1. Introduction
The Mackenzie River basin (MRB; Fig. 1) in northwestern Canada is a climatologically important region that exerts significant influences on both the global climate and the regional climates of North America. Being a large high-latitude continental region, it is one of the major heat and moisture sink regions for the global circulation; it is thus also an important source region of anticyclones (Ioannidou and Yau 2007) and cold continental polar air mass. Also, being located to the lee of the Rockies, it is one of the two source regions of lee cyclones in North America. In addition, the discharge from the Mackenzie River represents the largest North American source of freshwater for the Arctic Ocean (Milliman and Meade 1983), which could have significant consequences on the global ocean and atmospheric circulations.
The MRB is known to possess extremely variable climates (Szeto et al. 2007). In particular, as illustrated in Fig. 2 (see also Fig. 19 of Kistler et al. 2001), the region exhibits both the largest intraseasonal and interannual variability of winter temperatures in the world, suggesting that the cold-season climate of the region is extremely sensitive to variations in the large-scale circulation on both intraseasonal and interannual time scales. These characteristics of the region make predicting the climate for the region a particularly challenging task. For example, simulating the hydroclimate for the MRB is shown to be particularly problematic among the several major arctic basins considered in the modeling study of Finnis (2005). Persistent strong cold bias and associated negative impacts on simulated water and energy cycling in the region have also been found in recent climate simulations for the basin (MacKay et al. 2003; Szeto et al. 2008; see also section 4). Improving our physical understanding and modeling capability of climate variability for this northern region is a pressing challenge because most GCMs have predicted an amplified polar response to increase in greenhouse gases in the atmosphere (Serreze and Francis 2006) and the strongest warming signals have been observed in high-latitude continental regions, particularly the MRB (Serreze and Francis 2006; see also Fig. 3a).
A first step in improving climate predictions for the MRB is to examine the underlying mechanisms that cause the extreme winter temperature variability in this region. The purpose of the present study is to better understand these mechanisms through an assessment of the mean and anomalous atmospheric enthalpy budgets, and relate the temperature response of the MRB to variations in the large-scale circulation over the North Pacific. Implications of the results to improving the understanding and prediction of climate variability and change in the region will also be discussed.
2. Methodology and datasets
The 3D wind velocities and air temperatures from the 6-hourly and 2.5° resolution National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset (Kalnay et al. 1996; Kistler et al. 2001) were used to calculate the basin- and vertically integrated H-budget terms in Eq. (1), for the 30-yr period 1970–99. Both the NCEP–NCAR reanalysis data and climate indices used in the discussion of results are provided by the National Oceanic and Atmospheric Administration–Cooperative Institute for Research in Environmental Sciences (NOAA–CIRES) Climate Diagnostics Center (available online at http://www.cdc.noaa.gov/). Although there are newer global reanalysis datasets [e.g., the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40; Uppala et al. 2005)], it has been shown that their differences in winter tropospheric temperature and wind fields are quite negligible over the region of interest (Grotjahn 2007; Szeto et al. 2008) and both variables in the datasets are quite comparable to observations over the MRB (Szeto et al. 2008). Because both the wind field and temperatures are strongly nudged with observations in the analysis (Kistler et al. 2001), it is not a surprise to find good correspondence of these variables with observations. With the progressive incorporation of satellite data into the NCEP data assimilation system during the 1970s (Kistler et al. 2001), the 1970–99 period was chosen so that there is relative temporal homogeneity in the assimilated observations during the study period. Although the full assimilation of global satellite data did not happen until the late 1970s, the whole decade was included in this study because several major anomalous warm and cold winters in the MRB occurred during the 1970s (Table 1 and Fig. 3a).
All the budget terms, including the tendency term, are calculated at 6-hourly intervals before either the vertical integration or horizontal averaging. Given that we are only interested in the net
Physical processes that are responsible for the occurrence of extreme warm and cold winters in the region are investigated by examining the anomalous H budgets and corresponding large-scale atmospheric conditions that affected the basin during those periods. The five warmest and five coldest winters [based on the domain-averaged surface air temperatures for December–February (DJF)] for the MRB during the study period are listed in Table 1. These cases also represent the winters with basin-average temperature anomalies that exceed one interannual standard deviation of DJF surface air temperatures (σ = 2.3 K) for the basin during the study period. Although there is some variability among the large-scale circulations that characterized the different cases in each sample, the similarity of their large-scale conditions is evident in the similarity of the Pacific–North American (PNA) and ΔZ indices (see section 4b) that characterize the cases in the two samples. Anomalous H budgets and atmospheric parameters that are representative of extreme warm/cold winters in the region are obtained by compositing these variables over the five corresponding cases; the results are then intercompared to gain physical insights into the processes that produce the large observed interannual variability of winter temperatures.
3. Results
a. Mean H budgets and governing processes
The seasonal variations of the 30-yr mean component enthalpy budget terms are shown in Fig. 4a. Except for the summer, energy gained through net atmospheric transport (i.e., the combined horizontal advection and vertical terms in Fig. 4a) of dry static energy (DSE) is balanced by a net loss through atmospheric diabatic processes that occur within the basin. Since much of the region receives little solar radiation and relatively little precipitation during the cold season, the diabatic term is largely accounted for by thermal radiation to space through the top of the atmosphere, and to a lesser extent by the downward heat flux into the surface (Szeto et al. 2008). In the summer [June–August (JJA)], the basin experiences long daylight hours and receives ample precipitation (Szeto et al. 2008) and the region becomes a heat source for the large-scale flow (Fig. 4a). As expected, changes in energy storage are only significant during spring and autumn.
Located just downwind of the Pacific Ocean, climate in the MRB is strongly affected by large-scale atmospheric circulation features over the North Pacific, such as the prominent Aleutian low pressure system (AL; Fig. 5a). While the transport of energy into the continent during the cold season is largely accomplished by such circulation features, the atmospheric thermal response of the basin to the transport is affected significantly by the interactions of the circulation with its physical environment. This is particularly true for the MRB due to its immediate lee location to the Western Cordillera and its relative proximity to the Pacific Ocean. The combined land–sea contrast and mountain-blocking effects induce significant disturbance to the mean southwesterly flow to produce a stationary long wave with an upper-level ridge over the west coast of North America (Fig. 5a). The combined continentality and latitudinality effects also produce a mean southwest– northeast air temperature gradient over this part of the continent (Szeto et al. 2008). Consequently, although the coastal region is under the influence of a mean southwesterly flow that brings in warm and moist air from the North Pacific, the MRB is under the influence of a mean northwesterly flow (Fig. 5a), which is largely normal to the mean horizontal air temperature gradients throughout much of the troposphere. Hence, although magnitudes of the individual mean positive zonal and negative meridional heat transports are considerable, they largely cancel each other to produce a relatively weak net horizontal transport over the domain (Figs. 4a and 6a).
Strong horizontal flow convergence also occurs as the low-level onshore flow encounters abrupt changes in surface roughness and elevation (Fig. 5a) to produce enhanced updraft and precipitation (hence latent heat release) on the west-facing slopes of the coastal mountains (Figs. 7a and 8a). Because the flow from the warm ocean surface is loaded with moisture, winter precipitation at the coastal region of western Canada is comparable to that of either the equatorial or the storm-track regions (Fig. 8a). The maritime polar air mass thus undergoes strong modification by precipitating out a substantial portion of its moisture content as it crosses the Continental Divide, and the mountain barriers effectively enhance the transport of DSE into the basin by forcing latent heat to be released over the western slopes. The air that descends on the lee side is thus relatively warm and dry, and the DSE convergence into the MRB during the cold season is the strongest among the major high-latitude continental basins considered in Roads et al. (2002, see their Table 1 and Fig. 11). In fact, the cold-season H budgets in Fig. 4a show that when integrated over the vertical column, the warming from the adiabatic descent is far more important than the net horizontal heat transport in affecting the basin’s atmospheric enthalpy budgets. The warming in the subsiding air increases the air–land temperature contrasts, which in turn leads to enhanced heat loss to the underlying surface (Fig. 6a). The warming also enhances the horizontal temperature gradients between the region and its downstream vicinity, and the relatively warm air in the region is advected downwind through its eastern boundary (Fig. 9a). The enhanced low-level warm air advection (WAA) out of the basin is also reflected in the weakening of the basin-average meriodional cold air advection (CAA) and enhancement of the zonal WAA at the low levels (Fig. 6a).
b. Processes and H budgets for extreme warm/cold winters
Exceptionally warm winters in the 30-yr period were characterized by a strengthened AL and stronger-than-normal large-scale onshore/upslope flows into the continent (Figs. 5b and 7b). The enhanced southwesterly flow strengthened the mid- and upper-level high pressure ridge over the mountainous coastal region (Fig. 5b; see also Held and Ting 1990), which in turn intensified the northwesterly flow over the MRB. As a result, both the zonal WAA and meridional CAA increased (Figs. 4b and 6b) over the region but they largely negated each other to yield any significant change to the net basin-average transport (Fig. 4b). The strengthened onshore flow also induced powerful upward motions and condensation at the coastal regions (Figs. 7b and 8b) accompanied by enhanced subsidence and adiabatic warming over the MRB (Figs. 4b, 6b and 7b). The enhanced warming over the region was balanced by an increase in both radiational loss from the basin as well as sensible heat loss to the surface (Figs. 4b and 6b). The warmed basin also enhanced (reduced) WAA out of (into) the region through its eastern (western) boundary (Fig. 9b). At the same time, the increased temperature gradient between the warmed basin and the northern regions also enhanced CAA into the basin at its northern boundary (Fig. 9b).
The anomalously cold winters were associated with a weakened AL over the North Pacific (Figs. 5b). Such conditions are often accompanied by an increased occurrence of blocking high pressure systems over the North Pacific (Renwick and Wallace 1996), which diverted many of the Pacific cyclones to the west of Alaska and resulting in a split mean flow configuration on the seasonal time scale. One component of the flow was diverted northward into western Alaska, but it subsequently turned anticyclonically to bring cold air into the MRB through its northern boundary (Fig. 5c). The flow splitting also diminished the mean onshore wind (Fig. 5c) and weakened the long-wave ridge over the coastal region (Fig. 5c), thus reducing the mean northwesterly flow and associated mid- and upper-level meridional CAA and zonal WAA over the MRB (Fig. 6c). At the same time, the weakened onshore flow greatly reduced the mean updraft and precipitation in the coastal region (Figs. 5c, 7c and 8c) as well as diminishing the associated sinking and adiabatic warming on the lee side (Figs. 4c and 7c). Because of the cooler-than-normal condition, sensible heat loss to the surface was reduced, but atmospheric radiational cooling was significantly enhanced under general clear-sky conditions (Fig. 6c). At the same time, the weakened temperature contrast between the region and its northern environment reduced the mean mid- and upper-level CAA into the cold basin at its northern boundary while the mean WAA at its eastern (western) boundary was substantially reduced (enhanced; Fig. 9c). The net result was a marked reduction in mean meridional CAA (i.e., strongly positive υ-advection anomaly) and a slightly weakened mean zonal WAA into the basin (Figs. 4b and 6c).
In summary, heat transport into the northwest North American continent during the cold season is largely accomplished by dynamic features over the North Pacific. However, because of the high mountain barriers that separate the MRB and the North Pacific, horizontal temperature advection does not affect the energetics of the region directly. Instead, the onshore flow is adiabatically chilled in the forced ascent over the west slopes while at the same time it could be warmed through latent release from the orographic precipitation. The two effects largely negate each other in the strong upslope flows and the energy used to do work against the ambient air pressure and gravity is converted back into sensible heat when the air parcels descend and are warmed through adiabatic compression over the lee side. The mean horizontal temperature advection and diabatic cooling over the MRB are largely a passive response to the warming due to the mean descent. These effects are evident in the correlation between DJF H-budget terms and the mean surface temperatures or enthalpy for the region (Table 2). In particular, only the vertical term is positively and strongly correlated with either Ts or H. The net horizontal advection term is negatively correlated with Ts or H, reflecting the fact that the net positive horizontal heat advection into the MRB is reduced (enhanced) during anomalous warm (cold) periods over the basin. Hence, the net horizontal advection largely responds to, rather than forces, the temperature changes over the basin. As shown earlier, the atmospheric diabatic cooling increases during both anomalous warm and cold winters, hence the extreme low correlation of the term with Ts or H.
4. Discussion
The processes that govern the heat budgets over the MRB will be further analyzed in this section to gain insight into the causes for the extreme cold-season temperature variability in the area. The dynamic North Pacific is clearly a major source of temperature variations in the MRB, but the myriad North Pacific atmospheric flow features alone are not sufficient to account for the extreme temperature variability of the basin. For instance, such extreme variability does not occur over western Europe, which is also located just downwind of the equally dynamically active North Atlantic Ocean. Continentality, which promotes large regional temperature fluctuations, also does not explain the extreme variability in the MRB. For example, Siberia, which is farther from the ocean than the MRB, does not exhibit the same degree of extreme conditions (Fig. 2). Based on these arguments, we hypothesize that the extreme variability in the MRB is the collective result of vigorous upstream dynamic forcings, their interactions with the basin’s physical environment, and feedback processes that occur within the basin. These processes take place over the whole northeast Pacific–northwest North America sector that spans almost a third of the circumpolar region. To facilitate the analysis of the complex problem, we will take the reductionistic approach and break down the vast domain of interest into four key component regions: (i) a forcing region (FR), that is, the North Pacific which is the source region for the large-scale atmospheric circulation features that ultimately force temperature variations in the MRB; (ii) the response region (RR), which composes the MRB proper; and (iii) two interface zones (IZs) that separate and interface the FR and RR. The first IZ (IZ1) spans from the west coast of Canada to the Great Divide and the other IZ (IZ2) consists of the western Arctic Ocean located to the immediate west and north of Alaska. Typically, the mean flow from the FR frequently passes through IZ1 (IZ2) before entering the basin during anomalous warm (cold) winters in the MRB.
Synoptic-scale processes that play a fundamental role in affecting the energy transport and temperature response in the basin will be identified and discussed in sequence for the component regions. The aggregated effects of these short-period processes are then related to the longer-term circulation and heat transport features presented in the last section and a synthesis of these elements will be developed in section 5 to account for the extreme temperature variability in the MRB along with applications of the results to climate change and prediction issues.
a. The FR and variability of North Pacific atmospheric circulation
The variability of large-scale atmospheric circulation features over the North Pacific is characterized by patterns or modes of oscillations that occur over a wide range of temporal scales (e.g., Wallace and Gutzler 1981; Mantua et al. 1997; Overland et al. 1999). There is evidence that the longer-period climate modes often force or create large-scale conditions that modulate the frequency and phase of shorter-period variability patterns (e.g., Palmer 1998; Trenberth et al. 1998). For example, the positive (negative) phase of the PNA pattern (Wallace and Gutzler 1981) is known to be associated with increased frequency of cyclones (blocking events; Renwick and Wallace 1996) while increased frequency of positive (negative) PNA phase periods are in turn found to be associated with warm (cold) phases of ENSO (Mo et al. 1998; Trenberth et al. 1998) and Pacific decadal oscillation (PDO; Mantua et al. 1997).
The fundamental role played by synoptic-scale eddies in the energy transport at mid- and high latitudes is well known (Peixoto and Oort 1992). Located just downwind of the Pacific storm track, the transport of water and energy into the basin MRB is naturally affected most strongly by cyclonic activities over the North Pacific (Lackmann et al. 1998; Smirnov and Moore 2001), especially during the cold season when the baroclinic disturbances are active and frequent. Since the wave cyclones are fundamental features that are responsible for transporting energy into the region, temperature response in the MRB to variations in the North Pacific large-scale circulation is best understood through the ways by which the low-frequency modes modulate the occurrence frequency of shorter-period circulation features. For example, the typical warming (cooling) response of the MRB to a shift into a positive (negative) PNA period can be related to the increased frequency of cyclonic (blocking) activities over the North Pacific and associated enhanced (reduced) energy transport into the region, while the enhanced warming (cooling) typically found over the region during the warm (cold) phases of ENSO and PDO can in turn be linked with the associated increase in frequency of positive (negative) PNA events, and so on.
Such physical linkages between the North Pacific atmospheric circulation regimes and the temperature response in the MRB are reflected in the statistically significant correlations between winter temperatures in the basin and teleconnection indices that are commonly used to characterize the variability of these circulation patterns (Table 2). Note that the stronger correlations are established with shorter-period circulation patterns that are closely linked to the aggregated synoptic activities over the North Pacific on the seasonal or shorter time scale [e.g., the PNA and the Aleutian low index (ALI; Overland et al. (1999)]. Not surprisingly, very weak correlations are found for climate modes with weak links to the Pacific circulation such as the Arctic Oscillation/North Atlantic Oscillation (AO/NAO).
It is of interest to note that the stronger correlations are established with the vertically integrated enthalpy (i.e., with the mid- and upper-level temperatures) than with the surface temperatures. Regardless of the nature and detailed structure of the circulation features over the North Pacific, the transport of water and energy into the continent is ultimately accomplished by the shore-normal (onshore) flow component, in particular, by the ocean surface–conditioned lower-level flow, which is high in moist static energy content. Hence, lower-level temperatures in the MRB are affected most strongly by changes in the low-level onshore flows and the mechanical perturbations and thermal modifications they experience as they cross the Western Cordillera, which is what we will discuss next.
b. The IZ1 and coastal and orographic influences
In addition to the excitation of a stationary long wave from the impingement of the mean westerly into the Western Cordillera, the mountainous terrain that separates the MRB and the Pacific Ocean also exerts significant influences on the cyclonic systems that approach the northwest coast of North America. Although there were recent progresses in the theoretical investigation of orographic influences on the landfall of cyclones and their potential regenerations on the lee side (e.g., Bannon 1992; Davis 1997), many aspects of such events as summarized in Palmén and Newton (1969) still remain valid today. In their conceptual model, the cyclonic onshore flow strengthens as a Pacific cyclone approaches the coast of western Canada, and results in enhanced precipitation over the west slopes as the flow crosses the Rocky Mountains. A quasi-stationary surface trough would form as a result of the tropospheric warming by the enhanced descent over the lee side. If coupled to a mid- to upper-level feature such as a short-wave trough, a jet streak, or the decaying parent cyclonic system, this lee trough may become mobile and develop into a lee cyclone. Effects of the synoptic features developed over the basin will be discussed in the next section and we will focus on discussing the processes that affect the cross-barrier flows inside the IZ1 in this section.
The interaction of moist flow with orography and its consequences on precipitation enhancement is a complicated phenomenon governed by the interplay of factors such as the velocity, static stability, and shear structure that characterize the ambient flow, synoptic conditions, and the geometry of the orographic features (Lin 2005). Despite extensive research on the subject, many aspects of orographic precipitation remain poorly understood. For the present discussion, it suffices to note that due to the release of potential instability in the forced ascent, the orographic precipitation and associated latent heat release depend nonlinearly on the onshore flow configuration, such that latent heating at the coastal region could be significantly enhanced (significantly reduced or totally halted) during strong (weak) onshore flow situations. Through its amplifying and gating effects on latent heat release, the mountainous coastal region thus enlarges the difference of DSE transports into the MRB for various flow configurations over the North Pacific. For example, using the precipitation anomalies presented in Figs. 8b,c, difference in mean latent heating over the coastal region between warm and cold winters could exceed 60 W m−2 (∼0.5 K day−1; Fig. 10), which in turn contributed to the difference of over 25 W m−2 in basin-average combined vertical advection/adiabatic heating over the MRB on the seasonal time scale (Fig. 4b).
For obvious reasons, winter precipitation at the northern mountainous regions is neither well observed nor well simulated. To further examine the effects of the cross-barrier flow on the temperatures of the MRB, we will consider a new index ΔZ given by ΔZ = Z(A) − Z(B), where Z is geopotential height Z and A and B are two points within IZ1 such that B is north of A and the line joining A and B is roughly parallel to the alignment of the mountain barrier (Fig. 1). The ΔZ computed at the upper levels would give a measure of the mean geostrophic shore-normal flow strength within the coastal region. We will be more interested in the ΔZ computed at the lowest pressure levels, which gives a convenient measure of the magnitude of the mean near-surface cross-barrier flow speed. Note that it does not give an exact measure because the pressure level chosen for its computation might be below the physical surface over the mountainous region. Hence, large positive ΔZ would indicate strong near-surface cross-barrier flow into the continent and vise versa. Because of sloping effects within the region, larger positive ΔZ values at a location farther away from coast would indicate stronger mean cross-barrier flow, stronger mean ascent, and heavier precipitation, and hence stronger warming as they descent on the lee side. To verify these inferences, we computed the contemporaneous correlations between basin DJF T and different ΔZ evaluated at different distances from the coast and at different vertical levels near the surface (Table 2). The results show that the stronger correlations are established with ΔZ evaluated at the lowest levels (1000 or 925 hPa) and at a location near the exit of the IZ1, and hence lending support to the arguments. In addition, corrections established with ΔZs computed off the coast or over the interior basin (not shown) are also significantly lower than those established with ΔZs evaluated within IZ1. It is also evident that the correlations of Ts with ΔZ are somewhat stronger than those established with other teleconnection indices (Table 2), thus giving support to the argument that the low-level cross-barrier flow and the diabatic effects it experiences inside IZ1 exert the most direct effects on the low-level temperatures over the basin.
A corollary that follows from the above discussion is that (warm) cold bias in model simulations of cold-season temperatures over the MRB could result when the orographic precipitation over the mountainous coastal region is (over-) underpredicted. Simulations of water and energy processes in the MRB using the Canadian Regional Climate Model (CRCM; MacKay et al. 2003) were troubled by a cold bias in the modeled lower troposphere (MacKay et al. 2003; Szeto et al. 2008; see also Fig. 11a). This bias causes an overprediction of snow cover, which affects the surface energy balance, the timing and magnitude of spring runoff and soil recharge and, subsequently, the warm-season water cycling. Insight gained from the present study allows us to diagnose the cold bias as a consequence of the underprediction of orographic precipitation at the coastal region. In agreement with the theoretical prediction, the cold bias is the strongest during the cold season (Fig. 11a) when synoptic activities are frequent and the cross-barrier flow is strong, and it occurs mainly over the areas immediately downstream of coastal precipitation maximum and to the lee of the cordillera (Fig. 11b). Further diagnostics of the model results (not shown) confirmed that the CRCM precipitation over the coastal region was indeed substantially lower than either the analyzed or observed amounts (see also Mati 2006).
In addition, Szeto et al. (2008) found that, during the cold season over the MRB, there are large atmospheric moisture surpluses and energy deficits in the water and energy budgets assessed from the NCEP-R2 reanalysis (Kanamitsu et al. 2002). By the same token, these large residuals in balancing the budgets can be explained by an underprediction of orographic precipitation and associated latent heat release over the west slopes of the cordillera in the model used to produce the NCEP-R2 reanalysis. However, unlike the results from the CRCM, both the atmospheric temperature and humidity variables are strongly nudged with observations during the analysis, and therefore despite the potential model deficiencies, strong cold-season temperature or moisture biases are not evident in the R2 data (Szeto et al. 2008).
c. The RR and feedback processes within the MRB
While the processes that occur over the North Pacific and the coastal region play a dominant role in affecting the transport of energy into the MRB, feedback processes that take place within the basin can influence critically how the basin’s temperature responds to these forcings. For instance, latent heating in the ascending cross-barrier flow over the coastal region can only be “felt” over the MRB when the energy used to do work against the ambient air pressure and gravity are converted back into sensible heat when the air parcels descend and being compressed adiabatically over the lee side. Hence, any processes that affect the leeside subsidence can exert great effects on the temperatures over the basin. In addition, temperature advection by perturbation airflows and modifications to local radiative transfers by synoptic-scale processes could also affect significantly the temperature response.
As discussed earlier, a quasi-stationary shallow lee trough would form as a result of the warm descent over the basin. The low-level cyclonic perturbation flow induced by the surface trough could advect warm air into the basin from the south as well as enhancing the downslope descent at locations southwest of the trough. The low-level WAA could in turn destabilize the lower troposphere (i.e., weaken or destroy the surface-based low-level temperature inversions that are commonly observed over the basin during the winter), and hence create an environment that would facilitate further low-level descent and associated warming over the region. Enhanced low-level southerly WAA into the basin is evident in Figs. 6b and 9b. In addition, clouds that form as result of the WAA or lee cyclogenesis that sometimes develops from the lee troughs could reduce infrared radiative loss from the basin surface and thus contribute to the further warming of the basin’s surface and lower troposphere.
The results presented in section 3b show that cold winters in the MRB are often associated with a weakened AL and reduced mean onshore flow speed. Such conditions reflect the increased occurrence of blocking events over the Gulf of Alaska that divert the North Pacific cyclones into the western Arctic Ocean (i.e., the IZ2) via the Chukchi and Bering Seas. The enhanced upper-level WAA into IZ2 under such circumstances could induce development of deep warm-core anticyclones over the region. The increased frequency of warm-core anticyclones from this mechanism during negative PNA periods over the western Arctic Ocean was reported in Ioannidou and Yau (2007). Ioannidou and Yau also showed that these warm-core anticyclones often migrate southeastward into the MRB where they could transform into cold-core structures. In fact, detailed examination of the synoptic conditions that characterized the basin during the development of extreme cold episodes shows that many of these extreme events started with a ridge of high pressure that built into the area from Alaska.
Apart from the transport of cold air into the region (see the enhanced meridional CAA in Fig. 6c), the enhanced low-level anticyclonic flow into the basin could be forced to rise up the east slopes of the foothills to induce adiabatic cooling and subsequent condensation and light precipitation (e.g., ice crystal) that dries out the cold air. Curry (1983) theorized that longwave radiation loss from the thin clouds that form near the top of the inversion layer and the enhanced surface radiational loss under the dry and clear-sky conditions that follow the dissipation of the cloud through light precipitation could play an important role in the development of cold continental polar (cP) air mass and anticyclones. The forced gentle ascent of the anticyclonic flows up the east slopes provides a lifting mechanism that could initiate or accelerate the sequence of events hypothesized in Curry (1983). In addition, the resulting dense cold air could expand both horizontally and vertically, effectively reducing or eliminating the sloping effects of the mountains and thus restraining the descent and warming of the westerly flow that passed over the Rockies. Such cold-air damming phenomena and their associated impacts on the regional weather have been investigated extensively for southern regions (e.g., Bell and Bosart 1988, among others). The geographic and synoptic environments that characterize the basin suggest that these phenomena should occur frequently over the basin during the winter and they can exert significant influences on cold-season energy cycling in the MRB.
The MRB is a major source region of anticyclones and cP air mass during the winter (Ioannidou and Yau 2007; Szeto et al. 2007). It is found in previous studies (e.g., Tanaka and Milkovich 1990) that classical idealized theories of anticyclogenesis and cold airmass development (e.g., Wexler 1936; Curry 1983) could not accurately account for the observed development of these events in the close-by region of Alaska. We hypothesize that regional factors such as the topographic influences and the cold-air damming effects could complement the idealized static theories of Wexler and Curry to better account for the rapid development of these features and the transformation of warm-core anticyclones into cold-core structures over the basin.
The significant relationships between warm (cold) days and low (high) pressure over the MRB are evident in Fig. 12. When aggregated over a season, the frequent lee troughs and lee cyclones manifest themselves as a low-level pressure perturbation trough on the seasonal time scale for anomalous warm winters over the MRB (Fig. 5b). On the other hand, the high pressure systems and their associated perturbation flows that frequently affect the basin during cold episodes are manifested on the seasonal time scale as the anomalous low-level high pressure ridge over the region and the associated perturbation upslope flow and much reduced mean subsidence over the basin during extreme cold winters (Figs. 5c and 7c). Cao et al. (2001) conducted case studies of anomalous warm and cold winter months for the basin and they also identified the associations of stationary low and high pressure features with warm and cold periods on the monthly time scale. In addition, Stewart and Burford (2002) showed that, in agreement with the above discussions, clouds were more (less) widespread during warm (cool) winter periods when compared to average conditions. Further discussions of cloud radiative coupling processes that act to accentuate temperature responses in the region can be found in Szeto et al. (2007).
5. Synthesis, application to climate change, and final remarks
There have been extensive studies conducted on the atmospheric circulation features over the North Pacific and their interactions with the mountainous terrain of northwestern North America. This study provides fresh insight into the causes of extreme temperature variability in the MRB when these processes were examined in tandem with a detailed analysis of the atmospheric enthalpy budgets of the basin. Because of the high mountain barriers that separate the MRB from the Pacific Ocean, latent heat release over the west-facing slopes and processes that affect the warm descent over the lee side play a more critical role than the direct horizontal heat transport in governing the winter enthalpy budget for the basin. The results show that interactions of synoptic disturbances (that are of either Pacific or MRB origin) with the Western Cordillera can enlarge the differences in the indirect heat transport into the basin by the vertical motions for different synoptic situations over the North Pacific. Horizontal heat transports by the perturbation air flows and cloud radiative processes within the basin also act to accentuate the temperature perturbations induced by changes in the mean descent. Collectively, these processes enlarge the thermal response of the basin, and thereby increase the sensitivity of its winter temperatures, to changing upstream large-scale conditions. The observed extreme intraseasonal variability of winter temperatures in the basin can thus be understood as a consequence of the amplified sensitivity of the region’s temperatures to the highly variable synoptic conditions over the North Pacific. As longer-period climate variability can be described in terms of changes in occurrence frequency of shorter-period circulation patterns all the way down to the synoptic time scale, the extreme interannual temperature variability of the MRB can then be understood as a result of the enhanced sensitivity of the region’s seasonal temperatures to the low-frequency variability of atmospheric circulations through their attendant modulations effects on weather regimes over the North Pacific. A conceptual model that summarizes these processes is given in Fig. 13.
Several studies have reported significant recent winter warming of the MRB (e.g., Zhang et al. 2000; Serreze and Francis 2006). It is evident from Fig. 3a, however, that the linear trends noted in some of the previous studies are largely statistical artifacts of the “jump” in the mean basin temperatures as the basin responded to the well-documented shift in the PDO during the mid-1970s (e.g., Mantua et al. 1997). Using the argument put forward in this paper, the significant recent warming in the MRB can be interpreted as an amplified temperature response to the increased frequency of positive PNA patterns after the regime shift (Fig. 3b). There have been suggestions that, due to the nonlinear nature of the climate system, anthropogenic climate change will primary be manifest in terms of changes to the occurrence frequency of natural climate variability patterns (Palmer 1998). According to this nonlinear perspective of climate change, although the observed warming can be attributed to changes in natural climate variability modes, one cannot simply conclude that the warming was not due to anthropogenic forcing. In addition, given the sensitivity of the basin’s winter temperatures to climate variability and change, the MRB can serve as an effective “thermometer” for gauging anthropogenic climate change.
On the other hand, the sensitivity of the basin’s temperatures to synoptic-scale processes poses a particular challenge to model prediction of future temperature changes in the region. The results suggest that accurate model representations of the interactions of synoptic features with the Western Cordillera are critical for accurate predictions of long-term climate variability and change in the interior continent. In particular, the cold bias that has affected some climate models over the MRB was shown to be a result of underprediction of orographic precipitation over the mountainous coastal regions. Because of the strong coupling between water and energy processes in the region, the cold biases can contaminate other aspects of the simulated climate for the region and beyond. Careful examinations and validations of the simulated cold-season temperatures in the region are strongly recommended for all climate models which cover the MRB in their simulation domains.
Some of the conceptual understandings developed in this study to explain the temperature variability of the MRB, with appropriate modifications, should also be useful in the investigation of variability and change of water cycling processes such as precipitation and runoff in the region. For instance, the same orographic processes that enlarge the difference of dry static energy transports into the MRB for various upstream flow configurations could also reduce the difference in moisture influx into the basin through the effective orographic depletion of moisture from strong onshore flows, and hence mute the precipitation and subsequent hydrological responses in the basin to the large-scale forcings. Although there are many completing factors that complicate the problem, preliminary observed evidences that support such inferences are given in Woo et al. (2006). The significance of the ΔZ index and its potential application in dynamical–statistic seasonal temperature forecast for the region also has not been fully explored in this study. Given the importance of the MRB in affecting the weather and climate of other North American regions, further investigations of these issues are warranted.
Acknowledgments
The author would like to thank the Mackenzie GEWEX Study (MAGS) community for its support during the course of this work. The manuscript benefits substantially from comments by the anonymous reviewers. The CRCM simulations were conducted by Dr. Murray MacKay. Ms. H. Tran and Billy Szeto are acknowledged for their assistance in data processing and calculations. This work was financially supported by Environment Canada, the Panel on Energy Research and Development (PERD) and MAGS.
REFERENCES
Bannon, P. R., 1992: A model of Rocky Mountain lee cyclogenesis. J. Atmos. Sci., 49 , 1510–1522.
Bell, G. D., and L. F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev., 116 , 137–161.
Cao, Z., R. E. Stewart, and W. Hogg, 2001: Extreme winter warming events over the Mackenzie basin: Dynamic and thermodynamic contributions. J. Meteor. Soc. Japan, 79 , 785–804.
Curry, J., 1983: On the formation of continental polar air. J. Atmos. Sci., 40 , 2278–2292.
Davis, C. A., 1997: The modification of baroclinic waves by the Rocky Mountains. J. Atmos. Sci., 54 , 848–868.
Finnis, J., 2005: Representation of the atmospheric hydrologic cycle over the Arctic in CCSM3. Preprints, Eighth Conf. on Polar Meteorology and Oceanography, San Diego, CA, Amer. Meteor. Soc., CD-ROM, J1.4.
Grotjahn, R., 2007: Different data, different general circulations? A comparison of selected fields in NCEP/DOE AMIP-II and ECMWF ERA-40 reanalyses. Dyn. Atmos. Oceans, in press.
Held, I., and M. Ting, 1990: Orographic versus thermal forcing of stationary waves: The importance of the mean low-level wind. J. Atmos. Sci., 47 , 495–500.
Ioannidou, L., and M. K. Yau, 2007: Climatological analysis of northern hemisphere winter anticyclones. J. Geophys. Res., in press.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S-K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEP–DOE AMIP-II reanalysis (R-2). Bull. Amer. Meteor. Soc., 83 , 1631–1643.
Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50-Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247–267.
Lackmann, G. M., J. R. Gyakum, and R. Benoit, 1998: Moisture transport diagnosis of a wintertime precipitation event in the Mackenzie River basin. Mon. Wea. Rev., 126 , 668–691.
Lin, Y-L., 2005: Dynamics of orographic precipitation. McGraw-Hill 2005 Yearbook of Science and Technology, McGraw-Hill, 248–250.
MacKay, M. D., F. Seglenieks, D. Verseghy, E. D. Soulis, K. R. Snelgrove, A. Walker, and K. Szeto, 2003: Modeling Mackenzie basin surface water balance during CAGES with the Canadian Regional Climate Model. J. Hydrometeor., 4 , 748–767.
Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78 , 1069–1079.
Mati, I., 2006: A comparison study of two regional atmospheric models over the Mackenzie Basin. M.S. thesis, Dept. of Atmospheric and Oceanic Sciences, McGill University, 73 pp.
Milliman, J. D., and R. H. Meade, 1983: World-wide delivery of river sediment to the oceans. J. Geol., 91 , 1–21.
Mo, R., J. Fyfe, and J. Derome, 1998: Phase-locked and asymmetric correlations of the wintertime atmospheric patterns with the ENSO. Atmos.–Ocean, 36 , 213–239.
Overland, J. E., J. M. Adams, and N. A. Bond, 1999: Decadal variability of the Aleutian low and its relation to high-latitude circulation. J. Climate, 12 , 1542–1548.
Palmén, E., and C. W. Newton, 1969: Atmospheric Circulation Systems: Their Structure and Physical Interpretation. Academic Press, 603 pp.
Palmer, T. N., 1998: Nonlinear dynamics and climate change: Rossby’s legacy. Bull. Amer. Meteor. Soc., 79 , 1411–1423.
Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.
Renwick, J. A., and J. M. Wallace, 1996: Relationships between North Pacific wintertime blocking, El Niño, and the PNA pattern. Mon. Wea. Rev., 124 , 2071–2076.
Roads, J., M. Kanamitsu, and R. Stewart, 2002: CSE water and energy budgets in the NCEP–DOE reanalysis II. J. Hydrometeor., 3 , 227–248.
Serreze, M. C., and J. A. Francis, 2006: The Arctic amplification debate. Climatic Change, 76 , 241–264.
Smirnov, V., and G. W. K. Moore, 2001: Short-term and seasonal variability of the atmospheric water vapor transport through the Mackenzie River basin. J. Hydrometeor., 2 , 441–452.
Stewart, R. E., and J. E. Burford, 2002: On the features of clouds occurring over the Mackenzie River basin. J. Geophys. Res., 107 .4720, doi:10.1029/2001JD001559.
Szeto, K. K., R. E. Stewart, M. K. Yau, and J. Gyakum, 2007: Northern tales: A synthesis of MAGS atmospheric and hydrometeorological research. Bull. Amer. Meteor. Soc., 88 , 1411–1425.
Szeto, K. K., H. Tran, M. MacKay, R. Crawford, and R. E. Stewart, 2008: The MAGS water and energy budget study. J. Hydrometeor., in press.
Tanaka, H. L., and M. F. Milkovich, 1990: A heat budget analysis of the polar troposphere in and around Alaska during the abnormal winter of 1988/89. Mon. Wea. Rev., 118 , 1628–1639.
Trenberth, K. E., G. W. Branstator, D. Karoly, A. Kumar, N-C. Lau, and C. Ropelewski, 1998: Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophys. Res., 103 , 14291–14324.
Uppala, S. M., and Coauthors, 2005: The ERA-40 re-analysis. Quart. J. Roy. Meteor. Soc., 131 , 2961–3012.
Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109 , 784–812.
Wexler, H., 1936: Cooling in the lower atmosphere and the structure of polar continental air. Mon. Wea. Rev., 64 , 122–136.
Woo, M. K., R. Thorne, and K. K. Szeto, 2006: Reinterpretation of streamflow trends based on shifts in large-scale atmospheric circulation. Hydrol. Processes, 20 , 3995–4003.
Zhang, X., L. A. Vincent, W. D. Hogg, and A. Niitsoo, 2000: Temperature and precipitation trends in Canada during the 20th Century. Atmos.–Ocean, 38 , 395–429.
Geographic region of northwestern Canada with topography. Also shown are boundaries for the MRB (thin line), the domain for H-budget calculations (thick line), and the locations (A1–A3, B) used in the calculation of the ΔZ index as presented in section 4.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
(a) Interannual standard deviation of mean January air temperatures at 850 hPa and (b) mean intraseasonal standard deviation of daily January air temperatures at 850 hPa from the NCEP reanalysis data for the 1948–2005 period. The contour interval is 0.5 K and regions with values higher than 4 and 8 K are shaded in (a) and (b), respectively.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Interannual time series of (a) DJF basin-average surface air temperature (K) and (b) DJF PNA index for 1950–2005. Also shown are the linear trend lines (dashed) and the mean values for the periods before and after 1976.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
(a) Seasonal variability of 30-yr (1970–99) and domain-averaged H-budget components for the MRB.(b) Composite anomalous DJF H-budget components for the five warmest and five coldest winters.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Mean and anomalous DJF atmospheric conditions: (a) 30-yr mean geopotential height (m) at 1000 (solid) and 500 hPa (dashed); composite anomalous geopotential height (m) at 1000 (contours) and 500 hPa (shaded) for (b) the five warmest and (c) the five coldest winters.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
(a) Vertical profiles of 30-yr mean DJF H-budget components and composite budget anomaly profiles for (b) the five warmest and (c) the five coldest winters.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Mean and anomalous DJF atmospheric conditions: (a) 30-yr mean wind vectors at 1000 hPa and ω at 850 hPa (Pa s−1); composite anomalous wind vectors at 1000 hPa and ω at 850 hPa (Pa s−1) for (b) the five warmest and (c) the five coldest winters. Maximum magnitude of wind vectors equals 6.5 m s−1. Contour intervals are (a) 0.02 and (b), (c) 0.01 Pa s−1.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Mean and anomalous DJF atmospheric conditions: (a) Global Precipitation Climatology Project 1979–2000 mean precipitation rate (mm day−1, values higher than 5 mm day−1 are shaded); composite anomalous precipitation rate (mm day−1) and T (K, shaded) at 850 hPa for (b) the five warmest and (c) the five coldest winters. Precipitation contour intervals are (a) 0.5 and (b), (c) 0.3 mm day−1.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
(a) Vertical profiles of 30-yr DJF mean cross-boundary temperature advections at the four (west, east, south, and north) domain boundaries and corresponding composite T-advection anomaly profiles for (b) the five warmest and (c) the five coldest winters.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Difference in mean DJF atmospheric latent heating (K day−1) between the five warmest and five coldest winters.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Surface air temperature bias (K) in the CRCM obtained by using the Canadian Meteorological Centre analysis data as reference: (a) interannual variability of monthly basin-average bias and (b) spatial variability of 5-yr (1998–2002) mean January bias. The zero contour lines are highlighted in both (a) and (b).
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
Scatterplot of daily surface pressure anomaly at the center of the domain vs daily domain-average temperature anomalies for 1970–99.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
A schematic illustration of how interactions of synoptic-scale processes with the basin’s topographic features enhance the sensitivity of the MRB winter temperature (T) response to variations of the large-scale circulation over the North Pacific. Shown on the schematic are the pressure perturbation centers (H′ and L′), perturbation southerly (x) and northerly (.) flows and the associated WAA and CAA that are commonly found in the subregions (FR, IZ1, IZ2, and RR), which compose the northeast Pacific–northwest North America sector during anomalously warm and cold winters over the MRB.
Citation: Journal of Climate 21, 1; 10.1175/2007JCLI1583.1
The five coldest and five warmest winters and their corresponding DJF basin-average surface air temperature anomalies and PNA and ΔZ (at 1000 hPa and using points A2 and B in Fig. 1) indices for the 1970–99 period.
Contemporaneous correlations between DJF domain-average surface air temperature and vertically integrated enthalpy with various H-budget components and teleconnection indexes over the 30-yr period. (left to right) The three ΔZ values were calculated using points A1, A2, and A3, respectively, with B in Fig. 1. Correlations that are significant at the 95% level are highlighted bold.
