important. We examined trends regionally along
the margins of the ice sheet (Figs. 5 and 6) and find that trends can be
supported by reasonable inferences about regional differences in snowfall
frequency, melt duration and surface exposure of light-absorbing impurities
on albedo control. In west Greenland, from Humboldt Glacier to Jakobshavn
Glacier, substantial trends in NIR (B2, B5) albedo across a wide elevation
range indicate increased presence of melting conditions. Trends in visible
albedo are spatially correlated and most dominant at lower elevations where
melt accumulation of impurities, exposure of bare ice and algal growth
occurs. South of Jakobshavn on the western margin of the ice sheet, B1 and
B3 visible albedo are sharply declining but are accompanied by rising B5 NIR
albedo. We speculate that this signature could be evidence for substantial
surface exposure and accumulation of mineral impurities in this region of
GrIS. The spectra of most minerals exhibit higher NIR reflectance than bare
ice, leading to the potential for a trend toward increased NIR albedo while
visible albedo drops with high mineral content (Adams and Filice, 1967;
Painter et al., 2003; Bøggild et al., 2010; Casey et al., 2012; Tedesco
et al., 2013). The SE GrIS margin shows decreasing NIR albedo and only
isolated change in visible wavelengths at the lowest elevations. We
interpret this similarly to NW Greenland, only with much higher accumulation
in this region. NIR impacts of increased wet snow and ice presence dominate.
Higher accumulation drives the equilibrium line to a lower elevation, and
melt accumulation of light-absorbing impurities plays a substantial role in
lowering visible albedo below this elevation. Continuing around the
periphery of the northern GrIS, the remainder of the ice sheet margin from
approximately Scoresbysund to Humboldt Glacier, we see positive trends in NIR
albedo and no significant trends in short wavelength visible albedo. In this
region of low annual accumulation and long surface exposure times, this
signature appears to indicate increasing snowfall, which is confirmed by our
separate preliminary analysis of MAR data in northeastern Greenland showing
an increase in accumulation patterns starting in 2013 as well as GrIS
surface mass balance climate model results (e.g., Noël et al., 2015). A
very small addition of snowfall would cause more rapid surface burial and
result in a lower age of surface grains and higher NIR albedo. Ultimately,
each of these interpretations only clarifies what physical mechanisms would
be consistent with the spectral signature changes observed. These hypotheses
should be considered provisional and tested by in situ observation of snow
and ice properties, which may be guided by satellite-identified signals.
Future use of MODIS data
The MODIS record is a powerful tool for assessing surface reflectance and
albedo changes remotely. Our results indicate that future investigations
should use the latest data recalibration (currently C6) data, and
investigators should be aware of the limitations of the sensors, which we
here attempt to restate plainly for the community's benefit:
Absolute trends in reflectance and albedo on the order of 0.01 decade-1 are near
the limits of the sensor calibration accuracy. Though statistically
significant albedo trends of 0.01 or less TOA reflectance may exist over
some surfaces with particularly stable albedo, trends at this level should
be considered provisional and evaluated with great care, as they might not
reflect actual physical processes.
Calibration drift is band dependent. Small band-dependent calibration
degradations can be magnified in band ratio products, such as those used to
detect dust mineralogy or aerosols, indicating spurious trends.
Data limitations are greatest in recently collected data. Vicarious C6
calibration, based on pseudo-invariant Earth sites, may not fully capture
emerging trends in sensor degradation. Increasing Terra–Aqua discrepancies
appear since 2014 in C6 data.
When the sensors disagree in ways not explained by overpass time, MODIS Aqua
is likely to provide more stable data as Terra's calibration is expected to
continue to degrade in ways that will make it challenging to characterize.
Since C6 calibration is now vicarious (based on observation of
pseudo-invariant desert sites), it is likely that emerging trends in sensor
behavior will take some time to manifest in a statistically significant way
for calibration revisions.
Conclusions
MODIS C6 calibration revisions result in substantial modification of decadal
albedo trends on the GrIS reported by prior authors based on C5 data (e.g.,
Box et al., 2012; Stroeve et al., 2013; He et al., 2013; Polashenski et al.,
2015). Declining C5 surface reflectance trends that were particularly
pronounced in Terra's shortest wavelength bands are smaller or absent in C6
data. MODIS C6 surface reflectance and albedo data over dry snow areas of
the GrIS feature mostly small, non-statistically significant trends in
visible and NIR wavelengths. (The exceptions are the marginally statistically
significant B1 decline in BSA and WSA dry snow albedo and the statistically
significant B4 decline in BSA and WSA dry snow albedo.) These findings are
consistent with the recent study of Polashenski et al. (2015), which suggested
the dry snow albedo decline in C5 data would disappear in C6 after finding
no enhancement in light-absorbing impurity concentrations on the interior
Greenland Ice Sheet. The declining trends in wet snow and ice surface
reflectance and albedo, independently supported by evidence of increased
melt activity (Nghiem et al., 2012; Fausto et al., 2016b), remain
statistically significant in C6 data, though at lower magnitude.
An examination of spatial, wavelength-specific variability in C6 albedo
trends indicates several interesting attributes of GrIS albedo decline that
may motivate future work to better understand the mechanisms controlling
albedo feedbacks on the ice sheet. At higher elevations, patterns of NIR
albedo change, including increasing reflectance in NE Greenland and
declining reflectance in southern and western Greenland, highlight possible
regional changes in metamorphism, precipitation and surface constituents. In
the ablation zone, ratios of visible and NIR albedo trends suggest
that differences in snowfall frequency, melt duration and surface exposure of
light-absorbing impurities control recent albedo trends – with the net impact of
these mechanisms being complex and likely dependent on the location and the
seasonal timeframe chosen. Though the majority of albedo reduction occurs on
the GrIS in melt-impacted areas, these results may support a crucial role
for snow grain metamorphism in initiating (or preventing) feedbacks in dry
snow.
Melt-related albedo reductions continue to have the potential to trigger
significant ice–albedo feedbacks and accelerate melt and surface mass loss,
and, as a result, melt initiation remains a critical process. Our findings,
particularly those illustrating the regional complexity in spectral albedo
trends, highlight the need for future work on GrIS albedo to define and
differentiate the role of processes that control albedo decline.
The implications of this study extend beyond Greenland. The substantial
revision from C5 to C6 MODIS products impacts a broad array of
investigations. Conclusions based upon trends from C5 data, particularly
shorter wavelength band Terra data, should be re-examined for robustness
with C6 products. Future investigators should also note the limitations of
MODIS products. Investigators should use great caution in evaluating trends
of ∼ 0.01 decade-1 or smaller and note that C6 corrections may
not fully capture recent and emerging trends in sensor degradation,
particularly on the challenging-to-characterize Terra sensor.
MODIS data are freely available for download at the NASA US Geological Survey LP
DAAC (https://lpdaac.usgs.gov) and NSIDC (https://nsidc.org). MAR data are available for download at the NSF Arctic Data Center (https://arcticdata.io).
The authors declare that they have no conflict of interest.
Acknowledgements
We acknowledge funding from NSF grants ARC-1204145 and 1304807 and from NASA grants
NNX14AE72G, NNX14AD98G and NNX16AO75G. We thank NASA EOS, LP DAAC and NSIDC for
providing MODIS data and Crystal Schaaf and Qingsong Sun for providing an executable
file to grid recently released un-gridded MCD43 C5 data. We thank Marie Dumont for her service as editor as well as Jason Box and an anonymous
reviewer for constructive comments which improved this manuscript.
Edited by: Marie Dumont
Reviewed by: Jason Box and one anonymous referee
ReferencesAdams, J. B. and Filice, A. L.: Spectral reflectance 0.4 to 2.0 microns
of silicate rock powders, J. Geophys. Res., 72,
5705–5715, 10.1029/JZ072i022p05705, 1967.Alexander, P. M., Tedesco, M., Fettweis, X., van de Wal, R. S. W., Smeets, C. J. P. P., and van den Broeke, M. R.:
Assessing spatio-temporal variability and trends in modelled and measured Greenland Ice Sheet albedo (2000–2013), The Cryosphere, 8, 2293–2312, 10.5194/tc-8-2293-2014, 2014.Barnes, W. L., Pagano, T. S., and Salomonson, V. V.: Prelaunch
characteristics of the Moderate Resolution Imaging Spectroradiometer (MODIS)
on EOS-AM1, IEEE T. Geosci. Remote, 36, 4, 1088–1100, 10.1109/36.700993, 1998.
Bøggild, C. E., Brandt, R. E., Brown, K. J., and Warren, S. G.: The ablation
zone in northeast Greenland: ice types, albedos and impurities, J. Glaciol., 56, 101–113, 2010.Box, J. E., Fettweis, X., Stroeve, J. C., Tedesco, M., Hall, D. K., and Steffen, K.: Greenland ice sheet albedo feedback:
thermodynamics and atmospheric drivers, The Cryosphere, 6, 821–839, 10.5194/tc-6-821-2012, 2012.
Box, J. E.: Greenland Ice Sheet Mass Balance Reconstruction, Part II:
Surface Mass Balance (1840–2010), J. Climate, 26, 6974–6989, 2013.
Box, J. E., van As, D., and Steffen, K.: Greenland, Canadian and Icelandic
land ice albedo grids (2000–2016), Geological Survey of Denmark and
Greenland Bulletin, 38, 69–72, 2017.Casey, K. A., Kääb, A., and Benn, D. I.: Geochemical characterization of supraglacial debris via in situ
and optical remote sensing methods: a case study in Khumbu Himalaya, Nepal, The Cryosphere, 6, 85–100, 10.5194/tc-6-85-2012, 2012.Casey, K. A., Kaspari, S. D., Skiles, S. M., Kreutz, K., and Handley, M. J.: The
spectral and chemical measurement of pollutants on snow near South Pole,
Antarctica, J. Geophys. Res.-Atmos., 122,
10.1002/2016JD026418, 2017.Dumont, M., Brun, E., Picard, G., Michou, M., Libois, Q., Petit, J. R., Geyer,
M.,
Morin, S., and Josse, B.: Contribution of light-absorbing impurities in
snow to Greenland's darkening since 2009, Nat. Geosci., 7, 509–512,
10.1038/ngeo2180, 2014.Enderlin, E. M., Howat, I. M., Jeong, S., Noh, M. J., van Angelen, J. H., and van
den Broeke, M. R.: An improved mass budget for the Greenland ice sheet,
Geophys. Res. Lett., 41, 866–872, 10.1002/2013GL059010, 2014.Fausto, R. S., van As, D., Box, J. E., Colgan, W., and Langen, P. L.:
Quantifying the surface energy fluxes in South Greenland during the 2012
high melt episodes using in-situ observations, Front. Earth Sci.,
4, 82, 10.3389/feart.2016.00082, 2016a.Fausto, R. S., van As, D., Box, J. E., Colgan, W., Langen, P. L., and Mottram, R.
H.:
The implication of nonradiative energy fluxes dominating Greenland
ice sheet exceptional ablation area surface melt in 2012, Geophys. Res. Lett., 43, 2649–2658, 10.1002/2016GL067720, 2016b.Fettweis, X.: Reconstruction of the 1979–2006 Greenland ice sheet surface mass balance using the regional climate model MAR,
The Cryosphere, 1, 21–40, 10.5194/tc-1-21-2007, 2007.Fettweis, X., Tedesco, M., van den Broeke, M., and Ettema, J.: Melting trends over the Greenland ice sheet (1958–2009)
from spaceborne microwave data and regional climate models, The Cryosphere, 5, 359–375, 10.5194/tc-5-359-2011, 2011.Flannigan, M., Cantin, A. S., de Groot, W. J., Wotton, M., Newbery, A.,
and Gowman, L. M.: Global wildland fire season severity in the 21st
century, Forest Ecol. Manag., 294, 54–61,
10.1016/j.foreco.2012.10.022, 2013.
Franz, B. A., Kwiatkowska, E. J., Meister, G., and McClain, C. R.: Moderate
Resolution Imaging Spectroradiometer on Terra: limitations for ocean color
applications, J. Appl. Remote Sens., 2, 023525, 17 pp., 2008.Hall, D. K., Salomonson, V. V., and Riggs, G. A.: MODIS Terra Snow Cover Daily L3 Global 500m Grid,
Version 5, Boulder, Colorado, USA, NASA National Snow and Ice Data Center Distributed Active Archive Center, 10.5067/63NQASRDPDB0, 2006a.Hall, D. K., Salomonson, V. V., and Riggs, G. A.: MODIS Aqua Snow Cover Daily L3 Global 500m Grid, Version 5,
Boulder, Colorado, USA, NASA National Snow and Ice Data Center Distributed Active Archive Center, 10.5067/ZFAEMQGSR4XD, 2006b.Hall, D. K., Box, J. E., Casey, K. A., Hook, S. J., Shuman, C. A., and Steffen,
K.:
Comparison of satellite-derived and in-situ observations of ice and
snow surface temperatures over Greenland, Remote Sens. Environ.,
112, 3739–3749, 10.1016/j.rse.2008.05.007, 2008.Hall, D. K. and Riggs, G. A.: MODIS Terra Snow Cover Daily L3 Global 500m Grid, Version 6, Boulder, Colorado, USA,
NASA National Snow and Ice Data Center Distributed Active Archive Center, 10.5067/MODIS/MOD10A1.006, 2016a.Hall, D. K. and Riggs, G. A.: MODIS Aqua Snow Cover Daily L3 Global 500m Grid, Version 6, Boulder, Colorado, USA,
NASA National Snow and Ice Data Center Distributed Active Archive Center, 10.5067/MODIS/MYD10A1.006, 2016b.He, T., Liang, S., Yu, Y., Wang, D., Gao, F., and Liu, Q.: Greenland surface
albedo changes in July 1981-2012 from satellite observations, Environ. Res. Lett., 8, 044043,
10.1088/1748-9326/8/4/044043, 2013.Hodson, A. J., Nowak, A., Cook, J., Sabacka, M., Wharfe, E. S., Pearce, D. A.,
Convey, P., and Vieira, G.: Microbes influence the biogeochemical and optical
properties of maritime Antarctic snow, J. Geophys. Res.-Biogeo., 122, 1456–1470, 10.1002/2016JG003694, 2017.Howat, I. M., Negrete, A., and Smith, B. E.: The Greenland Ice Mapping Project (GIMP) land classification and surface
elevation data sets, The Cryosphere, 8, 1509–1518, 10.5194/tc-8-1509-2014, 2014.
Kohshima, S., Seko, K., and Yoshimura, Y.: Biotic acceleration of glacier
melting in Yala glacier, Langtang region, Nepal Himalaya, Proceedings of the
Kathmandu Symposium, November 1992, IAHS Publication 218, 1993.
Krabill, W., Abdalati, W., Frederick, E., Manizade, S., Martin, C., Sonntag,
J.,
Swift, R., Thomas, R., Wright, W., and Yungel, J.: Greenland Ice Sheet: High
Elevation Balance and Peripheral Thinning, Science, 289, 428–430, 2000.
Kwiatkowska, E. J., Franz, B. A., Meister, G., McClain, C. R., and Xiong, X.:
Cross calibration of ocean-color bands from Moderate-Resolution Imaging
Spectroradiometer on Terra platform, Appl. Opt., 47, 6796–6810, 2008.
LaChapelle, E. R.: Field Guide to Snow Crystals, University of
Washington Press, Seattle, WA, USA, 112 pp., 1969.Levy, R. C., Remer, L. A., Kleidman, R. G., Mattoo, S., Ichoku, C., Kahn, R., and Eck, T. F.: Global evaluation of the
Collection 5 MODIS dark-target aerosol products over land, Atmos. Chem. Phys., 10, 10399–10420, 10.5194/acp-10-10399-2010, 2010.Levy, R. C., Mattoo, S., Munchak, L. A., Remer, L. A., Sayer, A. M., Patadia, F., and Hsu, N. C.: The Collection 6 MODIS
aerosol products over land and ocean, Atmos. Meas. Tech., 6, 2989–3034, 10.5194/amt-6-2989-2013, 2013.Lyapustin, A., Wang, Y., Xiong, X., Meister, G., Platnick, S., Levy, R., Franz, B., Korkin, S., Hilker, T., Tucker, J., Hall, F.,
Sellers, P., Wu, A., and Angal, A.: Scientific impact of MODIS C5 calibration degradation and C6+ improvements,
Atmos. Meas. Tech., 7, 4353–4365, 10.5194/amt-7-4353-2014, 2014.Meister, G. and Franz, B. A.: Corrections to the MODIS Aqua calibration
derived from MODIS Aqua ocean color products, IEEE T.
Geosci. Remote, 52, 6534–6541,
10.1109/TGRS.2013.2297233, 2014.
Moody, E. G., King, M. D., Schaaf, C. B., Hall, D. K., and Platnick, S.:
Northern Hemisphere five-year average (2000–2004) spectral albedos of
surfaces in the presence of snow: statistics computed from Terra MODIS land
products, Remote Sens. Environ., 111, 337–345, 2007.Nghiem, S. V., Hall, D. K., Mote, T. L., Tedesco, M., Albert, M. R., Keegan, K.,
Shuman, C. A., DiGirolamo, N. E., and Neumann, G.: The extreme melt across the
Greenland ice sheet in 2012, Geophys. Res. Lett., 39, L20502,
10.1029/2012GL053611, 2012.Noël, B., van de Berg, W. J., van Meijgaard, E., Kuipers Munneke, P., van de Wal, R. S. W., and van den Broeke, M. R.:
Evaluation of the updated regional climate model RACMO2.3: summer snowfall impact on the Greenland Ice Sheet, The Cryosphere, 9, 1831–1844, 10.5194/tc-9-1831-2015, 2015.
Painter, T. H., Duval, B., Thomas, W. H., Mendez, M., Heintzelman, S., and Dozier,
J.:
Detection and quantification of snow algae with an airborne imaging
spectrometer, Appl. Environ. Microbiol., 67, 5267–5272, 2001.
Painter, T. H., Dozier, J., Roberts, D. A., Davis, R. E., and Green, R. O.:
Retrieval of subpixel snow-covered area and grain size from imaging
spectrometer data, Remote Sens. Environ., 85, 64–77, 2003.Painter, T. H., Barrett, A. P., Landry, C. L., Neff, J. C., Cassidy, M. P.,
Lawrence, C. R., McBride, K. E., and Farmer, G. L.: Impact of disturbed desert
soils on duration of mountain snow cover, Geophys. Res. Lett., 34,
L12502, 10.1029/2007GL030284, 2007.
Platnick, S., King, M., Meyer, K. G., Wind, G., Amarasinghe, N., Marchant,
B.,
Arnold, G. T., Zhang, Z., Hubanks, P. A., Ridgway, B., and Riedi, J.: MODIS
Cloud Optical Properties: User Guide for the Collection 6 Level-2
MOD06/MYD06 Product and Associated Level-3 Datasets, Version 1.0, 145 pp.,
2015.Polashenski, C. M., Dibb, J. E., Flanner, M. G., Chen, J. Y., Courville, Z. R.,
Lai, A. M., Schauer, J. J., Shafer, M. M., and Bergin, M.: Neither dust nor black
carbon causing apparent albedo decline in Greenland's dry snow zone:
Implications for MODIS C5 surface reflectance, Geophys. Res. Lett.,
42, 9319–9327, 10.1002/2015GL065912, 2015.Rignot, E., Velicogna, I., van den Broeke, M. R., Monaghan, A., and Lenaerts, J. T.
M.:
Acceleration of the contribution of the Greenland and Antarctic ice
sheets to sea level rise, Geophys. Res. Lett., 38, L05503,
10.1029/2011GL046583, 2011.Ryan, J. C., Hubbard, A., Irvine-Flynn, T. D., Doyle, S. H., Cook, J. M., Stibal,
M., and
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., 44, 6218–6225, 10.1002/2017GL073661, 2017.
Sayer, A. M., Hsu, N. C., Bettenhausen, C., Jeong, M. J., and Meister, G.:
Effect of MODIS Terra radiometric calibration improvements on Collection 6
Deep Blue aerosol products: Validation and Terra/Aqua consistency, J. Geophys. Res.-Atmos., 120, 12157–12174, 2015.Schaaf, C. and Wang, Z.: MCD43A3 MODIS/Terra+Aqua BRDF/Albedo Quality
Daily L3 Global – 500 m V006, NASA EOSDIS Land Processes Distributed Active
Archive Center (LP DAAC), USGS/Earth Resources Observation and Science
(EROS) Center, Sioux Falls, South Dakota, USA,
10.5067/MODIS/MCD43A3.006, 2015.
Soja, A. J., Tchebakova, N. M., French, N. H. F., Flannigan, M. D., Shugart, H.
H.,
Stocks, B. J., Sukhinin, A. I., Parfenova, E. I., Chapin III, F. S., and Stackhouse Jr., P. W.: Climate-induced boreal forest change: Predictions versus
current observations, Global Planet. Change, 56, 274–296, 2007.
Stroeve, J. C., Box, J. E., Gao, F., Liang, S., Nolin, A., and Schaaf, C.:
Accuracy assessment of the MODIS 16-day albedo product for snow: comparisons
with Greenland in situ measurements, Remote Sens. Environ., 94,
46–60, 2005.
Stroeve, J. C., Box, J. E., and Haran, T.: Evaluation of the MODIS (MOD10A1)
daily snow albedo product over the Greenland ice sheet, Remote Sens.
Environ., 105, 155–171, 2006.
Stroeve, J., Box, J. E., Wang, Z., Schaaf, C., and Barrett, A.: Re-evaluation
of MODIS MCD43 Greenland albedo accuracy and trends, Remote Sens.
Environ., 138, 199–214, 2013.Sun, J., Angal, A., Xiong, X., Chen, H., Geng, X., Wu, A., Choi, T., and Chu, M.:
MODIS RSB calibration improvements in Collection 6, Proc. SPIE 8528,
Earth Observing Missions and Sensors: Development, Implementation and
Characterization II, 85280N, 10.1117/12.979733, 2012.
Takeuchi, N.: Temporal and spatial variations in spectral
reflectance and characteristics of surface dust on Gulkana Glacier, Alaska
range, J. Glaciol., 55, 701–709, 2009.
Tedesco, M.: Greenland Daily Surface Melt 25km EASE-Grid,
(2001–2015), New York, NY, USA: City College of New York, The City
University of New York, Digital media, 2014.
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., and Box, J. E.: The role of albedo
and accumulation in the 2010 melting record in Greenland, Environ. Res. Lett., 6, 014005, 6 pp., 2011.
Tedesco, M., Foreman, C. M., Anton, J., Steiner, N., and Schwartzman, T.:
Comparative analysis of morphological, mineralogical and spectral properties
of cryoconite in Jakobshavn Isbræ, Greenland, and Canada Glacier,
Antarctica, Ann. Glaciol., 54, 147–157, 2013.Tedesco, M., Doherty, S., Warren, W., Tranter, M., Stroeve, J., Fettweis, X.,
and Alexander, P.: What darkens the Greenland Ice Sheet?, EOS, 96,
10.1029/2015EO035773, 2015.Tedesco, M., Doherty, S., Fettweis, X., Alexander, P., Jeyaratnam, J., and Stroeve, J.: The darkening of the Greenland
ice sheet: trends, drivers, and projections (1981–2100), The Cryosphere, 10, 477–496, 10.5194/tc-10-477-2016, 2016.Thomas J. L., Polashenski, C. M., Soja, A. J., Marelle, L., Casey, K. A., Choi, H. D., Raut, J.-C., Wiedinmyer, C., Emmons, L. K.,
Fast, J., Pelon, J., Law, K. S., Flanner, M. G., and Dibb, J. E.: Quantifying black carbon deposition over the Greenland ice sheet
from forest fires in Canada, Geophys. Res. Lett., 44, 10.1002/2017GL073701, 2017.Toller, G., Xiong, X., Sun, J., Wenny, B. N., Geng, X., Kuyper, J., Angal, A.,
Chen, H., Madhavan, S., and Wu, A.: Terra and Aqua moderate-resolution imaging
spectroradiometer collection 6 level 1B algorithm, J. Appl. Remote Sens., 7, 073557, 10.1117/1.JRS.7.073557, 2013.
van As, D. and Fausto, R. S.: Programme for monitoring of the Greenland
ice sheet (PROMICE): first temperature and ablation records, Geological
Survey of Denmark and Greenland, 23, 73–76, 2011.
van den Broeke, M., van As, D., Reijmer, C., and van de Wal, R.: Assessing
and improving the quality of unattended radiation observations in
Antarctica, J. Atmos. Ocean. Tech., 21, 1417–1431, 2004.van den Broeke, M., Bamber, J., Ettema, J., Rignot, E., Schrama, E., van de
Berg, W. J., van Meijgaard, E., Velicogna, I., and Wouters, B.: Partitioning
recent Greenland mass loss, Science, 326, 984–986,
10.1126/science.1178176, 2009.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., and Wouters, B.: On the recent contribution of the Greenland ice sheet to
sea level change, The Cryosphere, 10, 1933–1946, 10.5194/tc-10-1933-2016, 2016.Velicogna, I. and Wahr, J.: Time-variable gravity observations of ice
sheet mass balance: Precision and limitations of the GRACE satellite data,
Geophys. Res. Lett., 40, 3055–3063, 10.1002/grl.50527, 2013.
Vermote, E.: MODIS/Surface Reflectance 8-Day L3 Global 500 m SIN Grid
V005, NASA EOSDIS Land Processes Distributed Active Archive Center (LP
DAAC), USGS/Earth Resources Observation and Science (EROS) Center, Sioux
Falls, South Dakota, USA, 2007.Vermote, E.: MODIS Terra Surface Reflectance 8-Day L3 Global 500m SIN Grid V006, Sioux Falls, South Dakota, USA,
NASA EOSDIS US Geological Survey Land Processes Distributed Active Archive Center (LP DAAC), 10.5067/MODIS/MOD09A1.006, 2015a.Vermote, E.: MODIS Aqua Surface Reflectance 8-Day L3 Global 500m SIN Grid V006, Sioux Falls, South Dakota, USA, NASA
EOSDIS US Geological Survey Land Processes Distributed Active Archive Center (LP DAAC), 10.5067/MODIS/MYD09A1.006, 2015b.
Wang, D., Morton, D., Masek, J., Wu, A., Nagol, J., Xiong, X., Levy, R.,
Vermote, E., and Wolfe, R.: Impact of sensor degradation on the MODIS NDVI
time series, Remote Sens. Environ., 119, 55–61, 2012.Wang, W., Zender, C. S., van As, D., Smeets, P. C. J. P., and van den Broeke, M. R.: A Retrospective, Iterative,
Geometry-Based (RIGB) tilt-correction method for radiation observed by automatic weather stations on snow-covered surfaces:
application to Greenland, The Cryosphere, 10, 727–741, 10.5194/tc-10-727-2016, 2016.
Warren, S. G. and Wiscombe, W. J.: A Model for the Spectral Albedo of
Snow. II: Snow Containing Atmospheric Aerosols, J. Atmos.
Sci., 37, 2734–2745, 1980.
Wiscombe, W. J. and Warren, S. G.: A model for the spectral albedo of
snow. I: Pure snow, J. Atmos. Sci., 37, 2712–2733, 1980.
Wright, P., Bergin, M., Dibb, J., Lefer, B., Domine, F., Carman, T.,
Carmagnola, C., Dumont, M., Courville, Z., and Schaaf, C.: Comparing MODIS
daily snow albedo to spectral albedo field measurements in Central
Greenland, Remote Sens. Environ., 140, 118-129, 2014.
Xiong, X. and Barnes, W.: An overview of MODIS radiometric calibration
and characterization, Adv. Atmos. Sci., 23, 69–79, 2009.Xiong, X., Wu, A., Wenny, B. N., Madhavan, S., Wang, Z., Li, Y., Chen, N.,
Barnes, W. L., and Salomonson, V. V.: Terra and Aqua MODIS Thermal Emissive Bands
On-Orbit Calibration and Performance, IEEE T. Geosci.
Remote, 53, 5709–5721, 10.1109/TGRS.2015.2428198, 2015.Zennaro, P., Kehrwald, N., McConnell, J. R., Schüpbach, S., Maselli, O. J., Marlon, J., Vallelonga, P., Leuenberger, D.,
Zangrando, R., Spolaor, A., Borrotti, M., Barbaro, E., Gambaro, A., and Barbante, C.: Fire in ice:
two millennia of boreal forest fire history from the Greenland NEEM ice core, Clim. Past, 10, 1905–1924, 10.5194/cp-10-1905-2014, 2014.
Zhan, Y. and Davies, R.: Intercalibration of CERES, MODIS and MISR
reflected solar radiation and its application to albedo trends, J.
Geophys. Res.-Atmos., 121, 6273–6283, 2016.