Lake-Aggregate Mesoscale Disturbances. Part IV: Development of a Mesoscale Aggregate Vortex

Peter J. Sousounis Atmospheric, Oceanic and Space Sciences Department, University of Michigan, Ann Arbor, Michigan

Search for other papers by Peter J. Sousounis in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Many studies have noted that cyclone development in the Great Lakes region during winter is the result of strong diabatic heating and low-level destabilization from the lakes. The exact mechanisms, however, by which this heating and moistening lead to sea level pressure falls, and to weak cyclones over the lakes (e.g., mesoscale aggregate vortices), have not been investigated previously.

In this study, model output that includes all of the Great Lakes and none of the Great Lakes is analyzed to understand more completely the importance of synoptic-scale forcing, diabatic heating, and perturbation–synoptic-scale processes for the development of a mesoscale aggregate vortex over the region during a 48-h period between 0000 UTC 13 and 0000 UTC 15 November 1982. The analysis indicates that the sea level pressure falls and vortex development were not simply the hydrostatic result of heat from the Great Lakes “spreading” over a large region. Rather, the synoptic-scale flow contributed to vortex development during the first 24 h by providing strong cold northwesterly flow, which generated significant surface heat fluxes; and during the second 24 h by providing low-level warm advection and midlevel positive vorticity advection from southwesterly flow, which enhanced large-scale ascent and horizontal perturbation heat flux convergence near the surface. The eventual collocation of strong cyclonic perturbation southerly winds at 900 hPa, strong anticyclonic perturbation southerly winds at 700 hPa, and east–west-oriented isotherms in between greatly enhanced the warm advection and vortex development in the region. Finally, the intensifying cyclonic perturbation flow contributed significantly to surface sensible and latent heat fluxes and to further vortex development when it phased with the synoptic-scale flow at the surface.

The one case that has been examined does not likely serve as an explanation for all mesoscale aggregate vortices. More studies are needed to determine the climatology of these vortices that develop over the Great Lakes region in winter.

Corresponding author address: Dr. Peter J. Sousounis, Atmospheric Oceanic and Space Sciences Department, University of Michigan, Ann Arbor, MI 48109-2143.

Email: sousou@umich.edu

Abstract

Many studies have noted that cyclone development in the Great Lakes region during winter is the result of strong diabatic heating and low-level destabilization from the lakes. The exact mechanisms, however, by which this heating and moistening lead to sea level pressure falls, and to weak cyclones over the lakes (e.g., mesoscale aggregate vortices), have not been investigated previously.

In this study, model output that includes all of the Great Lakes and none of the Great Lakes is analyzed to understand more completely the importance of synoptic-scale forcing, diabatic heating, and perturbation–synoptic-scale processes for the development of a mesoscale aggregate vortex over the region during a 48-h period between 0000 UTC 13 and 0000 UTC 15 November 1982. The analysis indicates that the sea level pressure falls and vortex development were not simply the hydrostatic result of heat from the Great Lakes “spreading” over a large region. Rather, the synoptic-scale flow contributed to vortex development during the first 24 h by providing strong cold northwesterly flow, which generated significant surface heat fluxes; and during the second 24 h by providing low-level warm advection and midlevel positive vorticity advection from southwesterly flow, which enhanced large-scale ascent and horizontal perturbation heat flux convergence near the surface. The eventual collocation of strong cyclonic perturbation southerly winds at 900 hPa, strong anticyclonic perturbation southerly winds at 700 hPa, and east–west-oriented isotherms in between greatly enhanced the warm advection and vortex development in the region. Finally, the intensifying cyclonic perturbation flow contributed significantly to surface sensible and latent heat fluxes and to further vortex development when it phased with the synoptic-scale flow at the surface.

The one case that has been examined does not likely serve as an explanation for all mesoscale aggregate vortices. More studies are needed to determine the climatology of these vortices that develop over the Great Lakes region in winter.

Corresponding author address: Dr. Peter J. Sousounis, Atmospheric Oceanic and Space Sciences Department, University of Michigan, Ann Arbor, MI 48109-2143.

Email: sousou@umich.edu

1. Introduction

Cox (1917) was likely the first person to document that the Great Lakes region experiences an increase in the frequency and the intensity of synoptic-scale lows in winter. He and others since then have attributed these increases to the addition of sensible and latent heat and to destabilization from the Great Lakes during cold-air conditions.1 For example, Danard and Rao (1972), Danard and McMillan (1974), and Boudra (1981) found by performing numerical simulations of relatively strong synoptic-scale extratropical cyclones that heating from the Great Lakes aggregate reduces sea level pressure (SLP) by 4–7 hPa but that neither the position nor the circulation of these storms changes significantly.

Sousounis and Fritsch (1994, hereafter referred to as Part II) demonstrated using with-lake (WL) and no-lake (NL) numerical simulations that the Great Lakes can significantly affect the path and intensity of dynamically weak (e.g., 1–3 closed isobars at 4-hPa contour interval) highs and lows. A careful examination of their case study (cf. Fig. 1) shows that the low that developed in the WL simulation and that appeared in the observations around 0000 UTC 15 November 1982 over the Great Lakes was more likely the result of in situ development rather than the result of the attraction and deepening of a weak existing low that was approaching from the southwest.2 The NL simulation showed no such low over the Great Lakes.

Sousounis (1997, hereafter referred to as Part III) labeled the closed low that developed over the lakes between 0000 UTC 13 and 0000 UTC 15 November 1982 a mesoscale aggregate vortex (MAV) because of its size, location, and the fact that it developed in part from aggregate heating and moistening by the Great Lakes. This MAV was ∼800-km wide and ∼4-km deep. It had cyclonic flow from 1000 to 800 hPa and anticyclonic flow from 800 to 600 hPa. It reached maximum intensity northeast of the lakes region during prevailing southwesterly flow ahead of a weak upper-level shortwave, approximately 36 h after the lakes had warmed and moistened cold and dry northwesterly flow over a region that extended southeastward to the Mid-Atlantic coast. Figure 2 summarizes some of the MAV characteristics over the 48-h period of interest. A complete description is provided in Part III.3 Selected WL–NL analyses from that study show that SLP falls began immediately and extended eastward gradually during the 48-h period. Impressive development occurred at 850 hPa, where the thermal perturbation nearly quadrupled in size in a 24-h period. The regional precipitation pattern changed from individual-lake-scale precipitation near lakeshores to aggregate-lake-scale precipitation several hundred kilometers away from the lakes. Although in Part III the characteristics of the MAV were presented, its development was not discussed.

All of the above-mentioned studies have shown that the SLP reductions (perturbations) that develop when cold air flows across the warm Great Lakes are the result of sensible and latent heat being added from the lakes to the lower troposphere. However, none of the studies have examined exactly how the sensible and latent heat causes the SLP perturbations. Petterssen (1956); Danard and McMillan (1974); Reitan (1974); and Angel and Isard (1997) to name a few, have all attributed the SLP perturbations simply to low-level destabilization by the Great Lakes. These studies have suggested that the increased destabilization enhances ascent and reduces adiabatic cooling, which generates stronger pressure falls at the surface as divergence of relatively cold air aloft exceeds convergence of relatively warm air below. Petterssen and Calabrese (1959) estimated using simple hydrostatic reasoning that the Great Lakes can account for a 6–7-hPa SLP perturbation.

There is evidence, though, that more than just the simple addition of sensible and latent heat and/or destabilization is required to generate the SLP perturbations that are observed during dynamically weak conditions. For example, Fig. 3 shows average WL–NL surface (e.g., sensible and latent heat4) flux differences and 12-h 1000-hPa height falls for each of the four 12-h periods for the MAV described in Part III. The flux differences in the first 12-h period averaged ∼500 W m−2 over the upper lakes. Significant 1000-hPa height falls occurred during this period, when the perturbation deepened by more than 30 m over a region extending from Lake Superior southward to northern Lake Michigan. The strongest deepening occurred over the western lakes where the strongest surface WL–NL differences in synoptic-scale wind speed and air–lake temperature differences also existed. The second 12-h period was characterized by weaker synoptic-scale winds and surface flux differences of 200–300 W m−2 and a smaller area of much weaker height falls centered near Sault Ste. Marie, Michigan. During the third 12-h period, heat flux differences continued to decrease slightly to ∼100–200 W m−2 as synoptic-scale winds remained weak but height falls more than 30 m occurred again over a large area just northeast of Lake Huron. During the fourth 12-h period, surface flux differences increased over the lakes and height falls exceeding 30 m occurred again over a large region centered several hundred kilometers northeast of Lake Huron.

The fact that the MAV reached maximum intensity in the WL simulation more than 36 h after the lakes had been warming and moistening the air and after the strongest heat and moisture fluxes were present, the fact that the most significant 1000-hPa height falls occurred several hundred kilometers away from any Great Lake, and the fact that no equivalently strong circulation developed in the NL simulation suggest that diabatic heating, synoptic-scale forcing, and interactions between the developing MAV and the evolving synoptic-scale flow (e.g., MAV–synoptic-scale processes) were likely instrumental for development. Interactions between diabatic heating and synoptic-scale forcing (e.g., Kuo et al. 1991b; Stoelinga 1996) and interactions between synoptic-scale and sub-synoptic-scale processes (Rausch and Smith 1996) have been found to be significant in association with explosively deepening extratropical marine cyclones. Such process interactions for weak cyclones (e.g., MAVs) that develop over the Great Lakes in winter have not been examined.

This study will focus on understanding more completely how heating and moistening from the Great Lakes can develop weak cyclones (e.g., MAVs) over the region. Specific objectives are to assess 1) the role of synoptic-scale forcing; 2) the relative importance of diabatic heating and dynamical forcing, including contributions from MAV–synoptic-scale and MAV–MAV scale interactions; and 3) the influence of MAV development on the surface fluxes of heat and moisture. The case described in Part III is ideal for accomplishing the above objectives because the synoptic-scale forcing was relatively weak, and the evolution as shown in Figs. 2 and 3 suggests that diabatic heating was not solely responsible for development. A better understanding of the thermodynamic and dynamic processes that were responsible for the development of this particular MAV should provide a better understanding in general of how MAVs (weak lows) develop and how the Great Lakes impact regional weather during winter.

To accomplish our objectives, the existing model output from the simulations for that November 1982 case will be analyzed. The existing output will be used because the case was well simulated and because the appropriate model output exists at sufficiently high spacial (30 km) and temporal (3 h) resolution. Both the WL and NL simulations will be compared to assess the impacts of the lakes, because even a careful examination of the WL simulation by itself will not allow unambiguous assessment, owing to the fact that the aggregate effects are similar in scale to many synoptic-scale, dynamically forced features that cross the lakes. Additionally, it is more advantageous to compare the WL and NL simulations after 48 h of continuous simulation than it is, for example, to perform four separate sets of 12-h long WL and NL simulations over the 48-h period of interest, where the WL and NL simulations share similar initial conditions at the start of each set. The advantage of the former strategy is that it allows the accumulated effects of the lakes that are important for MAV development to be addressed, whereas the latter strategy precludes an accurate assessment because each successive NL simulation becomes progressively more contaminated with the effects of the lakes (cf. Monobianco 1989). The remainder of this paper is organized as follows. The evolving synoptic-scale environment is examined in section 2. The relative importance of diabatic heating and MAV–synoptic-scale and MAV–MAV-scale forcing mechanisms are examined in section 3. A summary and conclusions are presented in section 4.

2. The importance of synoptic-scale evolution on MAV development

The fact that temperature perturbations above 850 hPa did not really appear until nearly 24 h after the strongest WL–NL heat flux differences existed over the lakes (cf. Fig. 2) and until a weak synoptic-scale trough approached the region, and the fact that 1000-hPa height falls were all but suppressed temporarily between 12 and 24 h (cf. Fig. 3), suggest that the synoptic-scale flow affected MAV development. It is therefore useful to examine some kinematic aspects of the evolving background (e.g., NL) flow to better understand the synoptic-scale antecedent conditions for MAV development.

Table 1 indicates that distinct changes occurred in the synoptic-scale flow throughout the 48-h period. Specifically, relatively low values of static stability at the beginning increased during the first 15 h and then became steady for the remainder of the first 24-h period. The low static stability at the beginning was associated with a synoptic-scale low and an attendant cold front, which had crossed the Great Lakes just before the start of the 48-h period. The increasing static stability was the result of subsidence from a synoptic-scale high that was moving over the region during the first 24 h. The exiting low to the east and the approaching high to the west provided synoptic-scale cold advection with prevailing northwesterly flow between 1000 and 500 hPa. At 500 hPa, an approaching ridge provided negative vorticity advection. All of these factors: cold advection, negative vorticity advection, subsidence, and high static stability, are known to be detrimental for development of low pressure at the surface.

Interestingly, however, changes in the 1000-hPa height perturbations were anticorrelated with changes in these detrimental factors during this period. Specifically, the 1000-hPa perturbation heights developed rapidly from 0 to 12 h despite strong cold advection and strong negative vorticity advection, and then very slowly from 12 to 24 h despite weak cold advection and weak negative vorticity advection. Changes in the 1000-hPa perturbation heights were instead more correlated with changes in the synoptic-scale contribution5 to the surface heat fluxes as shown in Table 2. During the first 6 h, surface fluxes increased on average over the western lakes but then decreased during the next 18 h, because decreases in wind speed had greater impact than increases in lake–air temperature differences (that resulted from cold advection over the lakes).

The temporal variation of the surface fluxes matches very closely that of the 1000-hPa height tendencies (∂ZMAV/∂t)|1000 shown in Fig. 4. Thus, it appears that the synoptic-scale environment affected MAV development during this first 24-h period more by providing a decreasingly favorable thermally forced environment (e.g., decreasing surface fluxes and resultant diabatic heating) than by providing an increasingly favorable dynamically forced environment (e.g., decreasing cold advection and negative vorticity advection). It is likely, however, that the above-mentioned dynamical aspects of the synoptic-scale environment did restrict MAV development to a shallow layer near the surface.

During the second 24-h period, the synoptic-scale environment became more dynamically favorable for surface development. Specifically, as the high moved east of the region and as a weak trough approached from the west, the prevailing flow changed from northwesterly to southwesterly; warm advection replaced cold advection; and positive vorticity advection replaced negative vorticity advection. As a result, large-scale ascent replaced large-scale descent. Although the stability remained high during the first half of the second 24-h period, it decreased slightly during the second half. The synoptic-scale surface fluxes shown in Table 2 continued to decrease during most of the second 24-h period as the wind speed decreased and as warm advection overspread the region, but then increased slightly near the end—especially over Lakes Superior and Michigan, where cold air and strong northwesterly flow on the west side of the synoptic-scale trough had entered the region. The increased surface fluxes over Lakes Erie and Ontario were the result of increases in wind speed, which compensated for decreasing lake–air temperature differences due to southwesterly flow and warm advection.

The decrease in static stability and the development of large-scale ascent over the region provided an environment that was more favorable for MAV development. From a quasigeostrophic viewpoint it can be understood that the lower static stability, the positive vorticity advection, and the warm advection, allowed stronger synoptic-scale ascent to develop at 700 hPa. From a mass continuity standpoint, it can be understood that the enhanced synoptic-scale ascent at 700 hPa corresponded to enhanced low-level synoptic-scale horizontal convergence. The enhanced low-level convergence contributed to stronger horizontal flux convergences of warm perturbation air (e.g., − · [VSYNTMAV] > 0) as shown in Fig. 4, which increased warming at low levels, and which hence led hydrostatically to stronger perturbation height falls near the surface. Thus, in contrast to the first 24-h period when the thermodynamic forcing from the synoptic-scale flow was more effective than the dynamical forcing for affecting MAV development, during the second 24-h period it was the dynamical forcing from the synoptic-scale flow that was more effective than the thermodynamic forcing for MAV development.

The synoptic-scale analysis described in this section presents an incomplete picture of MAV development. For example, the correlation between − · [VSYNTMAV] > 0 and (∂ZMAV/∂t)|1000 suggests that the synoptic-scale flow acted in conjunction with the developing perturbation to result in further development.

3. The importance of MAV synoptic-scale processes and MAV-enhanced diabatic heating on MAV development

The relative importance of horizontal temperature advections, adiabatic cooling, and diabatic heating on the 1000-hPa height falls of the developing MAV can be assessed by considering the following height tendency equation
i1520-0493-126-12-3169-e1
where ZMAV|i is the perturbation height associated with the MAV at pressure level pi, TMAV is the perturbation temperature associated with the MAV, and the other variables have their usual meaning. Equation (1) is obtained from subtracting versions of the hypsometric formula that are valid for the WL and NL simulations. To evaluate (1), it is useful to rewrite the integrand on the right-hand side (rhs) as:
i1520-0493-126-12-3169-e2
where all variables in (2) have their usual meaning and the stability is defined as in Carlson (1991)—for example, Si = −(Ti/θi)∂θi/∂p. The term on the left-hand side of (2) is the local tendency of perturbation temperature. The first three terms on the rhs of (2) are the temperature changes resulting from the advection of perturbation temperature by the synoptic-scale wind, the advection of synoptic-scale temperature by the perturbation wind, and the advection of perturbation temperature by the perturbation wind. The next three terms represent adiabatic temperature changes resulting from perturbation vertical motion operating on synoptic-scale stability, synoptic-scale vertical motion operating on perturbation stability, and perturbation vertical motion acting on perturbation stability. The last term represents WL–NL diabatic heating differences that are occurring at the time.6 Note that a thermal dissipation term—for example, DTK2T/∂z2—is absent in (2). The magnitude of this term, which represents the difference in thermal dissipation between the WL and NL simulations, is relatively small—even in the boundary layer—compared to the other terms in (2).

The explicit partitioning in (2) is essential for accomplishing the objectives stated in the introduction. Assessment of the various terms will indicate the extent to which diabatic heating acted explicitly and implicitly [to cause MAV–MAV scale and MAV–synoptic-scale processes (SYN)] to produce the 1000-hPa height falls. For example, if the MAV–synoptic-scale processes (e.g., scale interactions) are not important, then all of the MAV–SYN terms would be relatively small compared to the MAV–MAV terms and compared to the diabatic heating term. Note that the present analysis method does not account for modifications of the background state because even large-scale modifications are considered to be part of the perturbation. But, the analysis method does account for the processes between the background state and those that are induced by the presence of the lakes. A more detailed discussion of the process versus scale perspective is provided in the appendix.

Equation (1) was solved for the height tendency (∂ZMAV/∂t)|pb at pb = 1000 hPa. To determine a suitable value for pt, height changes at various levels over 12 h periods were calculated. It was found that the height tendencies averaged over the region(s) of MAV development (cf. Figs. 5d, 8d, 11d, and 15d) at all levels above 600 hPa were less than 10% in magnitude of the height tendencies at 700, 800, 900, and 1000 hPa. It isappropriate to set (∂ZMAV/∂t)|600 ≈ 0, and to rewrite (1) as
i1520-0493-126-12-3169-e3

In the following analyses, (3) was evaluated by determining the local temperature tendency averages for each of four 100-hPa thick vertical layers. The contributions to the local temperature tendency for each layer and for each of the four 12-h time periods were determined as follows. First, the horizontal advection terms and adiabatic cooling terms were obtained at each mid level as temporal mean rates of change based on instantaneous rates of change at five times in a 12-h period (e.g., 0, 3, 6, 9, 12 h), time-averaged using the trapezoidal rule, and then converted to 12 h temperature changes. The vertical motion ω was assumed to be zero at 1000 hPa, but values at other levels were obtained kinematically7 by computing mean horizontal convergences over 100-hPa thick layers. The diabatic heating term QMAV/Cp was then computed at each mid level as a residual from (2) and interpreted to be the result of lake-related processes, because other WL–NL diabatic heating differences such as those from radiation were insignificant. This residual strategy was used because the Pennsylvania State University/National Center for Atmospheric Research PSU/NCAR mesoscale (MM4 and MM5) models do not output values of surface heat fluxes8 or diabatic heating rates (e.g., along with other variables such as wind components, temperature, and specific humidity) without user-implemented code modifications.9 Thus, there would not have been any real advantage in this respect to resimulating the case with a more recent version of the PSU/NCAR model such as MM5. Other mesoscale modeling studies such as Kuo et al. (1991a), Kuo et al. (1995), and Stoelinga (1996) have assessed the impacts of diabatic heating and surface fluxes implicitly by turning appropriate model switches “on” and “off” and then comparing the results.

The inaccessibility of model-generated diabatic heating output does not preclude the validity of the residual approach, especially when explicit output for all other significantly contributing terms (processes) exists at high spatial (30 km) and high temporal (3 h) resolution.10 A comparison of the computed height tendencies (using Eqs. (3) and (4) in the manner just described) to the height tendencies obtained as 12-h differences shown in Fig. 3 serves as an independent check on the applicability of using (1)–(3). A comparison of the diabatic heating contributions for the height tendencies to the surface heat fluxes shown in Fig. 3 serves as a partial check on the applicability of computing diabatic heating effects as residuals.

Figure 5 illustrates the contributions of the various terms in (2)–(3) to the 12-h height falls between 0 and 12 h, when a broad region of height falls extended across the entire lakes region and when the strongest height falls in excess of 30 m occurred along the western shores of northern lower Michigan, near Traverse City, Michigan. The height fall pattern in Fig. 5d compares favorably with the 12-h height changes at 1000 hPa shown in Fig. 3a. The slight differences are a result of using 100-mb thick layers and truncating the calculation at 600 hPa.

Inspection of the various contributions in Figs. 5a–c indicates that the largest 1000-hPa height falls were primarily the result of strong diabatic heating over Lakes Superior and Michigan. The heating was the result of strong, cold synoptic-scale northwesterly flow across the warm lake surfaces. This flow generated combined surface sensible and latent heat fluxes exceeding 700 W m−2 early in the period over northern Lake Superior (cf. Fig. 3a), where lake–air temperature and moisture differences were largest. The fact that the diabatic heating contribution shown in Fig. 5c compares favorably11 with the surface heat flux pattern shown in Fig. 3a, and that warming was greatest in the lowest 100-hPa layer and decreased monotonically to ∼800 hPa (cf. Fig. 6a), add credibility to the diabatic heating being calculated as a residual.

Although a strong temperature perturbation had developed rapidly at low levels as shown in Fig. 7a, the perturbation winds were still weak and subgeostrophic so that strong warm advection from −VSYN · TMAV was offset by strong cold advection from −VMAV · TMAV. Additionally, the temperature advection from −VMAV · TSYN was weak so that height falls from the sum of the three horizontal advection components were negligible (cf. Figs. 6a,b). Temperature changes from −ωMAVSMAV and −ωMAVSSYN (cf. Fig. 6c) essentially canceled, so that −ωSYNSMAV generated adiabatic cooling12 and height rises that offset much of the warming and height falls from diabatic heating between 900 and 800 hPa.

Figures 3b and 8d show that the strongest 1000-hPa height falls between 12 and 24 h were located near Sault Ste. Marie, Michigan, and were significantly less than they were during the previous 12 h. This lessening occurred as the synoptic-scale winds subsided, which reduced the surface fluxes (cf. Fig. 3b) and hence the contribution from diabatic heating (cf. Figs. 8c, 9a). The net contribution from the three horizontal advection components, however, was greater during the 12–24 h period despite the weaker winds (cf. Figs. 8a, 9a). This increase resulted primarily because warm advection was beginning to develop from −VMAV · TSYN, as perturbation southerly flow on the east side of the thermal perturbation began advecting the synoptic-scale temperature pattern (cf. Fig. 10b). The importance of the contribution from −VMAV · TSYN is that it was developmental rather than advective, like that from −VSYN · TMAV. The strong southerly perturbation flow was developing as a result of the east–west perturbation temperature gradient that had been strengthening since the previous 12-h period. Height rises from adiabatic cooling were less during this period primarily because cooling from −ωSYNSMAV had weakened (cf. Figs. 8b, 9a,c).

Figures 3c and 11d show that the 1000-hPa height falls shifted between 24 and 36 h to the northern portions of Lake Superior and Lake Huron with magnitudes that were characteristic of those during the first 12-h period. This northwest–southeast height fall pattern represented a change in the pattern that existed during the first 24 h when the strongest height falls were in a north–south direction. This changed height fall pattern likely resulted as the synoptic-scale flow shifted from northwesterly to westerly over the region and as the temperature perturbation expanded northeastward as shown in Figs. 2b and 13a.

The similar shapes and intensities of the local height tendency pattern in Fig. 11d, the contribution from horizontal temperature advection in Fig. 11a, and that from diabatic heating in Fig. 11c suggest that warm advection was becoming as significant as diabatic heating. Specifically, Figs. 12b and 13 show that strong southerly perturbation winds, which had intensified because of developing temperature and height perturbations, combined with increasingly westerly synoptic-scale winds and north–south synoptic-scale temperature gradients to provide increased warm advection from −VSYN · TMAV and from −VMAV · TSYN. Both of these processes contributed significantly in the area of strongest height falls. Warm advection over the same region from these same processes did not yet exist at higher levels (cf. Figs. 12b and 14) although anticyclonic flow had developed at 700 hPa as part of the perturbation outflow. The associated perturbation northerly flow that existed farther to the northeast as part of this outflow did result in cold advection from −VMAV · TSYN at that level and did restrict the northeastward extent of perturbation height falls at 1000 hPa. This restriction enhanced the perturbation height fall gradient along the east side, which enhanced the perturbation isallobaric wind, which later enhanced the southerly perturbation flow and hence the warm advection at lower levels.

The diabatic heating contribution to the height falls that existed during the 24–36 h period extended several hundred kilometers northeast from Lake Huron. This horizontal extension (cf. Fig. 11c) and its vertical distribution (cf. Fig. 12a) suggest that the heat and moisture that had been added and that were being added to the atmosphere at the surface over Lake Huron from southwesterly flow was being manifested as height falls from sensible and latent heating farther downwind (e.g., inland), where synoptic-scale warm advection and ascent were beginning to occur. Although the net contribution from adiabatic motions remained small (cf. Fig. 12c), the individual contributions increased. Specifically, strong adiabatic cooling from −ωMAV · SSYN indicated that perturbation ascent was strengthening.

Figures 3d and 15d both show that the strongest height falls between 36 and 48 h were located several hundred kilometers northeast of Lake Huron. The height falls during this period were the greatest in any of the four 12-h periods. The strong height falls were the result of warming that occurred at all levels between 600 and 1000 hPa (cf. Fig. 16a). The similar shapes and intensities of the local height tendency pattern in Fig. 15d, and that from horizontal temperature advection in Fig. 15a suggest that warm advection was the most significant process. Figures 16 and 17 show that the strong contribution from warm advection continued to be the result of increasingly strong low-level southerly perturbation flow on the east side of the MAV. However, the warm advection between 36 and 48 h differed from that between 24 and 36 h in at least four significant ways. First, the extent of the warm advection from −VMAV · TSYN was narrow, being only ∼200-km wide and extending along the northern shores of Lakes Superior and Huron from 24 to 36 h. From 36 to 48 h, the warm advection from −VMAV · TSYN was ∼500 km wide. Second, the magnitude of the contribution from −VMAV · TSYN was comparable only to that from −VSYN · TMAV at 900 hPa from 24 to 36 h but nearly twice as great from 36 to 48 h. Third, the warm advection from −VMAV · TSYN extended only over a shallow layer from 24 to 36 h but over a deep layer from 36 to 48 h. Fourth, the contribution from −VMAV · TMAV was negative before 36–48 h but positive from 36 to 48 h. The contribution at 700 hPa in Fig. 18 reveals that this deeper contribution was not the result of deeper cyclonic flow over the region. Rather, it was the result of southerly flow around the western edge of anticyclonic outflow located at 700 hPa, which had advected past (e.g., farther northeastward than) the lower-level cyclonic circulation because of stronger upper-level winds.

The contribution from adiabatic motions indicates that a net cooling and hence a net increase in 1000-hPa perturbation heights occurred from 36 to 48 h (cf. Fig. 15b). Figure 16c indicates that the adiabatic cooling was primarily the result of −ωMAV · SSYN. This cooling contribution completely overwhelmed the combined warming from −ωSYN · SMAV and −ωMAV · SMAV and demonstrates that the increased destabilization from the lake aggregate heating actually increased adiabatic cooling. This increased cooling adds evidence to the notion that the MAV did not develop simply because heat from the lakes destabilized the low levels and reduced adiabatic cooling.

The location and extent of the diabatic heating shown in Fig. 15c indicate that latent heating was likely being generated by convective-scale ascent from air that was sensibly heated and moistened as it moved across the lakes, and by strong MAV-scale ascent that was being generated by warm advection from −VMAV · TSYN (cf. Fig. 17b).13 The 700-hPa wind vectors that are valid at 42 h in Fig. 18 substantiate these heating mechanisms, although a detailed airstream trajectory analysis would be necessary to identify the relative contributions. Regardless of which of these mechanisms was more important, it is likely that the surface fluxes contributed at least indirectly by enhancing the warm advection and hence the large-scale ascent and height falls over the region. Additionally, the surface heat fluxes increased over the lakes because of the increased perturbation height falls and winds. The presence of MAV-enhanced surface fluxes suggests that a mechanism like wind-induced surface heat exchange (WISHE) may have played a role (cf. Emanuel 1995; Craig and Gray 1996). The WISHE hypothesis states that a perturbation intensifies because of a feedback that depends on the strength of the perturbation and on the strength of the surface heat and moisture fluxes. As the perturbation intensifies, stronger winds at the surface increase the surface heat exchange, which increases the strength of the perturbation.14

Evidence for this feedback mechanism exists in the current case. The evidence can be seen more clearly and heuristically by separating the synoptic-scale contribution from the MAV contribution to the surface fluxes. The contribution from the MAV to the WL surface fluxes relative to the synoptic-scale contribution can be assessed by considering from a Bulk Aerodynamic perspective that portion of the fluxes ΔFMAV that result from wind and temperature perturbations
i1520-0493-126-12-3169-e4
where TWLG and qWLS represent the lake surface temperature and saturated specific humidity in the WL simulation, TNLA and qNLA represent the surface air temperature and specific humidity in the NL simulation, and all other variables have their usual meaning. The modification to the sensible heat flux is a result of perturbation-altered wind speed |VWL| − |VNL| acting on the existing lake–air temperature difference TWLGTNLA and the actual wind speed |VWL| acting on the temperature perturbation TMAV. A similar interpretation exists for the last two terms in (4) regarding the latent heat flux. The consequence of (4) is that a perturbation that generates a circulation at the surface in the same sense as the synoptic-scale circulation (e.g., enhanced wind speed), or that exceeds the strength of the synoptic-scale circulation, will increase fluxes. A perturbation that generates a positive temperature modification will reduce fluxes because the lake–air temperature difference will have been reduced. A similar consequence exists for the last two terms in (4) regarding the latent heat flux.

Table 3 indicates the amount by which the MAV enhanced the surface fluxes during the 48-h period. From 0–12 h, the perturbation winds, temperatures, and specific humidities enhanced the combined surface sensible and latent heat fluxes in the WL simulation by 100–200 W m−2 over Lakes Superior, Michigan, and Huron. Almost no enhancement occurred over Lake Erie, while a reduction occurred over Lake Ontario because of significantly reduced lake–air temperature differences. For all five lakes, the MAV enhanced the surface fluxes by ∼25%. From 12 to 24 h, the MAV enhanced the surface fluxes over Lakes Superior and Michigan, but reduced them over Lakes Huron, Erie, and Ontario as the cyclonic perturbation circulation at the surface intensified. This pattern of MAV-altered fluxes resulted because the developing perturbation flow direction was correlated with the synoptic-scale flow direction over the western lakes but anticorrelated with it over the eastern lakes. Thus, |VWL| − |VNL| was large over the western lakes and small over the eastern lakes. For all five lakes, the MAV still accounted for an enhancement of the surface fluxes by ∼20%. From 24 to 36 h, the developing MAV reduced the surface fluxes over Lakes Superior and Michigan but enhanced them over Lakes Huron, Erie, and Ontario. This pattern occurred because by that time the synoptic-scale flow had become southwesterly so that the developing perturbation flow direction was correlated with the synoptic-scale flow direction over the eastern lakes but anticorrelated with it over the western lakes. Thus, |VWL| − |VNL| was large over the eastern lakes and small over the western lakes. The enhancement was especially evident over Lake Huron, where MAV-enhanced fluxes at 30 h were more than 100 W m−2 higher than those contributed from the synoptic-scale flow (∼50 W m−2). The surface fluxes over Lakes Erie and Ontario also increased considerably toward the end of the 24–36 h period. From 36–48 h, the strengthening perturbation southerly flow of the MAV continued to enhance surface fluxes. In general, the surface fluxes were nearly doubled by the developing perturbation during this last 12-h period. This increase was the largest of the four 12-h periods and resulted from a phasing between the cyclonic perturbation circulation and the synoptic-scale cyclonic flow (e.g., trough) over the region. That is, northwesterly synoptic-scale flow on the windward side of the weak trough was enhanced considerably by northwesterly flow on the windward side of the MAV and southerly synoptic-scale flow on the lee side of the weak trough was enhanced considerably by southerly flow on the lee side of the MAV.

The MAV-enhanced surface heat fluxes likely contributed directly and indirectly to perturbation height falls. Specifically, the relative increase in surface fluxes from 24–48 h over the eastern lakes enhanced the diabatic heating that contributed to height falls directly (i.e., hydrostatically). In turn, the diabatically heated air enhanced the perturbation temperature gradient, which led to increased warm advection from −VSYN · TMAV and therefore contributed to height falls indirectly (i.e., dynamically). The increased perturbation height falls then led to increased perturbation wind speeds, which contributed to increased diabatic heating from increased surface heat fluxes and also from increased warm advection via −VMAV · TSYN and from increased perturbation ascent.

4. Summary and conclusions

Previous studies (e.g., Petterssen 1956; Danard and McMillan 1974; Reitan 1974; Sousounis and Fritsch 1994; Angel and Isard 1997) have noted the impacts of the Great Lakes on developing cyclones that move through the region in winter. A study by Sousounis (1997) has described the characteristics of a weak low that developed from heating and moistening by all the lakes, which he called a mesoscale aggregate vortex (MAV). These and other studies have suggested that cyclone development in the Great Lakes region is the result of strong diabatic heating and low-level destabilization from the lakes. The exact mechanisms, however, by which this heating and moistening lead to the development of SLP falls, and the importance of diabatic heating relative to other processes, particularly for weak systems over the lakes, have not been investigated previously. In this study, detailed analyses of model output that included all of the Great Lakes (WL) and none of the Great Lakes (NL) for the November 1982 case described in Parts II and III were performed to understand more completely the specific roles that synoptic-scale forcing, diabatic heating, and MAV–synoptic-scale processes played in the in situ development of a MAV.

From 0–12 h, strong development at 1000 hPa (∼35 m 12 h−1) occurred over northern Lake Michigan and was primarily the result of diabatic heating as strong synoptic-scale northwesterly flow advected cold air over the Great Lakes. The developing MAV contributed ∼30% to the surface fluxes over the western lakes from northerly perturbation flow that was developing on the western side of the MAV because increases in wind speeds were more significant than decreases in lake–air temperature differences.

From 12–24 h, weak development at 1000 hPa (∼15 m 12 h−1) occurred despite decreased synoptic-scale cold advection, negative vorticity advection, subsidence, and constant static stability. This slowing occurred primarily because the synoptic-scale winds subsided, which reduced the diabatic heating. The slowing also occurred even though warm advection from −VMAV · TSYN was beginning to develop from enhanced southerly perturbation flow. The enhanced low-level cyclonic perturbation flow, superimposed on weak synoptic-scale northwesterly flow, resulted in enhanced surface heat fluxes over the western lakes but reduced heat fluxes over the eastern lakes.

From 24 to 36 h, strong development at 1000 hPa (∼30 m 12 h−1) occurred once again along the northern shore of Lake Huron as the synoptic-scale flow became southwesterly ahead of a weak trough. This change in the large-scale flow allowed synoptic-scale warm advection, positive vorticity advection, weak ascent, and low-level convergence to develop, although the static stability remained high. Perturbation southerly flow increased significantly during this period, which increased surface heat fluxes over the eastern lakes and which also increased warm advection and height falls from −VMAV · TSYN. These height falls were significant because they represented development and not just advection of the MAV.

From 36 to 48 h, strong development at 1000 hPa (∼35 m 12 h−1) continued several hundred kilometers to the northeast of Lake Huron. Warm advection from −VMAV · TMAV occurred for the first time. Additionally, warm advection from −VMAV · TSYN increased significantly on the east side, which likely resulted in latent heating from enhanced perturbation ascent. The increased depth of the warm advection was the result of low-level (900 hPa) cyclonic perturbation flow underneath upper-level (700 hPa) anticyclonic perturbation flow that was shifted farther downwind. The low-level perturbation cyclonic circulation was now in phase with the weak synoptic-scale trough over the region, which resulted in a ∼50% increase to the surface heat fluxes over the eastern Great Lakes and to further development of the MAV. The relationship between MAV-enhanced surface heat fluxes and an intensified MAV circulation suggests that a WISHE-like mechanism may have been present.

The following conclusions are made, based on the findings in this study. 1) The SLP falls that developed over the Great Lakes between 0000 UTC 13 and 0000 15 UTC November 1982 were not simply the hydrostatic result of heat from the Great Lakes “spreading” over a large region. 2) The evolution of the synoptic-scale flow, from northwesterly winds with strong cold advection to southwesterly winds with warm advection, was important for MAV development. 3) The eventual collocation of strong cyclonic perturbation southerly winds at 900 hPa, strong anticyclonic perturbation southerly winds at 700 hPa, and east–west-oriented synoptic-scale isotherms in between, was important for MAV development. 4) The superpositioning of MAV-scale cyclonic flow and synoptic-scale cyclonic flow from an approaching weak trough was important for significantly enhancing the surface sensible and latent heat fluxes and warm advection, and for further MAV development.

The one case that has been examined does not likely serve as an explanation for all MAVs. For example, it is not known how MAV development would have occurred had the synoptic-scale flow not become as favorable as it did. Development could have still occurred, albeit more slowly, even if synoptic-scale warm advection, positive vorticity advection, ascent, and decreased static stability had not overspread the region. Evidence for this likelihood comes from the fact that development did occur near the surface during the first two periods under unfavorable large-scale conditions. Also, although a ∼24-h lag was observed in this particular case between the strongest synoptic-scale surface heat fluxes existing over the region and the MAV reaching maturity, it is not certain whether the lag was necessary, or more specifically, whether the heat and moisture needed to be distributed over the region before development, or whether a shortwave at upper levels could generate a MAV without previously distributed heat and moisture but from concomitant sensible and latent heating. It is possible that MAV development from a shortwave (e.g., trough) and coincident lake-aggregate heating and moistening may involve a different sequence of events (processes). Such a case may occur from a true “Alberta Clipper” that propagates southeastward from Canada over the Great Lakes; although the flow at low and midlevels is cyclonic, winds are not so strong from the south ahead of the trough that warm air is overspreading the lakes. The cold westerly or southwesterly flow ahead of the shortwave may therefore serve to maintain direct diabatic heating and moistening at the same time that other dynamical processes are operating over the lakes.

Additional studies are needed to determine the climatology of the different types of weak lows that pass through the Great Lakes region in winter and the different types of MAVs that develop over the region. Additional studies are also needed to understand more completely the importance of the prevailing wind direction relative to the lake-aggregate geometry and how MAV circulations actually evolve from multiple, meso-β-scale adjacent circulations into a single, meso-α scale circulation. Further study is also needed to determine how kinetic energy is generated and transferred among the convective scale, the individual lake scale, and lake-aggregate scale. Such studies will likely require the analyses of WL, NL, and individual-lake simulations.

Acknowledgments

The author wishes to thank several anonymous reviewers for their many helpful suggestions and comments. This research was supported by NSF Grant ATM-9502009.

REFERENCES

  • Alpert, P., M. Tsidulko, and U. Stein, 1995: Can sensitivity studies yield absolute comparisons for the effects of several processes? J. Atmos. Sci.,52, 597–601.

  • Angel, J. R., and S. A. Isard, 1997: An observational study of the influence of the Great Lakes on the speed and intensity of passing cyclones. Mon. Wea. Rev.,125, 2228–2237.

  • Anthes, R. A., E.-Y. Hsie, and Y. H. Kuo, 1987: Description of the Penn State/NCAR Mesoscale Model Version 4 (MM4). NCAR Tech. Note NCAR/TN-282+STR, 66 pp. [Available from National Center for Atmospheric Research, P.O. Box 3000, Boulder, CO 80307.].

  • Boudra, D. B., 1981: A study of the early winter effects of the Great Lakes. Part I: Comparison of very fine scale numerical simulations with observed data. Mon. Wea. Rev.,109, 2507–2526.

  • Carlson, T. N., 1991: Mid Latitude Weather Systems. Harper-Collins Academic, 507 pp.

  • Cox, H. J., 1917: Influence of the Great Lakes upon movement of high and low pressure areas. Proc. Second Pan Amer. Sci. Congress,2 (2), 432–459.

  • Craig, G. C., and S. L. Gray, 1996: CISK or WISHE as the mechanism for tropical cyclone intensification. J. Atmos. Sci.,53, 3528–3540.

  • Danard, M. B., and G. V. Rao, 1972: Numerical study of the effects of the Great Lakes on a winter cyclone. Mon. Wea. Rev.,100, 374–382.

  • ——, and A. C. McMillan, 1974: Further numerical studies of the effects of the Great Lakes on winter cyclones. Mon. Wea. Rev.,102, 166–175.

  • Emanuel, K. E., 1995: The behavior of a simple hurricane model using a convective scheme based on subcloud layer entropy equilibrium. J. Atmos. Sci.,52, 3960–3968.

  • Kuo, Y.-H., and R. J. Reed, 1988: Numerical simulation of an explosively deepening cyclone in the eastern Pacific. Mon. Wea. Rev.,116, 2081–2105.

  • ——, ——, and S. Low-Nam, 1991a: Effects of surface energy fluxes during the early development and rapid intensification stages of seven explosive cyclones in the western Atlantic. Mon. Wea. Rev.,119, 457–476.

  • ——, M. A. Shapiro, and E. G. Donall, 1991b: The interaction between baroclinic and diabatic processes in a numerical simulation of a rapidly intensifying extratropical cyclone. Mon. Wea. Rev.,119, 368–384.

  • ——, J. R. Gyakum, and Z. Guo, 1995: A case of rapid continental mesoscale cyclogenesis. Part I: Model sensitivity experiments. Mon. Wea. Rev.,123, 970–997.

  • Monobianco, J., 1989: Explosive east coast cyclogenesis: Numerical experimentation and model-based diagnostics. Mon. Wea. Rev.,117, 2384–2405.

  • Nuss, W. A., and R. A. Anthes, 1987: A numerical investigation of low-level processes in rapid cyclogenesis. Mon. Wea. Rev., 115, 2728–2743.

  • Petterssen, S., 1956: Weather Analysis and Forecasting. Vol. 1. 2d ed. McGraw-Hill, 428 pp.

  • ——, and P. A. Calabrese, 1959: On some weather influences due to warming of the air by the Great Lakes in winter. J. Meteor.,16, 646–652.

  • Rausch R. L. M., and P. J. Smith, 1996: A diagnosis of a model-simulated explosively developing extratropical cyclone. Mon. Wea. Rev.,124, 875–904.

  • Reitan, C. H., 1974: Frequencies of cyclones and cyclogenesis for North America, 1951–1970. Mon. Wea. Rev.,102, 861–868.

  • Roebber, P. J., 1984: Statistical analysis and updated climatology of explosive cyclones. Mon. Wea. Rev.,112, 1577–1589.

  • Sousounis, P. J., 1997: Lake aggregate mesoscale disturbances. Part III: Description of a mesoscale aggregate vortex. Mon. Wea. Rev.,125, 1111–1134.

  • ——, and J. M. Fritsch, 1994: Lake aggregate mesoscale disturbances. Part II: A case study of the effects on regional and synoptic-scale weather systems. Bull. Amer. Meteor. Soc.,75, 1793–1811.

  • Stein, U., and P. Alpert, 1993: Factor separation in numerical simulations. J. Atmos. Sci.,50, 2107–2115.

  • Stoelinga, M. T., 1996: A potential-vorticity-based study of the role of diabatic heating and friction in a numerically simulated baroclinic cyclone. Mon. Wea. Rev.,124, 849–874.

APPENDIX

Process versus Scale Interpretations

The analyses in this study utilize comparisons between WL and NL simulations. The strategy of comparing a control numerical simulation (WL) to one where a process or a geographic feature in this case (the Great Lakes) has been eliminated (NL) has long been used to identify the impacts of that process (or feature). However, care must be taken when interpreting the results (cf. Stein and Alpert 1993; Alpert et al. 1995). In general, the differences between the two simulations are the result of direct effects from that process and interactions between that process (feature) and other processes or features—whether they are being evaluated explicitly or not. In general, there are several ways to interpret the difference (perturbation) that results from a comparison between a control simulation and a sensitivity simulation. For example, because the perturbation χ̂ = perturbation temperature, winds, moisture, etc., may consist of a variety of spatial scales, one may interpret it to consist of both large-scale and small-scale contributions: χ̂MAV = χ̂SYN + χ̂MESO. The former contributions have sometimes been interpreted to be those that alter the background (large-scale) flow so that the full interpretation of the difference between the two numerical simulations is that a small-scale perturbation is generated and that there are modifications to the background state, and that the small-scale perturbation may interact with the (modified) background state. Thus, the total flow, according to scales, may be expressed as χWL = χSYN + χ̂MESO where χSYN represents the (large-scale) background flow that has been modified by the lakes.

In the current study, however, the lake-aggregate perturbation (e.g., MAV) is interpreted as the difference between the situations that evolved in the WL and NL simulations, independent of the scales that were involved. Thus, regardless of scale, regardless of whether the differences resulted from direct processes (e.g., linear effects) or indirect processes (e.g., nonlinear interactions), and regardless of whether the differences are collocated with the lakes or far removed, the WL–NL differences are all attributable to the presence of the lakes and so the differences are all interpreted as part of the lake-aggregate perturbation. Thus, the total flow, according to processes, may be expressed as χWL = χNL + χ̂MAV, where χ̂MAV represents all of the lake-induced modifications (perturbations) regardless of scale. However, it will be noted that in the case (or assumption) that the WL–NL differences consist primarily of small-scale contributions (e.g., χ̂SYNχ̂MESO), then χWLχSYN + χ̂MAV. Additionally, the prognostic equation for χMAV may be written by subtracting the prognostic equations for χWL and χNL
i1520-0493-126-12-3169-ea1
The first term on the rhs of (1), −VMAV · χSYN, represents the time-dependent changes from advection of the background (e.g., SYN or NL) quantity by lake-aggregate- (e.g., MAV) induced wind changes. The second term, −VSYN · χMAV, represents the time-dependent changes that result from advection of the lake-aggregate-modified quantity by the background wind. The third term, −VMAV · χMAV, represents the time-dependent change that results from advection of the lake-aggregate-modified quantity by the lake-aggregate-modified wind. The fourth term represents net sources that contribute to the time-dependent change of the lake-aggregate-induced quantity. The above interpretation of (A1) does not account for the various scales that comprise the lake-induced responses, but it does account for the processes between the background state and that which is induced by the presence of the lakes.

Fig. 1.
Fig. 1.

SLP distribution (hPa—leading 10 omitted) for WL simulation at (a) 33 h, (b) 36 h, (c) 39 h, (d) 42 h, (e) 45 h, and (f) 48 h. Contour interval is 1 hPa. Gray “L” indicates the low, which approached from the southwest, that was present in the NL simulation but not in the WL simulation.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 2.
Fig. 2.

Selected characteristics of the mesoscale aggregate vortex identified from WL–NL differences. (a) Perturbation SLP (−4 hPa contour) at times shown. (b) Perturbation 850-hPa temperature (−2 °C contour) at times shown. (c) Perturbation 12-hourly precipitation amounts (1-mm contour) at times shown.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 3.
Fig. 3.

Twelve-hour averages of WL–NL combined surface sensible and latent heat flux differences (solid contours—100 W m−2 contour interval) and 12-hourly 1000-hPa WL–NL height fall differences (dashed contours—7.5-m contour interval) less than −7.5 m during (a) 0–12 h, (b) 12–24 h, (c) 24–36 h, and (d) 36–48 h. Flux differences exceeding 100 W m−2 are shaded.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 4.
Fig. 4.

Time series plots of average sensible and latent surface fluxes, 1000-hPa MAV height falls, and 900-hPa horizontal heat flux convergence.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 5.
Fig. 5.

MAV (e.g., WL–NL) height fall tendencies at 1000 hPa between 0 and 12 h from (a) horizontal advection, (b) adiabatic cooling, (c) diabatic heating, and (d) sum of contributions a–c. Shaded regions indicate areas with height fall tendencies in excess of −7.5 m 12 h−1. Unshaded regions enclosed by solid contours indicate height rises in excess of +7.5 m 12 h−1. Contours shown at ±7.5 m, ±15 m, and then every 15 m. The small square in panel d indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for 0–12 h period in Fig. 6. The large square in panel d indicates most of a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 0–12 h period.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 6.
Fig. 6.

Profiles from 600–1000 hPa (vertical axis) of various terms in (3) averaged over the region shown in Fig. 5d for 0–12 h. Horizontal axis shows 12-h temperature changes (°C). Terms include local temperature tendency (∂T/∂t), horizontal temperature advection (−V · T) as the sum of the three terms plotted in panel b, adiabatic cooling term (−) as the sum of the three terms plotted in panel c, and diabatic heating (Q/Cp). See text for interpretation of subscripts MAV and SYN.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 7.
Fig. 7.

Contributions to horizontal temperature advection at 900 hPa at 06 h from horizontal advection components (a) −VSYN · TMAV and (b) −VMAV · TSYN. The contour interval for temperature is 1 °C in (a) and 2 °C in (b). Wind vector length equal to grid separation is 15 m s−1 in (a) and 5 m s−1 in (b). See text for interpretation of subscripts MAV and SYN.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 8.
Fig. 8.

Similar to Fig. 5 but for 12–24 h. The small square in (d) indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for a 12–24 h period in Fig. 9. The large square in (d) indicates a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 12–24 h period.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 9.
Fig. 9.

Similar to Fig. 6 but for 12–24 h.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 10.
Fig. 10.

Similar to Fig. 7 but for 18 h.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 11.
Fig. 11.

Similar to Fig. 5 but for 24–36 h. The small square in (d) indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for a 24–36 h period in Fig. 12. The large square in (d) indicates a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 24–36 h period.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 12.
Fig. 12.

Similar to Fig. 6 but for 24–36 h.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 13.
Fig. 13.

Similar to Fig. 7 but for 30 h.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 14.
Fig. 14.

Similar to Fig. 8 but for 30 h at 700 hPa.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 15.
Fig. 15.

Similar to Fig. 5 but for 36–48 h. The small square in (d) indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for 36–48 h period in Fig. 16. The large square in (d) indicates most of a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 36–48 h period.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 16.
Fig. 16.

Similar to Fig. 6 but for 36–48 h.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 17.
Fig. 17.

Similar to Fig. 7 but for 42 h.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Fig. 18.
Fig. 18.

Similar to Fig. 7 but for 42 h at 700 hPa.

Citation: Monthly Weather Review 126, 12; 10.1175/1520-0493(1998)126<3169:LAMDPI>2.0.CO;2

Table 1.

Values for selected NL parameters averaged over 600 km × 600 km regions centered on MAV development at 3-h intervals.From left to right starting with the second column are 1000 hPa perturbation height changes (m 3h−1), 850-hPa temperature advection (°C 12 h−1), 500-hPa vorticity advection (s−1 12h−1), stability (°C Pa−1), vertical motion (x10−3 hPa s−1), 850-hPa wind speed (m s−1), and wind direction (degrees of azimuth). Averaging regions are shown in Fig. 5d (0–12 h), Fig. 8d (12–24 h), Fig. 11d (24–36 h), and Fig. 15d (36–48 h).

Table 1.
Table 2.

Lake-averaged values of synoptic-scale contributions to surface fluxes (W m−2) at 3-h intervals. Values reflect combined sensible and latent heat fluxes. Last column (all) represents a lake-size area-weighted average.

Table 2.
Table 3.

Lake-averaged values of WL contributions to surface fluxes (W m−2) at 3-h intervals. Values reflect combined sensible and latent heat fluxes. Italicized values represent the percentage by which the MAV enhanced the flux (e.g., 50% means that the MAV contributed half the total).

Table 3.

1

The increased destabilization explains why the Great Lakes region is a nonoceanic region where explosive cyclogenesis occurs with some comparable frequency (cf. Roebber 1984).

2

This latter interpretation appears to be more accurate because (a) a region of low pressure already existed over Lake Superior when the synoptic-scale low was approaching from the southwest, (b) the weak synoptic-scale low continued to weaken (not strengthen) as it approached the lakes from the southwest while the low over the Great Lakes continued to strengthen, and (c) the weakened synoptic-scale low maintained its identity in the form of a trough as it migrated across the Great Lakes region, which caused a southwestward and then a southward extension of the low that was developing over the Great Lakes.

3

A more detailed comparison between the WL simulation and observations at the surface and upper levels is provided in Part II.

4

The Bowen ratio B ∼ 1, so that sensible and latent heat fluxes contribute equally to the total.

5

The synoptic-scale fluxes were computed using a bulk-aerodynamic approach with NL winds and WL lake temperatures and humidities. This approach was necessary in order to extract the contribution in the WL simulation from the synoptic-scale winds. The bulk aerodynamic approach yields values that are nearly identical to those obtained with the Blackadar planetary boundary layer scheme (Anthes et al. 1987) for the wind speed, temperature, and humidity ranges for this case.

6

It is understood that all of the perturbation effects (e.g., the existence of the MAV) have resulted from the cumulative effects of diabatic heating.

7

For hydrostatic models, the vertical motion ω has to be determined kinematically because the vertical momentum equation reduces to the hydrostatic assumption and hence there is no equation for the time dependency of ω.

8

A new version of MM5 was released in January 1998 that provides a user option to output surface sensible and latent heat flux values.

9

One model that apparently does provide this information is the Limited Area Mesoscale Prediction System (LAMPS) model used by Rausch and Smith (1996). Even this model, however, did not provide sensible heating rates, which had to be estimated using other methods.

10

The only other WL–NL difference terms that were not accounted for explicitly in (2) are from radiational cooling and dissipation—both of which are negligibly small.

11

A combined surface sensible and latent heat flux of 400 W m−2 that decreases linearly to zero at the top of a 1500 m deep cloud filled boundary layer contributes ∼90 m 12 h−1 to the 1000-hPa height fall pattern.

12

The adiabatic cooling came from synoptic-scale subsidence and negative perturbation static stability.

13

Profiles of TMAV and ωMAV over this region (not shown) suggest that sensible heating QSENS ∝ −(∂ωT′/∂p) > 0 was occurring just above the level where the maximum temperature perturbation and maximum ascent existed, near 900 hPa. Latent heating near 800 hPa was likely present because of strong perturbation ascent.

14

Although WISHE theory has generally been used to explain tropical development, the importance of surface heat fluxes for explosive extratropical cyclone development along the east coast of the United States has been noted in many previous studies (Nuss and Anthes 1987; Kuo and Reed 1988; Monobianco 1989; Kuo et al. 1991a).

Save
  • Alpert, P., M. Tsidulko, and U. Stein, 1995: Can sensitivity studies yield absolute comparisons for the effects of several processes? J. Atmos. Sci.,52, 597–601.

  • Angel, J. R., and S. A. Isard, 1997: An observational study of the influence of the Great Lakes on the speed and intensity of passing cyclones. Mon. Wea. Rev.,125, 2228–2237.

  • Anthes, R. A., E.-Y. Hsie, and Y. H. Kuo, 1987: Description of the Penn State/NCAR Mesoscale Model Version 4 (MM4). NCAR Tech. Note NCAR/TN-282+STR, 66 pp. [Available from National Center for Atmospheric Research, P.O. Box 3000, Boulder, CO 80307.].

  • Boudra, D. B., 1981: A study of the early winter effects of the Great Lakes. Part I: Comparison of very fine scale numerical simulations with observed data. Mon. Wea. Rev.,109, 2507–2526.

  • Carlson, T. N., 1991: Mid Latitude Weather Systems. Harper-Collins Academic, 507 pp.

  • Cox, H. J., 1917: Influence of the Great Lakes upon movement of high and low pressure areas. Proc. Second Pan Amer. Sci. Congress,2 (2), 432–459.

  • Craig, G. C., and S. L. Gray, 1996: CISK or WISHE as the mechanism for tropical cyclone intensification. J. Atmos. Sci.,53, 3528–3540.

  • Danard, M. B., and G. V. Rao, 1972: Numerical study of the effects of the Great Lakes on a winter cyclone. Mon. Wea. Rev.,100, 374–382.

  • ——, and A. C. McMillan, 1974: Further numerical studies of the effects of the Great Lakes on winter cyclones. Mon. Wea. Rev.,102, 166–175.

  • Emanuel, K. E., 1995: The behavior of a simple hurricane model using a convective scheme based on subcloud layer entropy equilibrium. J. Atmos. Sci.,52, 3960–3968.

  • Kuo, Y.-H., and R. J. Reed, 1988: Numerical simulation of an explosively deepening cyclone in the eastern Pacific. Mon. Wea. Rev.,116, 2081–2105.

  • ——, ——, and S. Low-Nam, 1991a: Effects of surface energy fluxes during the early development and rapid intensification stages of seven explosive cyclones in the western Atlantic. Mon. Wea. Rev.,119, 457–476.

  • ——, M. A. Shapiro, and E. G. Donall, 1991b: The interaction between baroclinic and diabatic processes in a numerical simulation of a rapidly intensifying extratropical cyclone. Mon. Wea. Rev.,119, 368–384.

  • ——, J. R. Gyakum, and Z. Guo, 1995: A case of rapid continental mesoscale cyclogenesis. Part I: Model sensitivity experiments. Mon. Wea. Rev.,123, 970–997.

  • Monobianco, J., 1989: Explosive east coast cyclogenesis: Numerical experimentation and model-based diagnostics. Mon. Wea. Rev.,117, 2384–2405.

  • Nuss, W. A., and R. A. Anthes, 1987: A numerical investigation of low-level processes in rapid cyclogenesis. Mon. Wea. Rev., 115, 2728–2743.

  • Petterssen, S., 1956: Weather Analysis and Forecasting. Vol. 1. 2d ed. McGraw-Hill, 428 pp.

  • ——, and P. A. Calabrese, 1959: On some weather influences due to warming of the air by the Great Lakes in winter. J. Meteor.,16, 646–652.

  • Rausch R. L. M., and P. J. Smith, 1996: A diagnosis of a model-simulated explosively developing extratropical cyclone. Mon. Wea. Rev.,124, 875–904.

  • Reitan, C. H., 1974: Frequencies of cyclones and cyclogenesis for North America, 1951–1970. Mon. Wea. Rev.,102, 861–868.

  • Roebber, P. J., 1984: Statistical analysis and updated climatology of explosive cyclones. Mon. Wea. Rev.,112, 1577–1589.

  • Sousounis, P. J., 1997: Lake aggregate mesoscale disturbances. Part III: Description of a mesoscale aggregate vortex. Mon. Wea. Rev.,125, 1111–1134.

  • ——, and J. M. Fritsch, 1994: Lake aggregate mesoscale disturbances. Part II: A case study of the effects on regional and synoptic-scale weather systems. Bull. Amer. Meteor. Soc.,75, 1793–1811.

  • Stein, U., and P. Alpert, 1993: Factor separation in numerical simulations. J. Atmos. Sci.,50, 2107–2115.

  • Stoelinga, M. T., 1996: A potential-vorticity-based study of the role of diabatic heating and friction in a numerically simulated baroclinic cyclone. Mon. Wea. Rev.,124, 849–874.

  • Fig. 1.

    SLP distribution (hPa—leading 10 omitted) for WL simulation at (a) 33 h, (b) 36 h, (c) 39 h, (d) 42 h, (e) 45 h, and (f) 48 h. Contour interval is 1 hPa. Gray “L” indicates the low, which approached from the southwest, that was present in the NL simulation but not in the WL simulation.

  • Fig. 2.

    Selected characteristics of the mesoscale aggregate vortex identified from WL–NL differences. (a) Perturbation SLP (−4 hPa contour) at times shown. (b) Perturbation 850-hPa temperature (−2 °C contour) at times shown. (c) Perturbation 12-hourly precipitation amounts (1-mm contour) at times shown.

  • Fig. 3.

    Twelve-hour averages of WL–NL combined surface sensible and latent heat flux differences (solid contours—100 W m−2 contour interval) and 12-hourly 1000-hPa WL–NL height fall differences (dashed contours—7.5-m contour interval) less than −7.5 m during (a) 0–12 h, (b) 12–24 h, (c) 24–36 h, and (d) 36–48 h. Flux differences exceeding 100 W m−2 are shaded.

  • Fig. 4.

    Time series plots of average sensible and latent surface fluxes, 1000-hPa MAV height falls, and 900-hPa horizontal heat flux convergence.

  • Fig. 5.

    MAV (e.g., WL–NL) height fall tendencies at 1000 hPa between 0 and 12 h from (a) horizontal advection, (b) adiabatic cooling, (c) diabatic heating, and (d) sum of contributions a–c. Shaded regions indicate areas with height fall tendencies in excess of −7.5 m 12 h−1. Unshaded regions enclosed by solid contours indicate height rises in excess of +7.5 m 12 h−1. Contours shown at ±7.5 m, ±15 m, and then every 15 m. The small square in panel d indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for 0–12 h period in Fig. 6. The large square in panel d indicates most of a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 0–12 h period.

  • Fig. 6.

    Profiles from 600–1000 hPa (vertical axis) of various terms in (3) averaged over the region shown in Fig. 5d for 0–12 h. Horizontal axis shows 12-h temperature changes (°C). Terms include local temperature tendency (∂T/∂t), horizontal temperature advection (−V · T) as the sum of the three terms plotted in panel b, adiabatic cooling term (−) as the sum of the three terms plotted in panel c, and diabatic heating (Q/Cp). See text for interpretation of subscripts MAV and SYN.

  • Fig. 7.

    Contributions to horizontal temperature advection at 900 hPa at 06 h from horizontal advection components (a) −VSYN · TMAV and (b) −VMAV · TSYN. The contour interval for temperature is 1 °C in (a) and 2 °C in (b). Wind vector length equal to grid separation is 15 m s−1 in (a) and 5 m s−1 in (b). See text for interpretation of subscripts MAV and SYN.

  • Fig. 8.

    Similar to Fig. 5 but for 12–24 h. The small square in (d) indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for a 12–24 h period in Fig. 9. The large square in (d) indicates a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 12–24 h period.

  • Fig. 9.

    Similar to Fig. 6 but for 12–24 h.

  • Fig. 10.

    Similar to Fig. 7 but for 18 h.

  • Fig. 11.

    Similar to Fig. 5 but for 24–36 h. The small square in (d) indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for a 24–36 h period in Fig. 12. The large square in (d) indicates a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 24–36 h period.

  • Fig. 12.

    Similar to Fig. 6 but for 24–36 h.

  • Fig. 13.

    Similar to Fig. 7 but for 30 h.

  • Fig. 14.

    Similar to Fig. 8 but for 30 h at 700 hPa.

  • Fig. 15.

    Similar to Fig. 5 but for 36–48 h. The small square in (d) indicates a 300 km × 300 km region of developing MAV over which average profiles are computed for 36–48 h period in Fig. 16. The large square in (d) indicates most of a 600 km × 600 km region over which NL quantities are averaged in Table 1 for a 36–48 h period.

  • Fig. 16.

    Similar to Fig. 6 but for 36–48 h.

  • Fig. 17.

    Similar to Fig. 7 but for 42 h.

  • Fig. 18.

    Similar to Fig. 7 but for 42 h at 700 hPa.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 97 37 6
PDF Downloads 49 27 1