The Hydrologic Cycle
Monthly precipitation for the central Arctic Ocean based on data from the Russian North Pole manned camps with daily bias adjustments. Raw precipitation totals are shown along with the adjustments for winds and neglect of trace amounts. Adjusted precipitation is represented by the total length of the bars [from Yang, 1999, by permission of AGU]. The adjustments are large compared to the measured precipitation. Take home message: Arctic precipitation is hard to measure.
Distribution of precipitation measuring stations north of 40 N with at least ten years of record for the period 1960-1989, based on blending various data sets [by the authors]. The high-latitude network is quite sparse (another reason why Arctic precipitation is hard to measure).
Wyoming snow gauge Blowing snow is a major problem. Wyoming snow gauges are designed to provide accurate measurements of snowfall water equivalent. The one pictured above is located on the drainage divide of Imnaviat Creek, near Toolik Lake, on the North Slope of Alaska (photo by M. Serreze).
Double-fence intercomparison reference (DFIR) gauge Courtesy J. Walsh, Univ. IL Urbana Champaign
U.S. standard 8-inch gauge; Russian Tretyakov gauge Courtesy J. Walsh, Univ. IL Urbana Champaign
Mean precipitation north of 60 N for the four mid-season months, based on data from land stations and the Arctic Ocean with bias adjustments, the NCEP/NCAR reanalysis and satellite retrievals (over open ocean). The fields are based primarily on data for 1960-1989. Contours are at every 10 mm up to 80 mm and at every 50 mm (dashed) for amounts of 100 mm and higher [by the authors]. The Atlantic sector of the Arctic is quite wet. Precipitation over other parts of the Arctic is quite low. Some land areas are classified as polar desert. Precipitation peaks over the Atlantic sector during the cold months and in summer over most land areas.
Annual accumulation over the Greenland Ice Sheet in mm water equivalent. The contour intervals are 200 mm, but 100 mm if smaller than 400 mm and 600 mm if larger than 1000 mm [from Chen et al., 1997, by permission of AMS]. Due to orographic precipitation, accumulation along the southeast coast of Greenland locally exceeds 2000 mm. Accumulation over the northcentral part of the ice sheet is only 100-200 mm.
Precipitation frequency, the frequency of moderate to heavy precipitation, solid precipitation frequency and liquid precipitation frequency (in %), based on present weather reports in COADS records over the period 1950-1995. Maps are provided for January and July. See textbook for details [from Serreze et al., 1997, by permission of NSIDC, Boulder, CO]. Note the sharp contrast in the statistics between the cold and dry central Arctic Ocean and the warmer and wetter Atlantic-side of the Arctic.
The aerological approach W/ t = ET - P - Q W/ t = time change in precipitable water (column water vapor) ET = Evapotranspiration P = Precipitation Q = Vertically-integrated vapor flux divergence W/ t Q Rearrange: P- ET = - Q - W/ t P ET Key: While P and ET are hard to measure over large areas, we can get the net precipitation (P-ET) using atmospheric winds and humidities
Aerological estimates of mean annual precipitation minus evapotranspiration (P- ET) based on NCEP/NCAR data for the period 1970-1999 (mm). Contours are at every 100 mm up to 500 m (negative values dashed) and at every 200 mm for amounts of 600 mm and higher [by the authors]. Annual P-ET is positive everywhere except locally in the Norwegian Sea and southern Barents Sea where evaporation rates are high from autumn through spring (because of cold air overlying a fairly warm open ocean).
Aerological estimates of mean annual precipitation minus evapotranspiration (P-ET) based on NCEP/NCAR data for the period 1970-1999 (mm). Contours are at every 100 mm up to 500 m (negative values dashed) and at every 200 mm for amounts of 600 mm and higher [by the authors]. The annual map (previous slide) masks strong seasonal variations. Note the negative P-ET in July over some land areas. While precipitation peaks over these land areas in summer, this is countered by even stronger evapotranspiration.
Focus on the Arctic-draining rivers
Physiography of the Arctic lands, showing topography and major river systems [courtesy of R. Lammers, University of New Hampshire, Durham, NH].
The Arctic terrestrial drainage (all shaded regions) and its four largest individual watersheds (darker shading) [courtesy of A. Barrett, NSIDC, Boulder, CO].
The water balance of a watershed Inputs (I), outputs (O) and storage (S): I: Precipitation (P) Groundwater in (G in ) O: Evapotranspiration (ET) Groundwater out (G out ) River discharge (Q) Storage (S): In groundwater, rivers and lakes Dingman 2002, Fig. 2-3 S = P + G in (Q + ET + G out ) If we assume that G in and G out are negligible, and that for the long-term annual mean, S is zero, then: P = ET + Q, or ET = P - Q What can we usually measure? P: rain gauges Q: stream gauges ET: hard to get except local values G in : hard to get, assume zero G out : hard to get, assume zero S: often hard to get
Mean monthly precipitation (P), precipitation minus evapotranspiration (P-ET) and evapotranspiration (ET) for the four major Arctic-draining watersheds, based on data from 1960 through 1989 (mm). ET is calculated as a residual from P and P-ET [adapted from Serreze et al., 2003a, by permission of AGU]. Note now the peak P in summer is countered by ET
Mean evapotranspiration (ET) for the Arctic terrestrial drainage (mm) for alternate months, estimated as a residual from biasadjusted precipitation and aerological estimates of P-ET from NCEP/NCAR. Results are based on data from 1960 through 1999 [from Serreze et al., 2003a, by permission of AGU]. ET peaks in July.
P-ET anomalies by month and year over the Arctic terrestrial drainage from the JRA-25 atmospheric reanalysis. P-ET is highly variable and there is no obvious temporal trend. For years as a whole, P-ET was highest in 2007 and lowest in 1990 [from Serreze and Barrett and Slater, 2008].
Precipitation recycling What fraction of precipitation that falls within a watershed is due to water evapotranspirated from that watershed which then falls back within the watershed? From the formulation of Brubaker et al. (2003): P L /P = 1/(1+ 2.F + /ET.A) P = Total precipitation P L = Precipitation of local origin E T = Evapotranspiration A = Area of watershed F + = Vertically integrated vapor flux directed into the watershed (advective moisture term) F+ To get a high recycling ratio (P/P L ) you want a large evaporation rate and a small advective moisture term. P E T The ratio is very scale dependant. At the global scale, ALL precipitation is recycled! F+ Dingman 2002 Figure 2-3
Mean monthly precipitation recycling ratio for four domains chosen to approximate represent the Ob, Yenisey, Lena and Mackenzie basins [from Serreze et al., 2003a, by permission of AGU]. Typically 20-30% of precipitation falling in these watersheds during summer is recycled the precipitation falls, the water evaporates, then falls again in the same watershed.
River system for the Arctic terrestrial drainage based on the STN- 30 digital network and location of discharge gauging stations (dots) for basins with areas greater than 10,000 km 2 [from Lammers et al., 2001, by permission of AGU].
Number of gauges in the R-ArcticNET data base through time and the percentage of the drainage area monitored. Data are shown for North America, Eurasia and the Arctic drainage as a whole. Only gauges monitoring an area greater than 10,000 km 2 are used for the percentage of area monitored [adapted from Lammers et al., 2001, by permission of AGU]. The Eurasian network has degraded since about 1985.
Sources of runoff (mm/yr) reaching the Arctic Ocean Mean annual runoff for the Arctic drainage over the period 1960-1989. Calculations were made by subtracting all upstream discharge values from the discharge value at the next downstream gauge. Runoff was then obtain by dividing the these intersection discharges by the intersection drainage area. White areas within the southern Ob represent internal drainage basins. Large areas along the Arctic coast and the Canadian Arctic Archipelago (shown as white) are ungauged. The lines over the Arctic Ocean indicate sea basins into which different river systems drain. (From Lammers et al., 2001, by permission of AGU)
Mean hydrographs for the four major Arctic-draining rivers (at the gauge sites closest to the river mouths) expressed in terms of runoff (mm) [adapted from Serreze et al., 2003a, by permission of AGU]. Like here in Colorado, there is a June peak in runoff due to snowmelt. Because the Yenisey and Lena are underlain by continuous, impermeable permafrost, there is little infiltration. Hence the June peak is especially pronounced in these basins.
Mean surface salinity (psu) for summer [from Arctic Climatology Project,1998, NSIDC, Boulder CO]. Note the low salinities along coastal areas due to seasonal river runoff.
Snow water equivalent (SWE) and depth are hard to measure Data sources cited in the next slide: 1) MERRA atmospheric reanalysis - mean SWE for 2002-2009 (to match GRACE period) 2) GRACE (courtesy Sean Swenson, NCAR) - 2002-2009 average SWE (mm), derived from total satellite-derived gravity anomaly less soil moisture and hydro component anomalies computed by running the CLM model with observed forcing (station corrected NCEP) 3) ERA-I - ERA Interim atmospheric reanalysis 2002-2009, SWE (mm) 4) AMSR-E, 2002-2009, passive microwave SWE (mm), using Richard Kelly's algorithm (data stored at NSIDC) 5) ETAC-USAF (Environmental Technology Applications Center, US Air Force) station climatology - Foster & Davey, (1988). Based on station data going back to the 1950's. Data are DEPTH (m) 6) CLM2.0 (land surface) model forced with ERA-40 data (1979-2002) (only over Arctic catchments, not entire hemisphere) Data are DEPTH (m)
SWE and snow depth for March Figure and analysis by Andrew Slater, NSIDC
Monthly discharge (m 3 s -1 ) at the mouth of the Lena over the period 1936-1999. For each month, the plot shows the discharge for all years [from Yang et al., 2002, by permission of AGU]. While almost every year shows a June peak, monthly discharge from year to year can be highly variable.
Daily averaged precipitation over the Lena for 1980 and 1981 (solid line) along with the mean annual cycle (dotted line) over the period 1960-1991 [from Serreze and Etringer, 2003, by permission of John Wiley and Sons]. Extreme precipitation events can lead to strong variations in discharge, expecially in the smaller sub-basins of the large watersheds.
Water-year time series (mm) of runoff (R), precipitation (P), precipitation minus evapotranspiration (P-ET) and ET for the Lena. ET is calculated as a residual from P and P-ET. P is adjusted for estimated gauge undercatch and P- ET is from aerological estimates using NCEP/NCAR reanalysis data [adapted from Serreze et al., 2003a, by permission of AGU]. In this basin, where there is extensive permafrost that inhibits infiltration and promotes rapid movement of precipitation into river networks, water-year runoff and P-ET are strongly correlated. They are only weakly correlated in the other large basins.
The large-scale freshwater budget
The mean annual freshwater budget of the Arctic The Domain From Serreze et al., 2006