1. Introduction
Greenland is the world’s largest island, and the Greenland Ice Sheet (GrIS) is the Northern Hemisphere’s largest terrestrial permanent ice- and snow-covered area. Ice mass and snow cover serve as water reservoirs that are highly vulnerable to ongoing climatic variations and change (e.g., Hanna et al. 2005; Hinzman et al. 2005). The climate is changing: The average surface air temperature north of 60°N has increased by ∼0.09°C decade−1, and this change is conspicuous in winter months (e.g., Box 2002; Sturm et al. 2005). The climate has warmed substantially since the end of the Little Ice Age, and significantly in the last 30 yr (Serreze et al. 2000). This warming was accompanied by an increase in precipitation of ∼1% decade−1 (ACIA 2005). The Arctic is undergoing a system-wide response to climatic change, and the effect of a warmer and wetter climate on terrestrial cryospheric and hydrological processes and their components have already been documented on hemispheric, regional, local, and microscales (e.g., Serreze et al. 2000; Vorosmarty et al. 2001; Moritz et al. 2002; Hinzman et al. 2005; Mernild et al. 2007b, d).
Since Benson (1962) and Bauer (1968) first estimated GrIS mass balance components, a number of studies using a variety of methods (e.g., airborne and satellite laser altimetry, positive-degree and energy-balance models) have followed. Recent studies have documented GrIS mass-balance loss up to 238(±36) km3 yr−1 with an increasing trend of loss over the last several years (e.g., van de Wal 1996; Ohmura et al. 1999; Reeh et al. 1999; Janssens and Huybrechts 2000; Church et al. 2001; Mote 2003; Hanna et al. 2005; Chen et al. 2006; Luthcke et al. 2006; Ramillien et al. 2006; Velicogna and Wahr 2006). Modeling studies have shown that every 1-K rise in surface air temperature produces 20%–50% more Greenland ice melt (Oerlemans 1991; Braithwaite and Olesen 1993; Janssens and Huybrechts 2000; Hanna et al. 2005). Available satellite data show an attendant 47% K−1 increase in GrIS snowmelt extent (Abdalati and Steffen 1997b; Hanna et al. 2005), so a predicted rise in air temperature of 2–5 K would approximately double melt rates and increase water storage losses. These changes in storage affect runoff to the Arctic Ocean, the only ocean with a contributing land area greater than its surface area (Barry and Serreze 2000). Previous GrIS runoff estimates detected GrIS mass losses via runoff. For example, Janssens and Huybrechts (2000) showed losses of 281 km3 yr−1 (1953–2003), Mote (2003) showed 278 km3 yr−1 (1988–99), Hanna et al. (2005) showed 324 km3 yr−1 (1993–98) and 372(±37) km3 yr−1 (1998–2003), and Box et al. (2006) found 373 km3 yr−1 (1998–2004). Together, these sources indicate a trend of increasing GrIS runoff through the last decades. Changes in freshwater runoff to the ocean (or more specifically to the Greenland–Iceland–Norwegian Seas) play an important role in determining the global ocean thermohaline circulation, salinity, ice sea dynamics (Broecker et al. 1985; Broecker and Denton 1990; Su et al. 2006), global sea level rise (Dowdeswell et al. 1997; ACIA 2005; Box et al. 2006), and plans for hydroelectric power schemes (Hock and Jansson 2005; Mernild and Hasholt 2006), as well as the influx of sediment and nutrients to the ocean (Rysgaard et al. 2003; Hasholt et al. 2006).
Rough terrain, harsh climatic conditions, and remote location are commonly cited reasons for lack of knowledge and data for Greenland. Logistical constraints make it difficult to collect extensive observations of snow distribution, sublimation (surface and blowing-snow), evaporation, and snow and glacier-melt observations; collecting runoff measurements has typically been considered impossible. Only a few quality observations related to the spatial and temporal distributions of snow have been reported. Furthermore, the use of gauging stations that underestimate solid precipitation amounts, scattered Arctic meteorological stations, and limited winter and summer GrIS mass-balance observations produce sparse and unreliable data related to the spatial and temporal distributions of snow precipitation, sublimation, and surface melt across much of the GrIS. Such key climate-system components are essential to hydrological research efforts, and there is a clear need to explore issues associated with data sparseness and modeling capabilities.
This study attempts to improve our quantitative understanding of GrIS surface melt distributions and its related water balance components, particularly changes in surface mass balance (SMB) and freshwater runoff. The goal of this study was to apply a well-tested approach—a state-of-the-art modeling system, SnowModel (Liston and Elder 2006a; Mernild et al. 2006b)—to Greenland, including the GrIS. SnowModel was first tested at a local-to-regional scale using independent in situ observations from two long-term automatic meteorological and hydrometric monitoring catchments located in East Greenland between the GrIS and the ocean (Fig. 1). The model configuration was then adjusted to run over all of Greenland and tested using independent GrIS meteorological observations, satellite images, and equilibrium line altitude (ELA) studies. We performed model simulations for a 10-yr period, (1995–2005) with the following objectives: 1) assess MicroMet–SnowModel meteorological driving data against independent observations; 2) compare year-round simulated snow evolution components (snow accumulation, snow redistribution by wind, surface and blowing-snow sublimation, evaporation, and snow and ice melt) with independent in situ observations from the Mittivakkat and Zackenberg catchments, East Greenland; 3) quantify the yearly maximum and the 1995–2005 interannual variability in the GrIS surface melt cover (and the nonmelt area in the GrIS interior); 4) estimate and analyze the GrIS water balance components, including the SMB and GrIS mass balance; 5) simulate the interannual variability in GrIS specific runoff and the runoff separation to the GrIS western and eastern drainage areas; and 6) calculate the GrIS runoff contribution to global sea level rise.
2. Study area
a. Physical settings and climate
Greenland is roughly 2600 km long from the northernmost point at Cape Morris Jesup (83°N) to the southern tip at Cape Farwell (60°N). The island is dominated by the largest ice sheet in the Northern Hemisphere, the GrIS (1.834 × 106 km2), which covers approximately 85% of the island. The ice sheet’s maximum altitude is more than 3200 m MSL. (Fig. 1). The maximum width of the ice-free land strip is 200 km. The majority of the land strip is mountainous and includes numerous marginal glaciers and ice caps, and a number of fjords that reach the interior.
The Mittivakkat (Ammassalik Island, SE Greenland, 65°N) and the Zackenberg (NE Greenland, 74°N) catchments are the only two areas on the East Greenland land strip permanently instrumented for automatic collection of meteorological, hydrometric, and snow monitoring (Fig. 1a). Additional glacier observations regularly occur at Mittivakkat glacier (Knudsen and Hasholt 2004; Mernild et al. 2006b, 2008). These catchments are not connected to the GrIS. The Mittivakkat catchment is 18.4 km2, characterized by strong topographic relief, and ranges in elevation from 0 to 973 m MSL. Roughly 78% (14.4 km2) of the catchment is covered by parts of the Mittivakkat Glacier, a temperate glacier ranging from approximately 160 to 930 m MSL in elevation (Mernild et al. 2006b).
The Zackenberg catchment covers 512 km2 and is characterized by high-relief mountainous landscapes. Its elevation ranges from 0 to 1450 m MSL, from wide valleys to extensive glaciated plateaus mainly above 1000 m MSL. Roughly 20% (101 km2) of the catchment is covered by glaciers.
The climate in Greenland is arctic; that is, the average air temperature for the warmest month is below 10°C everywhere—except for the fjords in the south that fall into the subarctic zone where temperatures dip only slightly below this limit (Born and Böcher 2001). In the northern parts of the GrIS, winter air temperatures can drop below −70°C, while on the East Greenland land strip, summer temperatures can briefly rise above 25°C (Mernild et al. 2007a). The mean annual air temperature (MAAT) varies from 1.3° to −16.9°C from south to north. The simulation data period (1995–2005) shows a MAAT warming of ∼1.8°C (based on data from the 10 coastal meteorological stations; Fig. 1 and Table 1, stations 16–25). In southern and southeastern Greenland, the annual precipitation is ∼2500 mm w.eq. yr−1 (where w.eq. means water equivalent), while the northern areas receive little precipitation (Ohmura and Reeh 1991; Born and Böcher 2001; Serreze and Barry 2005). Many of the island’s characteristics cause considerable contrast in its weather conditions, including complex coastal topography, elevation, distance from the coastal area, marginal glaciers and ice caps, and the GrIS, which makes the climate vary considerably even over short distances. Temperature inversions are a common feature for Greenland coastal areas (Mernild et al. 2007a, c; Hansen et al. 2008) and for the GrIS (Putnins 1970).
3. SnowModel
a. SnowModel description
SnowModel (Liston and Elder 2006a) is a spatially distributed snowpack evolution modeling system specifically designed to be applicable over the wide range of snow landscapes, climates, and conditions found around the world. It is made up of four submodels: MicroMet defines the meteorological forcing conditions (Liston and Elder 2006b); EnBal calculates the surface energy exchanges, including melt (Liston 1995; Liston et al. 1999); SnowPack simulates snow depth and water-equivalent evolution (Liston and Hall 1995); and SnowTran-3D is a blowing-snow model that accounts for snow redistribution by wind (Liston and Sturm 1998; 2002; Liston et al. 2007). While other distributed-snow models exist (e.g., Tarboton et al. 1995; Marks et al. 1999; Winstral and Marks 2002), the SnowTran-3D component allows application in Arctic, alpine (i.e., above treeline), and prairie environments that compose 68% of the seasonally snow-covered areas in the Northern Hemisphere (Liston 2004). SnowModel also simulates snow-related physical processes at spatial scales ranging from 5 m to global and temporal scales ranging from 10 min to a whole season. Simulated processes include 1) accumulation and loss from snow precipitation, blowing-snow redistribution, and sublimation; 2) loading, unloading, and sublimation within forest canopies; 3) snow-density evolution; and 4) snowpack ripening and melt. SnowModel was originally developed for glacier-free landscapes. For glacier surface mass balance studies on eastern Greenland, SnowModel was modified to simulate glacier-ice melt after winter snow accumulation had ablated (Mernild et al. 2006b, 2007c).
1) MicroMet
MicroMet is a quasi–physically based meteorological distribution model (Liston and Elder 2006b) designed specifically to produce the high-resolution meteorological forcing distributions (air temperature, relative humidity, wind speed, wind direction, precipitation, solar and longwave radiation, and surface pressure) required to run spatially distributed terrestrial models over a wide range of landscapes in a physically realistic manner. MicroMet uses elevation-related interpolations to modify air temperature, humidity, and precipitation following Kunkel (1989), Walcek (1994), Dodson and Marks (1997), and Liston et al. (1999). Temperature and humidity distributions are defined to be compatible with the observed lapse rates. Wind flow in complex topography is simulated following Ryan (1977) and Liston and Sturm (1998). Solar radiation variations are calculated using elevation, slope, and aspect relationships (Pielke 2002). Incoming longwave radiation is calculated while taking into account cloud cover (Walcek 1994; Liston and Elder 2006b) and elevation-related variations following Iziomon et al. (2003). Precipitation is distributed following Thornton et al. (1997). In addition, any data from more than one location, at any given time, are spatially interpolated over the domain using a Gaussian distance-dependent weighting function and interpolated to the model grid using the Barnes objective analysis scheme (Barnes 1964, 1973; Koch et al. 1983). Liston and Elder (2006b) and Liston et al. (2007) performed a rigorous validation of MicroMet using various observational datasets, data denial, and geographic domains. Further, MicroMet has been used to distribute observed and modeled meteorological variables over a wide variety of landscapes in the United States—Colorado (Greene et al. 1999), Wyoming (Hiemstra et al. 2002, 2006), Idaho (Prasad et al. 2001), and Arctic Alaska (Liston et al. 1999, 2002, 2007; Liston and Sturm 1998, 2002); Norway—Svalbard and central Norway (Bruland et al. 2004); East Greenland (Hasholt et al. 2003; Mernild et al. 2006a, b, 2007c); and near-coastal Antarctica (Liston et al. 1999).
2) EnBal
3) SnowPack
SnowPack is a single-layer, snowpack-evolution and runoff–retention model that describes snowpack changes in response to precipitation and melt fluxes defined by MicroMet and EnBal (Liston and Hall 1995; Liston and Elder 2006a). Its formulation closely follows Anderson (1976). In SnowPack, the density changes with time in response to snow temperature and the weight of the overlying snow (Liston and Elder 2006a). A second density-modifying process results from snow melting. The melted snow reduces the snow depth and percolates through the snowpack. If the snow temperature is below freezing, any percolating/liquid water refreezes and is stored in the snow (in the “pores”) as internal refreezing. When saturated snow density is reached, assumed to be 550 kg m−3 (Liston and Hall 1995), actual runoff occurs. This provides a method of accounting for heat and mass transfer processes, such as snowpack ripening, during spring melt. The density of new snow from additional accumulation is defined following Anderson (1976) and Liston and Hall (1995). Static-surface (nonblowing snow) sublimation calculated in EnBal is used to adjust the snowpack depth; blowing-snow sublimation is calculated in SnowTran-3D (Liston and Elder 2006a).
4) SnowTran-3D
b. SnowModel input
To solve this system of equations, SnowModel requires spatially distributed fields of topography, and land-cover and meteorological data (air temperature, relative humidity, wind speed, wind direction, and precipitation), obtained from stations located within the simulation domain. For this study, data are obtained from 26 meteorological stations (Fig. 1b and Table 1). Sixteen stations, operated by the Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, Boulder, Colorado, are located on the GrIS at altitudes from 283 to 3208 m MSL. Four were located along the ice sheet crest (2500–3200 m MSL), eight stations are close to the 2000-m contour line (1800–2500 m MSL), and four stations are positioned in the ablation area (280–1200 m MSL). The Danish Meteorological Institute (DMI) operates nine, and the GeoBasis program, in cooperation with the Danish National Environmental Research Center and the Department of Geography and Geology, University of Copenhagen, operates one peripheral low-elevation station located below 110 m MSL. Simulations were performed on a daily time step. Admittedly, snow and ice melt and blowing snow are threshold processes that may not be accurately represented by this time step; however, computational constraints prohibited higher temporal-resolution simulations. For the SnowModel test areas—the Mittivakkat and the Zackenberg catchments—the simulations span the 10-yr period from 1995 through 2005, and the start and end of a year are designated as 1 September and 31 August of the next year to appropriately separate the accumulation and ablation components of the glacier mass balance annual cycle. For the GrIS, water balance components were simulated based on the calendar year for better comparison with previous studies.
Greenland topographic data for the model simulations were provided by Bamber et al. (2001) who applied “correction” elevations derived by satellite imagery to an existing radar-altimetry digital elevation model (DEM). The image-derived correction was determined from a high-resolution (625 m) grid of slopes inferred from the regional slope-to-brightness relationship of 44 Advanced Very High Resolution Radiometer (AVHRR) images covering all of Greenland (Scambos and Haran 2002). For the model simulations, this DEM was aggregated to a 5-km grid-cell increment and clipped to yield a 2830 × 1740 km2 simulation domain that encompassed all of Greenland. The GrIS terminus was confirmed or estimated by using aerial photos and maps (1:250000 Geodetic Institute, Denmark). For the SnowModel test areas (Mittivakkat and the Zackenberg catchments) a 100-m grid-cell increment DEM was used (for further detail see Mernild et al. 2006a, b, 2007c), to capture small-scale features such as drifts. Relatively finescale features are absent from the coarser GrIS-scale simulations at 5-km resolution.
Each grid cell within the domains was assigned a U.S. Geological Survey (USGS) Land Use/Land Cover System class (Fig. 1b) according to the North American Land Cover Characteristics Database, version 2.0 [available online at the USGS Earth Resources Observation and Science (EROS) Data Center’s Distributed Active Archive Center Web site: http://edcdaac.usgs.gov/glcc/na_int.html]. The snow-holding depth (the snow depth that must be exceeded before snow can be transported by wind) and canopy gap fraction (the solar radiation reaching the snow surface below the canopy) were assumed to be constant during the 10-yr simulation period (Table 2). The albedo was assumed to be 0.8 for snow. Realistically, snow albedo changes with time and surface characteristics (Pomeroy and Brun 2001); thus, the model will likely underestimate the energy available for surface melting. When the snow is melted, GrIS surface ice conditions are used. User-defined constants for SnowModel are shown in Table 2 [for parameter definitions, see Liston and Sturm (1998, 2002) and Liston and Elder (2006a)]. All fjord and ocean areas within the domain were excluded from model simulations (Fig. 1b).
Solid and liquid precipitation measurements at the DMI meteorological stations (Fig. 1b and Table 1; stations 16–18 and 20–25) were calculated from Helman–Nipher shield observations corrected according to Allerup et al. (1998, 2000). Solid (snow) precipitation was calculated from snow-depth sounder observations (Fig. 1b and Table 1; station 19) after the sounder data noise was removed; these data are assumed to be accurate within ±(10%–15%) (Mernild et al. 2007c). The snow-depth sounder observations were fractionated into liquid (rain) precipitation and solid (snow) precipitation at different air temperatures based on observations from different locations on Svalbard (Førland and Hanssen-Bauer 2003). For air temperatures below −1.5°C, sounder data were considered to represent solid precipitation and for temperatures above 3.5°C precipitation is considered liquid; for temperatures between these limits, the snow and rain fraction is calculated by linear interpolation. Snow-depth increases at relative humidity <80% and at wind speed >10 m s−1 were removed to better distinguish between the proportions of real snow accumulation based on precipitation events and blowing snow redistribution (Mernild et al. 2007c). Remaining snow-depth increases were adjusted using a temperature-dependent snow density (Brown et al. 2003) and an hourly snowpack settling rate for estimating the mm w.eq. (Anderson 1976).
Temperature inversions with cold, low clouds or sea fog coming from the ocean dominate the coastal climate (approximately 300–400 m MSL). This study’s use of data from meteorological stations located both in low-lying coastal areas and on the GrIS contributes to a more detailed understanding of the altitudinal air temperature distribution within the simulation domain. Unfortunately, this information is not detailed enough to provide a full understanding of the inversion height, strength, and thickness on Greenland. Average monthly lapse rates (1997–2005) based on air temperature observations from eight different transects all around Greenland were used as a model input (Fig. 1, Table 3). Transects lay between low-lying meteorological stations located in the land strip area almost at sea level and stations on the GrIS. The minimum monthly lapse rate of −8.26°C km−1 occurred in February, and the maximum (−5.77°C km−1) occurred in June. The low winter lapse rate is followed by a high standard deviation and vice versa for the summer (Table 3), due to the relatively cold and variable winter temperatures at the GrIS interior.
To assess the performance of upscaled SnowModel–MicroMet distributed meteorological data, simulated meteorological data were tested against observations not used in MicroMet. The Swiss Camp station (Table 1), located on the GrIS (Fig. 1), was used for comparisons spanning 1995–2005. The validation station was located 40 km from the nearest station (JAR1) used in MicroMet to drive SnowModel.
c. SnowModel validation
Few quality observations for spatial in situ snow evolution, snow and ice surface melt, and glacier net mass balance are available. SnowModel accumulation and ablation routines were tested by visual inspection, cumulative values, and simple linear regression (Tables 4 –6). Unfortunately, the only available observations for validation were collected outside the GrIS, from two well-instrumented and reliable long-term automatic stations collecting meteorological and hydrometric data in the Mittivakkat and Zackenberg catchments of East Greenland. These catchments supplied independent, in situ observations on glacier surface mass balance, snow-depth distribution, and snow cover extent used to validate SnowModel snow accumulation and ablation routines on local-to-regional scales before the routines were upscaled for all of Greenland, including the GrIS.
1) Glacier surface mass balance observations at Mittivakkat
Modeled end-of-winter (31 May) spatial SWE depths and end-of-summer ablation (31 August) were validated against observed SWE depths (winter glacier mass balance) and ablation values (summer glacier mass balance) from the 14.4 km2 Mittivakkat glacier test area (Table 4). The validation was conducted for the 10-yr period 1995/96–2004/05 at the end of May and the end of August. During these field campaigns, snow depth, snow density, and ablation from snow and glacier ice were measured using cross-glacier stake lines spaced approximately 500 m apart; the the stakes in each line were set 50–100 m apart for snow accumulation (in total 230 measurements) and 200–250 m apart for snow and ice ablation measurements (in total 60 measurements). The accuracy levels of the observed winter and summer mass balances are each assumed to be within ∼15%; however, larger errors might occur, especially in glacier areas with many crevasses (Knudsen and Hasholt 1999, 2004; Mernild et al. 2006a).
2) Snow observations at Zackenberg
Approximately 2000 end-of-winter snow-depth measurements were made at the end of May or beginning of June for 2004 and 2005 in the 16.8 km2 Zackenberg valley site (Table 5; Mernild et al. 2007c). Snow depth was measured approximately every 25–30 m using a global positioning system (GPS) MagnaProbe (Snow-Hydro, Fairbanks, Alaska; information online at www.snowhydro.com), a device that records snow depth and location. Average total snowpack snow density was measured at 40–50 different places in the valley each year (Mernild et al. 2007c). Spatial observations of average SWE depth were used to validate SnowModel winter components.
3) Photographic snow-cover observations at Zackenberg Valley
Zackenberg snow cover distributions have been observed from 1 June through the ablation period (1 June–31 August) by photographs taken once every day at solar noon (Table 5). Traditionally, determining snow-cover extent has been based on point measurements; however, such measurements will not always detect significant area snow-cover variations. Therefore, since 1995, a digital camera has been placed on a hillside 477 m MSL overlooking the Zackenberg Valley (16.8 km2) taking daily oblique photos to quantify the evolution of spatial snow-cover distributions. These photographs were transformed into digital orthophotos and used to perform snow-cover mapping (for technical specifications see Hinkler et al. 2003). Snow-cover distribution for the test area was converted into depletion curves to illustrate the daily ablation from 1 June, typically illustrating a laterally reversed S-shaped curve as a result of gradual snow-cover decrease. Depletion curves based on daily values from 1996 through 2005 were used to validate the SnowModel summer components. For brevity, the data in Table 6 only show results from the 1st, 10th, and 20th days in each month during the ablation period.
4) Surface melt observed from satellite images
Detection of surface melt at large spatial scales is effectively accomplished by using satellite microwave data. The daily GrIS snowmelt extent is mapped (25-km grid-cell increment) using passive microwave satellite observations that discriminate wet from dry snow (Fig. 3; Abdalati and Steffen 1997a). The criterion for melt is 1% mean liquid water content by volume in the top meter of snow. The center part of the GrIS is the area where the melting threshold of the cross-well ground-penetrating radar (XGPR) microwave algorithm did not show any melt. The end-of-summer maximum observed spatial surface melt distribution at the GrIS was used to validate SnowModel melt simulations (Fig. 3).
4. Results and discussion
Validations of MicroMet-simulated GrIS meteorological data indicate substantial correlation with independent observed GrIS meteorological data from the Swiss Camp (Fig. 2). Critical MicroMet-generated air temperature, relative humidity, and precipitation values account for 84%, 63%, and 69%, respectively, of the variance in the observed 1995–2005 daily averaged dataset. The wind speed has less strong correlations, but the results remain respectable (>50% variance) for representations of GrIS meteorological processes. While this validation is limited because it employs only one independent station, a rare commodity, it indicates that MicroMet satisfactorily represents GrIS conditions.
SnowModel was chosen for this study because of its robustness and ease of implementation over new simulation domains. This model demands rather limited input data, an important consideration in areas like Greenland, for which data are sparse due to rough terrain, harsh climatic conditions, and its remote location. It appears that our choice of a simple methodology provided estimates of the GrIS surface melt distribution and related water balance components that agree well with observed values and previous studies. Nevertheless, it is important to keep in mind the limitation for SnowModel results when tested against observations collected from the strip of land surrounding the GrIS and not from the GrIS itself. SnowModel tests were conducted both for 100-m and 5-km grid cells, showing acceptable results (Tables 4 –6).
Table 4 presents the modeled winter mass balance for the accumulation period (September–May), the modeled summer mass balance for the ablation period (June–August), and the modeled mass-balance data (100-m grid cell) for the Mittivakkat Glacier test area for the years 1995/96–2004/05. The average modeled winter mass balance was 1207(±168) mm w.eq. (Table 4). This corresponds well with the observed winter mass balance of 1228(±197) mm w.eq., or a 21-mm w.eq. (∼2%) difference. The modeled winter mass balance shows significant correlation with observed values (1995–2005): R2 = 0.92, p < 0.01 (where p is the level of significance), covering a maximum variation between modeled and observed values of 160 mm w.eq. for 2002/03 (Table 4). The average modeled summer mass balance was −1915(±407) mm w.eq., which corresponds to an observed summer mass balance of −1904(±485) (Table 4), or an 11 mm w.eq. (approximately <1%) difference. The maximum variation between modeled and observed values was 220 mm w.eq. for 2002/03 (Table 4). The modeled summer mass balance shows significant correlation with observed values (1995–2005): R2 = 0.95, p < 0.01. The average modeled net mass balance was −698(±492) mm w.eq. yr−1, an underestimation of 88 mm w.eq. yr−1 compared to the observations (∼11% difference). The average simulated mass loss was less than the observed value. The modeled net mass balance shows significant correlation with observed values (1995–2005): R2 = 0.93, p < 0.01. The maximum yearly net mass balance difference between the observed and modeled values was 320 mm w.eq. yr−1 for 2002/03.
At the Zackenberg test area, the end-of-winter SWE depth (31 May) was modeled for 2003/04 and 2004/05 (100-m grid cell) and yielded average SWE depths of 207- and 166-mm w.eq., respectively (Table 5). The average modeled and observed end-of-winter SWE depths indicate a maximum SWE depth difference of 14 mm w.eq., or ∼6% (Table 5). SnowModel over- and under-performed randomly in response to both elevation and topographic influences–characteristics (ridge and hills), and to finescale snow-depth variations not captured by the 100-m DEM. Our analysis of the snow-cover extent in the Zackenberg test area for the ablation period (June–August 1996–2005) inferred inter- and intra-annual variations. The modeled snow-cover extent (100-m grid cell) shows significant correlation with observed values (based on time-lapse photography) for the ablation periods (R2 = 0.99, p < 0.01; Table 6), even though the maximum variation between the modeled and observed snow cover extents through the ablation periods was 8%, or approximately 1.3 km2 (at 10 June 2004; Table 6). Snow-cover extent is a product of both snow accumulation and ablation processes (phase-change processes like evaporation, sublimation, and melting). Within SnowModel, SnowTran-3D simulates spatial snow deposition patterns in response to erosion and deposition, and EnBal calculates the energy flux available for snowmelt. Table 6 illustrates the modeled snow-cover variation through June–August 1996–2005, showing that 80%–100% of the test area was snow covered on 1 June, 50% of the snow cover extent had melted away by mid-June to early July, and 95% melted by early July to late August. The model variations were very similar to the recorded observations.
All three tests of the SnowModel winter and summer snow-evolution components developed for the Mittivakkat and Zackenberg catchments indicated good agreement between observed and simulated values (100-m grid cell). Further, the snow validation, while not from areas on the GrIS, do indicate that the SnowModel results are representing key physical snow accumulation and ablation processes and that the models can calculate reasonable estimates of the mass balance on a finescale (100-m grid cell).
Figure 3 plots both the spatially satellite-observed melt and modeled end-of-summer snow and ice surface melt (any melt amount) and nonmelt extents for the GrIS from 1995 to 2005. There is a high degree of similarity in the nonmelt distributions. In some areas the discrepancy between modeled and satellite-observed melt and nonmelt boundaries can be up to 160 km (1996), especially in northeastern Greenland, where the distance between meteorological stations is great. This discrepancy might also be due to the 1-day simulation time step; hourly variations in surface melt are not represented in the simulations, the fixed albedo used for snow and ice, and temporal and spatial uncertainties in satellite observations. Modeled nonmelt areas of the GrIS are, on average, underestimated by ∼3% (1995–2005; see Fig. 4), confirming the robustness of the ablation processes in SnowModel. The modeled GrIS surface melt area is, therefore, on average overestimated by ∼29 000 km2 yr−1 when compared with satellite observations. Observed interannual variability ranges from ∼68 000 km2 (or ∼6%) in 2004 to <8000 km2 (or <1%) in 1996 and 2005; these 2 yr represent the extreme low (2005) and high (1996) nonmelt areas (Fig. 4). Simulated interannual variability for the nonmelt area agrees (R2 = 0.96) with the observations, illustrating that the nonmelt area can vary from year to year from as high as 71% (1996) to as low as 50% (2005). On average, the simulated nonmelt area decreased ∼6% in size from 1995 through 2005 (R2 = 0.09; p < 0.25), indicating an increasing GrIS surface melt area, due to a significant average increase (R2 = 0.76, p < 0.01) in the annual temperature anomaly of ∼1.8°C (Table 7). For 2005, the modeled surface melt occurred at elevations as high as 2950 m MSL.
The melt index (defined as the melting area above the 2000-m GrIS contour line times the number of melting days) was further used to map snow-melt changes (Fig. 5). The index varies from 0.42 × 106 km2 × days (1996) to 3.31 × 106 km2 × days (1999), on average increasing ∼0.57 × 106 km2 × days from 1995 through 2005 (R2 = 0.07, p < 0.25). The low 1996 melt index indicates good agreement with the high observed nonmelting area (71%) for 1996 (Figs. 3 and 4). The smallest nonmelting area (50%) and highest temperature anomaly (1.12°C; Table 7) occurred in 2005; however, the 2005 melt index (2.75 × 106 km2 × days) is only second largest after 1999 (3.29 × 106 km2 × days; Fig. 5). The trends in modeled melt-index results are consistent with values found by Tedesco (2007) (Fig. 5).
Figure 6 illustrates the ELA for the western and eastern GrIS regions along a latitude line from 60° to 81°N. The ELA is defined as the elevation where the SMB equals zero. Therefore, the ELA provides a useful metric for the accumulation and the ablation’s net influence on the SMB. Regional variations between the western and eastern parts of the GrIS are due to changes in local topography. On the western GrIS, the ELA varies from 810 m MSL (81°N) to 1640 (63°N), averaging 1260 m MSL (Fig. 6a), and on the eastern GrIS it varies from 600 m MSL (81°N) to 1400 (69°N), averaging 1130 m MSL (Fig. 6b). The modeled ELA is lower with increased latitude (Figs. 6a and 5b), which is consistent with the parameterization of Zwally and Giovinetto (2001) (also in Box et al. 2004). The trend in average ELA from 1995 through 2005 is shown on Fig. 6c. The lowest average modeled ELA occurs in 1996 (western GrIS, 670 m MSL; eastern GrIS, 550 m MSL), a year with extensive observed nonmelt area. In contrast, the highest ELA developed in 2005 (western GrIS, 1690 m MSL; eastern GrIS, 1610 m MSL), a year with the smallest nonmelt area (Figs. 3 and 4). The general trend for 1995–2005 is an increase in average ELA in the western GrIS of ∼42 m MSL yr−1 (R2 = 0.25, p < 0.10) and of ∼45 m MSL yr−1 (R2 = 0.26, p < 0.10) in the eastern GrIS (Fig. 6c).
Sublimation can play an important role in the high-latitude hydrological cycle during the year. Previous Mittivakkat Glacier studies (Hasholt et al. 2003; Mernild et al. 2006b, 2008), Zackenberg glacier studies (Mernild et al. 2007c), and GrIS studies (Box and Steffen 2001) have all shown that as much as 12%–23% of the annual precipitation may be returned to the atmosphere by sublimation. In Arctic North America, studies by Liston and Sturm (1998, 2004), Essery et al. (1999), and Pomeroy and Essery (1999) indicate that 5%–50% of the annual solid precipitation was returned to the atmosphere by sublimation. For the GrIS (1995–2005), modeled annual sublimation averaged 28(±3) mm w.eq. yr−1, which equaled 52(±6) km3 yr−1, or ∼10% of the solid precipitation input for the GrIS (Tables 7 and 10). SnowModel results were slightly lower than Box and Steffen’s (2001) observed GrIS values of 62(±23) to 120(±65) km3 yr−1. In our GrIS simulation domain, low air temperatures coincide with high relative humidity, and, therefore, sublimation has played a lesser role in the surface high-latitude water budget.
Table 7 presents the surface modeled water balance components [Eq. (3)] for the GrIS from 1995 through 2005. The SMB is governed by accumulation (snow precipitation) and by ablation (evaporation, sublimation, and runoff). Net snow accumulation occurs over the GrIS interior while net surface ablation dominates the terminus/low-lying parts of the GrIS (Figs. 3 and 5). The interannual variability in precipitation and ablation causes sizeable SMB fluctuations with correlations of R2 = 0.44, p < 0.01, and R2 = 0.83, p < 0.01, respectively (Table 7). SMB fluctuations were largely tied to changes in ablation processes, mainly runoff. In 1998 the SMB was 1 mm w.eq. yr−1 (Table 7), because of high ablation (325 mm w.eq. yr−1), of which 262 mm w.eq. yr−1 was runoff. The same year featured a −0.18 temperature anomaly and relatively low precipitation (326 mm w.eq. yr−1). Other relatively low SMB values are found in 1995, 2002, 2003, 2004, and 2005, and are also due to a high runoff-related flux into the ocean. The absolute maximum SMB of 169 mm w.eq. yr−1 occurred in 1996 due to high precipitation, 376 mm w.eq. yr−1, and concurrent low ablation (207 mm w.eq. yr−1, where 151 mm w.eq. was runoff). The Table 7 estimated SMB conditions are in agreement with Hanna et al. (2005, 2008) and Box et al. (2006), indicating that SMB on average is 11% (9 mm w.eq. yr−1 or 17 km3 yr−1) lower than the Box et al. (2006) (1995–2004) simulated values. Integrated over the GrIS, the 11-yr precipitation rate indicates a nonsignificant decreasing trend of ∼7 mm w.eq. The ablation increase averages ∼40 mm w.eq. (R2 = 0.18, p < 0.10), and runoff alone increases by ∼34 mm w.eq. (R2 = 0.12, p < 0.25) (Table 7). As the precipitation decreases combined with an increasing runoff, the net effect of these parameters indicates an increasing average SMB loss on 47 mm w.eq. for 1995–2005 (R2 = 0.12, p < 0.25). The mean annual runoff of 211(±31) mm w.eq. yr−1, equals 392(±58) km3 yr−1 and a specific runoff of 6.7 l s−1 km−2 yr−1 (Table 8). The calculated runoff was similar to those estimated by Hanna et al. (2005), 372 km3 yr−1 for the period 1998–2003, and by Box et al. (2006), 396 km3 yr−1 (1995–2004). Our simulated runoff was 46 km3 yr−1 (11%) higher than Hanna et al.’s (2005) estimated runoff for the same period (1998–2003), and 3 km3 yr−1 (<1%) lower than Box et al’s (2006) estimated runoff for the period 1995–2004.
The SnowModel-simulated runoff was spatially separated into a western and an eastern GrIS drainage area contribution. Around 58% of the runoff drains from the western GrIS area, 227(±32) km3 yr−1, and 42% from the eastern GrIS area, 165(±26) km3 yr−1 (Fig. 1, Table 9). Serreze et al. (2006) reported values of freshwater export from the Arctic Ocean to the Greenland Sea of 4700 km3 yr−1 (2300 km3 yr−1 as sea ice and 2400 km3 yr−1 as upper-ocean freshwater). Using these values as the most reliable estimate for freshwater export to the Greenland Sea, the eastern GrIS and the total GrIS runoff amounts of 165 km3 yr−1 and 392 km3 yr−1 (1995–2005) contribute ∼4% and ∼8% of the total freshwater input to the Greenland Sea, respectively. From the standpoint of a global eustatic sea level rise, the 1995–2005 GrIS runoff contribution is ∼1.1 mm w.eq. yr−1 (Table 8).
Table 10 shows the GrIS water balance (1995–2005), including values for SMB, bottom melting [yielding an estimated 17 mm w.eq. yr−1, or 32 km3 yr−1; values from Church et al. (2001)], and iceberg calving [yielding an estimated 191 mm w.eq. yr−1, or 357 km3 yr−1 (1996–2005); values from Rignot and Kanagaratnam (2006)]. SnowModel is a surface model and only produces first-order effects of climate change; it does not include glacio–hydro-dynamic routines. This study suggests (Table 10) that the GrIS is losing mass: 133 mm w.eq. yr−1, or 257 km3 yr−1 of GrIS mass was lost on average during 1995–2005. Losses ranged from 83 (1996) to 394 km3 yr−1 (1998). Our results span the 80 km3 yr−1 overall GrIS volume loss during 1997–2003 estimated by airborne laser altimetry (Krabill et al. 2000, 2004; Thomas et al. 2006) and the mass losses of 111–248 km3 yr−1 for 2002–2006 generated by the Gravity Recovery and Climate Experiment (GRACE) results (Chen et al. 2006; Luthcke et al. 2006; Ramillien et al. 2006; Velicogna and Wahr 2006).
The disparity between the SnowModel-simulated surface melt extent and the passive microwave satellite-observed melt extent can be used as a guide to understanding where additional meteorological stations might be deployed within the simulation domain to improve the model simulations. This disparity can also be useful in developing model routines for simulating the temperature inversion layer, which is a common Arctic feature. Air temperature inversion test simulations will first be conducted on the Mittivakkat catchment before routines are automated, and upscaled for greater catchments, for example, the GrIS.
5. Summary and conclusions
This study presents simulations of the GrIS surface melt area and water-balance components for the period 1995–2005. Our SnowModel simulations have been validated against independent in situ observations (accumulation and ablation observations) made on the land between the GrIS and the ocean in eastern Greenland. This simulated GrIS series yielded useful insights into the present conditions on the ice sheet and the interannual variability of water-balance components. There is a high degree of agreement between these GrIS simulations and the recorded observations, and both indicate an increasing surface melt area during the simulation period. Further, simulation values for GrIS surface change, runoff, SMB, and GrIS loss are in line with previous modeling and satellite studies. Runoff increased over the simulation period, while a reduction in the GrIS mass balance occurred. The runoff has likely been a factor in global sea level rise, contributing ∼1.1 mm w.eq. yr−1 to the eustatic sea level rise (not considering ocean loss by evaporation or the contribution from thermal expansion).
Acknowledgments
This work was supported by grants from the University of Alaska Presidential IPY Postdoctoral Foundation, and the University of Alaska Fairbanks (UAF) Office of the Vice Chancellor for Research and conducted during the first author’s IPY postdoctoral fellowship at the UAF. A special thanks goes to the Cooperative Institute for Research in the Atmosphere (CIRA), Colorado State University, for hosting the first author in February and October 2007; to Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado at Boulder, for hosting the first author from November 2007 through February 2008; and to the Faculty of Science, Hokkaido University, Sapporo, Japan, for hosting the first author from April through July 2008. Furthermore, the authors thank President’s Professor of Climate Change and Chief Scientist John Walsh, International Arctic Research Center (IARC), UAF, for his review of the paper. A special thanks to Dr. Theodore Scambos, CIRES, University of Colorado, for providing a satellite picture and the Greenland Digital Elevation Model.
REFERENCES
Abdalati, W., and Steffen K. , 1997a: Snowmelt on the Greenland Ice Sheet as derived from passive microwave satellite data. J. Climate, 10 , 165–175.
Abdalati, W., and Steffen K. , 1997b: The apparent effects of the Mt. Pinatubo eruption on the Greenland ice sheet melt extent. Geophys. Res. Lett., 24 , 1795–1797.
ACIA, 2005: Arctic Climate Impact Assessment. Cambridge University Press, 1042 pp.
Allerup, P., Madsen H. , and Vejen F. , 1998: Estimating true precipitation in arctic areas. Proc. Nordic Hydrological Conf., Helsinki, Finland, Nordic Hydrological Programme Rep. 44, 1–9.
Allerup, P., Madsen H. , and Vejen F. , 2000: Correction of precipitation based on off-site weather information. Atmos. Res., 53 , 231–250.
Anderson, E. A., 1976: A point energy balance model of a snow cover. NOAA Tech. Rep. NWS 19, 150 pp.
Bamber, J., Ekholm S. , and Krabill W. , 2001: A new, high-resolution digital elevation model of Greenland fully validated with airborne laser altimeter data. J. Geophys. Res., 106B , 6733–6746.
Barnes, S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3 , 396–409.
Barnes, S. L., 1973: Mesoscale objective analysis using weighted time series observations. NOAA Tech. Memo. ERL NSSL-62, National Severe Storms Laboratory, Norman, OK, 60 pp.
Barry, R. G., and Serreze M. C. , 2000: Atmospheric components of the arctic ocean freshwater balance and their interannual variability. The Freshwater Budget of the Arctic Ocean, E. L. Lewis et al., Eds., Kluwer Academic, 45–56.
Bauer, A., 1968: Nouvelle estimation du bilian de masses de L’Inlandsis du Groenland. Deep-Sea Res., 14 , 13–17.
Benson, C. S., 1962: Stratigraphic studies in the snow and firn of Greenland Ice Sheet. U.S. Army SIPRE Research Rep. 70, 93 pp.
Born, E. W., and Böcher J. , 2001: The Ecology of Greenland. Ministry of Environment and Natural Resources, Nuuk, Greenland, 429 pp.
Box, J. E., 2002: Survey of Greenland instrumental temperature records: 1973–2001. Int. Climatol., 22 , 1829–1847.
Box, J. E., and Steffen K. , 2001: Sublimation estimates for the Greenland ice sheet using automated weather station observations. J. Geophys. Res., 106 , D24. 33965–33982.
Box, J. E., Bromwich D. H. , and Bai L-S. , 2004: Greenland ice sheet surface mass balance 1991–2000: Application of Polar MM5 mesoscale model and in situ data. J. Geophys. Res., 109 .D16105, doi:10.1029/2003JD004451.
Box, J. E., and Coauthors, 2006: Greenland ice sheet surface mass balance variability (1988–2004) from calibrated Polar MM5 output. J. Climate, 19 , 2783–2800.
Braithwaite, R. J., and Olesen O. B. , 1993: Seasonal variation of ice ablation at the margin of the Greenland ice sheet and its sensitivity to climate change, Qamanarassup sermia, West Greenland. J. Glaciol., 39 , 267–274.
Broecker, W. S., and Denton G. H. , 1990: The role of ocean–atmosphere reorganization in glacial cycles. Quat. Sci. Rev., 9 , 305–341.
Broecker, W. S., Peteet D. M. , and Rind D. , 1985: Does the ocean–atmosphere system have more than one stable mode of operation. Nature, 315 , 21–26.
Brown, R. D., Brasnett B. , and Robinson D. , 2003: Gridded North American monthly snow depth and snow water equivalent for GCM evaluation. Atmos.–Ocean, 41 , 1–14.
Bruland, O., Liston G. E. , Vonk J. , Sand K. , and Killingtveit A. , 2004: Modelling the snow distribution at two high-Arctic sites at Svalbard, Norway, and at a sub-Arctic site in central Norway. Nord. Hydrol., 35 , 191–208.
Chen, J. L., Wilson C. R. , and Tapley B. D. , 2006: Satellite gravity measurements confirm accelerated melting of Greenland ice sheet. Science, 313 , 1958–1960.
Church, J. A., Greory J. M. , Huybrechts P. , Kuhn M. , Lambeck C. , Nhuan M. T. , Qin D. , and Woodworth P. L. , 2001: Change in sea level. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 639–694.
Dodson, R., and Marks D. , 1997: Daily air temperature interpolation at high spatial resolution over a large mountainous region. Climate Res., 8 , 1–20.
Dowdeswell, J. A., and Coauthors, 1997: The mass balance of circum-Arctic glaciers and recent climate change. Quat. Res., 48 , 1–4.
Essery, R. L. H., Li L. , and Pomeroy J. W. , 1999: A distributed model of blowing snow over complex terrain. Hydrol. Processes, 13 , 2423–2438.
Førland, E. J., and Hanssen-Bauer I. , 2003: Climate variations and implications for precipitations types in the Norwegian Arctic. Norwegian Meteorological Institute Rep. 24/02, 21 pp.
Greene, E. M., Liston G. E. , and Pielke R. A. , 1999: Simulation of above treeline snowdrift formation using a numerical snowtransport model. Cold Reg. Sci. Technol., 30 , 135–144.
Hanna, E., Huybrechts P. , Janssens I. , Cappelen J. , Steffen K. , and Stephens A. , 2005: Runoff and mass balance of the Greenland ice sheet: 1958–2003. J. Geophys. Res., 110 .D13108, doi:10.1029/2004JD005641.
Hanna, E., and Coauthors, 2008: Increased runoff from melt from the Greenland Ice Sheet: A response to global warming. J. Climate, 21 , 331–341.
Hansen, B. U., and Coauthors, 2008: Present day climate at Zackenberg. High-Arctic Ecosystem Dynamics in a Changing Climate: Ten Years of Monitoring and Research at Zackenberg Research Station,Northeast, Greenland, H. Meltofte et al., Eds., Advances in Ecological Research, Vol. 40, Elsevier, 111–149.
Hasholt, B., Liston G. E. , and Knudsen N. T. , 2003: Snow distribution modelling in the Ammassalik region, southeast Greenland. Nord. Hydrol., 34 , 1–16.
Hasholt, B., Bobrovitskaya N. , Bogen J. , McNamara J. , Mernild S. H. , Milbourn D. , and Walling D. E. , 2006: Sediment transport to the Arctic Ocean and adjoining cold oceans. Nord. Hydrol., 37 , 413–432.
Hiemstra, C. A., Liston G. E. , and Reiners W. A. , 2002: Snow redistribution by wind and interactions with vegetation at upper treeline in the Medicine Bow Mountains, Wyoming. Arct. Antarct. Alp. Res., 34 , 262–273.
Hiemstra, C. A., Liston G. E. , and Reiners W. A. , 2006: Observing, modelling, and validating snow redistribution by wind in a Wyoming upper treeline landscape. Ecol. Modell., 197 , 35–51.
Hinkler, J., Pedersen S. B. , Rasch M. , and Hansen B. U. , 2003: Automatic snow cover monitoring at high temporal and spatial resolution, using images taken by a standard digital camera. Int. J. Remote Sens., 23 , 4669–4682.
Hinzman, L. D., and Coauthors, 2005: Evidence and implications of recent climate change in northern Alaska and other Arctic regions. Climatic Change, 72 , 251–298.
Hock, R., and Jansson P. , 2005: Modelling glacier hydrology. Encyclopedia of Hydrological Science, M. Andersen, Ed., John Wiley, 1–9.
Iziomon, M. G., Mayer H. , and Matzarakis A. , 2003: Downward atmospheric longwave irradiance under clear and cloudy skies: Measurement and parameterization. J. Atmos. Sol.-Terr. Phys., 65 , 1107–1116.
Janssens, I., and Huybrechts P. , 2000: The treatment of meltwater retention in mass-balance parameterisation of the Greenland Ice Sheet. Ann. Glaciol., 31 , 133–140.
Knudsen, N. T., and Hasholt B. , 1999: Radio-echo sounding at the Mittivakkat Gletscher, southeast Greenland. Arct. Antarct. Alp. Res., 31 , 321–328.
Knudsen, N. T., and Hasholt B. , 2004: Mass balance observations at Mittivakkat Glacier, southeast Greenland 1995–2002. Nord. Hydrol., 35 , 381–390.
Koch, S. E., DesJardins M. , and Kocin P. J. , 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor., 22 , 1487–1503.
Krabill, W. E., and Coauthors, 2000: Greenland ice sheet: High-elevation balance and peripheral thinning. Science, 289 , 428–430.
Krabill, W., and Coauthors, 2004: Greenland Ice Sheet: Increased coastal thinning. Geophys. Res. Lett., 31 .L24402, doi:10.1029/2004GL021533.
Kunkel, K. E., 1989: Simple procedures for extrapolation of humidity variables in the mountainous western United States. J. Climate, 2 , 656–669.
Liston, G. E., 1995: Local advection of momentum, heat, and moisture during the melt of patchy snow covers. J. Appl. Meteor., 34 , 1705–1715.
Liston, G. E., 2004: Representing subgrid snow cover heterogeneities in regional and global models. J. Climate, 17 , 1381–1397.
Liston, G. E., and Hall D. K. , 1995: An energy-balance model of lake-ice evolution. J. Glaciol., 41 , 373–382.
Liston, G. E., and Sturm M. , 1998: A snow-transport model for complex terrain. J. Glaciol., 44 , 498–516.
Liston, G. E., and Sturm M. , 2002: Winter precipitation patterns in arctic Alaska determined from a blowing-snow model and snow-depth observations. J. Hydrometeor., 3 , 646–659.
Liston, G. E., and Sturm M. , 2004: The role of winter sublimation in the Arctic moisture budget. Nord. Hydrol., 35 , 325–334.
Liston, G. E., and Elder K. , 2006a: A distributed snow-evolution modeling system (SnowModel). J. Hydrometeor., 7 , 1259–1276.
Liston, G. E., and Elder K. , 2006b: A meteorological distribution system for high-resolution terrestrial modeling (MicroMet). J. Hydrometeor., 7 , 217–234.
Liston, G. E., Winther J-G. , Bruland O. , Elvehøy H. , and Sand K. , 1999: Below surface ice melt on the coastal Antarctic ice sheet. J. Glaciol., 45 , 273–285.
Liston, G. E., Winther J-G. , Bruland O. , Elvehøy H. , Sand K. , and Karlöf L. , 2000: Snow and blue-ice distribution patterns on the coastal Antarctic ice sheet. Antarct. Sci., 12 , 69–79.
Liston, G. E., McFadden J. P. , Sturm M. , and Pielke R. A. Sr., 2002: Modeled changes in arctic tundra snow, energy, and moisture fluxes due to increased shrubs. Global Change Biol., 8 , 17–32.
Liston, G. E., Haehnel R. B. , Sturm M. , Hiemstra C. A. , Berezovskaya S. , and Tabler R. D. , 2007: Simulating complex snow distributions in windy environments using SnowTran-3D. J. Glaciol., 53 , 241–256.
Luthcke, S. B., and Coauthors, 2006: Recent Greenland ice mass loss by drainage system from satellite gravity observations. Science, 314 , 1286–1289.
Marks, D., Domingo J. , Susong D. , Link T. , and Garen D. , 1999: A spatially distributed energy balance snowmelt model for application in mountain basins. Hydrol. Processes, 13 , 1935–1959.
Mernild, S. H., and Hasholt B. , 2006: Climatic control on river discharge simulations from the Mittivakkat Glacier catchment, Ammassalik Island, SE Greenland. Nord. Hydrol., 37 , 327–346.
Mernild, S. H., Hasholt B. , and Liston G. E. , 2006a: Water flow through Mittivakkat Glacier, Ammassalik Island, SE Greenland. Dan. J. Geogr., 106 , 25–43.
Mernild, S. H., Liston G. E. , Hasholt B. , and Knudsen N. T. , 2006b: Snow-distribution and melt modeling for Mittivakkat Glacier, Ammassalik Island, SE Greenland. J. Hydrometeor., 7 , 808–824.
Mernild, S. H., Hasholt B. , Kane D. L. , Hansen B. U. , Jakobsen B. H. , and Knudsen N. T. , 2007a: Climate, glacier mass balance, and runoff 1993–2005, and in a long term perspective (106 year), Mittivakkat Glacier catchment, Ammassalik Island, SE Greenland. Hydrol. Res., 39 , 239–256.
Mernild, S. H., Hasholt B. , and Liston G. E. , 2007b: Climatic control on river discharge simulations, Zackenberg River drainage basin, NE Greenland. Hydrol. Processes, 22 , 1932–1948. doi:10.1002/hyp.6777.
Mernild, S. H., Liston G. E. , and Hasholt B. , 2007c: Snow-distribution and melt modeling for glaciers in Zackenberg River drainage basin, NE Greenland. Hydrol. Processes, 21 , 3249–3263. doi:10.1002/hyp.6500.
Mernild, S. H., Sigsgaard C. , Rasch M. , Hasholt B. , Hansen B. U. , Stjernholm M. , and Petersen D. , 2007d: Climate, river discharge and suspended sediment transport in the Zackenberg River drainage basin and Young Sund/Tyrolerfjord, northeast Greenland, 1995–2003. Carbon Cycling in Arctic Marine Ecosystems: Case Study Young Sound, S. Rysgaard and R. N. Glud, Eds., Meddr. om Grønland—BioScience, Vol. 58, 24–43.
Mernild, S. H., Liston G. E. , Kane D. L. , Hasholt B. , and Knudsen N. T. , 2008: Spatial snow distribution, runoff, and mass balance modelling for entire Mittivakkat Glacier (1998–2006), Ammassalik Island, SE Greenland. Dan. J. Geogr., 108 , 121–136.
Moritz, R. E., Bitz C. M. , and Steig E. J. , 2002: Dynamics of recent climate change in the Arctic. Science, 297 , 1497–1502.
Mote, T. L., 2003: Estimation of runoff rates, mass balance, and elevation changes on the Greenland ice sheet from passive microwave observations. J. Geophys. Res., 108 .4052, doi:10.1029/2001JD002032.
Oerlemans, J., 1991: The mass balance of the Greenland ice sheet: Sensitivity to climate change as revealed by energy-balance modelling. Holocene, 1 , 40–49.
Ohmura, A., and Reeh N. , 1991: New precipitation and accumulation maps for Greenland. J. Glaciol., 37 , 125. 140–148.
Ohmura, A., Calanca P. , Wild M. , and Anklin M. , 1999: Precipitation, accumulation and mass balance of the Greenland Ice Sheet. Z. Gletscherkunde Glazialgeol., 35 , 1–20.
Pielke R. A. Sr, , 2002: Mesoscale Meteorological Modeling. Academic Press, 676 pp.
Pomeroy, J. W., and Essery R. L. H. , 1999: Turbulent fluxes during blowing snow: Field test of model sublimation predictions. Hydrol. Processes, 13 , 2963–2975.
Pomeroy, J. W., and Brun E. , 2001: Physical properties of snow. Snow Ecology: An Interdisciplinary Examination of Snow-Covered Ecosystems, H. G. Jones et al., Eds., Cambridge University Press, 45–126.
Prasad, R., Tarboton D. G. , Liston G. E. , Luce C. H. , and Seyfried M. S. , 2001: Testing a blowing snow model against distributed snow measurements at Upper Sheep Creek. Water Resour. Res., 37 , 1341–1357.
Putnins, P., 1970: The climate of Greenland. Climates of the Polar Regions, S. Orvig, Ed., World Survey of Climatology, Vol. 12, Elsevier, 3–112.
Ramillien, G., Lombard A. , Cazanave A. , Ivins E. R. , Llubes M. , Remy F. , and Biancala R. , 2006: Interannual variations of the mass balance of the Antarctic and Greenland ice sheet from GRACE. Global Planet. Change, 53 , 198–208.
Reeh, N., Mayer C. , Miller H. , Thomson H. H. , and Weidick A. , 1999: Present and past climate control on fjord glaciations in Greenland: Implications for IRD-deposition in the sea. Geophys. Res. Lett., 26 , 1039–1042.
Rignot, E., and Kanagaratnam P. , 2006: Changes in the velocity structure of the Greenland Ice Sheet. Science, 311 , 986–990.
Ryan, B. C., 1977: A mathematical model for diagnosis and prediction of surface winds in mountainous terrain. J. Appl. Meteor., 16 , 1547–1564.
Rysgaard, S., Vang T. , Stjernholm M. , Rasmussen B. , Windelin A. , and Kiilsholm S. , 2003: Physical conditions, carbon transport, and climate change impacts in a northeast Greenland fjord. Arct. Anarct. Alp. Res., 35 , 301–312.
Scambos, T., and Haran T. , 2002: An image-enhanced DEM of the Greenland Ice Sheet. Ann. Glaciol., 34 , 291–298.
Serreze, M. C., and Barry R. G. , 2005: The Arctic Climate System. Cambridge Atmospheric and Space Science Series, Cambridge University Press, 424 pp.
Serreze, M. C., and Coauthors, 2000: Observational evidence of recent change in the northern high-latitude environment. Climatic Change, 46 , 159–207.
Serreze, M. C., and Coauthors, 2006: The large-scale freshwater cycle of the Arctic. J. Geophys. Res., 111 .C11010, doi:10.1029/2005JC003424.
Sturm, M., Schimel J. , Michaelson G. , Welker J. M. , Oberbauer S. F. , Liston G. E. , Fahnestock J. , and Romanovsky V. E. , 2005: Winter biological processes could help convert Arctic tundra to shrubland. Bioscience, 55 , 17–26.
Su, F., Adam J. C. , Trenberth K. E. , and Lettenmaier D. P. , 2006: Evaluation of surface water fluxes of the pan-Arctic land region with a land surface model and ERA-40 reanalysis. J. Geophys. Res., 111 .D05110, doi:10.1029/2005JD006387.
Tarboton, D. G., Chowdhury T. G. , and Jackson T. H. , 1995: A spatially distributed energy balance snowmelt model. Biogeochemistry of Seasonally Snow-Covered Catchments, K. A. Tonnessen, M. W. Williams, and M. Tranter, Eds., IAHS Publ. 228, 141–155.
Tedesco, M., 2007: A new record in 2007 for melting in Greenland. Eos, Trans. Amer. Geophys. Union, 88 , 383.
Thomas, R. E., Krabill F. W. , Manizade S. , and Martin C. , 2006: Progressive increase in ice loss from Greenland. Geophys. Res. Lett., 33 .L10503, doi:10.1029/2006GL026075.
Thornton, P. E., Running S. W. , and White M. A. , 1997: Generating surfaces of daily meteorological variables over large regions of complex terrain. J. Hydrol., 190 , 214–251.
Van de Wal, R. S. W., 1996: Mass balance modelling of the Greenland Ice Sheet: A comparison of an energy balance and a degree-day model. Ann. Glaciol., 23 , 36–45.
Velicogna, I., and Wahr J. , 2006: Acceleration of Greenland ice mass loss in spring 2004. Nature, 443 , 329–331.
Vorosmarty, C. J., and Coauthors, 2001: The hydrological cycle and its role in Arctic and global environmental change: A rationale and strategy for synthesis study. Arctic Research Consortium of the United States, Fairbanks, AK, 84 pp.
Walcek, C. J., 1994: Cloud cover and its relationship to relative humidity during a spring midlatitude cyclone. Mon. Wea. Rev., 122 , 1021–1035.
Winstral, A., and Marks D. , 2002: Simulating wind fields and snow redistribution using terrain-based parameters to model snow accumulation and melt over a semi-arid mountain catchment. Hydrol. Processes, 16 , 3585–3603.
Zwally, J. H., and Giovinetto M. B. , 2001: Balance mass flux and ice velocity across the equilibrium line in drainage systems of Greenland. J. Geophys. Res., 106 , 33. 717–728.
Greenland simulation domain: (a) topography (500-m contour interval), a division of the GrIS into a western and an eastern drainage area based on surface topography, and the location of the Mittivakkat catchment and the Zackenberg catchment; and (b) the location of the coastal and GrIS meteorological tower stations, designation of snow–ice and vegetation–rock–water surface cover, and the eight air temperature lapse rate transects between the following meteorological stations: 22 and 9, 21 and 12, 20 and 4, 19 and 10, 18 and 5, 25 and 1, 15 and 4, and 23 and 6.
Citation: Journal of Hydrometeorology 9, 6; 10.1175/2008JHM957.1
A comparison between daily observed meteorological data [(a) mean wind speed, (b) mean air temperature, (c) mean relative humidity, and (d) precipitation and SnowModel–MicroMet-simulated meteorological data for the Swiss Camp on the GrIS (1995–2005) (for station info, see Table 1). Only precipitation values >1 mm w.eq. were included.
Citation: Journal of Hydrometeorology 9, 6; 10.1175/2008JHM957.1
The 1995–2005 Greenland melt extent based on satellite observations [the surface melt zone (lightest area), where summer warmth turns snow and ice around the edges of the ice sheet into slush and ponds of meltwater], and the 1995–2005 Greenland SnowModel-simulated surface melt zone (lightest area). The observed melt extent is based on satellite data provided by CIRES.
Citation: Journal of Hydrometeorology 9, 6; 10.1175/2008JHM957.1
Time series for the SnowModel-simulated and satellite-observed nonmelt areas located at the inner part of the GrIS from 1995 through 2005. Maximum and minimum percentages of the SnowModel-simulated nonmelt area in relation to the total GrIS area are illustrated.
Citation: Journal of Hydrometeorology 9, 6; 10.1175/2008JHM957.1
Simulated melt index above 2000 m MSL on the GrIS including the trendline. The melt index is defined as the melting area times the number of melting days. The unit on the abscissa is km2 × days in millions. The total GrIS area equal to or greater than 2000 m in elevation is 1 084 317 km2.
Citation: Journal of Hydrometeorology 9, 6; 10.1175/2008JHM957.1
SnowModel-simulated average ELA for the (a) western GrIS, (b) eastern GrIS (1995–2005) including estimated the average ELA from Zwally and Giovinetto (2001), and (c) average western and eastern GrIS ELA for the simulation period.
Citation: Journal of Hydrometeorology 9, 6; 10.1175/2008JHM957.1
Meteorological input data for the Greenland SnowModel simulations. Meteorological station data on the GrIS (stations 1–15, and 26) were provided by the Steffen Research Group at CIRES, coastal meteorological station data (stations 16–18 and 20–25) by the DMI, and the Zackenberg meteorological station (station number 19) by the Danish Polar Center (DPC), the Greenland Survey (ASIAQ), the GeoBasis (Danish National Environmental Research Center, NERI), and the Department of Geography and Geology, University of Copenhagen.
User-defined constants used in the SnowModel simulations[see Liston and Sturm (1998) for parameter definitions].
Mean monthly air temperature lapse rates and standard deviation based on data from eight transects laid between meteorological stations in the Greenland coastal area and on the GrIS (from 1997 through 2005). See Fig. 1b for transect locations.
Validation of SnowModel simulations: observed and modeled winter, summer, and net glacier mass balance from the Mittivakkat Glacier, Ammassalik Island (SE Greenland; 65°N) from 1995/96 to 2004/05. Validation was done for both 100-m (test area, 14.4 km2) and 5-km grid-cell increments (25.0 km2). Winter mass balance observations are carried out in late May and in early June, and summer mass balance observations in late August, while modeled winter values are taken on 31 May and summer values on 31 August. Both R2 and p are estimated between the observed and modeled values. Observed data are based on information from previous studies by Knudsen and Hasholt (2004) and Mernild et al. (2006a).
Validation of SnowModel simulations: observed and modeled winter snow depths from the Zackenberg valley (NE Greenland; 74°N) from 2003/04 to 2004/05. Validation was done for both 100-m (test area, 16.8 km2) and 5-km grid-cell increments (25.0 km2). Snow depth (SWE) observations are carried out in late May and in early June, while modeled snow-depth values are determined on 31 May.
Validations of SnowModel routines were done on a daily basis. Here, the observed and modeled snow cover extents (%) are shown every 10th day through the ablation period (from 1 Jun through 31 Aug) from the Zackenberg valley (NE Greenland; 74°N) from 1996 through 2005. Validations were done for both 100-m (test area, 16.8 km2) and 5-km grid-cell increments (25.0 km2). Observed snow cover is based on data from J. Hinkler et al. (2007; personal communication). Both R2 and p are estimated between observed and modeled values.
Table. 6. (Continued)
Surface water balance elements: corrected precipitation (P), modeled evaporation (E), modeled sublimation (SU), modeled runoff (R), and storage (ΔS) (also referred as SMB) for the GrIS from 1995 through 2005 (change in storage is calculated by the residual value), and the annual temperature anomaly.
Specific runoff (L s−1 km−2) and contribution from the GrIS to the global sea level change from 1995 through 2005. The specific runoff values do not include hydroglacio processes such as the sudden release of bulk water.
A separation of the SnowModel-simulated GrIS runoff into western and eastern drainage areas showing contributions to the Arctic Ocean from 1995 through 2005.
Average water balance components for the GrIS from 1995 through 2005. The change in storage is calculated by the residual value.
