CHAPTER 3
RESULTS
3.1 Overview
The numerical sensitivity simulations presented in this chapter investigate the role of snow
cover and boundary-layer clouds on the 1-6 February 2013 persistent CAP intensity and vertical
structure below about 1000 m AGL. Above that level, large-scale synoptic scale forcing is the
dominant driver (e.g., warm air advection, subsidence warming, mid-level clouds) of these
episodes, and does not vary between simulations.
The analysis and validation of the simulated CAPs in this study focuses on the vertical
characteristics of the CAP, rather than surface characteristics. Pollutant mixing depth and
transport within these complex, multiday CAPs depends on the atmospheric stablility and flows
within the lowest several 100 meters of the boundary-layer in most cases. A common problem in
air quality modeling and model validation of persistent cold air pools and a motivating factor in
conducting this research is the relative uselessness of commonly available surface wind
observations at point locations in determining advective fluxes within surface mixed layers of
varying depth, wind speed, and wind direction (Avey, personnal communication 2014). The
modeling results presented here provide the first coherent view of the spatiotemporal
characteristics of terrain- and thermally-induced flows within the basin.
This Chapter is organized as follows: First, an observed overview of the evolution of the 1-6
February 2013 persistent CAP in the Uintah Basin is presented in Section 3.2, followed by
analysis of the BASE WRF model simulation in Section 3.3. The sensitivity of the simulated
CAP to variations in cloud type and snow cover are presented in sections 3.3 and 3.4,
respectively. The mesoscale flows important to pollutant transport within the basin are discussed
in section 3.5.
3.1 1-6 February 2013 Persistent Cold Air Pool
A deep upper-level trough and associated midlatitude cyclone moved through the Great
Basin between 28-30? January 2013, bringing very cold air aloft (700 mb temperatures ~-20 ᵒC)
and 1-3 cm? of light snowfall on top of a ~10-20 cm base across most of the Uintah Basin.
Following the upper-level trough passage, 1-6 February was dominated by upper level ridging
over the western US and surface high pressure in the Great Basin (Fig. 3.0). This pattern placed
northeastern UT in a region of large-scale subsidence and mid-leveling warming as noted by the
subsidence inversion in Fig. 3.1a. With warm air aloft (700 mb temperatures between ~-7 and 0
ᵒC), quiescent surface weather conditions, fresh snow cover, cold low-level air in place, and
sufficient incoming solar insolation to drive photochemistry, the stage was set for a persistent
CAP event with high wintertime ozone concentrations in the Uintah Basin.
The surface temperatures from -5 to -15 °C observed at 1800 UTC (11 MST) 2 February
2013 (Fig. 2.1) provide an indication of the extent of the cold-air trapped in the lowest elevations
of the Uintah Basin during this period. Above-freezing temperatures at higher elevations are
common. At this time, ice fog and low stratus was observed in the lowest reaches of the basin
with relative humidity values generally in the 85-90% range (reflecting saturation with respect to
ice) while values drop to less than 50% on the slopes (not shown).
The vertical profile of temperature, dew point temperature, and wind from rawinsondes
released at Roosevelt on 2, 4, and 6 February 2013 at 1800 UTC are shown in Fig. 3.1. A
shallow layer of fog is evident on 2 and 4 February below 850 hPa and is capped by a strong
inversion extending upwards to 700 hPa. Weak easterly winds within the inversion are present
with stronger westerly winds above. The capping inversion between 800 and 700 hPa (500 and
1500 m AGL) remained almost constant between the 2 and 4th February, with slight warming at
this level noted on 6 February associated with a descending subsidence inversion induced by
downsloping northwesterly flow over the high Uintah Mountains. Below 825 hPa (400 m AGL),
the thickness and spatial extent of the surface-based clouds and fog within the Basin (and
possibly other unknown factors) resulted in day to day variations in the mixed-layer depth during
the period. Between 2 and 4 of February, the cloudy mixed-layer depth was observed to deepen
from ~150 to ~300 m AGL, followed by subsequent lowering of the mixed layer depth to ~150
m on 5 and 6 February associated with a decrease in low clouds during this period and the
passage of a weak upper-level trough (not shown). This deepening of the mixed layer when the
low clouds and fog in the basin peaked (in spatial extent and temporal occurrence) during the
middle of the episode and the subsequent lowering of the mixed layer when the clouds decreased
later on was likely a critical factor in modulating the boundary-layer ozone concentrations noted
during this period.
The spatial coverage of snow and cloud on 2 February 2013 is shown in a snow-cloud
product from NASA SPoRT in Fig. 3.2a. This mid-day image depicts the lower elevations of the
Uintah Basin covered in snow with less snow in its far western extremity (as mentioned in the
section 2.3). Fog and stratus are confined to the lowest elevations of the basin. An earlier VIIRS
nighttime microphysics RGB product from SPoRT (0931 UTC 2 February, Fig 3.2b) helps to
detect low clouds and discriminate particle phase by combining data from the 3.9, 10.8, and 12.0
micron infrared channels. In this image, liquid-phase low stratus and fog are evident in terms of
aqua/green colors in southern ID and portions of western and central UT while ice-phase stratus
and fog in the Uintah Basin are evident by the yellow/orange colors.
Figure 3a presents the time evolution of surface ozone measured at Ouray, Roosevelt, and
Vernal during 1-6 February 2013. The concentrations were noted to exceed EPA standards of 75
ppbv over XX% of the time during this event. Figure 3.3b presents the time evolution of aerosol
backscatter, low clouds, and an estimate of the depth of the aerosol layer from the Roosevelt
laser ceilometer during the 1-6 February period. Fewer aerosols were observed on 1 February
followed that evening by the development of ice fog evident in Fig. 3.2b. Then, a semi-regular
pattern developed over the next several days with shallow nighttime fog and low clouds thinning
by mid-day and followed by a deeper layer of aerosols in the afternoon that quickly collapses at
sunset. The ceilometer backscatter data also corroborates other observations that the fog and low
cloud occurrence in the basin peaked on the 3 and 4 of February. Significant hoar frost was
observed on trees and other surfaces after sunrise with light accumulations of snow crystals
falling out of the ice clouds in Roosevelt later in the morning. The high levels of aerosol
backscatter on 5-6 February diminished abruptly during late afternoon on 6 February as a result
of a weak weather system ; however elevated ozone levels continued until 9 February, when a
stronger cold front (?) moved through the region.
3.2 1-6 February 2013 BASE Model Simulation
As discussed in Section 3.1,, the large-scale synoptic scale forcing of the cold-air pool from
above the boundary layer was captured well by the BASE model simulation. Simulated potential
temperature profiles above about 500 m AGL agree well with observations (Fig. 3.6). However,
the BASE model simulation of the persistent CAP results in unrealistically deep cloudy mixed
layers (Fig. 3.6).
The liquid-cloud layer starts out as fog during the first couple days of the simulation, before
lifting into a low stratus layer after 1800 UTC 2 February. While the fog/stratus does break up
on most afternoons, its persistence greatly hinders cooling overnight, disrupts the strength of the
near-surface layers of the CAP, and results in sub-cloud mixed layers that are 2-4 times deeper
than those observed (Fig 3.6). Comparison in Fig. 3.6 of potential temperature vertical profiles
taken at Roosevelt at 1800 UTC on 1-6 February to those simulated by the BASE model
illustrate these discrepancies. For example, the depth of the mixed layer beneath the strong stable
layer is nearly twice as high as observed on 1 and 6 February and the near-surface temperatures
are much lower than observed (see also Fig. 3.5). The low temperatures on these two days can
be attributed to the overproduction of fog/stratus by the WRF model, resulting in decreased
surface insolation. As evident from the aerosol backscatter (Fig. 3.3) and soundings at 1800
UTC (Fig 3.1, 1 Feb not shown), no clouds were present by mid-day on those two days. In
contrast, temperatures in the boundary layer are generally 2-5 °C too high through the depth of
the PBL from the 2nd through the 5th with PBL heights too deep by 200 m or more as a result of
simulating liquid clouds as opposed to the observed ice clouds. The strength of the stability
within the inversion layer atop the surface mixed layer is handled relatively well by WRF. This
is due to strong radiative cooling at the top of the cloud layer produced by the model and results
in a lapse rate similar to what was observed.
Examination of 2-m temperatures from the BASE model simulation shows unrealistic spatial
patterns for much of the period associated with the cloud overproduction (compare Fig. 3.4a to
Fig. 2.1). Warmer air in the BASE simulation is found at low elevations of the basin, surrounded
by a ring of colder air midway between the basin floor and mountain crests. The warm core is
related to downwelling longwave radiation originating from the dense low-level liquid clouds
formed by the WRF microphysics parameterization scheme that were not present in the basin.
Figure 3.5 shows the temporal evolution of the temperature bias (model – observed) at the 6
surface stations in the center of the basin that are highlighted in Figs. 1.5b and 2.1. Large warm
biases are evident from the 2nd to the 5th with the overall bias of 1.7 °C when averaged over the
entire simulation and a mean absolute error of 3.3 °C (Table 3.1).
The structure of the boundary layer shows extensive variation over the course of the BASE
simulation (Fig. 3.7). Early in the period on 1 February, the persistent CAP initially forms with
an inversion base near 1600-1700 m mean sea level (MSL) (Fig. 3.7a). It then deepens through
the day on 2 February to a base near 2100 m at 0600 UTC 3 February (Fig. 3.7b) before settling
near 2000 m by 1800 UTC 4 February (Fig. 3.7c). The CAP then lowers back down to the 17001800 m range from 0000 UTC 6 February through the end of the simulation (Fig. 3.7d). A
similar pattern is also depicted in the time-height plot of potential temperature at Horsepool in
Fig. 3.8a. Shallow mixed layers are present each afternoon for a few hours near 0000 UTC,
when potential temperature is nearly constant with height. The complex evolution in CAP depth
is continually modulated by synoptically driven mid-level flow across the top of the CAP near
2300 m. The variations in westerly flow and wind stress atop the CAP lead to a “sloshing”
behavior with inhomogeneities in the CAP depth and inversion base across much of the basin.
One example of this is shown by the temporary partial mix-out of the CAP near Starvation
Reservoir along the western slope of the Uintah Basin at 0600 UTC on 4 February. Figure 3.9a
shows a narrow band of increased westerly to northwesterly flow at 2300 m over the western
portion of the basin. A cross-section of potential temperature from west to east through the
center of the basin at the same time is shown in Fig. 3.9b. By 0600 UTC on 4 February, the
westerly downslope winds have eroded the CAP and pushed it east of Starvation Reservoir
(vertical line labeled “STA” in Fig. 3.9b) with the inversion lowering to around 1800 m, notably
depressed from the eastern half of the basin. Just two hours later, by 0800 UTC, the CAP
rebounds westward past Starvation Reservoir and the inversion base quickly rises to near 2000
m, approximately leveling out with the rest of the basin and the 280 K contour (not shown).
3.3 Sensitivity to Cloud Type
As discussed in the previous section, the default settings in the Thompson microphysics
scheme (BASE) over-produce dense, liquid-phase low clouds and fog that notably alter the
radiation budget within the basin, and consequently, the strength of the CAP. The extensive fog
and stratus, particularly during the overnight hours, result in excessive longwave radiation that
inhibits low-level cooling and forms a region of warm air in the center of the basin during much
of the CAP period. For this reason, modifications to the Thompson microphysics scheme were
employed in the FULL model run to examine the sensitivity of the CAP simulation to cloud type.
The intent was to force the WRF model to produce and maintain clouds dominated by ice-phase
particles.
The 2-m temperatures at 1800 UTC 2 February from the FULL simulation (Fig. 3.4b) are
much lower in the center of the basin and appear closer to those observed (Fig. 2.1). Although
the bias and mean errors of the FULL simulation are sharply reduced relative to the 6-station
sample (Table 3.1), ice clouds form on the 1st and continue on the 6th, times without clouds
observed, which results in lower temperatures than observed (Fig. 3.5). The depth of the surface
mixed layer in the FULL simulations tends to be closer to those observed as well (Fig. 3.6),
particularly during the heart of the CAP from 2-5 February (Figs. 3.6b-e). Additional differences
in FULL are displayed in the potential temperature time-height plot Fig. 3.8b. FULL has a
slightly shallower and colder boundary layer than BASE (Fig. 3.8a) throughout the entire
simulation period. A small increase in stability is also noted in FULL, with a tighter potential
temperature gradient within the inversion layer above the CAP. These differences in the vertical
structure of the FULL simulation relative to the BASE simulation yield a more realistic
deepening (2-4 February) and subsequent lowering (5-6 February) of the top of the shallow
mixed-layer during the cold air pool event, which was a key factor in modulating the high ozone
concentrations during this episode.
Temperatures at 2-m averaged over the entire simulation period demonstrate the colder CAP
produced in the FULL simulation (compare Fig. 3.10a to Fig. 3.10b). The difference between
those two fields indicates an average decrease in temperature of 1.25 to 1.75 °C throughout much
of the basin (Fig. 3.11a).
These improved 2-m temperatures are related to the change from liquid to ice cloud particles.
Snapshots of the cloud characteristics at 0600 UTC 5 February (Fig. 3.12) reflect similar total
cloud amounts and coverage. The BASE run is dominated by liquid-phase particles, or “cloud
water” (Fig. 3.12c), while the FULL run is dominated by ice-phase particles, or “cloud ice” (Fig.
3.12d). Since the liquid clouds produced in BASE are generally stratus while the ice clouds
produced in FULL are primarily surface-based fog, occasionally periods of greater horizontal
cloud coverage are found in BASE as the stratus is able to extend outwards away from the center
of the basin. Considering the sedimentation of cloud ice is turned off in the FULL simulation,
there is greater cloud mass in that run relative to the BASE simulation. However, the liquid
clouds in the BASE run produce 70-80 W m-2 of downwelling longwave radiation, while the ice
clouds in the FULL run produce only 40-70 W m-2 over the same region at 0600 UTC on 5
February even with the extra cloud mass (compare Fig. 3.12e to Fig. 3.12f). Averaged over the
entire 6-day period, downwelling longwave radiation from the liquid clouds is 10-20 W m-2 more
than from the ice clouds (Fig. 3.11b), which is consistent with the elevated temperatures over the
entire period as well (Fig. 3.11a). The greatest difference in 2-m temperature is at the low
elevations in center of the basin, while the greatest difference in longwave radiation is mid-way
up the basin slope along the periphery of cloud cover. Table 3.2 demonstrates this behavior with
mean differences below specific elevation contours inside the basin. The pattern indicates that
the stratus clouds in BASE typically had greater spatial coverage than the ice fog in FULL,
leading to large differences in longwave radiation near the edge and outside of the ice clouds.
3.4 Sensitivity to Snow Cover3.4.1 No Snow
As mentioned in the introduction, the lack of snow during the 2012 winter led to high surface
temperatures, deep afternoon mixed layers, and low ozone concentrations (Lyman and Shorthill
2013). We examine the impact of snow cover by comparing a simulation for the 1-6 February
2013 period with no snow below 2000 m (NONE) to the FULL simulation. Returning to Fig.
3.5, it is clear that the lack of snow raises the surface temperatures by as much as 10 °C during
the night relative to the 6-station sample with the average difference for the 6-day period of 7.6
°C (Table 3.3). Minimal 2-m temperature differences are evident outside of the basin. The
boundary layer structure in NONE simulated at Roosevelt (Fig. 3.6) shows higher temperatures
and 100-200 m larger depths from that observed as well as the FULL simulation. As a further
example, potential temperature profiles from FULL and NONE are plotted for Ouray from 0900
UTC to 1800 UTC on 3 February (Fig. 3.13). Extensive tethersonde observations at this location
provide an indication of the boundary layer structure there at 1500 and 1800 UTC (Schnell et al.
2014). Potential temperatures in the PBL are 7-8 °C warmer in NONE with 200-250 m deeper
CAP than FULL.
Several interrelated processes contribute to the considerably higher low-level
temperatures in NONE than FULL. First, the decrease in surface albedo when the snow is
removed from the basin floor results in greater absorption of solar radiation and leads to a
warmer boundary layer in NONE. Second, the sensitivity of the CAP to ice-phase microphysics
is minimized in NONE since the warmer boundary layer over the bare ground/vegetation is not
cold enough to nucleate cloud ice (i.e., -12 °C). Finally, the development of liquid-phase stratus
as opposed to ice-phase fogs overnight leads to higher temperatures due to increased longwave
radiation at the surface.
Differences in vertical profile structure between FULL and NONE greater than 1 km above
the basin floor were imperceptible, indicating that the absence/presence of snow cover only
impacts atmospheric characteristics within the PBL. This can be seen in Fig. 3.8c, where NONE
shows very similar characteristics to FULL (Fig. 3.8b) above 2400 m MSL. However, the
NONE simulation has a much warmer PBL that is generally deeper than FULL. The strong
inversion above the CAP effectively isolates the thermal and stability differences to inside the
CAP itself, leaving the outside regions unaffected. In addition, the stability within the inversion
at the top of the CAP in the NONE simulation was often weaker than that evident in the FULL
simulation (compare Fig. 3.8c to Fig. 3.8b). This weakened stability allows further eastward
penetration of downsloping westerly flow from the Wasatch Range into the western periphery of
the CAP. Boundary layer depth and stability within the inversion are both important factors that
play a role in air quality. Deeper PBLs generally allow for pollutants to mix through a larger
volume of air resulting in decreased concentrations compared to shallower PBLs. Similarly, a
weaker and less stable inversion at the top of the PBL would typically allow for greater mixing
in the lowest layers of the inversion, helping to decrease pollutant concentrations. Accounting
for these factors, one would expect the PBL structure in NONE to allow for improved air quality
over the FULL case.
3.4.2 No Western Snow
A simulation with snow removed from the western corner of the basin (NW) was completed
to emulate situations where downsloping westerly flow has melted the snowpack there, similar to
early February 2014. The removal of snow cover in NW proved to have very little impact on the
overall CAP structure. Modeled 2-m temperatures and vertical structure within the PBL show
strong similarities to FULL (see Figs. 3.5 and 3.6). 2-m temperatures, on average, are within
about 0.2 °C of FULL, as shown in Table 3.3.
Nonetheless, notable differences in 2-m
temperatures and wind flow patterns are detected near the snow boundary (compare Fig. 3.10b to
Fig. 3.10c). The lower albedo and increased solar absorption of the region where snow is
removed in NW leads to 2 m temperatures that average 2.5-3 °C higher than the same region in
FULL. Temperature differences over areas that remain snow-covered in NW are generally less
than 0.25 °C higher than FULL. These patterns hold true for both daytime and nighttime
periods. As a consequence of this low-level temperature gradient, a thermally driven circulation
is produced across the snow boundary. The 10-m zonal wind component over the bare ground in
NW is 1.5-2 m s-1 more negative (easterly) than the same region in FULL when averaged over
the entire simulation (Fig. 3.14). Furthermore, this “snow breeze” extends several km away from
the snow boundary in both directions indicating that the circulation has significant horizontal
extent. Beyond the thermal and wind flow differences near the snow/no snow border, NW
exhibited striking similarity to FULL.
3.5 Primary Flow Features
Several flow features were detected in the FULL simulation that likely play an important role
in transporting pollutants within the CAP and ultimately impacting air quality. The FULL
simulation is used for analysis here since it most accurately represents the evolution of the 1-6
February CAP.
3.5.1 Synoptic Westerlies and Terrain-Flow Intrusions
The first flow feature is westerly winds along the top of the inversion and occasional
intrusion of this westerly flow into the CAP. Broad upper level ridging is often coincident with
CAP formation and persistence in the Uintah Basin. When the ridge axis is nearby, the synoptic
pattern frequently places 2-3 km level? winds with a westerly component over the basin and near
the top of the CAP and inversion layer. Wind speeds vary with the local pressure gradient, and
when speeds and stabilities are appropriate, downsloping flow from the Wasatch Range can lead
to enhanced mixing on the western end of the Uintah Basin. During these partial mix-out
periods, warmer, cleaner air from above the CAP can be entrained into the CAP leading to
temporary improvements in air quality and a decrease in local stability. Section 3.2 and Fig. 3.9
display an example of this phenomenon. The westerly flow atop the inversion may impact CAP
evolution during non-mix-out periods, as well. A tilted CAP structure is often the result of a
balance between pressure gradients and wind stress, which can then be disrupted by variations in
wind speed (Lareau and Horel 2014). These disruptions may produce a gravity current-like
behavior as the CAP rebounds, causing relatively large changes in depth (a few hundred meters)
within just a few hours. Downsloping flow with a northerly component off the high Uintah
Mountains has also been observed to be associated with a notable strengthening of the inversion
top in the Uintah Basin due to subsidence warming of flow descending in the wake of the
mountains (not shown).
3.5.2 Elevated Easterlies
The second feature is the presence of elevated easterly flow within the strongly stable
inversion layer at the top of the CAP. Figure 3.15a shows the time-averaged zonal wind
component for the entire 6-day simulation along the west-east cross section shown in Fig. 1.5b.
This figure highlights the synoptic westerly flow generally above 2100 m MSL, and the easterly
flow that is present a few hundred meters above the basin floor. The core of the easterly flow
coincides with the strongest stability in the basin and lies between 1800-2000 m MSL. Although
this feature is relatively weak (~ 0.5 m s-1), it is persistent enough to appear as a coherent spatial
pattern when averaged over the 6-day period. Even weaker flow is noted in the lowest 100 m
above the basin floor within the CAP, indicating that the mean flow is near zero for that region.
A similar region of weak flow is noted between 2000 and 2100 m MSL near the top of the
inversion layer and in the transition zone between the elevated easterly flow and the synoptic
westerly flow above.
The mean flow is also broken up into daytime and nighttime periods, as shown in Figs. 3.16a
and 3.17a, respectively. For this analysis, daytime is defined at 0800-1700 local time. Breaking
the time-average into day-night periods shows that the elevated easterly flow has a clear diurnal
variation in magnitude. During the daytime, the core of the easterly flow is stronger (~1 m s-1)
and the west-east spatial extent is greater (compare Fig. 3.15a to Fig. 3.16a). At night, the
easterly flow exhibits a smaller core (at 0.5 m s-1) than the overall average with a decrease in
spatial extent, particularly in the western portion of the basin (compare Fig. 3.15a to Fig. 3.17a).
This behavior indicates that thermal gradients are likely the main driver of the elevated flow
pattern. The daytime plot also shows that the easterly flow is more elevated in the basin center,
with very weak winds in the lowest 200 m. At night, however, the easterly flow extends down to
the surface in the basin center (compare Fig. 3.16a to Fig. 3.17a).
Although the bulk of the elevated easterly flow is within the inversion layer above the CAP,
it is still an important variable for the air quality picture within the basin. Pollutants and ozone
precursors are often able to mix into the lower layers of stratification (Schnell et al. 2014) from
weak turbulence and entrainment (Cai and Luhar 2002; Salmond 2005). During the daytime
heating of shallow near-surface convective mixed layers or within sub-cloud mixed layers at
night, pollutants from overlying stable layers can be be dispersed to the surface through
entrainment and weak turbulence (Cai and Luhar 2002; Salmond 2005). The ozone precursors
that are able to leak into the easterly flow layer would then be transported downstream to
portions of the basin with potentially lower concentrations. Observational data from UBOS
2012-2013 indicate that there are high ozone concentrations throughout the basin and removed
from the maximum in precursor concentrations in the main gas field. This implies that precursor
concentrations are being advected to different regions of the basin, where ozone is produced
locally through photochemical processes (Karion et al. 2014).
3.5.3 Diurnal Flows
The third feature is the diurnal pattern of upslope/downslope flows that is typically seen in
mountain-valley systems. During the nighttime (Fig. 3.17a), drainage flows are depicted by
FULL with light westerly winds in the lowest 100 m on the west slope of the basin, while light
easterly winds exist on the east slope. This pattern is generally reversed during the daytime (Fig.
3.16a), but the upslope flow on the western side of the basin is weaker and connected to the
elevated easterly flow around 1700-1800 m MSL. Upslope flow on the east slope is more welldefined and on the order of 100 m deep. Further evidence of diurnal flow patterns exists in Fig.
3.18, where mean wind direction is shown in map-view for both daytime and nighttime periods.
The daytime plot (Fig. 3.18a) demonstrates upslope flow within the CAP, indicated by black
vectors, where even local ridgelines inside the basin impact flow directions. Outside of the CAP,
to the north and west of the Uintah Basin, synoptic west-northwesterly flow and its interaction
with local topography dominates (white vectors). This exhibits how the strong stability above the
CAP is able to effectively shield the basin interior from synoptic flows, allowing for thermallydriven terrain circulations. The nighttime plot (Fig. 3.18b) depicts drainage flows toward the
center of the basin and ultimately down the Green River into Desolation Canyon to the
southwest.
Diurnal upslope/downslope flows likely to contribute the spatial variations of
precursors/ozone within the basin. Nocturnal drainage flows may lead to building pollutant
concentrations in the lower elevations at the center of the basin. Upslope flow during the
daytime may help to mix the lower atmosphere by spreading pollutants away from the basin
center.
3.5.4 Differences in Snow vs. No Snow
Removal of snow cover in NONE resulted in similar flow patterns to FULL, but there were a
number of differences in magnitude and structure of the elevated easterlies. Core easterly winds
in FULL were typically 0.25 m s-1 stronger and 100 m higher than NONE. The transition from
easterly flow in the inversion layer to westerly flow aloft was also 100-150 m higher in FULL.
At first glance, the more elevated nature of the easterly flow in FULL is a little puzzling. One
might expect that the colder, shallower PBL produced in FULL would lead to a lower height of
this easterly flow regime. However, since the top of the strong inversion layer is typically at a
similar elevation in FULL and NONE, it appears that the weaker stability in NONE allows the
westerly flow above the CAP to mix further down into the inversion layer. The effect of the
westerly flow above the CAP in offsetting the easterly flow regime also apparently offsets the
stronger thermal forcing one would expect in NONE given the lack of snow cover. This creates a
shallower easterly flow layer with slightly weaker winds by an average of 0.25 m s -1. The region
of strongest easterly flow is shifted westward in NONE, with a broader area extending to the
surface on the western slope compared to FULL. All differences in flow patterns are confined
within the lowest 1 km above the basin floor and any disparities above 2400 m MSL are
negligible. This is further evidence that the removal of snow cover only affects the near-surface
atmosphere and PBL.
Daytime plots (Fig. 3.16) show greater spatial variation in flow patterns between FULL and
NONE than the overall mean plots. Both simulations had daytime core easterly winds of 1.1 m
s-1, which is notably stronger than the mean flow speeds for all hours. However, the strongest
easterly flow in FULL was 400-500 m above the center of the basin, while in NONE it was
within 100 m of the surface and midway up the western slope of the basin. The height of the
transition from easterly to westerly flow was approximately 100 m lower in NONE, similar to
the overall mean. The west-east extent of elevated easterlies in FULL nearly spans the entire
basin, while the easterlies in NONE are primarily confined to the western half of the basin and
extend down to the surface. In fact, the eastern third of the basin experiences daytime westerly
flow in NONE, where elevated easterlies exist in FULL.
The easterly flow pattern is much weaker during nighttime hours for both FULL and NONE
(Fig. 3.17) with core winds of 0.55 and 0.35 m s-1, respectively. Like the daytime and overall
means, the transition height to westerly flow is about 100 m lower in NONE. The easterly flow
pattern is generally displaced eastward at night when compared to daytime, leaving weak
westerly flow over the western one-third of the basin for both simulations. In the center of the
basin, FULL has easterly flow down to the basin floor of 0.3 m s -1, while NONE has westerly
flow of 0.1 m s-1 undercutting elevated easterlies in the lowest 80 m.
Interestingly, this
difference implies that drainage flows on the east side of the basin extend into the basin center in
FULL, while drainage flows from the west side extend into the basin center in NONE.
An examination of flow patterns in the FULL and NONE simulations indicate there are
several differences in magnitude and structure, while many of the same general trends are
present. These differences are due to the complex influences of snow cover on CAP strength and
structure. The resulting flow patterns indicate that the presence of snow cover likely impacts
mixing, transport of pollutants, and pollutant concentrations within the basin.