• Allan, C. J., and Roulet N. , 1994: Runoff generation in zero order Precambrian shield catchments: The stormflow response of a heterogeneous landscape. Hydrol. Processes, 8 , 369388.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Amiro, B. D., and Wushcke E. E. , 1987: Evapotranspiration from a boreal forest drainage basin using an energy balance/eddy correlation technique. Bound.-Layer Meteor., 38 , 125139.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bello, R. L., and Arama A. , 1999: Rainfall interception in lichen canopies. Climatol. Bull., 23 , 7478.

  • Blanford, J. H., and Gay L. W. , 1992: Tests of a robust eddy correlation system for sensible heat flux. Theor. Appl. Climatol., 46 , 5360.

  • Buttle, J. M., and Sami K. , 1992: Testing the groundwater ridging hypothesis of streamflow generation during snowmelt in a forested catchment. J. Hydrol., 135 , 5372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Drury, M., and Lewis T. , 1981: Thermal studies in the Lac du Bonnet batholith at Pinawa, Manitoba. Atomic Energy of Canada Limited Tech. Rep. II 5000, 14 pp.

    • Search Google Scholar
    • Export Citation
  • Farouki, O., 1981: Thermal Properties of Soils. CRREL Monogr., No. 81, Cold Regions Research and Engineering Laboratory, United States Army Corps of Engineers, 136 pp.

    • Search Google Scholar
    • Export Citation
  • Fitzjarrald, D. R., and Moore K. E. , 1994: Growing season boundary layer climate and surface exchanges in a subarctic lichen woodland. J. Geophys. Res., 99 , 18991917.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garratt, J. R., 1975: Limitations of the eddy correlation technique for the determination of turbulent fluxes near the surface. Bound.-Layer Meteor., 8 , 255259.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jarvis, P. G., and McNaughton K. G. , 1986: Stomatal control of transpiration. Adv. Ecol. Res., 15 , 149.

  • Jarvis, P. G., Massheder J. M. , Hale S. E. , Moncrieff J. B. , Rayment M. , and Scott S. L. , 1997: Seasonal variation of carbon dioxide, water vapor, and energy exchanges of a boreal black spruce forest. J. Geophys., Res., 102 , 2895328966.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kershaw, G. P., and Rouse W. R. , 1971: Studies on lichen dominated systems. I: The water relations of Cladonia alpestris in spruce lichen woodland in northern Ontario. Can. J. Bot, 49 , 13891399.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lafleur, P. M., 1992: Energy balance and evapotranspiration from a subarctic forest. Agric. For. Meteor., 58 , 163175.

  • Lafleur, P. M., and Adams P. , 1986: The radiation budget of a subarctic woodland canopy. Arctic, 39 , 172176.

  • Lafleur, P. M., and Schreader C. P. , 1994: Water loss from the floor of a subarctic forest. Arct. Alpine Res., 26 , 152158.

  • Lafleur, P. M., Rouse W. R. , and Carlson D. W. , 1992: Energy balance differences and hydrologic impacts across the northern treeline. Int. J. Climatol., 12 , 193203.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lafleur, P. M., McCaughey J. H. , Joiner D. W. , Bartlett P. A. , and Jelinski D. E. , 1997: Seasonal trends in energy, water and carbon dioxide at a northern boreal wetland. J. Geophys. Res., 102 ((D24),) 2900929020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McCaughey, J. H., Lafleur P. M. , Joiner D. W. , Bartlet P. A. , Costello A. M. , Jelinski D. E. , and Ryan M. G. , 1997: Magnitudes and seasonal patterns of energy, water and carbon exchanges at a boreal young jack pine forest in the BOREAS northern study area. J. Geophys. Res., 102 , 2899729007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moore, C. J., 1986: Frequency response corrections for eddy correlation systems. Bound.-Layer Meteor., 37 , 1735.

  • Moore, K. E., Fitzjarrald D. R. , Wofsy S. C. , Daube B. C. , Munger J. W. , Bakwin P. S. , and Crill P. , 1994: A season of heat, water vapor, total hydrocarbon, and ozone fluxes at a subarctic fen. J. Geophys. Res., 99 , 19371952.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moore, K. E., Fitzjarrald D. R. , Sakai R. K. , and Freedman J. M. , 2000: Growing season water balance at a boreal jack pine forest. Water Resour. Res., 36 , 483493.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peters, D. L., Buttle J. M. , Taylor C. H. , and LaZerte B. D. , 1995: Runoff production in a forested shallow soil Canadian Shield basin. Water Resour. Res., 31 , 12911304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Petrone, R. M., Rouse W. R. , and Marsh P. , 2000: Comparative surface energy budgets in western and central subarctic regions of Canada. Int. J. Climatol., 20 , 11311148.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Petzold, D. E., and Rencz A. N. , 1975: The albedo of selected subarctic surfaces. Arct. Alpine Res., 13 , 287293.

  • Phillips, D., 1990: The Climates of Canada. Environment Canada, 176 pp.

  • Rouse, W. R., and Bello R. L. , 1985: Impact of Hudson Bay on the energy balance in the Hudson Bay Lowlands and the potential for climate modification. Atmos.–Ocean, 23 , 375392.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rouse, W. R., Bello R. L. , and Lafleur P. , 1997: The low arctic and subarctic. The Surface Climates of Canada, W. Bailey, T. Oke, and W. R. Rouse, Eds., McGill Queen's University Press, 198–221.

    • Search Google Scholar
    • Export Citation
  • Schuepp, P. H., Leclerc M. Y. , Macpherson J. I. , and Desjardins R. L. , 1990: Footprint prediction of scalar fluxes from analytical solutions of the diffusion equation. Bound.-Layer Meteor., 50 , 355373.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and Coauthors. 1997: BOREAS in 1997: Experiment overview, scientific results, and future directions. J. Geophys. Res., 102 , 2873128769.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Singh, B., and Taillifer R. , 1986: The effect of synoptic scale advection on the performance of the Priestley–Taylor evaporation formula. Bound.-Layer Meteor., 36 , 367382.

    • Search Google Scholar
    • Export Citation
  • Spence, C., 2000: The effect of storage on runoff from a headwater subarctic shield basin. Arctic, 53 , 237247.

  • Szeto, K., 1998: Low level inversions, precipitation recycling and air-land interactions over the Mackenzie Basin. Proc. Fourth Scientific Workshop for the Mackenzie GEWEX Study (MAGS), Environ. Can., 27–28.

    • Search Google Scholar
    • Export Citation
  • Thorne, G. A., Laporte J. M. , Clarke D. , and Davison C. C. , 1999: Water budget of an upland outcrop recharge area in granitic rock terrane of southeastern Manitoba, Canada. Proc. 12th Int. Northern Research Basins Symposium and Workshop, Reykjavik, Iceland, Working Group on Northern Research Basins, 317–330.

    • Search Google Scholar
    • Export Citation
  • Wedel, J. H., Smart A. , and Squires P. , 1990: An Overview Study of the Yellowknife River Basin, NWT. Inland Waters Directorate, Environment Canada, 82 pp.

    • Search Google Scholar
    • Export Citation
  • Wight, J., 1973: Aspects of evaporation and evapotranspiration in the water balance of Baker Creek, near Yellowknife, NWT,. M.S. thesis, Dept. of Geography, University of Alberta, 329 pp.

    • Search Google Scholar
    • Export Citation
  • Wright, R. K., 1981: The water balance of a lichen tundra underlain by permafrost. McGill Subarctic Research Paper 33, Centre for Northern Studies, McGill University, 110 pp.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Photo of Skeeter Lake upland with a location map. The arrow points to the climate tower. The lightest gray tones are exposed bedrock (B). The wetland (PW) is located behind the tower. Black spruce (BS) and mixed stands (MW) are scattered throughout the site. White arrows denote direction of surface water flow. Predominant winds come from the northeast, which is the upper-right-hand corner of the picture

  • View in gallery

    Monthly (top) air temperature and (bottom) precipitation. The white circles represent observations made in 1999 at Skeeter Lake, and the white boxes are for 2000 observations. The black squares are Environment Canada 1961–90 averages for Yellowknife, with the bars representing ± 1 std dev

  • View in gallery

    (a) Cumulative precipitation and (b) cumulative change in soil moisture storage from the beginning of the study period on the Skeeter Lake upland during 1999 and 2000

  • View in gallery

    Cumulative spring snowmelt and runoff at Skeeter Lake in 1999 and 2000

  • View in gallery

    Half-hourly values of Q* measured over exposed bedrock and two soil-covered land cover types on 17 Jun 1999

  • View in gallery

    Five-day running means of ground heat fluxes during the 2000 growing season

  • View in gallery

    The seasonal energy budget pattern for (top) 1999 and (bottom) 2000. The values presented are 5-day running means of the average daily flux

  • View in gallery

    Average diurnal energy budgets for common months between 1999 and 2000 during the study period

  • View in gallery

    The seasonal pattern of evaporation, surface resistance, and vapor pressure deficit. Values shown are 5-day running means; 1999 values are represented by the heavy line with black boxes, and 2000 values are the light line with white boxes

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 120 115 3
PDF Downloads 18 16 2

The Energy Budget of Canadian Shield Subarctic Terrain and Its Impact on Hillslope Hydrological Processes

View More View Less
  • 1 School of Geography and Geology, McMaster University, Hamilton, Ontario, Canada
© Get Permissions
Full access

Abstract

The objective of the study is to determine the local processes that control the partitioning of the energy budget of shield terrain in the western Canadian subarctic. The magnitude of the spring snowmelt and its potential to flood exposed bedrock portions of the landscape control the energy budget in the early part of the summer. In wet years, Bowen ratios are low and increase over the growing season. The high latent heat fluxes early in the year are promoted by water pooled in bedrock depressions and stored in the shallow soil. The high evaporation rates deplete moisture storage by the end of July after which latent heat fluxes decrease so that Bowen ratios exceed unity until the end of the growing season. This regime differs from other subarctic terrain types with similar vegetation. Exposed and shallow Precambrian bedrock keeps water close to the topographic surface and available for evaporation. The low surface resistance of ponded water on the bedrock surface and high vapor pressure deficits are possible causes for the high evaporation rates in June and July. However, in drier years, when ponded snowmelt water is minimal, evaporation is small and sensible heat dominates the early summer energy budget to a much larger degree than is observed elsewhere in the subarctic. It then becomes a very arid landscape. High evaporation-to-precipitation ratios throughout the summer are an important feature of the western Canadian Shield subarctic region, and this feature has significant hydrological implications. It is normal for a moisture deficit to be created each summer. The magnitude of this moisture deficit is the primary control on hillslope runoff response to precipitation later in the summer and autumn.

Current affiliation: Atmospheric and Hydrological Sciences Division, Environment Canada, Yellowknife, Northwest Territories, Canada

Corresponding author address: Christopher Spence, Northern Section, Atmospheric and Hydrological Sciences Div., Environment Canada, Suite 301, 5204 50th Ave., Yellowknife, NT X1A 1E2, Canada. Email: chris.spence@ec.gc.ca

Abstract

The objective of the study is to determine the local processes that control the partitioning of the energy budget of shield terrain in the western Canadian subarctic. The magnitude of the spring snowmelt and its potential to flood exposed bedrock portions of the landscape control the energy budget in the early part of the summer. In wet years, Bowen ratios are low and increase over the growing season. The high latent heat fluxes early in the year are promoted by water pooled in bedrock depressions and stored in the shallow soil. The high evaporation rates deplete moisture storage by the end of July after which latent heat fluxes decrease so that Bowen ratios exceed unity until the end of the growing season. This regime differs from other subarctic terrain types with similar vegetation. Exposed and shallow Precambrian bedrock keeps water close to the topographic surface and available for evaporation. The low surface resistance of ponded water on the bedrock surface and high vapor pressure deficits are possible causes for the high evaporation rates in June and July. However, in drier years, when ponded snowmelt water is minimal, evaporation is small and sensible heat dominates the early summer energy budget to a much larger degree than is observed elsewhere in the subarctic. It then becomes a very arid landscape. High evaporation-to-precipitation ratios throughout the summer are an important feature of the western Canadian Shield subarctic region, and this feature has significant hydrological implications. It is normal for a moisture deficit to be created each summer. The magnitude of this moisture deficit is the primary control on hillslope runoff response to precipitation later in the summer and autumn.

Current affiliation: Atmospheric and Hydrological Sciences Division, Environment Canada, Yellowknife, Northwest Territories, Canada

Corresponding author address: Christopher Spence, Northern Section, Atmospheric and Hydrological Sciences Div., Environment Canada, Suite 301, 5204 50th Ave., Yellowknife, NT X1A 1E2, Canada. Email: chris.spence@ec.gc.ca

1. Introduction

Energy fluxes from exposed bedrock must play an important role in the overall subarctic Canadian Shield energy budget because rock outcrops occupy up to 30% of surface cover. Such a high percentage of exposed Precambrian bedrock makes the subarctic shield substantially different from other subarctic regions. The objectives of this study are to document the significance of these differences to the growing-season energy budget and to the processes that control the partition of energy over shield terrain and to determine whether shield environments have significantly different energy budgets than nonshield environments. The field research has been pursued in the Yellowknife River basin, which is located in the central Mackenzie River basin of Canada. In this paper, the term “shield” refers to the typical terrain and vegetation of the Canadian Shield in subarctic regions, which is exemplified in Fig. 1.

Previous energy budget studies in the subarctic Canadian Shield have been pursued mainly in eastern parts of the shield in Québec–Labrador, which is a wetter environment. Singh and Taillifer (1986) identified the existence of regional-scale advection on the energy balance of subarctic forest. Lafleur and Adams (1986) examined the effect of the open canopy on the radiation budget. Fitzjarrald and Moore (1994) presented results from Schefferville, Québec, that showed notably high Bowen ratios. In the only other study in the Yellowknife region, Wight (1973) implies that the presence of exposed Precambrian bedrock in the subarctic shield may result in distinctive partitioning of the energy budget over the growing season. Wight's results differed substantially from those found in similar vegetation in the boreal forest and in nonshield subarctic open woodland regions (Sellers et al. 1997; Rouse et al. 1997). Amiro and Wuschke (1987) speculate that exposed bedrock, with its high thermal conductivity, higher net radiation, and little to no latent heat flux, should have a large ground heat flux, but, because exposed bedrock tends to occupy small areas, basinwide values of ground heat flux remain relatively small. Wight (1973) noted that because bedrock experiences little to no evaporative flux it would have a disproportionately large sensible heat flux, some of which must be advected to the surrounding environment.

There is considerable variability in reported Canadian Shield precipitation:runoff ratios ranging from 30% to 69% (Allan and Roulet 1994; Thorne et al. 1999). There have been studies that have identified saturation overland flow as the predominant hillslope runoff process acting in the Canadian Shield (Buttle and Sami 1992; Peters et al. 1995), but there have been no interannual studies that attempt to explain the year-to-year variability in the runoff response at the hillslope scale.

2. Study site

The study site is located 100 km north of Yellowknife, Northwest Territories, Canada, at 63°35.5′N, 113°53.5′W (Fig. 1). The study site is an upland that is the main source area for a small but beautiful headwater lake, unofficially known as Skeeter Lake. Exposed bedrock ridges to the north, east, and west surround a shallow valley 700 m wide by 1200 m long. Most of the 20 m of relief at the site is around the bedrock ridges. The slope averages 2%.

The vegetation is typical of the subarctic shield. The surface cover is not homogenous but includes stands of black spruce (26%), mixed stands of spruce and aspen (26%), a peat wetland (20%), and exposed bedrock (28%). Black spruce (Picea mariana) dominates the arboreal vegetation, but there is a significant portion of trembling aspen (Populus tremuloides). Maximum tree heights are less than 4 m, with an average height of 3 m. The tree canopy is open, with average tree spacing estimated at about 4 m. This spacing allows for a significant understory dominated by dwarf birch (Betula glandulosa), labrador tea (Ledum groenlandicum), and blueberry (Vaccinium augustifolium). Exposed bedrock is scattered throughout, with much of it in the southern and eastern portions of the upland. Lichens (fruticose, foliose, and crustose forms including Cladina spp. and Cladonia spp.) dominate the sparse vegetation cover on exposed bedrock. Soils occupy some shallow (<20 cm) depressions in the exposed bedrock. There are two patches of sandy and cobbley soils derived from glaciofluvial deposits in the southern reaches of the upland. The depth of these deposits is not known, but, given the bedrock slope in the area, it could be up to 10 m. The soils in the wetland include a moss mat (Sphagnum spp.) up to 30 cm deep above organic soils with an estimated average depth of 1 m. There have been no mineral soils observed below the organic soils, which are underlain by granitic bedrock.

The area's climate is characterized by short, cool summers, with a July daily average temperature of 16°C, and long, cold winters, with a January daily average temperature of −29°C. Summer wind speeds average 3.5 m s–1 and dominantly blow from the south (Phillips 1990). The region averages about 280 mm of precipitation, with approximately 55% of that falling as snow. Convective cells produce much of the summer precipitation, and, as such, summer rainfall is highly variable from year to year. As the jet stream moves south over the region in September, conditions are often cool and damp. If this shift begins earlier into August, the annual precipitation will tend to be higher than normal because of the extended period during which rainfall occurs before freeze-up.

Permafrost in this region is classified as widespread and discontinuous (Prowse and Ommanney 1990). Observations suggest that the soils in the wetland completely thaw every year but that permafrost may occur in the two patches of sandy soils because of their greater depth. The annual snow cover begins in October and disappears near the end of April or beginning of May when a quick and intense spring thaw normally occurs (Wedel et al. 1990).

Snow meltwater is routed from the ridges through the upland by two channels on the edges of the wetland. These become poorly defined in places of bedrock outcrops that promote swaths of sheetflow. Near the south side of the upland, the runoff is confined into one outlet, which follows a steep channel exiting into Skeeter Lake (Fig. 1). No runoff occurs from the upland after snowmelt during the summer months, but some may occur in the autumn during wet years.

3. Methods

The energy budget at the terrestrial surface is
QQhQeQg
where Q* is net radiation, Qh is sensible heat, and Qg is the ground heat flux. Latent heat Qe can be estimated by rearranging the energy budget equation to
QeQQhQg
Energy budget components Q*, Qh, and Qg were measured continuously on the tower and integrated for half-hourly periods. Sensible heat flux was measured at 6.5-m height using a robust eddy correlation system similar to that described by Amiro and Wuschke (1987). A 25-μm copper–constantan unshielded single-junction thermocouple was used to measure temperature, and an R. M. Young Company vertical propeller anemometer was used to measure vertical wind speed. Because a propeller anemometer tends to underestimate higher-frequency eddies (Garratt 1975), the corrections of Moore (1986) and Blanford and Gay (1992) were applied to estimates of sensible heat flux, which resulted in a +31% correction to Qh. This correction is of similar magnitude to the necessary corrections in previous studies by Blanford and Gay (1992), Amiro and Wuschke (1987), and Petrone et al. (2000).

Net radiation was measured with a Kipp and Zonen, Inc., NR Lite net radiometer mounted 6 m above the surface on a meteorological tower. Immediately prior to field use, the net radiometer was factory calibrated and is expected to have an accuracy of ± 10%. Given that the sensible heat measurements are a composite value from the tower footprint that changes with wind speed and direction and the net radiometer measures a disk beneath the tower, the two cannot be used together to solve for latent heat flux because they are measuring two different areas. To address this problem, coincident net radiation measurements using another factory calibrated Kipp and Zonen NR Lite were made over each land cover type for 24-h periods. Spatial differences in Q* are very small. The root mean difference between daily values of tower and mobile measurements over the different terrain types was 8.3%. Correlation coefficients between each land cover's net radiation flux and the tower measurement were determined. The hourly footprint of the tower was estimated using the methods described in Schuepp et al. (1990). The proportion of each land cover type along the footprint was determined using a classified Landsat Thematic Mapper image (30-m resolution) of the Yellowknife River basin from the Northwest Territories Centre for Remote Sensing. Using the hourly footprint and the correlations between the main tower and each land cover type, hourly land cover–weighted net radiation values were calculated for a period of 21 weeks from May to October of 1999. Hourly-weighted and tower values were compared, and an empirical equation was derived that was used to determine weighted Q* values beyond the original comparison period.

Ground temperatures were measured half hourly using thermistor strings with thermistors placed at the surface and at depths of 0.05, 0.10, 0.25, 0.5, 0.75, and 1.0 m in both soil-covered and exposed bedrock areas. The bedrock thermistor string was placed in a hole created with an Atlas Copco, Inc., Pionjär rock drill and backfilled with sawdust. Ground heat flux Qg was calculated for each segment within the thermistor string using the Fourier heat flow equation
i1525-7541-3-2-208-e3
and was totaled to determine Qg for the entire 1-m depth. In Eq. (3), K is the thermal conductivity and T is the ground temperature (°C) at the specified depth z (m). The thermal conductivity of the bedrock is treated as constant at 3.4 W (m°C)–1 (Drury and Lewis 1981), and the thermal conductivity of the soil as a function of the proportions of soil, air, water, and ice is given by (Farouki 1981) as
i1525-7541-3-2-208-e4
where f is the fraction of air (a), water (w), ice (i), and soil (s) in the soil column. The water content was determined using time-domain reflectometery (TDR) measurements. Ice content was determined by the difference in soil moisture values immediately prior to freeze up the autumn before and the measured TDR value during the spring. Measurements of soil-covered and exposed bedrock areas were weighted by percentage coverage of each to estimate the ground heat flux within the tower footprint. A measurement error of 30% in the soil moisture measurement results in an estimated error of ± 11% in the ground heat flux. The maximum error of the latent heat flux, calculated as a residual, is ± 21% as the sum of the error in the other energy flux measurements.

A 6-m meteorological tower was employed to measure the atmospheric variables of air temperature, relative humidity, and wind speed (all at 2-m height), as well as soil moisture at three depths (2, 10, and 20 cm) using Campbell Scientific, Inc., TDR sensors. The TDR sensors were calibrated against destructive soil samples, but, with only a 0.70 variance (r2) value between the soil samples and the TDR sensors, there still remained a 30% error in the soil moisture measurements. All the foregoing measurements were taken every 30 s and were averaged over half-hour periods. Destructive soil sampling was used for gravimetric soil moisture measurements, to calibrate the TDR probes, and to determine physical properties of the soil. Wind direction was measured hourly at an Environment Canada remote climate station on an island in Lower Carp Lake, 2 km from the study site. The Environment Canada station is on an unobstructed ridge. Observations suggest that wind directions are rarely different between the island tower and the study site. Rainfall volume across the site was measured using three Meteorological Service of Canada Type-B rain gauges. Rainfall intensity was recorded hourly using a tipping bucket wired to a datalogger (Campbell Scientific CR10X). Daily snowmelt was estimated by measuring the lowering of the snowpack and the snow density at the top of the snowpack. Runoff was measured continuously at a 90° V-notch weir at the outlet of the upland.

To provide some insight as to the processes that control evaporation E at this site, the evaporation rate can be compared with the surface resistance rs . Surface resistance can be calculated using the Penman–Monteith equation (Jarvis and McNaughton 1986):
i1525-7541-3-2-208-e5
where λ is the latent heat of vaporization, Δ is the slope of the saturation pressure temperature curve, ρ is air density, cp is the heat capacity of air, γ is the psychrometric constant, D is the vapor pressure deficit, and ra is the aerodynamic resistance.

4. Results

The study period extended from 1 May to 1 October 1999 and 12 May to 2 August 2000. The period of record is uninterrupted, except for 7–13 June and 21 June–13 July 1999, during which air temperature and relative humidity data were not logged because caribou chewed through those sensors' wires. Figure 2 compares the precipitation and temperature record with normal conditions for the city of Yellowknife for the period of record. May, June, and July precipitation and temperature in this subarctic region are normally 84 mm and 11°C. Rainfall was below normal at 50 and 64 mm in 1999 and 2000, respectively. Average temperatures for the same 3-month period were 9° and 11°C in 1999 and 2000, respectively. Prolonged dry periods at the beginning of both summers resulted in a decline in soil moisture that slowed at the end of July (Fig. 3). Rain at the end of the summer of 1999 and the snowmelt of 2000 replenished some of the water lost, but over the entire study period there was a net soil moisture loss of 73 mm.

In 1999, snow meltwater flooded into the bedrock depressions because there was little available soil storage at the site after a wet autumn in 1998. Despite a similar volume of snowmelt in 2000 (Fig. 4), much of the water went to replenishing soil moisture storage (Fig. 3b), reducing the fraction of snow meltwater that ran off from 0.76 in 1999 to 0.58 in 2000. The ponding on the bedrock associated with the larger runoff in 1999 was subdued in 2000, so much of the exposed bedrock began the summer dry.

If it is assumed that incoming long- and shortwave radiation are equal over the upland, the differences in Q* between land cover types will be due to differences in albedo and surface temperature (Fig. 5). Petzold and Rencz (1975) report that exposed bedrock and wetlands have similar albedos at 0.09 and 0.109 and that subarctic mixed vegetation has an albedo of 0.214. The lower albedos of the bedrock and wetlands result in higher net radiation during daylight hours (Fig. 5). On 17 June 1999, when the measurements depicted in Fig. 5 were made, nighttime surface temperatures of the exposed bedrock and soil cover were 12° and 6°C, respectively. The higher bedrock surface temperature resulted in higher outgoing longwave radiation and, in turn, lower nighttime Q*.

The land cover distribution along the tower footprint changes with wind direction (Table 1). If Wight (1973) is correct in his assumption that exposed bedrock has an enhanced sensible heat flux, one would expect Bowen and Qh/Q* ratios to be higher when winds blow from the west, where bedrock dominates the land cover. Average hourly Bowen ratios of 1.1 on 5 and 6 August 1999 when winds came from the west were not significantly different than those of 1.3 on 1 and 2 August 1999 when winds came from the southwest. Average Qh/Q* ratios were almost identical between the two periods at 0.53 and 0.55, respectively. The similarity between the turbulent fluxes may, however, be a function of the dryness of the entire landscape at the beginning of August of 1999 (Fig. 3b). Immediately after the 2000 freshet, exposed bedrock surfaces were dry and soil-covered zones were wet. At this time, Bowen ratios and Qh/Q* responded to changes in wind direction. The daily Bowen and Qh/Q* ratios on 16 June 2000 of 1.93 and 0.56, respectively, when the wind came from the west, declined on 17 June 2000 to 1.2 and 0.44, respectively, when the wind shifted to the southwest. These results suggest there are subtle changes in the sensible heat flux sensed by the tower with changes in the proportion of exposed bedrock in the tower footprint. There were much more significant differences in ground heat flux between soil and exposed bedrock (Fig. 6).

Average daily Q* and Qe increased from 1 May to a maximum on 20 June (Fig. 7). After the summer solstice, Q* declined gradually until the end of September at which time it approached zero. Latent heat declined after the summer solstice to the end of July, remained steady through the month of August, and declined further in September because of low radiation input and cool and wet conditions. Sensible heat fluxes remained relatively steady throughout the season and exceeded latent heat from mid-July to the end of September. Positive Qg was significant early in the summer but was small once the ground had warmed by the end of June. Negative Qg in September augmented Qh so that it equaled or exceeded Q* through the second half of the month. Although Q* in 2000 had a similar pattern to that observed in 1999, Qe remained low until the end of the study period (Fig. 7). Also, Qh was much larger and Qg slightly larger than in 1999.

Net radiation peaks near midday (1130–1400 MDT) during every month in the study period (Fig. 8). There is a slight delay until the sensible heat peaks in the afternoon (1400 MDT). Latent heat rises quickly in the morning to a maximum from 1000 to noon MDT and then recedes slowly through sunset. The ground warms during the day (1200–1500 MDT) and releases a small amount of heat overnight.

Net radiation was larger in May of 2000 than in May of 1999, but the situation was reversed in June. Net radiation was similar in July of both years (Fig. 8). Monthly Bowen, Qe/Q*, and Qh/Q* ratios differed significantly between 1999 and 2000 (Table 2). Bowen ratios were less than unity through the early summer of 1999 and increased to the end of September. Monthly average Bowen ratios in 2000 were consistently greater than unity. May and June Qe/Q* ratios also suggest that more energy was directed to evaporation in 1999 than in 2000. May Qg/Q* values were similar in 1999 and 2000, but June and July values were larger in 2000.

5. Interannual differences

In 1999, after all of the snowpack except for late-lying drifts had melted, meltwater remained flowing through the upland until mid-June (Fig. 4). As water levels receded, some areas became detached from the runoff, which left much water ponded in depressions created by the bedrock microtopography. As a result, initial evapotranspiration rates were high and there was a general decrease in rates and increase in Bowen ratios during the long dry period beginning after the summer solstice in June and persisting through July. At the beginning of this period, surface resistances were lower than those observed by Lafleur (1992) and vapor pressure deficits were high (Fig. 9) so that evaporation rates exceeded those observed at other subarctic sites (Jarvis et al. 1997; Lafleur 1992; Wright 1981). It was not until the bedrock surfaces dried that a loss in soil moisture was observed (Fig. 3). As the soil dried, latent heat flux decreased while vapor pressure deficits remained high but surface resistance increased, suggesting there was some physiological response by the vegetation or a physical barrier in the soil that prevented a loss of moisture. Lafleur (1992) observed that much of the evapotranspiration from subarctic vegetation originates at the forest floor. Sphagnum peat does not have a large degree of physiological control on evaporation, so it is expected there was a physical control on evaporation from the lichen and moss cover. Bello and Arama (1989) and Lafleur and Schreader (1994) have observed high evaporation rates over wet lichen and moss, and Kershaw and Rouse (1971) note that dry lichen and moss cover is an effective mulch in preventing evapotranspiration from the soil.

In 2000, the snowpack was almost gone before a mid-May snowstorm (Table 2). This snowstorm added an additional 25.5 mm of snow water equivalent, which melted in two days. Even with the extra snow, hillslope runoff ceased early and there was less ponding and fewer soil-covered areas filled with water, so the site began the summer much drier than in 1999. Despite high vapor pressure deficits, surface resistance was high because of the dry conditions, so evapotranspiration rates in the early summer were low (Fig. 9). These rates increased and Bowen ratios decreased after considerable rainfall in late June.

For the growing season, Qe/Q* values and evapotranspiration rates in 1999 were much higher than in 2000 and indeed higher than for most other subarctic terrain types (Table 3; Lafleur 1992; Jarvis et al. 1997; Lafleur et al. 1997; Wright 1981). Low β and high Qe/Q* in 1999 at first suggest the Skeeter Lake upland is a relatively wet subarctic environment, but high β and low Qe/Q* in 2000 imply the site is a more arid environment than has been observed previously in subarctic Canada. The differences between 1999 and 2000 were substantially greater than annual differences found by Lafleur (1992) in wetland open canopy black spruce tamarack forest, for comparable precipitation amounts. Most of the studies of subarctic energy budgets have been performed in wetter climates [e.g., Schefferville, Quebec (Wright 1981; Fitzjarrald and Moore 1994)], or in terrain with deeper soils [e.g., Churchill, Manitoba (Lafleur et al., 1992)], or a combination of both [e.g., Thompson, Manitoba (Lafleur et al. 1997)]. This provides a moisture buffer in most subarctic environments that is not present in the northwestern subarctic Canadian Shield. The lack of a buffer creates a situation in which the landscape has the potential for much larger energy budget interannual variability than do other subarctic landscapes.

6. Site hydrological behavior and energy-budget feedbacks

The magnitude of the spring freshet controls the summer energy budget. When the snowmelt magnitude is large, as in 1999, the subarctic shield landscape is prone to flooding in flat-lying areas such as the Skeeter Lake study site. Soil depths are shallow, and much water can spill onto surrounding exposed bedrock. When conditions are as they were in June of 1999, latent heat fluxes are large because much of the net radiation goes to evaporating the water ponded on the exposed bedrock. The 2000 snowmelt was more subdued and created conditions in which the bedrock and soils remained dry and warmed more, augmenting ground and sensible heat fluxes. The low storage capacity of the soil cover does not allow much water to remain in the landscape from year to year, so that the early-summer energy budget is highly variable and depends on the magnitude of the snowmelt.

It is expected that a storage deficit will develop every year on subarctic shield slopes. In Schefferville for which the climate is much wetter, Moore et al. (1994) observed a summer storage deficit. Summer rainfall in the Yellowknife region is recirculated water from convective cells (Szeto 1998), so a positive feedback system may exist in which wet conditions (i.e., heavy ponding) enhance summer storm activity, which reduces summer storage deficits. The effect of climate variability will be to determine the magnitude of the deficit. Wet years may see a small deficit, which will increase the likelihood of runoff events in the autumn. During a normal year, the storage deficit may prevent a runoff response from autumn rains. The large moisture deficit in a dry year may have implications for the following snowmelt, because meltwater will be directed to replenishing storage, lowering the size of the spring freshet. Spence (2000) illustrated that headwater lake storage deficits are important in controlling the magnitude of the spring freshet at the basin scale; the results from this study imply that storage deficits are important at the hillslope scale.

7. Conclusions

The objective of this study was to measure the growing-season energy budget and to determine the local processes that control the partition of energy over a subarctic shield land surface. There are important feedbacks between site hydrological processes and energy budgets in the subarctic Canadian Shield. The growing-season energy budget differs from that found in other investigations over subarctic vegetation in its large interannual variability due to its susceptibility to the volume of the spring freshet. After a large snowmelt when water is plentiful, Bowen ratios are low and evapotranspiration rates are high at the beginning of the summer, but the situation is reversed after a smaller snowmelt event that does not flood exposed bedrock. Under these conditions, the surface energy balance indicates greater aridity than for any other subarctic surfaces that have been studied and documented in the research literature. High latent heat fluxes and low soil storage capacity consistently create a moisture deficit that controls hillslope runoff response through the summer and autumn.

Acknowledgments

The authors thank Jennifer Dougherty, Shawne Kokelj, Claire Oswald, Mark Dahl, and Andrea Czarnecki of, or formerly of, Atmospheric and Hydrologic Sciences, Environment Canada; Bob Reid of Water Resources Division, Indian and Northern Affairs Canada; and Iain Stewart for their assistance in the field. Thanks to Cindy Taylor of the NWT Centre for Remote Sensing for providing the LANDSAT TM land cover data. This paper was improved by the comments of three reviewers. This work was funded by Environment Canada and the Mackenzie GEWEX Study. Special thanks go to Shauna Sigurdson and Jesse Jasper of Environment Canada for supporting this research. The authors would be happy to be involved in any collaborative work that would compare these data with other sites or that would use these data for comparison with and/or evaluation of atmospheric model output. The authors can be contacted at the address provided at the beginning of the paper.

REFERENCES

  • Allan, C. J., and Roulet N. , 1994: Runoff generation in zero order Precambrian shield catchments: The stormflow response of a heterogeneous landscape. Hydrol. Processes, 8 , 369388.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Amiro, B. D., and Wushcke E. E. , 1987: Evapotranspiration from a boreal forest drainage basin using an energy balance/eddy correlation technique. Bound.-Layer Meteor., 38 , 125139.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bello, R. L., and Arama A. , 1999: Rainfall interception in lichen canopies. Climatol. Bull., 23 , 7478.

  • Blanford, J. H., and Gay L. W. , 1992: Tests of a robust eddy correlation system for sensible heat flux. Theor. Appl. Climatol., 46 , 5360.

  • Buttle, J. M., and Sami K. , 1992: Testing the groundwater ridging hypothesis of streamflow generation during snowmelt in a forested catchment. J. Hydrol., 135 , 5372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Drury, M., and Lewis T. , 1981: Thermal studies in the Lac du Bonnet batholith at Pinawa, Manitoba. Atomic Energy of Canada Limited Tech. Rep. II 5000, 14 pp.

    • Search Google Scholar
    • Export Citation
  • Farouki, O., 1981: Thermal Properties of Soils. CRREL Monogr., No. 81, Cold Regions Research and Engineering Laboratory, United States Army Corps of Engineers, 136 pp.

    • Search Google Scholar
    • Export Citation
  • Fitzjarrald, D. R., and Moore K. E. , 1994: Growing season boundary layer climate and surface exchanges in a subarctic lichen woodland. J. Geophys. Res., 99 , 18991917.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garratt, J. R., 1975: Limitations of the eddy correlation technique for the determination of turbulent fluxes near the surface. Bound.-Layer Meteor., 8 , 255259.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jarvis, P. G., and McNaughton K. G. , 1986: Stomatal control of transpiration. Adv. Ecol. Res., 15 , 149.

  • Jarvis, P. G., Massheder J. M. , Hale S. E. , Moncrieff J. B. , Rayment M. , and Scott S. L. , 1997: Seasonal variation of carbon dioxide, water vapor, and energy exchanges of a boreal black spruce forest. J. Geophys., Res., 102 , 2895328966.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kershaw, G. P., and Rouse W. R. , 1971: Studies on lichen dominated systems. I: The water relations of Cladonia alpestris in spruce lichen woodland in northern Ontario. Can. J. Bot, 49 , 13891399.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lafleur, P. M., 1992: Energy balance and evapotranspiration from a subarctic forest. Agric. For. Meteor., 58 , 163175.

  • Lafleur, P. M., and Adams P. , 1986: The radiation budget of a subarctic woodland canopy. Arctic, 39 , 172176.

  • Lafleur, P. M., and Schreader C. P. , 1994: Water loss from the floor of a subarctic forest. Arct. Alpine Res., 26 , 152158.

  • Lafleur, P. M., Rouse W. R. , and Carlson D. W. , 1992: Energy balance differences and hydrologic impacts across the northern treeline. Int. J. Climatol., 12 , 193203.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lafleur, P. M., McCaughey J. H. , Joiner D. W. , Bartlett P. A. , and Jelinski D. E. , 1997: Seasonal trends in energy, water and carbon dioxide at a northern boreal wetland. J. Geophys. Res., 102 ((D24),) 2900929020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McCaughey, J. H., Lafleur P. M. , Joiner D. W. , Bartlet P. A. , Costello A. M. , Jelinski D. E. , and Ryan M. G. , 1997: Magnitudes and seasonal patterns of energy, water and carbon exchanges at a boreal young jack pine forest in the BOREAS northern study area. J. Geophys. Res., 102 , 2899729007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moore, C. J., 1986: Frequency response corrections for eddy correlation systems. Bound.-Layer Meteor., 37 , 1735.

  • Moore, K. E., Fitzjarrald D. R. , Wofsy S. C. , Daube B. C. , Munger J. W. , Bakwin P. S. , and Crill P. , 1994: A season of heat, water vapor, total hydrocarbon, and ozone fluxes at a subarctic fen. J. Geophys. Res., 99 , 19371952.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moore, K. E., Fitzjarrald D. R. , Sakai R. K. , and Freedman J. M. , 2000: Growing season water balance at a boreal jack pine forest. Water Resour. Res., 36 , 483493.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peters, D. L., Buttle J. M. , Taylor C. H. , and LaZerte B. D. , 1995: Runoff production in a forested shallow soil Canadian Shield basin. Water Resour. Res., 31 , 12911304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Petrone, R. M., Rouse W. R. , and Marsh P. , 2000: Comparative surface energy budgets in western and central subarctic regions of Canada. Int. J. Climatol., 20 , 11311148.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Petzold, D. E., and Rencz A. N. , 1975: The albedo of selected subarctic surfaces. Arct. Alpine Res., 13 , 287293.

  • Phillips, D., 1990: The Climates of Canada. Environment Canada, 176 pp.

  • Rouse, W. R., and Bello R. L. , 1985: Impact of Hudson Bay on the energy balance in the Hudson Bay Lowlands and the potential for climate modification. Atmos.–Ocean, 23 , 375392.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rouse, W. R., Bello R. L. , and Lafleur P. , 1997: The low arctic and subarctic. The Surface Climates of Canada, W. Bailey, T. Oke, and W. R. Rouse, Eds., McGill Queen's University Press, 198–221.

    • Search Google Scholar
    • Export Citation
  • Schuepp, P. H., Leclerc M. Y. , Macpherson J. I. , and Desjardins R. L. , 1990: Footprint prediction of scalar fluxes from analytical solutions of the diffusion equation. Bound.-Layer Meteor., 50 , 355373.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and Coauthors. 1997: BOREAS in 1997: Experiment overview, scientific results, and future directions. J. Geophys. Res., 102 , 2873128769.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Singh, B., and Taillifer R. , 1986: The effect of synoptic scale advection on the performance of the Priestley–Taylor evaporation formula. Bound.-Layer Meteor., 36 , 367382.

    • Search Google Scholar
    • Export Citation
  • Spence, C., 2000: The effect of storage on runoff from a headwater subarctic shield basin. Arctic, 53 , 237247.

  • Szeto, K., 1998: Low level inversions, precipitation recycling and air-land interactions over the Mackenzie Basin. Proc. Fourth Scientific Workshop for the Mackenzie GEWEX Study (MAGS), Environ. Can., 27–28.

    • Search Google Scholar
    • Export Citation
  • Thorne, G. A., Laporte J. M. , Clarke D. , and Davison C. C. , 1999: Water budget of an upland outcrop recharge area in granitic rock terrane of southeastern Manitoba, Canada. Proc. 12th Int. Northern Research Basins Symposium and Workshop, Reykjavik, Iceland, Working Group on Northern Research Basins, 317–330.

    • Search Google Scholar
    • Export Citation
  • Wedel, J. H., Smart A. , and Squires P. , 1990: An Overview Study of the Yellowknife River Basin, NWT. Inland Waters Directorate, Environment Canada, 82 pp.

    • Search Google Scholar
    • Export Citation
  • Wight, J., 1973: Aspects of evaporation and evapotranspiration in the water balance of Baker Creek, near Yellowknife, NWT,. M.S. thesis, Dept. of Geography, University of Alberta, 329 pp.

    • Search Google Scholar
    • Export Citation
  • Wright, R. K., 1981: The water balance of a lichen tundra underlain by permafrost. McGill Subarctic Research Paper 33, Centre for Northern Studies, McGill University, 110 pp.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Photo of Skeeter Lake upland with a location map. The arrow points to the climate tower. The lightest gray tones are exposed bedrock (B). The wetland (PW) is located behind the tower. Black spruce (BS) and mixed stands (MW) are scattered throughout the site. White arrows denote direction of surface water flow. Predominant winds come from the northeast, which is the upper-right-hand corner of the picture

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 2.
Fig. 2.

Monthly (top) air temperature and (bottom) precipitation. The white circles represent observations made in 1999 at Skeeter Lake, and the white boxes are for 2000 observations. The black squares are Environment Canada 1961–90 averages for Yellowknife, with the bars representing ± 1 std dev

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 3.
Fig. 3.

(a) Cumulative precipitation and (b) cumulative change in soil moisture storage from the beginning of the study period on the Skeeter Lake upland during 1999 and 2000

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 4.
Fig. 4.

Cumulative spring snowmelt and runoff at Skeeter Lake in 1999 and 2000

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 5.
Fig. 5.

Half-hourly values of Q* measured over exposed bedrock and two soil-covered land cover types on 17 Jun 1999

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 6.
Fig. 6.

Five-day running means of ground heat fluxes during the 2000 growing season

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 7.
Fig. 7.

The seasonal energy budget pattern for (top) 1999 and (bottom) 2000. The values presented are 5-day running means of the average daily flux

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 8.
Fig. 8.

Average diurnal energy budgets for common months between 1999 and 2000 during the study period

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Fig. 9.
Fig. 9.

The seasonal pattern of evaporation, surface resistance, and vapor pressure deficit. Values shown are 5-day running means; 1999 values are represented by the heavy line with black boxes, and 2000 values are the light line with white boxes

Citation: Journal of Hydrometeorology 3, 2; 10.1175/1525-7541(2002)003<0208:TEBOCS>2.0.CO;2

Table 1.

Proportions of land cover along the tower footprint for cardinal compass directions. All values are in percentages

Table 1.
Table 2.

Monthly energy budget ratios, average Bowen ratios, evaporation rates, and precipitation

Table 2.
Table 3.

Average growing season β and Qe/Q* results from the Canadian Shield and subarctic locations. Table lists wettest locales and conditions first, then drier examples with each subsequent row down the table

Table 3.
Save