Abstract

The magnitude and timing of water delivery in two large northern basins are analysed to clarify where runoff is generated and how their rivers acquire comparable regimes (or seasonal rhythms) of flow. These two rivers, the Mackenzie in Canada and the Yenisei in Russia, traverse similar latitudes, physiographic provinces, vegetation zones and climatic regions. Within the basins, mountainous terrain and high-precipitation sections usually yield large runoff, but low runoff comes from the plains, low plateaus and areas of aridity. Winter runoff is commonly low and snowmelt is responsible for annual peak runoff in most parts of these basins, though rainfall is a prominent runoff source in southern Yenisei. Many rivers in the drainage networks display a seasonal pattern that suggests the dominance of snowmelt to produce a spring freshet followed by a general decline in summer that diminishes to winter low flows. Regulation of reservoir outflow greatly distorts the natural flow regime. Yet, along the main river downstream of the reservoirs, the influx of tributary discharge can dilute such human influence. To truly understand how water is produced and transferred in large northern rivers, the spatial and temporal complexity of flow-generation mechanisms and storage effects need to be unravelled.

INTRODUCTION

Several mega-basins in North America and Eurasia discharge directly or indirectly northward, bringing substantial amounts of continental water to freshen the seawater of the Arctic Ocean. The importance of this river water in affecting the temperature and salinity of the surface ocean-water layer, sea ice growth and decay, and thermohaline circulation of the ocean have been discussed in the literature (e.g. Aagaard & Carmack 1989; Peterson et al. 2002; Carmack et al. 2016; Park et al. 2017). Since these northern rivers of sub-continental scale traverse multiple latitudinal and altitudinal zones, the magnitude and timing of water delivery vary greatly within and without the river system (McClelland et al. 2006; Overeem & Syvitski 2010; Rood et al. 2017). To know how large rivers attain their discharge magnitude and how they acquire their regime of flow, it is necessary to analyse the contribution of runoff from various parts of the basin and the modification of streamflow regime along the drainage network.

Two mega-basins are chosen for this comparative study of their hydrological behaviour. These are the Mackenzie River Basin in Canada and the Yenisei River Basin in Russia (abbreviated as MRB and YRB in the rest of this paper). Together, their rivers provide a substantial amount of fresh water that directly enters the Arctic Ocean (Aagaard & Carmack 1989). They are selected as northern basins in two separate continents, yet both with comparable ranges in climate, topography and vegetation, and with many areas underlain by permafrost (McClelland et al. 2004; Serreze et al. 2006; Muskett & Romanovsky 2009). Using these basins as examples, this study considers where and when runoff is produced and relates it to the setting of various hydro-physiographical regions. We examine the amount and seasonality of water delivery from the drainage networks and analyse how the rhythm of river flow is modified as water moves downstream. Since the two selected basins have physical settings shared by other major northern basins, information obtained from this investigation is pertinent to others that discharge to the polar sea.

DATA SOURCES

This paper utilizes discharge data available from public domain websites. Discharge data were extracted from the Environment and Climate Change Canada Historical Hydrometric Data web site (HYDAT, https://wateroffice.ec.gc.ca/mainmenu/historical_data_index_e.html) for the MRB and R-ArcticNet version 4.0 (Lammers et al. 2001, updated, http://www.r-arcticnet.sr.unh.edu/v4.0/index.html) for YRB. Climate data were obtained from the Meteorological Service of Canada (http://climate.weather.gc.ca/historical_data/search_historic_data_e.html) and from the Daily Temperature and Precipitation Data for 518 Russian Meteorological Stations available from the Carbon Dioxide Information Analysis Centre (Bulygina & Razuvaev 2012). The contributing drainage areas above each station were delineated manually using sub-basins obtained from the HydroBASINS and HydroSHEDS projects (Lehner & Grill 2013). The spatial precipitation plots were generated using the Global Precipitation Climatology Project Version 2.3 Combined Precipitation Dataset Data, which combines observations and satellite precipitation data (Adler et al. 2003).

For clarification, we consider discharge as the rate that water passes a particular point of a river (in m3 s−1). Discharge is used to calculate runoff (depth in mm) and flow (volume in km3). The terms runoff and flow are conventionally considered to be synonymous. For this study, however, runoff is regarded as the depth of water produced from a drainage area during a particular time interval (in mm), and flow is the amount of water delivered by a river over a particular time period (in km3). This distinction is to take into consideration the effect of storage along river channels that modifies the amount and timing of water arriving from the land (i.e. runoff) before being released downstream through river flow. Flow contribution is computed from a sub-basin or from an area between two gauging stations (the inter-station area).

HYDROLOGICAL SETTING

Physical environment

The two basins cover large areas: 1.8 million km2 for MRB and 2.55 million km2 for YRB, spanning 45 °N to 70 °N. Figure 1 shows the positions of these basins in the circumpolar setting. The Mackenzie and the Yenisei rivers have built up deltas at their mouths, with the rivers branching into many distributaries that drain extensive wetlands. In this study, we deal with areas upstream of the lowest hydrometric stations that collect most of the flows in MRB (at the village of Arctic Red River) and in YRB (at the town of Igarka). In so doing, we exclude the flows in the deltas and from some mountain areas in the far north. The major hydro-physiographical provinces in these basins are shown in Figure 2.

Figure 1

Topography of Mackenzie River Basin and Yenisei River Basin (top), physiographic regions of circumpolar areas with outlines of the study basins (centre), and distribution of continuous and discontinuous permafrost (bottom).

Figure 1

Topography of Mackenzie River Basin and Yenisei River Basin (top), physiographic regions of circumpolar areas with outlines of the study basins (centre), and distribution of continuous and discontinuous permafrost (bottom).

Figure 2

Mackenzie and Yenisei River Basins showing their hydro-physiographic regions, rivers and lakes. Also shown are the hydrometric stations mentioned in this study.

Figure 2

Mackenzie and Yenisei River Basins showing their hydro-physiographic regions, rivers and lakes. Also shown are the hydrometric stations mentioned in this study.

Inland of the Mackenzie Delta, the topographical alignment of MRB is predominantly north–south. Western MRB comprises the Western Cordilleras, with sub-parallel ranges that reach over 2,500 m, with some small ice caps and valley glaciers at high elevations. In the middle are the Interior Plains with subdued relief, rising gradually from the Mackenzie Delta to about 400 m at the southern end of MRB. Small lakes abound, occupying hollows left by Pleistocene glaciation and former fluvio-lacustrine processes, or developed as thermokarst lakes in permafrost terrain. In the east lies the rolling terrain of the Canadian Shield, with outcrops of Precambrian rocks forming rounded hills, and soil-filled valleys that often contain wetlands and small lakes (Spence & Woo 2003). These valleys follow major lines of bedrock structural weakness and the resulting drainage network is often contorted. The areas that contribute to Canadian Shield river flow depend on how well the network is connected. As most basins consist of chains of lakes and rivers, where the level of a major lake drops below its outlet elevation during periods of drought, cessation of its outflow would disrupt connectivity within the network.

In contrast to the north–south alignment of the MRB, the mountains and high plateau of YRB have a preferred east–west orientation. Inland from the coastal zone is the Central Siberian Plateau, which can be distinguished into an upland and a low plateau section. The Putorana Mountains are part of the upland that rises sharply from the Yenisei River, with its valley stretching to the plains in the Ob River Basin to the west. The mountain tops are relatively flat with an average elevation of about 700 m, and the northwestern edge is even higher, reaching 1,500 m. Southwest of the Putorana Mountains, the Yenisei valley narrows as it intersects the Yenisei Ridge, which drops from heights of 600–800 m northeastward to the lower Siberian Plateau. The elevation and position of Putorana Mountains and the Yenisei Ridge have effects on their hydrological behaviour in that they provide orographic uplift for the oncoming wind to deposit precipitation. They are therefore treated as parts of the upland. South of the Plateau are mountain chains and basins, the highest being the Sayan Mountains and the Tannu Ola in the west, reaching 2,500 m. Topographic depressions enclosed by the mountain ridges often exceed 500–600 m at their lowest elevation. These depressions and many mountain slopes in southern YRB (in Mongolia and Russia) are predominantly arid areas that would produce low runoff.

As both basins extend from cold temperate to the Arctic domain, they have comparable latitudinal ranges in vegetation. The southern end of YRB is situated in Mongolia, mostly being semi-arid grassland and shrubland. The southern tip of MRB is mainly grassland, with agricultural practices at limited locations. North of the grasslands lies the boreal and subarctic forests, which are similar to the taiga of YRB, but large tracts of these MRB forest zones are occupied by wetlands. Beyond the forests of both basins lies the tundra. In mountainous and upland areas, elevation gives rise to vertical zonation, with forests at the lower elevations and alpine tundra and rock outcrops exposed above the alpine treeline. The alpine treeline drops to lower elevations at the higher latitudes.

Continuous and discontinuous permafrost together occupy about half of MRB and YRB, mainly in the north and at high altitudes (Figure 1) although sporadic permafrost can be found further south. The southern limit of continuous permafrost in MRB corresponds roughly with its northern treeline, but not so for YRB where continuous permafrost extends well within the taiga. In small catchments, the presence of permafrost gives rise to quick and sharp response to snowmelt and rainfall inputs (Slaughter et al. 1983), but the influence of permafrost on discharge in large basins is not well quantified as yet (McClelland et al. 2004).

Atmospheric moisture and precipitation

Atmospheric moisture that yields precipitation in the two basins is either advected from outside or originates from within. For MRB, the westerly airflow acquires high moisture content from the Pacific Ocean, but this airflow is obstructed by the lofty Cordilleras and sheds much of its moisture, giving rise to relatively dry conditions to the eastern footslopes. However, cyclogenesis can occur on the leeward side of the Cordilleras to enrich local precipitation (Szeto et al. 2008a). In summer, westerly flows are subdued and air masses originating from the Gulf of Mexico, and sometimes from California, introduce wet conditions to MRB (Liu et al. 2008). In the fall, cyclones travelling with the returning Arctic front bring in moisture to raise autumnal precipitation. For YRB, winter depressions that originate from the Atlantic are rejuvenated by crossing the open water of the Barents Sea and travel along the Arctic front to bring moisture to its northern sector (Shahgedanova 2003). Further south, prolonged radiation loss and extreme coldness intensify high pressure over southern Central Siberia and Kazakhstan, and anticyclonic flows deflect the depressions to leave most of YRB dry and cold. Weakening of the high pressure centre in the spring allows the polar front to bring moisture to the southern fringe of YRB. However, the high mountains hinder the moisture of tropical origin from reaching other parts of the basin. For both MRB and YRB, interaction of airflow with topography strongly influences the distribution of precipitation.

In addition to moisture derived from external sources, recycling is a mechanism that supplies moisture to the atmosphere in the warm season (Szeto et al. 2008b). Evaporation from lakes and wetlands in the basins provides much of the recycled moisture. Recycling is enhanced by mountain-plains circulation. In this process, heating over mountain slopes induces advection of moisture from the plains to the foothills to support convective circulation. Both MRB and YRB receive 20–25% of their precipitation from recycling (Szeto et al. 2008b).

Precipitation is related to atmospheric moisture in the basins. The precipitation patterns of MRB and YRB are shown in Figure 3. Rainfall constitutes a large portion of annual precipitation while snow represents about 25% of total precipitation at most places. Winter (Nov–Apr) is the season of low precipitation for both basins. However, the snow seldom melts in the winter and the snow-cover duration increases northward and with elevation (R. Brown, cited as Figure 4.1 in Woo 2012). MRB has snow on the ground for 150 days in the southern plains and 270 days at its Arctic sea coast. The northern half of YRB has over 200 snow-covered days and the duration reaches nine months at the Arctic coast, while the snow duration in the south varies considerably with topography, lasting from 50 to more than 100 days. Spatially, precipitation is particularly variable in mountain regions. In general, higher amounts fall on high altitudes and large differences exist between slopes. The Sayan Mountains and Tannu-Ola ridge of YRB exhibit pronounced contrasts in precipitation between windward and leeward slopes. Topographic depressions and deep valleys are precipitation shadow areas, many of which are semi-arid. Cyclones from the Atlantic bring high precipitation to the Yenisei Mountains and Putorana Plateau in northern YRB, yielding precipitation to the western parts of the plateaus, but decreasing eastward in the adjacent Lena River basin (Onuchin & Burenina 2010). Similarly, considerable quantities of snow and rain are frequently deposited on high altitudes of the Cordilleras of MRB, sufficient at places to nourish some glaciers; but precipitation diminishes greatly in the precipitation shadow areas at the eastern foothills downwind of the mountain chains (Hydrological Atlas of Canada 1978). The least amount of precipitation reaches the lowlands and low plateaus.

Figure 3

Annual precipitation of MRB and YRB, with hyetographs for selected stations to show contrasts in magnitude and timing of monthly precipitation in different parts of the basins.

Figure 3

Annual precipitation of MRB and YRB, with hyetographs for selected stations to show contrasts in magnitude and timing of monthly precipitation in different parts of the basins.

Figure 4

(LEFT): Runoff of Mackenzie River Basin.

Figure 4

(LEFT): Runoff of Mackenzie River Basin.

Storage

Storage modifies the timing of river flow response to water inputs (in liquid and solid phases). Situated at continental high latitudes, the storage of snow and ice is an important consideration. Snowfall accumulates in winter to build up a seasonal snow cover that undergoes re-distribution, mainly through drifting by wind and interception by vegetation. For both basins, snowmelt and associated runoff is a major event in the spring, and streamflow is often accompanied by the breakup of river ice to generate annual peak flows.

Rivers and lakes usually acquire an ice cover in the winter. Not only does the ice represent one form of water storage, but the presence of ice in channels retards river flow to result in hydraulic storage (Prowse & Carter 2002). This feature, commonly revealed as an abrupt drop in the winter hydrograph, is insignificant in magnitude compared with the storage of winter snow. In the thaw season, wetlands provide summer storage through their numerous hollows that retain water and their vegetation that impedes water movement. Wetlands are extensively developed on the Interior Plains of MRB but they are more localized in YRB.

The rates of groundwater recharge and storage in large northern basins are poorly known, although the mechanisms are understood. Groundwater sustains base flow of large rivers such as those cited in this paper, but most smaller ones would cease to flow. It can produce icing (or ‘naled’) that is another form of seasonal storage in river channels. Karst terrains in limestone areas are especially favourable for groundwater discharge, sometimes maintaining perennial springs to feed the river system. The presence of permafrost restricts or confines groundwater storage to the seasonally thawed active layer and to the intra-permafrost and sub-permafrost ‘taliks’ (unfrozen zones in permafrost area), while the imperviousness of frozen materials enhances near-surface runoff. Ground ice in permafrost is a long-term storage but a warming climate and disturbance resulting from geomorphic processes or human activities lead to permafrost thaw and the release of water from ground ice storage to affect river flow (e.g. Walvoord & Striegl 2007; St. Jacques & Sauchyn 2009).

Lakes and wetlands are particularly effective for surface water storage. Water influx is first retained and then released gradually as relatively more uniform and delayed outflow. In addition to numerous ponds and small lakes, several of the world's largest lakes are found in YRB (Lake Baikal) and in MRB (Lakes Athabasca, Great Slave and Great Bear). Sizeable human-made lakes (reservoirs) have also been created in both basins and they are a significant component of the hydrological landscape (Peters & Prowse 2001; Yang et al. 2004; Jaguś et al. 2015). YRB has a longer history and larger number of reservoirs than MRB (Adam et al. 2007).

RUNOFF GENERATION

Runoff generated from the drainage areas is controlled by water gains from precipitation and losses to evapotranspiration, while land storage (through surface detention and groundwater exchange) governs the timing of water release to the rivers. Figure 4 presents the pattern of runoff in MRB for the period 1973–2015, mapped using data from sub-basins with areas of <100,000 km2 to provide fine spatial resolution. The discharge for sub-basins of such size is not easily obtainable for YRB and therefore runoff is calculated using the discharge of headwater catchments or by considering the difference in discharge for the inter-station area between two adjacent stations along a river (Figure 5). Stations that lie immediately below reservoirs are excluded because their flows strongly reflect the influence of storage rather than runoff contribution from the land area of the basin. We used two different periods (1971–2006 for Angara and 1981–2005 for upper Yenisei) that have more data available to us. Discrepancies in dates are expected to affect the mean runoff values, yet the averages thus obtained give a general picture of runoff contribution from various parts of the basin.

Figure 5

(RIGHT): Runoff of Yenisei River Basin. Monthly runoff from inter-station areas strongly affected by reservoirs is not shown.

Figure 5

(RIGHT): Runoff of Yenisei River Basin. Monthly runoff from inter-station areas strongly affected by reservoirs is not shown.

Annually, the western mountains of MRB with high precipitation consistently produce the highest runoff among all regions, generally >400 mm. Within this region, the southern zone usually has the largest runoff; it exceeds 500 mm, supported by rainfall, snowmelt and glacier melt at high altitudes. Runoff diminishes in the foothill areas, dropping to the lowest amount of <100 mm in the southern Interior Plains where summer evaporation is particularly intense. The Shield region has 100–200 mm of annual runoff, values that are intermediate between the mountains and the plains. Regarding the YRB, large annual runoff comes from the western uplands and from the southern mountains, but the latter area has a complex relationship with topography. Large contrasts exist between the rivers fed by high rainfall and ample snowmelt water (e.g. Nizhnyaya Tunguska and Yenisei above Nikitino) and those in the rain shadow areas (e.g. Selanga River), and annual runoff can range from <50 mm to >250 mm. Elsewhere, the high plateau yields high runoff of >200 mm while less comes from the lower plateau areas (125–200 mm), and low precipitation and high evaporation in the southern end of YRB result in <100 mm/year of runoff.

Seasonally, winter is the period of low runoff, as indicated by the November to March values (Figures 4 and 5). Air temperatures are exceedingly low (e.g. within MRB, the January average is −20 °C at Fort McMurray, 59 °39′N, and −27 °C at Norman Wells, 65 °17′N; for YRB, the January average is −19 °C at Irkutsk, 52 °15′N, and −27 °C at Igarka, 67 °47′N). The basins are subject to seasonal frost, and winter runoff is provided mainly by groundwater and in some cases, by discharge from lakes and reservoirs (Woo & Thorne 2014). At low temperatures, winter precipitation is stored and does not melt until spring. Snowmelt first comes in late April or May in the southern parts of both basins and on slopes below the high plateau in YRB, then proceeds to other basin areas in June and July. Meltwater is responsible for generating high runoff in most areas, but summer rainfall and glacier melt in restricted localities also augment runoff. Evaporation increases in the summer and the lowest runoff comes from southern YRB and from the plains of MRB (Woo & Thorne 2016).

The general runoff patterns of MRB and YRB are in agreement with the findings of Lammers et al. (2001). They found that for 1960–89, large amounts of runoff come from mountainous areas in Siberia and the Canadian Rockies, and in areas with frequent cyclonic activities. Low runoff occurs on continental areas leeward of mountain ranges, including the Selenga basin and the western Canadian Arctic drainage.

RIVER FLOW

Flow regime and storage function

The seasonal rhythm, or the regime, of flow at a particular station along the main river is largely controlled by two water sources: the discharge coming from the upstream station and the contribution from tributaries that enter the main river below the upstream station. The flow is also modified by channel storage along the drainage network.

For both MRB and YRB, the most common natural flow rhythm is the nival regime in which the melting of winter snow in conjunction with river ice breakup gives rise to annual high flow, usually starting in the southern regions where the snow melts early (e.g. Taseeva and Podkamennaya Tunguska rivers have high flow in May, but for Nizhnyaya Tunguska River further north, it occurs in June). Snowmelt high flow is followed by a general recession in the summer, ending with low flow in the winter. Glaciers and late-lying snow in northern MRB prolongs high flow. Although southern YRB has glaciers that would continue to yield melt runoff in the summer, the storage function of a series of reservoirs downstream of Tannu Ola and Sayan mountains mutes the hydrograph rises produced by glacier melt discharge and summer rain. The scarcity of late-lying snow in northern YRB, and insufficient water to supplement flow, results in a steeper recession in the YRB than the MRB hydrographs (Figure 6). Where both winter snow storage and summer rain are important water sources, river flow exhibits a pluvio-nival regime with mixed or dual peak periods (e.g. the Yenisei at Kyzyl, shown in Figure 6(b)). Where rainfall becomes increasingly dominant, the pluvial regime prevails, with high flows responding to summer rain. On a local scale, flow regimes can be modified by other considerations such as glacier melt or substantial evapotranspiration loss from wetlands that draws down the water level to dampen the flow response to summer rain events (Woo 2012).

Figure 6

Amalgamation of flow in large river systems as shown by the 2003 daily discharge of rivers in (a) the Mackenzie Basin from Peace River at Hudson Hope to Artic Red River, (b) the Yenisei Basin from Kyzyl to Igarka.

Figure 6

Amalgamation of flow in large river systems as shown by the 2003 daily discharge of rivers in (a) the Mackenzie Basin from Peace River at Hudson Hope to Artic Red River, (b) the Yenisei Basin from Kyzyl to Igarka.

Lake storage can have pronounced effect on river flow. For large natural lakes including the Athabasca, Great Slave and Great Bear in MRB and Lake Baikal in YRB, the retention and subsequent release functions significantly impact the timing and magnitude of their outflow, producing a prolacustrine flow regime. Reservoirs as artificial lakes significantly modify the natural flow regime of a river (Vyruchalkina 2004; Stuefer et al. 2011). This study concerns only the operational (excluding the construction and filling) phase when discharge fluctuates from day to day to suit the need for power production. Since water is not for consumptive use, flow regulation alters the seasonal rhythm of flow (Yang et al. 2004) and has a less serious effect on the flow amount totaled over an extended time period, such as a year (Adam et al. 2007).

Flow amalgamation

Along the main trunks of Mackenzie and Yenisei rivers, there is a downstream amalgamation of flow regimes and an integration of flow volume, as their tributaries enter the main stems. Figure 6(a) traces the downstream integration of flow along these mega-rivers. Williston Reservoir on Peace River in MRB distorts the original pre-dammed nival regime, which is restored by the natural inflow downstream. After joining the Athabasca River at Lake Athabasca, it becomes the Slave River (at Fitzgerald in Figure 6(a)) before entering Great Slave Lake. The river that flows out from this lake is the Mackenzie River, which displays a prolacustrine regime as shown at Strong Point. However, in meeting the Liard River, the pronounced nival flow regimen of this tributary overwhelms the lake effect on Mackenzie flow. This flow pattern is conveyed all the way downstream to Arctic Red River station before the Mackenzie branches into its delta. In a similar fashion (Figure 6(b)), the pluvio-nival flow regime found at Kyzyl is altered by reservoir operation downstream and is changed again at Yeniseysk where it is joined by the Angara River, itself having a combination of flows from such unregulated rivers as the Taseeva and the highly regulated Angara above Boguchany. Further downstream, the Yenisei receives natural inflows from the Podkamennaya Tunguska and Nizhnyaya Tunguska rivers, which further adjusts the regime of the Yenisei River at Igarka. As a consequence of the merging of flows in the drainage networks, the regime of Mackenzie and Yenisei rivers at their mouths is an amalgamation of different flow patterns. Superficially, the resulting configuration of seasonal flow takes on the appearance of a nival regime despite the mixed origins of the flow (i.e. not solely attributable to snowmelt).

Downstream transformation of regime for a mega-river depends on the hydrograph shape of the main stem, which usually has larger flow than its tributaries; and the strength of the signal conveyed by the incoming tributaries that reinforces or weakens the flow rhythm of the main trunk. One measure to assess similarity of flow regimes is to correlate the hydrographs between two stations. In this, we correlate monthly flow of individual years at the lowest stations in the basins (i.e. Mackenzie River at Arctic Red River; Yenisei River at Igarka) with their major contributing river sections. For each year of record, 12 pairs of monthly values are used and the correlation coefficients (r values) are plotted in Figure 7 (note that depending on the years of record available, the number of r values differs for different stations). Tight clustering of high r values for many years indicates similarity of flow regimes, but wide scattering of r values suggests inconsistency in the influence of a tributary on the flow regime of the main river. Anomalously high and low r values occur in some years. These are attributable to departures of seasonal precipitation from its normal pattern or human modification of the natural regime. Correlation of the tributaries with the main station generally increases for stations further downstream, and that is due to similarity between the natural nival regimes of these northern tributaries and the apparent nival regime exhibited at the main river outlets. For MRB, significant correlations are found between the flow at Arctic Red River station and its three principal tributaries (Slave, Liard and Mackenzie at Strong Point), with the correlation increasing downstream (except for Great Bear River with prolacustrine regime, and Peace River where the flow is human-modified). The high correlation between Athabasca River in the south and Arctic Red River station in the north results from the glacial flow regime of Athabasca River with late-season high flow imitating the large flow in northern rivers sustained by late-lying snowmelt. In the YRB, correlation also increases downstream, but to a lesser extent and significance. Here, the flow in the southern portion of the basin (Selegna, Angara, and Yenesei Rivers), where several of the major reservoirs in the YRB are situated (McClelland et al. 2004), is largely unrelated to the flow at Igarka. However, correlations improve for the northern portion of the basin below Yeniseysk where there is less human interference with the natural flow rhythm.

Figure 7

Correlation coefficients of monthly flows for (a) Mackenzie at Arctic Red River and (b) Yenisei at Igarka, with their contributing tributaries. Also shown are the significance levels of p < 0.05 and p < 0.01, for 10 degrees of freedom.

Figure 7

Correlation coefficients of monthly flows for (a) Mackenzie at Arctic Red River and (b) Yenisei at Igarka, with their contributing tributaries. Also shown are the significance levels of p < 0.05 and p < 0.01, for 10 degrees of freedom.

Flow magnitudes

In order to compare the flows of the Mackenzie (at Arctic Red River) and Yenisei (at Igarka) rivers, Figure 8 presents their monthly means as well as the contribution from their major sub-basins and inter-station sections along the main trunks. Note that these are averages for the periods after major dams were built and the flow of individual years can depart substantially from these values. On an annual basis, the mean flow of the Mackenzie at Arctic Red River is 290 km3 and for the Yenisei at Igarka it is 610 km3. Total flow is larger for YRB because it has higher mean runoff and considerably bigger basin area (173 mm/year runoff from 1.68 million km2 for MRB and 249 mm/year runoff from 2.44 million km2 for YRB).

Figure 8

Monthly flow contribution from major tributaries or inter-station areas to the total flow of the Mackenzie and Yenisei rivers.

Figure 8

Monthly flow contribution from major tributaries or inter-station areas to the total flow of the Mackenzie and Yenisei rivers.

For convenience of discussion, flow condition is partitioned into three seasons (Table 1): winter (November to April) of low flow; snowmelt period (May to July) with high flow; and summer (August to October) when the flow generally recedes from the spring high. A steep post-winter hydrograph rise comes in May for both basins when snowmelt commences, and the flow peaks in June. The Yenisei high flow is more sharply spiked than the Mackenzie due to prolonged high runoff from late-lying snow/glaciers in the MRB, with June alone accounting for 33% of the annual Yenisei flow, compared with 18% for the Mackenzie. For the high-flow months of May–July, the corresponding percentages are 57% (Yenisei) and 47% (Mackenzie). The decline of flow in late summer is more rapid for the Yenisei, with its August–October flow being 20% of the annual total, in contrast to 31% for the Mackenzie. The approximately six-month long winter is the low flow season, when only 23% (for YRB) and 21% (for MRB) of the annual total leaves these basins.

Table 1

Flow contribution from sub-basins and inter-station areas to seasonal flow of Mackenzie and Yenisei rivers

 Flow in km3
 
Flow as % of seasonal total
 
 Nov–Apr May–Jul Aug–Oct Annual Nov–Apr May–Jul Aug–Oct Annual 
Mackenzie River Basin
 
 Athabasca (133) 3.85 9.52 5.69 19.06 6.2 6.9 6.2 6.6 
 Peace (293) 25.73 25.71 14.59 66.03 41.5 18.8 16.0 22.7 
 Liard (275) 9.34 47.07 23.05 79.45 15.1 34.3 25.2 27.4 
 Great Bear (146) 7.98 4.31 4.54 16.83 12.9 3.1 5.0 5.8 
 Slave (180) 8.66 2.39 9.10 20.15 14.0 1.7 10.0 6.9 
 Mid-Mackenzie (389) 5.17 15.22 14.37 34.76 8.3 11.1 15.7 12.0 
 Lower Mackenzie (264) 1.24 32.83 20.05 54.12 2.0 24.0 21.9 18.6 
Seasonal Total 61.96 137.05 91.39 290.40     
Yenisei River Basin
 
 Selenga (360) 2.99 9.92 11.18 24.09 2.1 2.9 9.1 4.0 
 Taseeva (127) 3.43 13.12 7.03 23.58 2.4 3.8 5.7 3.9 
 Angara main trunk (553) 50.52 29.72 18.65 98.89 35.8 8.6 15.2 16.2 
 Podkamennaya Tunguska (232) 6.53 41.15 8.91 56.59 4.6 11.9 7.3 9.3 
 Nizhnyaya Tunguska (447) 5.53 86.93 21.23 113.69 3.9 25.2 17.3 18.7 
 Yenisei at Kyzyl (115) 4.93 17.54 9.97 32.24 3.5 5.0 8.1 5.3 
 Yenisei reservoir section (245) 35.73 14.53 14.81 65.07 25.3 4.2 10.8 10.7 
 Mid-Yenisei (128) 7.78 35.05 7.25 50.08 5.5 10.2 7.2 8.2 
 Lower Yenisei (233) 23.53 97.41 23.43 144.37 16.7 28.2 19.1 23.7 
Seasonal Total 140.97 345.17 122.46 608.60     
 Flow in km3
 
Flow as % of seasonal total
 
 Nov–Apr May–Jul Aug–Oct Annual Nov–Apr May–Jul Aug–Oct Annual 
Mackenzie River Basin
 
 Athabasca (133) 3.85 9.52 5.69 19.06 6.2 6.9 6.2 6.6 
 Peace (293) 25.73 25.71 14.59 66.03 41.5 18.8 16.0 22.7 
 Liard (275) 9.34 47.07 23.05 79.45 15.1 34.3 25.2 27.4 
 Great Bear (146) 7.98 4.31 4.54 16.83 12.9 3.1 5.0 5.8 
 Slave (180) 8.66 2.39 9.10 20.15 14.0 1.7 10.0 6.9 
 Mid-Mackenzie (389) 5.17 15.22 14.37 34.76 8.3 11.1 15.7 12.0 
 Lower Mackenzie (264) 1.24 32.83 20.05 54.12 2.0 24.0 21.9 18.6 
Seasonal Total 61.96 137.05 91.39 290.40     
Yenisei River Basin
 
 Selenga (360) 2.99 9.92 11.18 24.09 2.1 2.9 9.1 4.0 
 Taseeva (127) 3.43 13.12 7.03 23.58 2.4 3.8 5.7 3.9 
 Angara main trunk (553) 50.52 29.72 18.65 98.89 35.8 8.6 15.2 16.2 
 Podkamennaya Tunguska (232) 6.53 41.15 8.91 56.59 4.6 11.9 7.3 9.3 
 Nizhnyaya Tunguska (447) 5.53 86.93 21.23 113.69 3.9 25.2 17.3 18.7 
 Yenisei at Kyzyl (115) 4.93 17.54 9.97 32.24 3.5 5.0 8.1 5.3 
 Yenisei reservoir section (245) 35.73 14.53 14.81 65.07 25.3 4.2 10.8 10.7 
 Mid-Yenisei (128) 7.78 35.05 7.25 50.08 5.5 10.2 7.2 8.2 
 Lower Yenisei (233) 23.53 97.41 23.43 144.37 16.7 28.2 19.1 23.7 
Seasonal Total 140.97 345.17 122.46 608.60     

Numbers in brackets are drainage areas in thousands of km2.

Within MRB and YRB, the contribution of their sub-catchments to total outflow changes during the year (Figure 8 and Table 1). In the low-flow period, reservoirs (on Peace River in MRB, and series of reservoirs on the Angara and upper Yenisei in YRB) are the main providers of total basin flow. However, their absolute magnitudes are modest and stable. Outside of the winter season, other basin areas yield much higher runoff so that the lake and reservoir contributions become proportionally overwhelmed. This is especially notable in the principal snowmelt period. The flow produced by snowmelt runoff takes time to travel down river, frequently detained by river ice blockage. A delay in the arrival of upstream-generated flow gives rise to ‘negative flow contribution’ at the lower station. For example, early snowmelt in southern YRB often does not reach Lower Yenisei in May, so that the flow of the Yenisei River at Igarka is lower than further upstream, to result in an artificial ‘negative’ contribution from this inter-station section. Similarly, a negative contribution is found at the Peace River mouth in December due to delayed arrival of reservoir water released from the upper course. Summer months have the highest precipitation of the year but much of it is lost to evaporation, as experienced in southern Interior Plains of MRB and the Selenga and Kyzyl basins of YRB. Flow contributions from highlands decline in summer, and the recession from their spring flow is more rapid in northern YRB than in the northern mountains of MRB.

On the whole, rivers in mountains and high plateaus produce the largest flow. They include part of the Athabasca, the Liard and lower Mackenzie basin, which together provide almost half of the total Mackenzie River flow. In YRB, the lower Yenisei (from the confluence with Podkamennaya Tunguska River to Igarka) and Nizhnyaya Tunguska River yield 24% and 19% respectively of the Yenisei total. A notable departure of mountainous terrain as a high-flow zone is the arid range and basin topography in southern YRB, where the Yenisei above Kyzyl, the Taseeva River and the Selenga River each produce only 4% of the Yenisei flow. Regulated rivers yield moderately high annual flow: the Peace River gives 23% of total Mackenzie flow; the Angara produces 16%, although the reservoir section along the Yenisei River provides only 11% of total Yenisei flow. Rivers in the remaining parts of the two basins, consisting of large natural lakes, plains, the low plateau of Central YRB, and rolling areas such as the Canadian Shield, are responsible for 20–25% of the flows in MRB and YRB.

DISCUSSION AND CONCLUSION

The Mackenzie River basin (MRB) in Canada and the Yenisei River basin (YRB) in Russia are significant suppliers of freshwater to the Arctic Ocean. Their combined annual outflow of 900 km3 represents nearly one-fifth of the total freshwater received by the polar ocean (which is about 4,900 km3/year, estimated using values from Serreze & Barry (2005)). These basins have similar hydrological settings, spreading across major hydro-physiographical provinces from the cold temperate, through the subarctic to the Arctic regions. Their topography is similarly diverse. Both basins receive their atmospheric moisture from two sources: advected from outside and recycled from within. Westerly airflow brings in moisture from the Pacific for the MRB and from the Atlantic for the YRB, and Arctic air comes to both basins. Topography interacts with large-scale atmospheric flows, either blocking them or facilitating their movement, to influence the regional climates while enhancing moisture recycling through mountain-plains circulation.

Unlike small northern catchments where one can detect the role of permafrost in promoting fast and ample surface runoff, such effects cannot be identified in large catchments where the factors of topography, climate and lake storage co-vary with or negate the influence of permafrost. For MRB and YRB, high runoff comes from their mountains with continuous or discontinuous permafrost, and low runoff expectedly comes from areas with low precipitation and high summer evaporation, as in their southern sectors. Large lakes provide a notable storage function that detains high inflow and sustains moderate outflow in the winter months. Large reservoirs have been built in both MRB and YRB for hydro-power production, which significantly alters the natural flow regimes.

The seasonal rhythm of discharge of the Mackenzie and the Yenisei rivers appears similar, peaking in spring to early summer and diminishing to their annual minima in winter. This pattern appears superficially to resemble a nival regime with the high flow suggestive of meltwater release from the snow accumulated over their long winter. Such an apparent nival regime is not entirely related to spring snowmelt, but is the product of integration of flow contribution from their sub-basins and of flow modification along their drainage networks.

MRB and YRB have hydro-physiographical attributes typical of river basins in the circumpolar region (Figure 1), despite differences in detail. Some basins may have more extensive wetlands (e.g. Ob), more permafrost (e.g. Lena) or more intense usage of reservoir water (e.g. Ob). Overall, knowledge of spatial variations in runoff generation and flow contribution from different parts of a mega-basin can be helpful to future investigations for isolating human and natural influences and for identifying areas in the basin that are sensitive to expectant climatic, environmental and land use changes. The general conclusions drawn from this study regarding runoff generation, flow contribution and river discharge characteristics are of relevance to other northern mega-basins.

ACKNOWLEDGEMENTS

We thank the two referees for carefully going through the manuscript and for their helpful comments.

REFERENCES

REFERENCES
Aagaard
K.
&
Carmack
E. C.
1989
The role of sea ice and other fresh water in the Arctic circulation
.
Journal of Geophysical Research
94
(
C10
),
14485
14498
.
Adam
J. C.
,
Haddeland
I.
,
Su
F.
&
Lettenmaier
D. P.
2007
Simulation of reservoir influences on annual and seasonal streamflow changes for the Lena, Yenisei, and Ob’ rivers
.
Journal of Geophysical Research
112
,
D2411
.
doi:10.1029/2007JD008525
.
Adler
R. F.
,
Huffman
G. J.
,
Chang
A.
,
Ferraro
R.
,
Xie
P.
,
Janowiak
J.
,
Rudolf
B.
,
Schneider
U.
,
Curtis
S.
,
Bolvin
D.
,
Gruber
A.
,
Susskind
J.
&
Arkin
P.
2003
The version 2 global precipitation climatology project (GPCP) monthly precipitation analysis (1979–present)
.
Journal of Hydrometeorology
4
,
1147
1167
.
Bulygina
O. N.
&
Razuvaev
V. N.
2012
Daily Temperature and Precipitation Data for 518 Russian Meteorological Stations
.
Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, US Department of Energy
,
Oak Ridge, Tennessee
,
USA
.
doi:10.3334/CDIAC/cli.100
.
Carmack
E. C.
,
Yamamoto-Kawai
M.
,
Haine
T. W. N.
,
Bacon
S.
,
Bluhm
B. A.
,
Lique
C.
,
Melling
H H.
,
Polyakov
L. V.
,
Straneo
F.
,
Timmermans
M.-L.
&
Williams
W. J.
2016
Freshwater and its role in the Arctic Marine System: Sources, disposition, storage, export, and physical and biogeochemical consequences in the Arctic and global oceans
.
Journal of Geophysical Research: Biogeosciences
121
,
675
717
.
doi:10.1002/2015JG003140
.
Hydrological Atlas of Canada
1978
Ministry of Supply and Services
,
Ottawa
,
Canada
.
Lammers
R. B.
,
Shiklomanov
A. I.
,
Vörösmarty
C. J.
,
Fekete
B. M.
&
Peterson
B. J.
2001
Assessment of contemporary Arctic river runoff based on observational discharge records
.
Journal of Geophysical Research
106
(
D4
),
3321
3334
.
Liu
J. L.
,
Stewart
R. E.
&
Szeto
K. K.
2008
Water vapor fluxes over the Canadian prairies and the Mackenzie River Basin
. In:
Cold Region Atmospheric and Hydrologic Studies. The Mackenzie GEWEX Experience. Vol. 1: Atmospheric Dynamics
(
Woo
M.K.
, ed.).
Springer
,
Berlin
,
Germany
. pp.
109
125
.
McClelland
J. W.
,
Holmes
R. M.
,
Peterson
B. J.
&
Stieglitz
M.
2004
Increasing river discharge in the Eurasian Arctic: consideration of dams, permafrost thaw, and fires as potential agents of change
.
Journal of Geophysical Research
109
,
D18102
.
doi:10.1029/2004JD004583
.
McClelland
J. W.
,
Déry
S. J.
,
Peterson
B. J.
,
Holmes
R. M.
&
Wood
E. F.
2006
A pan-Arctic evaluation of changes in river discharge during the latter half of the 20th century
.
Geophysical Research Letters
33
,
L06715
.
doi:10.1029/2006GL025753
.
Muskett
R. R.
&
Romanovsky
V. E.
2009
Groundwater storage changes in Arctic permafrost watersheds from GRACE and in situ measurements
.
Environmental Research Letters
4
,
045009
.
doi:10.1088/1748-9326/4/4/045009
.
Onuchin
A.
&
Burenina
T.
2010
Climate and geographic patterns of spatial distribution of precipitation in Siberia
. In:
Environmental Change in Siberia: Earth Observation, Field Studies and Modelling
(
Balzter
H.
, ed.).
Advances in Global Change Research 40
,
Springer
,
Dordrecht
,
Netherlands
, pp.
193
210
.
Overeem
I.
&
Syvitski
J. P. M.
2010
Shifting discharge peaks in Arctic rivers, 1977–2007
.
Geografiska Annaler: Series A, Physical Geography
92
(
2
),
285
296
.
doi:10.1111/j.1468-0459.2010.00395
.
Park
H.
,
Yoshikawa
Y.
,
Yang
D.
&
Oshima
K.
2017
Warming water in Arctic terrestrial rivers under climate change
.
Journal of Hydrometeorology
18
,
1983
1995
.
Peters
D. L.
&
Prowse
T. D.
2001
Regulation effects on the lower Peace River, Canada
.
Hydrological Processes
15
,
3181
3194
.
Peterson
B. I. R.
,
Holmes
R. M.
,
McClelland
J. W.
,
Vorosmarty
C. J.
,
Lammers
R. B.
,
Shiklomanov
A. I.
,
Shiklomanov
I. A.
&
Rahmstorf
S.
2002
Increasing river discharge to the Arctic Ocean
.
Science
298
,
2171
2173
.
Rood
S. B.
,
Kaluthota
S.
,
Philipsen
L. J.
,
Rood
N. J.
&
Zanewich
K. P.
2017
Increasing discharge from the Mackenzie River system to the Arctic Ocean
.
Hydrological Processes
31
,
150
160
.
doi:10.1002/hyp.10986
.
Serreze
M. C.
&
Barry
R. G.
2005
The Arctic Climate System
.
Cambridge University Press
,
Cambridge
,
UK
.
Serreze
M. C.
,
Barrett
A. P.
,
Slater
A. G.
,
Woodgate
R. A.
,
Aagaard
K.
,
Lammers
R. B.
,
Steele
M.
,
Moritz
R.
,
Meredith
M.
&
Lee
C. M.
2006
The large-scale freshwater cycle of the Arctic
.
Journal of Geophysical Research
111
,
C11010
.
doi:10.1029/2005JC003424
.
Shahgedanova
M.
2003
Climate at present and in the historical past
. In:
The Physical Geography of Northern Eurasia
(
Shahgedanova
M.
, ed.).
Oxford University Press
,
Oxford
,
UK
, pp.
70
102
.
Slaughter
C. W.
,
Hilgert
J. W.
&
Culp
E. H.
1983
Summer streamflow and sediment yield from discontinuous-permafrost headwaters catchments
. In:
Proceedings of the 4th International Conference on Permafrost
,
Fairbanks, Alaska
.
National Academies Press
,
Washington, DC
,
USA
, pp.
1172
1177
.
Spence
C.
&
Woo
M. K.
2003
Hydrology of subarctic Canadian shield: soil-filled valleys
.
Journal of Hydrology
279
,
151
166
.
St. Jacques
J.-M.
&
Sauchyn
D. J.
2009
Increasing winter baseflow and mean annual streamflow from possible permafrost thawing in the Northwest Territories, Canada
.
Geophysical Research Letters
36
,
L01401
.
doi:10.1029/2008GL035822
.
Stuefer
S.
,
Yang
D. Q.
&
Shiklomanov
A.
2011
Effect of streamflow regulation on mean annual discharge variability of the Yenisei River
.
IAHS-AISH Publication
346
,
27
32
.
Szeto
K. K.
,
Stewart
R. E.
,
Yau
M. K.
&
Gyakum
J.
2008a
The Mackenzie climate system: a synthesis of MAGS atmospheric research
.
In: Cold Region Atmospheric and Hydrologic Studies. The Mackenzie GEWEX Experience. Vol. 1: Atmospheric Dynamics
(
Woo
M. K.
, ed.).
Springer
,
Berlin
,
Germany
, pp.
23
50
.
Szeto
K. K.
,
Liu
J. L.
&
Wong
A.
2008b
Precipitation recycling in the Mackenzie and three other major river basins
. In:
Cold Region Atmospheric and Hydrologic Studies. The Mackenzie GEWEX Experience. Vol. 1: Atmospheric Dynamics
(
Woo
M. K.
, ed.).
Springer
,
Berlin
,
Germany
, pp.
137
154
.
Vyruchalkina
T. Yu.
2004
Lake Baikal and the Angara River before and after the construction of reservoirs
.
Water Resources Research
31
,
484
489
.
Woo
M. K.
2012
Permafrost Hydrology
.
Springer
,
Heidelberg
,
Germany
.
Woo
M. K.
&
Thorne
R.
2014
Winter flows in the Mackenzie drainage system
.
Arctic
67
,
238
256
.
Yang
D. Q.
,
Ye
B. S.
&
Kane
D. L.
2004
Streamflow changes over Siberian Yenisei River Basin
.
Journal of Hydrology
296
,
59
80
.