Fig. 14.1 Mean annual P-E for the region north of 50° N from ERA-40, based on aerological calculations over the period 1979-2001
freshening from river runoff, P-E over the Arctic Ocean, and from other minor terms, but also for the large freshwater inflow into the Arctic Ocean via Bering Strait (Fig. 14.2). This represents 30% of annual freshwater input and manifests several processes, including (on the large scale) differences in P-E over the North Pacific and North Atlantic, associated with upper-ocean salinity differences that help maintain a gradient in sea surface height, and discharge from the Yukon river
Fig. 14.2 Mean annual freshwater budget for the Arctic, based on a reference salinity of 34.8. (Adapted from Serreze et al. 2006.) The ocean domain (area of 9.6 x 106 km2) is defined by lines across Fram Strait, from Svalbard to northern Scandinavia, across the Bering Strait, and along the northern coast of the Canadian Arctic Archipelago. The land region draining into this ocean domain (area of 15.8 x 106 km2) was defined using a digital river network. The atmospheric box combines the land and ocean domains. The boxes for land and ocean are sized proportional to their areas. All transports are in units of km3 per year. Stores are in km3. These are based on the best available estimates drawn from recent literature or computed as part of the Serreze et al. (2006) study. This includes data from the ERA-40 reanalysis
Fig. 14.2 Mean annual freshwater budget for the Arctic, based on a reference salinity of 34.8. (Adapted from Serreze et al. 2006.) The ocean domain (area of 9.6 x 106 km2) is defined by lines across Fram Strait, from Svalbard to northern Scandinavia, across the Bering Strait, and along the northern coast of the Canadian Arctic Archipelago. The land region draining into this ocean domain (area of 15.8 x 106 km2) was defined using a digital river network. The atmospheric box combines the land and ocean domains. The boxes for land and ocean are sized proportional to their areas. All transports are in units of km3 per year. Stores are in km3. These are based on the best available estimates drawn from recent literature or computed as part of the Serreze et al. (2006) study. This includes data from the ERA-40 reanalysis that appears as part of the seasonal Alaskan Coastal Current (Woodgate et al. 2006). Another aspect of atmospheric forcing is that ocean transports are sensitive to the regional wind field. This has been well documented for the Fram Strait ice flux (Vinje 2001) and Bering Strait inflow (Aagaard et al. 1985; Woodgate et al. 2005).
This paper reviews the atmospheric branch of the Arctic hydrologic cycle with a focus on the integrating theme of P-E. We examine seasonal and spatial patterns of P and E, and how these are expressed as net precipitation. Aspects of freshwater storage are addressed, as well as observed variability and projected future states of the freshwater system. Use is made of output from the European Centre for Medium Range Weather Forecasts (ECMWF) ERA-40 reanalysis, results from land surface models (LSMs), time series of observed precipitation, and findings from recent published studies.
Atmospheric reanalyses such as ERA-40 (Uppala et al. 2005) and from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) (Kalnay et al. 1996; Kistler et al. 2001) are retrospective forms of numerical weather prediction, whereby gridded fields of atmospheric and surface variables are compiled using fixed versions of a forecast/data assimilation system. ERA-40 provides 6-hourly fields on a grid of approximately 125 km from September 1957 through July 2002.
Fields of free-air variables, such as tropospheric pressure heights, winds, and humidity, are compiled by assimilating observations within a short term atmospheric forecast. These "blended" products are generally the most reliable aspects of reanalysis. Fields of surface variables such as precipitation, evaporation and other terms of the surface energy budget, do not involve blending with observations. In an operational setting (i.e., in routine weather forecasting), the forecast/data assimilation system is constantly refined to improve forecast skill. This can lead to non-climatic jumps and trends in archived fields. By using fixed systems, archives from reanalysis are more consistent, but temporal inconsistencies are still present due to changes in observing networks (e.g., rawinsonde and satellite data bases).
Atmospheric reanalyses allow for a full accounting of the atmospheric hydrologic budget. P-E can be obtained in two ways. The first is from the model forecasts of P and E. The second and preferred method (Cullather et al. 2000; Rogers et al. 2001) is the aerological approach. Consider an atmospheric column, extending from the surface to the top of atmosphere. Its water budget can be expressed as:
where dW/dt represents the change in precipitable water (W) in the atmosphere (the water depth of the vapor in the column), and -V^Q is the convergence of the vertically integrated horizontal water vapor flux Q. In the aerological approach, P-E is obtained by adjusting the vapor flux convergence by the tendency in precipitable water. For long-term annual means and assuming a steady-state, the tendency term can be dropped, so that net precipitation equals the vapor flux convergence. Small effects of phase transformations in the atmosphere represented by clouds, as well as convergence of water in liquid and solid phases, are ignored. A number of other studies (e.g. Walsh et al. 1994; Gober et al. 2003) have used the aerological approach to assess P-E averaged for large domains (such as the region north of 70° N) using data from rawinsonde profiles. An advantage of using reanalysis is that one can obtain gridded fields of P-E.
High-latitude precipitation fields from ERA-40 are known to be greatly improved over those from NCEP/NCAR, and, at least for most regions, capture observed interannual variability, although the model has generally less precipitation than observations (Serreze et al. 2005; Betts et al. 2003). Evaporation estimates from ERA-40 seem reasonable, at least for land (Slater et al. 2007). However, P-E based on the aerological method and from the forecasts of P and E are not in balance, with lower P-E in the latter. This results primarily from nudging the model humidity toward observations. For annual means over the period 1979-1993, Cullather et al. (2000) cite an imbalance over the polar cap (the region north of 70° N) of 50 mm for ERA-15 and 82 mm for NCEP/NCAR. For ERA-40, Serreze et al. (2006) calculate a smaller imbalance of 15 mm. These issues should be kept in mind when interpreting our results.
Our approach is to view aerological P-E as the best representation of truth, and then assess its components using ERA-40 fields of P and E. This recognizes that surface observations of P and E are insufficient to obtain gridded fields over the entire north polar region. The primary focus is on the period 1979-2001 for which fields are most reliable due to the wealth of satellite data for model assimilation. However, data back to 1958 are used to examine longer-term variations in aerological P-E. To complement ERA-40, estimates of zonally averaged P for the region 55-85° N are examined for 1900-2004, based on land station records contained in the Global Precipitation Climatology Centre (GPCC) database (http://gpcc.dwd.de).
Snowpack water equivalent (SWE) over the terrestrial drainage is a key aspect of the hydrologic system. Surface observations are insufficient to compile gridded fields, and those based on satellite remote sensing are of questionable fidelity. We use estimates of SWE (seasonal storage of P-E) based on averaging output from five different land surface models (LSMs), each driven with ERA-40 inputs for 1979-2001 (precipitation, temperature, low level humidity and winds, and downwelling solar and longwave radiation). The five model average should give a better representation of SWE than from any one model.
Was this article helpful?