Abstract

We analyzed modeled river runoff variations west of the Andes Cordillera’s continental divide for 1979/80–2013/14 (35 years). Our foci were annual runoff conditions, runoff origins (rain, snowmelt, and glacier ice), and runoff spatiotemporal variability. Low and high runoff conditions were defined as occurrences that fall outside the 10th (low values) and 90th (high values) percentile values of the period of record. SnowModel and HydroFlow modeling tools were used at 4-km horizontal grid increments and 3-h time intervals. NASA Modern-Era Retrospective Analysis for Research and Applications (MERRA) datasets were used as atmospheric forcing. This modeling system includes evaporation and sublimation from snow-covered surfaces, but it does not take into account evapotranspiration from bare and vegetation-covered soils and from river and lake surfaces. In general for the Andes Cordillera, the simulated runoff decreased before 1997 and increased afterward. This could be due to a model precipitation artifact in the MERRA forcing. If so, this artifact would influence the number of years with low runoff values, which decreased over time, while the number of high runoff values increased over time. For latitudes south of ~40°S, both the greatest decrease in the number of low runoff values and the greatest increase in high runoff values occurred. High runoff values averaged 84% and 58% higher than low values for nonglacierized and glacierized catchments, respectively. Furthermore, for glacierized catchments, 61% and 62% of the runoff originated from rain-derived runoff during low and high runoff extreme years, respectively; 28% and 30% from snowmelt-derived runoff; and 11% and 8% from glacier-ice-melt-derived runoff. As the results could be MERRA dependent, more work with other precipitation forcings and/or in situ measurements is needed to assess whether these are real runoff behaviors or artifacts.

Abstract

SnowModel, a physically based snow-evolution modeling system that includes four submodels—MicroMet, EnBal, SnowPack, and SnowTran-3D—was used to simulate variations in Greenland [including the Greenland Ice Sheet (GrIS)] surface snow and ice melt, as well as water balance components, for 1995–2005. Meteorological observations from 25 stations inside and outside the GrIS were used as model input. Winter and summer mass balance observations, spatial snow depth observations, and snowmelt depletion curves derived from time-lapse photography from the Mittivakkat and Zackenberg glacierized catchments in East Greenland were used to validate the performance of SnowModel. Model results compared well with observed values, confirming the robustness of the model. The yearly modeled GrIS interior nonmelt area differs from satellite observations by a maximum of ∼68 000 km² (or ∼6%) in 2004, and the lowest uncertainties (<8000 km², or <1%) occur for the years with the smallest (2005) and most extensive (1996) nonmelt areas. Modeled surface melt occurred at elevations reaching 2950 m MSL for 2005, while the equilibrium line altitude (ELA) fluctuates from 1640 to 600 m MSL. The modeled interannual variability in the nonmelt area also agrees with observation records (R ² = 0.96), yielding simulated GrIS nonmelt covers of 71% for 1996 and 50% for 2005. On average, the simulated nonmelt area decreased ∼6% from 1995 to 2005; this trend is similar to observed values. An average surface mass balance (SMB) storage of 138(±81) km³ yr⁻¹, a GrIS loss of 257(±81) km³ yr⁻¹, and a runoff contribution to the ocean of 392(±58) km³ yr⁻¹ occurred for the period 1995–2005. Approximately 58% and 42% of the runoff came from the GrIS western and eastern drainage areas, respectively. The modeled average specific runoff from the GrIS was 6.71 s⁻¹ km⁻² yr⁻¹, which, over the simulation period, represents a contribution of ∼1.1 mm yr⁻¹ to global sea level rise.

Abstract

A physically based snow-evolution modeling system (SnowModel) that includes four submodels—the Micrometeorological Model (MicroMet), EnBal, SnowPack, and SnowTran-3D—was used to simulate five full-year evolutions of snow accumulation, distribution, sublimation, and surface melt on the Mittivakkat Glacier, in southeast Greenland. Model modifications were implemented and used 1) to adjust underestimated observed meteorological station solid precipitation until the model matched the observed Mittivakkat Glacier winter mass balance, and 2) to simulate glacier-ice melt after the winter snow accumulation had ablated. Meteorological observations from two meteorological stations were used as model inputs, and glaciological mass balance observations were used for model calibration and testing of solid precipitation observations. The modeled end-of-winter snow-water equivalent (w.eq.) accumulation increased with elevation from 200 to 700 m above sea level (ASL) in response to both elevation and topographic influences, and the simulated end-of-summer location of the glacier equilibrium line altitude was confirmed by glaciological observations and digital images. The modeled test-period-averaged annual mass balance was 150 mm w.eq. yr⁻¹, or ∼15%, less than the observed. Approximately 12% of the precipitation was returned to the atmosphere by sublimation. Glacier-averaged mean annual modeled surface melt ranged from 1272 to 2221 mm w.eq. yr⁻¹, of which snowmelt contributed from 610 to 1040 mm w.eq. yr⁻¹. The surface-melt period started between mid-May and the beginning of June, and lasted until mid-September; there were as many as 120 melt days at the glacier terminus. The model simulated a Mittivakkat Glacier recession averaging −616 mm w.eq. yr⁻¹, almost equal to the observed −600 mm w.eq. yr⁻¹.

Abstract

This study evaluated the spatial variability of trends in simulated snowpack properties across the Rio Grande headwaters of Colorado using the SnowModel snow evolution modeling system. SnowModel simulations were performed using a grid resolution of 100 m and 3-hourly time step over a 34-yr period (1984–2017). Atmospheric forcing was provided by phase 2 of the North American Land Data Assimilation System, and the simulations accounted for temporal changes in forest canopy from bark beetle and wildfire disturbances. Annual summary values of simulated snowpack properties [snow metrics; e.g., peak snow water equivalent (SWE), snowmelt rate and timing, and snow sublimation] were used to compute trends across the domain. Trends in simulated snow metrics varied depending on elevation, aspect, and land cover. Statistically significant trends did not occur evenly within the basin, and some areas were more sensitive than others. In addition, there were distinct trend differences between the different snow metrics. Upward trends in mean winter air temperature were 0.3°C decade⁻¹, and downward trends in winter precipitation were −52 mm decade⁻¹. Middle elevation zones, coincident with the greatest volumetric snow water storage, exhibited the greatest sensitivity to changes in peak SWE and snowmelt rate. Across the Rio Grande headwaters, snowmelt rates decreased by 20% decade⁻¹, peak SWE decreased by 14% decade⁻¹, and total snowmelt quantity decreased by 13% decade⁻¹. These snow trends are in general agreement with widespread snow declines that have been reported for this region. This study further quantifies these snow declines and provides trend information for additional snow variables across a greater spatial coverage at finer spatial resolution.

Abstract

The distribution of terrestrial surface runoff to Ilulissat Icefjord, west Greenland, is simulated for the period 2009–13 to better emphasize the spatiotemporal variability in freshwater flux and the link between runoff spikes and observed hydrographic conditions at the Greenland Ice Sheet tidewater glacier margins. Runoff model simulations were forced with automatic weather station data and verified against snow water equivalent depth, equilibrium line altitude, and quasi-continuous salinity and temperature observations obtained by ringed seals. Instrumented seals provide a novel platform to examine the otherwise inaccessible waters beneath the dense ice mélange within the first 0–10 km of the calving front. The estimated mean freshwater flux from land was 70.6 ± 4.2 km³ yr⁻¹, with an 85% contribution of ice discharge from Jakobshavn Isbrae (also known as Sermeq Kujalleq), 14% from runoff, and the remaining 1% from precipitation on the fjord surface area, subglacial geothermal melting, and frictional melting due to basal ice motion. Runoff was simulated to be present from May to November and to vary spatially according to glacier cover and individual catchment size. Salinity and temperature observations correlate (significantly) with simulated runoff for the upper part of both the main fjord and southern fjord arm. Also, at the tidewater glacier margins in the northern and southern arm of Ilulissat Icefjord, salinity changes in the upper water column (upper 50 m) are significant after temporal spikes in runoff during late summer, while small-amplitude runoff variability during the recession of runoff did not create a clear signal in observed salinity variability. Also, in the southern arm near the glacier margin (between 100- and 150-m depth), the heterogeneous distribution in salinity could be because of the mixing of meltwater going upward from passing the grounding line. The effect of runoff spikes on observed salinity is less pronounced near the ice margin of Jakobshavn Isbrae than in the north and south arms.

Abstract

In the Arctic, where wind transport of snow is common, the depth and insulative properties of the snow cover can be determined as much by the wind as by spatial variations in precipitation. Where shrubs are more abundant and larger, greater amounts of drifting snow are trapped and suffer less loss due to sublimation. The snow in shrub patches is both thicker and a better thermal insulator per unit thickness than the snow outside of shrub patches. As a consequence, winter soil surface temperatures are substantially higher, a condition that can promote greater winter decomposition and nutrient release, thereby providing a positive feedback that could enhance shrub growth. If the abundance, size, and coverage of arctic shrubs increases in response to climate warming, as is expected, snow–shrub interactions could cause a widespread increase (estimated 10%–25%) in the winter snow depth. This would increase spring runoff, winter soil temperatures, and probably winter CO₂ emissions. The balance between these winter effects and changes in the summer energy balance associated with the increase in shrubs probably depends on shrub density, with the threshold for winter snow trapping occurring at lower densities than the threshold for summer effects such as shading. It is suggested that snow–shrub interactions warrant further investigation as a possible factor contributing to the transition of the arctic land surface from moist graminoid tundra to shrub tundra in response to climatic warming.

Abstract

A regional atmospheric model, the HIRHAM4 regional climate model (RCM) using boundary conditions from the ECHAM5 atmosphere–ocean general circulation model (AOGCM), was downscaled to a 500-m gridcell increment using SnowModel to simulate 131 yr (1950–2080) of hydrologic cycle evolution in west Greenland’s Kangerlussuaq drainage. Projected changes in the Greenland Ice Sheet (GrIS) surface mass balance (SMB) and runoff are relevant for potential hydropower production and prediction of ecosystem changes in sensitive Kangerlussuaq Fjord systems. Mean annual surface air temperatures and precipitation in the Kangerlussuaq area were simulated to increase by 3.4°C and 95 mm water equivalent (w.eq.), respectively, between 1950 and 2080. The local Kangerlussuaq warming was less than the average warming of 4.8°C simulated for the entire GrIS. The Kangerlussuaq SMB loss increased by an average of 0.3 km³ because of a 0.4 km³ rise in precipitation, 0.1 km³ rise in evaporation and sublimation, and 0.6 km³ gain in runoff (1950–2080). By 2080, the spring runoff season begins approximately three weeks earlier. The average modeled SMB and runoff is approximately −0.1 and 1.2 km³ yr⁻¹, respectively, indicating that ∼10% of the Kangerlussuaq runoff is explained by the GrIS SMB net loss. The cumulative net volume loss (1950–2080) from SMB was 15.9 km³, and runoff was 151.2 km³ w.eq. This runoff volume is expected to have important hydrodynamic and ecological impacts on the stratified salinity in the Kangerlussuaq Fjord and on the transport of freshwater to the ocean.

Abstract

Meteorological forcing data are necessary to drive many of the spatial models used to simulate atmospheric, biological, and hydrological processes. Unfortunately, many domains lack sufficient meteorological data and available point observations are not always suitable or reliable for landscape or regional applications. NOAA’s Local Analysis and Prediction System (LAPS) is a meteorological assimilation tool that employs available observations (meteorological networks, radar, satellite, soundings, and aircraft) to generate a spatially distributed, three-dimensional representation of atmospheric features and processes. As with any diagnostic representation, it is important to ascertain how LAPS outputs deviate from a variety of independent observations. A number of surface observations exist that are not used in the LAPS system, and they were employed to assess LAPS surface state variable and precipitation analysis performance during two consecutive years (1 September 2001–31 August 2003). LAPS assimilations accurately depicted temperature and relative humidity values. The ability of LAPS to represent wind speed was satisfactory overall, but accuracy declined with increasing elevation. Last, precipitation estimates performed by LAPS were irregular and reflected inherent difficulties in measuring and estimating precipitation.

Abstract

Thirteen regional climate model (RCM) simulations of June–July 1993 were compared with each other and observations. Water vapor conservation and precipitation characteristics in each RCM were examined for a 10° × 10° subregion of the upper Mississippi River basin, containing the region of maximum 60-day accumulated precipitation in all RCMs and station reports.

All RCMs produced positive precipitation minus evapotranspiration (P − E > 0), though most RCMs produced P − E below the observed range. RCM recycling ratios were within the range estimated from observations. No evidence of common errors of E was found. In contrast, common dry bias of P was found in the simulations.

Daily cycles of terms in the water vapor conservation equation were qualitatively similar in most RCMs. Nocturnal maximums of P and C (convergence) occurred in 9 of 13 RCMs, consistent with observations. Three of the four driest simulations failed to couple P and C overnight, producing afternoon maximum P. Further, dry simulations tended to produce a larger fraction of their 60-day accumulated precipitation from low 3-h totals.

In station reports, accumulation from high (low) 3-h totals had a nocturnal (early morning) maximum. This time lag occurred, in part, because many mesoscale convective systems had reached peak intensity overnight and had declined in intensity by early morning. None of the RCMs contained such a time lag. It is recommended that short-period experiments be performed to examine the ability of RCMs to simulate mesoscale convective systems prior to generating long-period simulations for hydroclimatology.