1
Numerical sensitivity simulations of persistent cold air
2
pools associated with elevated wintertime ozone in the
3
Uintah Basin, Utah
4
5
E. M. Neemann1, E. T. Crosman1, J. D. Horel1, and L. Avey2,*
6
[1]{Department of Atmospheric Sciences, University of Utah, Salt Lake City, Utah}
7
[2]{Utah Division of Air Quality, Salt Lake City, Utah}
8
Correspondence to: E. M. Neemann (erik.neemann@utah.edu)
9
10
Abstract
11
High ozone concentrations associated with a winter cold-air pool during the Uintah Basin
12
Winter Ozone Study are investigated. Field campaign observations combined with numerical
13
simulations are analyzed in Utah’s Uintah Basin from 1–6 February 2013 when ozone
14
concentrations exceeded 150 ppb. Cold-air pool sensitivity to cloud microphysics and snow
15
cover variations within the Weather Research and Forecasting model simulations are
16
examined, along with their impact on air quality in Community Multi-Scale Air Quality
17
model simulations.
18
compared to liquid-dominant clouds through increased nocturnal cooling and decreased
19
longwave cloud forcing. The presence of snow cover also strengthens cold-air pool structure
20
by lowering near-surface air temperatures and increasing boundary layer stability due to
21
reduced absorbed solar insolation by the high-albedo snow surface.
22
increases ozone levels by enhancing solar radiation available for photochemical reactions.
23
Flow features affecting Uintah Basin cold-air pools that affect pollutant mixing and air
24
quality within the basin are studied, including: penetration of clean air into the basin from 1
Ice-dominant clouds are found to enhance cold-air pool strength
Snow cover also
1
across the surrounding mountains, elevated easterlies within the inversion layer, and diurnal
2
upslope and drainage flows.
3
1
4
High concentrations of ozone are known to have an adverse impact on human health,
5
including respiratory irritation and inflammation, reduced lung function, aggravated asthma,
6
and long-term lung damage (Lippmann 1993; Bell et al. 2004). Ozone is formed through
7
photochemical reactions of pollutants emitted from industrial sources and vehicles known as
8
“ozone precursors”.
9
organic compounds (VOCs) (Pollack et al. 2013). Once thought to primarily be an urban,
10
summertime problem (due to the high insolation required for photochemical reactions), high
11
ozone levels have recently been detected during the wintertime in snow-covered rural basins
12
with significant industrial fossil fuel extraction activities (Schnell et al. 2009). The presence
13
of snow cover greatly increases the surface albedo and enhances the actinic flux (quantity of
14
light available to molecules) in the near-surface atmosphere leading to photolysis rates
15
notably larger (~50%) than those observed in summer (Schnell et al. 2009). In addition, the
16
shallow and highly stable boundary layer often observed during the wintertime in snow-
17
covered rural basins further exacerbates the problem by trapping the high ozone
18
concentrations in the lowest several hundred meters of the atmosphere. A schematic of this
19
typical setup is shown in Fig. 1.1.
20
High levels of ozone were first detected in Northeast Utah’s Uintah Basin in 2009, when 8-hr
21
average ozone concentrations were over 100 ppb (Lyman et al. 2014). This value was well
22
above the EPA’s National Ambient Air Quality Standard (NAAQS) of 75 ppb (U.S. EPA
23
2013), and far above the background levels of ozone near the earth’s surface that typically
24
range between 20–45 ppb (U.S. EPA 2006). Fossil fuel production has increased in the
25
Uintah Basin over the last several years and will likely continue to increase. In March 2014,
26
there were over 9,400 producing wells in the basin and over 3,600 additional permit
Introduction
These precursors are typically nitrogen oxides (NOX) and volatile
2
1
applications since the beginning of 2012.
These wells, and associated ozone precursor
2
sources, are currently operating throughout the basin with oil wells primarily clustered in the
3
west and gas wells clustered in the east.
4
Extensive scientific research has been conducted in the Uintah Basin to better understand the
5
wintertime rural ozone problem both chemically and meteorologically during the past several
6
winters (Lyman and Shorthill 2013; Stoeckenius and McNally 2014). These field campaigns
7
point to the importance of nitrous acid and formaldehyde in the surface snow pack for ozone
8
production. Meteorologically, the importance of snow cover, cloud cover, inter-basin flows,
9
and terrain-flow interactions in the evolution of the shallow, stable boundary layers associated
10
with these high wintertime ozone episodes are recognized as being important but have not
11
been well documented.
12
The presence of snow cover in combination with a stable, persistent boundary layer cold-air
13
pool (CAP) is integral to these wintertime high ozone events (Schnell et al. 2009).
14
Considerable year-to-year variations in seasonal snow cover exist in the Uintah Basin,
15
depending on whether or not strong early-winter storms deposit sufficient snow to last into
16
February or new snow is deposited during that month. These variations in annual snow cover
17
lead to similar variations in the occurrence of high ozone events (Lyman and Shorthill 2013).
18
For instance, snow cover was absent from the basin during February 2012 and much of
19
February 2014 and ozone levels remained low during those months, while February 2013 saw
20
extensive snow cover and several high ozone events. The variations in snow cover and their
21
impact on ozone concentrations is demonstrated in Table 1.
22
A persistent CAP is defined here following Lareau et al. (2013) as a stagnant, stable layer of
23
air confined in a topographical depression (with warmer air typically aloft) that lasts for more
24
than a diurnal cycle. They are most often a wintertime phenomena, with their onset coincident
25
with midlevel warming and removal coincident with midlevel cooling (Reeves and Stensrud
26
2009).
Persistent CAPs are also often associated with low clouds, fog, freezing
3
1
precipitation, and hazardous ground and air travel. Their impacts on local weather and
2
particulate air pollution in urban basins have been extensively studied (Whiteman et al. 2001;
3
Malek et al. 2006; Silcox et al. 2012; Lareau et al. 2013; Lareau 2014; Lareau and Horel
4
2014). Further observational and numerical studies have been conducted on CAPs for a
5
variety of idealized (Zangl 2005a; Katurji and Zhong 2012; Lareau 2014) and actual
6
topographic basins (Whiteman et al. 2001; Clements et al. 2003; Zangl 2005b; Billings et al.
7
2006; Reeves & Stensrud 2009; Reeves et al. 2011; Lareau et al. 2013; Lareau and Horel
8
2014).
9
Relatively few numerical studies, however, have examined the impact of clouds and cloud
10
microphysics on CAP formation and evolution. Zangl (2005a) indirectly examined the effect
11
of cloud particle phase on the formation of extreme CAPs in the Gstettneralm sinkhole,
12
Austria. He found that an efficient drying mechanism was required for extreme CAPs to
13
develop. Without one, the formation of fog and its accompanying longwave radiation would
14
prevent the low-level cooling necessary for a strong CAP. Zangl found that the most efficient
15
process for drying the atmosphere was through the formation of cloud ice and its eventual
16
sedimentation. Moreover, the lower vapor pressure over ice compared to water allowed for
17
preferential growth of cloud ice at the expense of cloud water. This accelerated the drying
18
process, with even greater acceleration possible if cloud ice was able to grow into snow,
19
which has a greater fall speed. Simulations dominated by cloud ice were found to dry out the
20
atmosphere more effectively than simulations with a mixture of cloud ice and cloud water,
21
translating to greater cooling and lower temperatures in the ice-dominant case.
22
The impact of the snow-albedo feedback on surface temperatures of interior continental
23
locations during wintertime is known to be high. However, in remote locations such as the
24
Uintah Basin, where snow cover is typically very thin (~5–10 cm) and spatially and
25
temporally variable, accurately assessing snow mass or water equivalent for input into
26
numerical models can be difficult (Jeong et al. 2013).
27
wintertime persistent CAP formation and maintenance, relatively few numerical studies have4
Despite their importance in
1
examined in detail the impact of snow cover on CAPs. Billings et al. (2006) studied the
2
impact of snow cover on a CAP in the Yampa Valley, CO. Although unable to produce a
3
temperature inversion, simulations with snow cover produced an isothermal profile that more
4
closely aligned with observations than snow-free simulations. The snow-free simulations
5
were incapable of producing the CAP and had the highest temperatures located on the valley
6
floor. They concluded that the primary influence of snow cover was to increase the surface
7
albedo and decrease the sensible heat flux. Snow cover also increased the latent heat flux for
8
certain locations due to melting and evaporation. Zangl (2005a) also investigated the role of
9
snow cover on the nocturnal formation of extreme CAPs.
He found that a small heat
10
conductivity of the ground was important for efficient cooling and was most readily
11
accomplished by the presence of fresh snowfall. Comparison between a snow-covered and
12
grass-covered sinkhole floor suggested that the larger surface heat capacity of the grass floor
13
resulted in more gradual cooling, a smaller afternoon-morning temperature difference, and
14
weaker static stability than simulations with a snow-covered floor. Interestingly, the grass
15
case produced no cloud cover in the center of the basin, while snow cases did.
16
The topography of the Uintah Basin is highly conducive to the formation of strong CAPs in
17
the presence of snow cover (Fig. 2). It is a large, deep, nearly bowl-like depression bounded
18
to the north by the Uinta Mountains, the west by the Wasatch Range, and the south by the
19
Tavaputs Plateau. Covering over 15,000 km2, the lowest elevations in the center of the basin
20
are below 1450 m and the terrain extends above 1950 m on all sides, with ridgelines reaching
21
well above 2500 m to the north, west, and south. The only outlet below 1950 m along the
22
basin’s perimeter is the deep, narrow Desolation Canyon where the Green River drains south
23
into East-Central Utah. As a winding, narrow river canyon, this outlet provides very little
24
airmass exchange between the Uintah Basin and Eastern Utah. The synergy between the
25
confining topography, winter snow cover, and surface high pressure often result in a stagnant,
26
stably stratified environment with little horizontal mixing. At the same time, vertical mixing
27
is inhibited by the existence of a strong inversion atop the shallow, cold planetary boundary
5
1
layer (PBL) that effectively shields the lower elevations of the basin from weak, transient
2
synoptic weather disturbances.
3
The detailed meteorology associated with wintertime high ozone episodes in the Uintah Basin
4
is not well-understood and only limited observational data sets are available at isolated
5
locations within the vast basin. While it is recognized that surface high pressure, strong
6
temperature inversions, light winds, and the presence of snow cover are associated with high
7
ozone episodes, detailed flow patterns and thermodynamic structure in the basin have been
8
only partially explored. A number of permanent surface weather stations exist within the
9
basin at various elevations, but upper level sounding data are not collected routinely. Hence,
10
limited documentation exists regarding the highly stratified boundary layer in the basin during
11
CAPs or the mesoscale terrain- and thermally-driven flows that likely play an important role
12
in the mixing of pollutants within the basin during these episodes. Understanding the ozone
13
concentrations within the basin requires adequate meteorological information to drive air
14
chemistry models capable of simulating the complex photochemical processes that lead to
15
degraded air quality. The size, depth, and shape of the Uintah Basin provide an appropriate
16
scale for mesoscale models and the frequent occurrence of persistent CAPs during the winter
17
season presents numerous cases for examination. The local topography and scale of the basin
18
potentially leave it more protected from synoptic flow interactions than persistent CAPs
19
observed in other locales (Zangl 2005b; Lareau et al. 2013; Lareau and Horel 2014).
20
The purpose of this study is to use atmospheric and air quality numerical models to simulate a
21
high ozone episode from 1–6 February 2013 during the Uintah Basin Winter Ozone Study
22
(UBWOS). We use the Weather Research and Forecasting (WRF) model to examine the
23
sensitivity of CAP structure to variations in snow cover, specification of snow albedo, and
24
cloud microphysics, for this case. We also investigate important wind flow regimes of the
25
CAP and the impact of snow cover on simulated air quality. Section 2 will describe data
26
sources, model setup, and modifications used in the simulations, and Sect. 3 will discuss
27
CAP evolution and modeling results. Section 4 will illustrate the sensitivity of simulated6
1
ozone concentrations during this period to surface and atmospheric forcings supplied to the
2
Community Multi-Scale Air Quality Model (CMAQ). Conclusions and additional discussion
3
will be provided in Sect. 5.
4
2
5
2.1 Observational data and control run WRF model setup
6
A description of the extensive instrumentation deployed for the intensive field campaign is
7
detailed by Stoeckenius and McNally (2014). For model validation, six surface stations with
8
complete records were selected as representative of the conditions in the core of the basin (see
9
Fig. 2). We model the evolution of a persistent cold air pool in the Uintah Basin that lasted
10
from 31 January to 10 February 2013. The simulations were run for 144 hours from 0000
11
UTC 1 February 2013 to 0000 UTC 7 February 2013. This time frame represents the
12
strongest period of the CAP and was one of the most heavily-instrumented periods during the
13
2013 UBWOS. Some of the highest-measured ozone concentrations during the winter season
14
were observed during this period, with readings along a mobile transect of over 150 ppb taken
15
near Ouray during the afternoon of 6 February 2013.
16
2.1.1 WRF domain and parameterization schemes
17
A summary of the WRF model setup in this study is given in Table 2. We use WRF version
18
3.5 and the Advanced Research WRF core. The WRF model is nonhydrostatic, with a
19
pressure-based, terrain-following (eta) vertical coordinate system. Simulations herein used 41
20
eta levels with the lowest 20 model levels within approximately 1 km of the terrain surface.
21
Three telescoping, one-way nested domains were employed to place the highest-resolution
22
nest over the Uintah Basin, with grid spacings of 12, 4, and 1.33 km, respectively (Fig. 2a).
23
Operational North American Mesoscale Model (NAM) analyses were used to initialize
24
atmospheric and land surface variables (except for snow variables, see Sect. 2.1.2) as well as
25
provide the lateral boundary conditions for the outer domain at 6-hour intervals for the
26
duration of the 6-day simulations.
Data and methods
7
1
2.1.2 Prescribing initial WRF snow cover in Uintah Basin
2
The available NAM snow cover analyses were found to poorly reflect the observed shallow
3
snow layer in the Uintah Basin during February 2013. While the NAM analyses represented
4
the spatial coverage of snow during the 1–6 February 2013 period fairly well, they
5
overestimated snow depth and snow water equivalent (SWE) within the basin and
6
underestimated them at higher elevations. In order to better represent the actual snow surface
7
conditions, an “idealized” layer of snow was specified in the WRF initialization fields based
8
on elevation, snow depth and SWE in a manner similar to Alcott and Steenburgh (2013). This
9
prescribed snow cover was determined using all available observations:
Snowpack
10
Telemetry; National Operational Hydrologic Remote Sensing Center analyses; Moderate
11
Resolution Imaging Spectroradiometer imagery; and manual and automated observations
12
from Community Collaborative Rain, Hail, and Snow Network, and field campaign locations.
13
Figure 2.3 shows the specified relationship between snow depth, SWE and elevation relative
14
to observations and NAM analyses, which demonstrates the over- (under-) estimation of snow
15
by the NAM at low (high) elevations. The prescribed snow cover was applied within all
16
model domains with no snow cover outside of the Uintah Basin below an elevation of 2000 m
17
and a 17 cm snow depth from the basin floor up to an elevation of 2000 m. Above 2000 m,
18
the snow depth was elevation-dependent, increasing to 100 cm for elevations at 2900 m or
19
higher.
20
In addition to poor representation of snow depth and SWE, the NAM analyses underestimated
21
snow albedo relative to observed shortwave radiation measurements at Horsepool and
22
Roosevelt. For example, albedo averaged from 1 January–2 March 2013 at Horsepool was
23
0.82 (Roberts et al. 2014), which is roughly 0.17 higher than the NAM analyses. Very low
24
temperatures combined with repeated light rime deposition onto the snow surface during
25
many nights apparently maintained the highly reflective surface. Hence, the snow albedo
26
variable in WRF was modified to be 0.82 inside the basin, although the mean albedo at
8
1
Horsepool during the 1–6 February period was actually a bit higher (~0.84). Furthermore,
2
based on visual observations of the snow covering nearly all of the sparse vegetation in the
3
basin during the 1–6 February period, changes were made to the WRF vegetation parameter
4
table for the two dominant vegetation/land use types: “shrubland” and “cropland/grassland
5
mosaic”. For these vegetation types, 20 kg m-2 of SWE was allowed to fully cover the
6
vegetation in the Noah land surface model. The combination of increasing the snow albedo
7
and modifying the vegetation parameter table enabled the model surface to attain the high
8
albedos observed during the field campaign (Fig. 4a and 4b). The NAM analyses were used
9
for all other land-surface variables in the snow-free region of these simulations, with surface
10
albedo being determined by the National Land Cover Database 2006 (Fry et al. 2011).
11
2.2 Numerical sensitivity studies
12
A number of sensitivity tests were conducted to study the impact of variations in cloud type
13
and snow cover on persistent cold air pools in the Uintah Basin. These tests are summarized
14
in Table 3.
15
As will be shown in Sect. 3, ice fog and low stratus were commonly observed throughout the
16
basin during this event. However, simulations with the default WRF model (referred to as the
17
BASE run) tended to develop liquid-phase clouds. In order to test the sensitivity of the
18
Uintah Basin CAP to ice vs. liquid phase cloud particles, modifications were made to the
19
Thompson microphysics scheme to produce fog/low clouds that were ice-phase dominant.
20
This was accomplished by editing the Thompson code in the lowest 15 model layers (~500 m)
21
to change the treatment of cloud ice by turning off cloud ice sedimentation and the
22
autoconversion of cloud ice to snow. These changes allowed low-level cloud ice to remain
23
suspended and thrive through vapor deposition due to the lower vapor pressure over ice
24
compared to water.
25
forcing than liquid-dominant clouds (Shupe and Intrieri 2004), allowing for stronger CAP
26
formation, shallower PBLs, and lower low-level temperatures with smaller bias/errors.
We anticipated that ice-dominant clouds would have less radiative
9
1
As discussed in Sect. 1, large variations of snow cover have been observed from winter to
2
winter in the Uintah Basin. To examine the sensitivity of the conditions in the basin to snow
3
cover, numerical simulations were run for the 1–6 February period using the same model
4
configuration but varying the spatial extent of the snow cover.
5
prescribed snow cover throughout the entire basin as discussed in Section 2.1.2 are shown in
6
Fig. 4c (BASE and FULL). In the snow cover sensitivity simulation (NONE), snow was
7
removed from the basin for elevations below 2000 m (Fig. 4d); similar to what was observed
8
during February 2012 and late February 2014. A fourth simulation (NW) was run with snow
9
cover removed from the western section of the basin (west of -110.4° W) for elevations below
10
2100 m (not shown). This partial erosion of the snowpack is often observed after periods of
11
downsloping flow over the Wasatch Range melt the snow in the western basin, while
12
snowpack further east is undisturbed.
13
Several additional numerical simulations were conducted while searching for the best WRF
14
model setup (not shown). These included the use of various microphysics and PBL schemes,
15
prescribed cloud droplet concentrations, and modified initial/boundary condition files.
16
Microphysics parameterizations tested included the WRF single moment 3-class and
17
Morrison schemes. Additional modifications to the Thompson scheme were also explored
18
before implementing the changes discussed above. The results from various microphysics
19
testing converged to 3 general outcomes for the 1–6 February period: (1) a CAP dominated by
20
dense liquid-phase stratus clouds; (2) a CAP with clear skies through the entire simulation;
21
and (3) a CAP dominated by low-level ice fog. Outcome (2) was deemed unrepresentative of
22
observed conditions, and outcomes (1) and (3) are analogs for the BASE and FULL
23
simulations discussed further in this study. In addition to the MYJ, the Yonsei University
24
(YSU), Asymmetric Convective Model, Grenier-Bretherton-McCaa, and Bretherton-Park
25
PBL schemes were tested. The YSU scheme with the Jimenez surface layer formulation and
26
updated stability functions (Jimenez et al. 2012) was also considered, but did not show
27
improved results over the original YSU scheme. The MYJ was chosen since it best
The simulations with
10
1
represented the combination of moisture, stability, and temperature characteristics that were
2
observed in the Uintah Basin for the simulated period. Simulations were conducted with
3
cloud droplet concentrations prescribed for maritime (100 x 106 m-3), continental (300 x 106
4
m-3), and polluted continental (1000 x 106 m-3) situations. However, changing the prescribed
5
cloud droplet concentrations had no perceptible impact on CAP formation, so the default
6
value in the Thompson scheme was used (100 x 106 m-3).
7
initial/boundary conditions were tested with relative humidity reduced by 5, 20, and 40% to
8
determine what effect it would have on cloud cover within the CAP. Reducing relative
9
humidity was noted to delay the onset and decrease the coverage of clouds/fog for the first
10
day or two of the simulation. Nonetheless, the final 3–4 days of the simulation converged on
11
similar results for all cases, so the original NAM analysis data was used.
12
3
13
3.1 Overview
14
The numerical sensitivity simulations presented in this Sect. investigate the role of snow
15
cover and boundary-layer clouds on the intensity and vertical structure of the 1–6 February
16
2013 CAP. The focus lies on the vertical profiles of temperature, wind, and moisture within
17
the first km above the surface since there is little sensitivity to snow cover and low clouds at
18
higher levels. In addition, pollutant mixing depth and transport within the CAP depends on
19
the atmospheric stability and flows in the lowest km above the surface. A common limitation
20
of many air quality modeling studies is an over-reliance on surface wind observations at
21
scattered locations to validate the advective fluxes of pollutants within CAPs (L. Avey, Utah
22
Division of Air Quality, personal communication 2014).
23
This Sect. is organized as follows: First, an overview of the observed evolution of the 1–6
24
February 2013 CAP in the Uintah Basin is presented in Sect. 3.2, followed by analysis of the
25
BASE WRF model simulation in Sect. 3.3. The sensitivity of the simulated CAP to variations
26
in cloud type and snow cover are presented in Sect. 3.3 and 3.4, respectively. Then, the
Finally, NAM analysis
Results
11
1
FULL WRF model simulation is used in Sect. 3.5 to diagnose the mesoscale flows that may
2
be important to pollutant transport within the basin.
3
3.2 1–6 February 2013 cold-air pool
4
A deep upper-level trough and associated midlatitude cyclone moved through the Great Basin
5
from 28–30 January 2013, bringing very cold air aloft (700 hPa temperatures ~-20 °C) and 1–
6
5 cm of light snowfall on top of a ~10–20 cm base across most of the Uintah Basin.
7
Following the upper-level trough passage, 1–6 February was dominated by upper level
8
ridging over the western US and surface high pressure in the Great Basin. This pattern placed
9
northeastern UT in a region of large-scale subsidence and mid-level warming. With warm air
10
aloft (700 hPa temperatures between ~-7 and 0 °C), quiescent surface weather conditions,
11
fresh snow cover, cold low-level air in place, and sufficient incoming solar insolation to drive
12
photochemistry, the stage was set for a persistent CAP event and development of high ozone
13
concentrations in the Uintah Basin.
14
Overnight low temperatures were often -15 to -18 °C, providing an indication of the extent of
15
the cold-air trapped in the lowest elevations of the Uintah Basin during this period. Above-
16
freezing temperatures at higher elevations were common. Ice fog and low stratus were
17
observed in the lowest reaches of the basin often breaking up during the late morning and
18
afternoon hours. Selected vertical profiles of temperature, dew point temperature, and wind at
19
Roosevelt from rawinsondes released daily at midday (1800 UTC) during the CAP are shown
20
in Fig. 5. A shallow mixed layer with high relative humidity near the surface (with thin ice
21
fog and stratus observed intermittently at the time) is capped by increasing temperatures on 4
22
February below 1800 m (Fig. 5a and 5c). The strong inversion extends upwards to 2750 m
23
with decreasing moisture aloft. Weak easterly winds of 2-3 m s-1 within the inversion layer
24
near 2000 m give way to westerly winds of ~10-12 m s-1 near the top of the inversion layer
25
(Fig. 5e and 5g). The mixed layer is shallower on the 5th (Fig. 5b) with thinner ice fog and
12
1
stratus observed at the time of the launch. Weak easterly winds are present again near 2000 m
2
with westerly winds of 7-9 m s-1 in the upper reaches of the capping inversion layer.
3
The spatial coverage of snow and cloud on 2 February 2013 is provided in Fig. 6a by the
4
snow-cloud product supplied by the NASA Short-term Prediction Research and Transition
5
Center (SPoRT). This mid-day image depicts the lower elevations of the Uintah Basin
6
covered in snow with less snow in its far western extremity. Fog and stratus are confined to
7
the lowest elevations of the basin. An earlier (0931 UTC 2 February) Visible Infrared
8
Imaging Radiometer Suite (VIIRS) nighttime microphysics RGB product from SPoRT (Fig.
9
6b) helped detect low clouds at night and discriminate particle phase by combining data from
10
the 3.9, 10.8, and 12.0 micron infrared channels. In this image, liquid-phase low stratus and
11
fog are represented by aqua/green colors in southern ID and portions of western and central
12
UT while the yellow/orange colors are typically associated with ice-phase stratus and fog.
13
The elevation dependence of the fog/stratus is evident in Fig. 6b by the cloud tendrils
14
extending up the river valleys within the basin.
15
Figure 7a presents the time evolution of surface ozone at selected locations in the basin during
16
the CAP. The concentrations exceeded the EPA standard of 75 ppb at all of these locations
17
on many of the days. While ozone at Vernal tended to drop to background concentration
18
levels at night, ozone at Horsepool and Ouray (one of the lowest locations in the basin)
19
remained above the standard during nearly all hours after the 3rd of February.
20
Figure 7b presents the time evolution of aerosol backscatter, low clouds, and an estimate of
21
the depth of the aerosol layer from the Roosevelt laser ceilometer during the 1–6 February
22
period.
23
development of ice fog evident as well in Fig. 6b. Then, a semi-regular pattern developed
24
over the next several days with shallow nighttime fog and low clouds thinning by mid-day
25
and followed by a deeper layer of aerosols in the afternoon that quickly collapsed at sunset.
26
The ceilometer backscatter data also corroborates other observations that the fog and low
Fewer aerosols were observed on 1 February followed that evening by the
13
1
cloud occurrence in the basin peaked during 3–4 February. During that time, significant hoar
2
frost was observed on trees and other surfaces after sunrise with light accumulations of snow
3
crystals falling out of the ice clouds in Roosevelt later in the morning. The high levels of
4
aerosol backscatter on 5–6 February diminished abruptly during late afternoon on 6 February
5
as a result of a weak weather system; however elevated ozone levels continued until 9
6
February, after which a stronger weather system with sufficient cold-air advection aloft to
7
destabilize the column moved through the region.
8
3.3 1–6 February 2013 BASE model simulation
9
Evaluation of the BASE model simulation on the 1.33 km inner grid confirms that the WRF
10
model captures the salient temperature, moisture and wind features of the CAP episode
11
throughout the 1–6 February period above ~500 m AGL.
12
potential temperature profiles from the BASE simulation at elevations above that level tend to
13
agree well with the observed mid-day profiles at Roosevelt (Fig. 5).
14
However, the BASE simulation exhibits unrealistically deep surface-based mixed layers mid-
15
day (Fig. 5) and unrealistic spatial patterns in 2-m temperature (not shown), which will both
16
be shown in the next subsection to be related to unrealistically thick layers of liquid fog and
17
stratus. While the simulated fog and stratus tends to dissipate during most afternoons, those
18
clouds greatly hinder cooling overnight, disrupt the strength of the near-surface layers of the
19
CAP, and often result in sub-cloud mixed layers that are 2–4 times deeper than those observed
20
and surface temperatures that are too warm. Table 4 shows the 2-m temperature error
21
statistics (model – observed) for all 4 model simulations, computed from the 6 representative
22
surface stations in the center of the basin that are highlighted in Fig. 2b. A notable warm bias
23
of 1.65 °C and mean absolute error of 3.3 °C are present in the BASE simulation.
24
3.4 Sensitivity to cloud type
25
The default settings in the Thompson microphysics scheme used in the BASE simulation
26
will be shown in this section to lead to dense, liquid-phase low clouds and fog that notably14
For example, the simulated
1
alter the radiation budget within the basin, and consequently, the strength of the CAP. The
2
extensive fog and stratus, particularly during the overnight hours, result in excessive
3
longwave radiation that inhibits low-level cooling and forms a region of warm air in the
4
center of the basin during much of the CAP period. Straightforward modifications to the
5
Thompson microphysics scheme that are employed in the FULL model run make it possible
6
to examine the sensitivity of the CAP simulations to cloud type. As detailed in Table 3, the
7
FULL simulation has cloud ice sedimentation and cloud ice autoconversion to snow turned
8
off in the lowest 15 model levels. The intent of these modifications is to force the WRF
9
model to produce and maintain clouds dominated by ice-phase particles.
10
Returning to Fig. 5, consider the differences between the mid-day potential temperature
11
profiles at Roosevelt of the FULL and BASE simulations relative to the observed profiles.
12
While the FULL profiles are not necessarily in perfect agreement with the observed ones,
13
they tend to show lower temperatures near the surface and shallower surface-based mixed
14
layers. As discussed in the previous section, the differences in potential temperature between
15
the FULL and BASE simulations become small above ~500 m AGL. The differences in the
16
vertical structure of the FULL simulation relative to the BASE simulation yield a more
17
realistic deepening (2–4 February) and subsequent lowering (5–6 February) of the top of the
18
shallow mixed-layer during the CAP (not shown).
19
Further, the bias and mean errors of the FULL simulation are sharply reduced relative to the
20
6-station sample (Table 4). Temperatures at 2-m averaged over the entire simulation period
21
demonstrate the colder CAP produced in the FULL relative to the BASE simulation (Fig. 8).
22
The difference between those two fields indicates an average decrease in temperature of ~1.5
23
°C throughout the interior of the basin with negligible differences elsewhere (Fig. 9a).
24
These improved 2-m temperatures are related to the change from liquid to ice cloud particles.
25
Snapshots of the cloud characteristics at 0600 UTC 5 February (Fig. 10) reflect similar total
26
cloud amounts and coverage. The BASE run is dominated by liquid-phase particles, or
15
1
“cloud water” (Fig. 3.10c), while the FULL run is dominated by ice-phase particles, or “cloud
2
ice” (Fig. 3.10d). Since the liquid clouds produced in BASE are generally stratus while the
3
ice clouds produced in FULL are primarily surface-based fog, occasionally periods of greater
4
horizontal cloud coverage are found in BASE as the stratus is able to extend outward away
5
from the center of the basin. Considering that the sedimentation of cloud ice is turned off in
6
the FULL simulation, there is greater cloud mass in that run relative to the BASE simulation.
7
However, the liquid clouds in the BASE run produce 70–80 W m-2 of downwelling longwave
8
radiation, while the ice clouds in the FULL run produce only 40–70 W m-2 over the same
9
region at the same time, despite the extra cloud mass (compare Fig. 10e to Fig. 10f).
10
Averaged over the entire 6-day period, downwelling longwave radiation from the liquid
11
clouds is 10–20 W m-2 more than from the ice clouds (Fig. 9b), which is consistent with the
12
elevated temperatures over the entire period as well (Fig. 9a). The greatest difference in 2-m
13
temperature is at the low elevations in center of the basin, while the greatest difference in
14
longwave radiation is mid-way up the basin slope along the periphery of cloud cover. Hence,
15
the stratus clouds in the BASE simulation typically had greater spatial coverage than the ice
16
fog in the FULL run, which leads to larger differences in longwave radiation in locales where
17
liquid clouds are present and ice clouds are not.
18
3.5 Sensitivity to snow cover
19
As mentioned in the Sect. 1, the lack of snow during the 2012 winter led to high surface
20
temperatures, deep afternoon mixed layers, and low ozone concentrations (Lyman and
21
Shorthill 2013). We examine the impact of snow cover by comparing a simulation for the 1–
22
6 February 2013 period with no snow below 2000 m (NONE) to the FULL simulation.
23
Returning to Fig. 5, the lack of snow raises the mid-day surface temperatures by as much as
24
10 °C and the depth of the mixed layer is generally much greater than that observed the
25
during February 2013 CAP. The differences in potential temperature at Roosevelt tend to
26
become negligible above a few hundred meters from the surface.
16
1
Not surprisingly, the 2-m temperature biases relative to the 6 surface stations are very large,
2
particularly at night, with the average difference for the 6-day period being 7.6 °C (Table 4).
3
In addition, the mean 2-m temperatures from the NONE simulation are much higher than
4
those apparent from the BASE or FULL simulations (Fig. 8).
5
Several interrelated processes contribute to the higher low-level temperatures in the NONE
6
simulation relative to the FULL simulation. First, when the snow is removed from the basin
7
floor, the thermal conductivity of the land surface increases, and the decrease in surface
8
albedo results in greater absorption of solar radiation. These combined land surface effects
9
lead to a warmer boundary layer.
Second, the sensitivity of the CAP to ice-phase
10
microphysics is minimized in the NONE simulation since the warmer boundary layer over the
11
bare ground/vegetation is too warm (i.e., higher than -12 °C) to nucleate cloud ice. Finally,
12
liquid-phase stratus develops in the NONE simulation as opposed to ice-phase fogs overnight,
13
which leads to higher temperatures due to increased longwave radiation at the surface.
14
3.6 Flow features
15
While the observations collected during the UBWOS field campaigns are the most extensive
16
available to date for studying the thermodynamic and dynamic conditions in the Uintah Basin
17
(Lyman and Shorthill 2013; Stoeckenius and McNally 2014), the majority of them consist of
18
enhanced surface observations throughout the basin combined with limited profiles at a few
19
locations (e.g., Horsepool, Ouray, and Roosevelt). While far from adequately resolving CAP
20
characteristics, the FULL simulation provides additional information on the four-dimensional
21
fields of temperature, wind, and moisture that helps identify relevant physical processes.
22
Several flow features were detected in the FULL simulation that could be partially validated
23
using the limited available data that likely play an important role to transport pollutants within
24
the CAP and ultimately affect air quality.
25
3.6.1 Clean-air intrusions into the basin
17
1
The structure of the CAP varies extensively over the course of the FULL simulation. Cross
2
sections suggest that the CAP is initially confined to elevations below 1800 m MSL before it
3
deepens to a base near 2000 m early on 3 February (not shown). By midday on 4 February,
4
the inversion base retreats below 2000 m, and eventually lowers to ~1800 m from early on 6
5
February through the end of the simulation.
6
The CAP is continually modulated by synoptically-driven mid-level flow atop the CAP,
7
forcing it to “slosh” back and forth within the basin.
8
surmounting the surrounding terrain from nearly every direction from the southwest to the
9
north. Downsloping flows mixing higher potential temperature and cleaner air downward
10
into the basin are common and their impact depends on the stability and strength of the flow
11
across the upwind barriers.
12
component across the high Uintah Mountains during the 2013 winter, a notable strengthening
13
of the inversion top due to subsidence warming of flow descending in the lee of the mountains
14
was evident in the Uintah Basin (not shown).
15
The CAP may become displaced or tilted through hydrostatic and dynamic processes, which
16
can then be disrupted by changes in wind speed above the CAP (Lareau and Horel 2014).
17
These disruptions may produce a gravity current-like behavior as the CAP rebounds, causing
18
relatively large changes in depth (a few hundred meters) within just a few hours. Figure 11
19
shows an example of this behavior. Strong westerly flow crossing the mountain barrier to the
20
west of the basin at 0600 UTC 4 February is highlighted by a narrow band of increased
21
westerly to northwesterly flow at 2.3 km MSL over the western portion of the basin (Fig.
22
11a). The cross section of potential temperature from west to east through the center of the
23
basin at the same time is shown in Fig. 11b. The westerly downslope winds have eroded and
24
tilted the CAP, pushing it east of Starvation Reservoir (vertical line labeled “STA” in Fig.
25
11b). The CAP is depressed to ~1750 m in the western basin, much lower than in the eastern
26
half of the basin. The FULL simulation suggests that weakening westerly winds over the
27
next several hours lead to the CAP rebounding westward past Starvation Reservoir with the18
Ridging aloft can lead to flow
For example, when the cross-barrier flow had a northerly
1
inversion base quickly rising to ~1900 m, roughly level with the rest of the basin (not shown).
2
3.6.2 East-west cross basin transport
3
Easterly flow immediately above the shallow mixed layer is evident in the mid-day soundings
4
at Roosevelt on a number of days (Fig. 5). The ceilometer data at Roosevelt (Fig. 7a) as well
5
as ozone tethersonde observations at Ouray (Schnell et al. 2014) suggest that aerosols, ozone
6
precursors, and ozone extend upward into this layer of easterly flow likely as a result of weak
7
turbulence and entrainment (Cai and Luhar 2002; Salmond 2005). The ozone precursors from
8
eastern basin source regions that are able to leak into the easterly flow layer may then be
9
transported westward to portions of the basin that have more limited precursor sources,
10
allowing ozone production to take place more widely (Karion et al. 2014).
11
Figure 12 shows the time-averaged zonal wind component from the FULL simulation along
12
the cross section shown in Fig. 2b, split into daytime and nighttime periods. Synoptic
13
westerly flow dominates above 2200 m MSL with easterly flow present a few hundred meters
14
above the basin floor. The core of the easterly flow coincides with the strongest stability in
15
the basin and lies between 1800–2000 m MSL. Although this feature is relatively weak (~1 m
16
s-1 during the day, 0.5 m s-1 at night), it is persistent enough to appear as a coherent spatial
17
pattern when averaged over the 6-day period. During the day, the core of the easterly flow is
18
more intense aloft, and the west-east spatial extent is greater (compare Fig. 12a to Fig. 12b).
19
At night, the easterly flow exhibits a weaker and more regional core shifted to the eastern
20
portion of the basin and extending down to the surface (Fig. 12b).
21
3.6.3 Diurnal valley and slope flows
22
Figure 12 suggests both additive and destructive interactions between the cross-basin elevated
23
easterly
24
downvalley/downslope flows. While basin-scale thermal gradients likely drive the elevated
25
easterly flow, those gradients are at times in concert with and at other times interfering with
26
more localized thermal gradients within drainages and along slopes.
flows
and
near-surface
daytime
upvalley/upslope
and
nighttime
19
1
During the night (Fig. 12b), drainage flows are evident by light westerly winds in the lowest
2
100 m on the west side of the basin in combination with light easterly winds on the east side.
3
This pattern reverses during the day (Fig. 12a); however, the cross-basin easterlies appear to
4
accentuate the upvalley/upslope winds at ~1800 m MSL. As with any basin or mountain
5
range, the diurnal flow patterns within the Uintah Basin are complex. An examination of
6
mean wind direction during the day (not shown) highlights areas of upslope easterly flow
7
within the CAP in the western half of the basin. Outside of the CAP, however, to the north
8
and west of the basin, synoptic west-northwesterly flow and its interaction with local
9
topography dominate. This demonstrates how the strong stability above the CAP is able to
10
effectively shield the basin interior from synoptic flows, allowing for thermally-driven terrain
11
circulations.
12
3.6.4 Effects of snow cover on terrain-flow interactions
13
The sensitivity of terrain-flow interactions to the presence or absence of snow cover in the
14
Uintah Basin is briefly examined here. Comparison of the cross sections of time-averaged
15
zonal winds from the FULL (Figs. 12a and 12b) and NONE (Figs. 12c and 12d) simulations
16
are consistent with earlier results: the removal of snow cover only affects the near-surface
17
atmosphere below the capping inversion. The weaker stability within the capping inversion in
18
the NONE simulation likely allows the synoptic-scale westerlies to extend further down
19
toward the basin floor. This extension appears to diminish the intensity of the easterly winds
20
within the lower reaches of the inversion layer that would be expected in NONE given the
21
lack of snow cover. Comparable differences are evident during the day (Figs. 12a and 12c)
22
and night (Figs. 12b and 12d) with weaker and lower elevation easterly flow aloft when snow
23
cover is removed. However, the intensity of the upvalley/upslope and downvalley/downslope
24
flows near the surface remains largely the same and is actually increased during the day on
25
the western side of the basin in the NONE simulation.
26
4 Ozone and air quality
20
1
4.1 1–6 February 2013 air quality overview
2
The January–March 2013 period featured seven extended periods of high ozone
3
concentrations in the Uintah Basin associated with persistent cold air pools (Stoeckenius and
4
McNally 2014). An upper-level trough and associated cyclone traversed the region between
5
28 and 30 January 2013 (as discussed in Section 3.2), effectively cleaning out the basin of
6
pollutants that had accumulated since mid-January and returning ozone concentrations to
7
background levels. However, the persistent CAP that set in on 1 February led to building
8
pollutant concentrations over the next several days. Fig. 7 shows the observed increase in
9
ozone at selected locations. Concentrations started out relatively low on 1 February (~20 to
10
60 ppb) and gradually built to a maximum of 154 ppb at Ouray on 6 February. A key
11
characteristic of ozone concentrations in many areas of the Uintah Basin is the maintenance of
12
high levels of ozone overnight far above background concentrations. A mobile transect of the
13
basin during the afternoon of 6 February shows elevated ozone concentrations in all locations,
14
but the highest values (up to 151 ppb) were generally seen in lower elevations and river
15
valleys (Fig. 13), particularly in the southeastern quadrant of the transect near Horsepool and
16
Ouray.
17
Data collected from ozonesondes and tethersondes from February 2013 show that the vertical
18
extent of maximum ozone concentrations was typically limited to 1700 m MSL and below, or
19
in the lowest 200–300 m of the boundary-layer (Schnell et al. 2014).
20
concentrations was noted above this level, with ozone concentrations returning to background
21
levels above 1900 m MSL (Karion et al. 2014).
22
4.2 Sensitivity of ozone concentrations to snow cover
23
While ozone concentrations in the Uintah Basin are known to be strongly controlled by the
24
presence of snow cover, there is uncertainty regarding what extent is due to the direct effects
25
of higher surface albedo on photolysis relative to the indirect meteorological effects
26
elucidated in Sect. 3: reduced near-surface temperatures; shallower mixed layer; and
A gradient in
21
1
possibly enhanced east-west cross-basin transport a few hundred meters above the surface.
2
Further, there are additional possible indirect effects such as chemical reactions in the surface
3
layers of the snow leading to critical precursor chemical compounds (Roberts et al. 2014).
4
We describe here some preliminary results that address the combined sensitivity to increased
5
surface albedo as well as meteorological changes in the boundary layer that result from the
6
presence of snow cover.
7
WRF output from the entire 6-day FULL and NONE simulations was imported into the Utah
8
Division of Air Quality’s (UDAQ) version of the Community Multi-Scale Air Quality
9
(CMAQ) model. The CMAQ model couples the meteorological data from WRF with an
10
emissions inventory from the Uintah Basin developed by UDAQ and chemistry-transport and
11
photochemical subsystems to simulate concentrations for a variety of chemical compounds
12
and pollutants (Byun and Schere 2006). The emissions inventory used in this study contained
13
the most complete and accurate data available for the Uintah Basin. However, the inventory
14
was prepared to represent emissions from 2011 based on growth of oil & gas activities since
15
2006 (Barickman et al. 2014). Data gathered during recent UBWOS field campaigns, along
16
with additional projects, will be incorporated into an emissions inventory update in 2014.
17
Since the emissions inventory and CMAQ are available from UDAQ on the 4 km grid, that
18
model was forced with WRF data from the 4 km nest shown in Fig. 2a. The objective of this
19
phase of the study is to simply assess the sensitivity of ozone concentration to snow cover
20
during a CAP. The potential shortcomings of driving CMAQ from imperfect atmospheric
21
information and emissions inventories as well as the limitations of CMAQ are not addressed.
22
The mean ozone concentrations near the surface averaged over the 6 afternoons (1100 to 1700
23
MST) from 1–6 February 2013 are generally 15–30% greater when the CMAQ model is
24
forced by the snow cover and meteorological conditions in the FULL simulation compared to
25
the absence of snow in the basin and concomitant conditions in the NONE simulation (Fig.
26
14a and 14b). Ozone concentrations simulated by the CMAQ model are highest in the
27
southeastern portion of the basin where the emission of ozone precursors (NOX and VOCs)22
1
from the emissions inventory is greatest (Barickman et al. 2014). The region where average
2
concentrations are greater than 75 ppb is ~6 times larger in the FULL simulation than that in
3
the NONE simulation. In addition, the peak ozone concentration simulated in the FULL case
4
is 16 ppb higher than that from the NONE case (Table 5) and the timing and magnitude of the
5
peak value on 6 February in the FULL case is comparable to that observed (see Figs. 7 and
6
13). Higher levels of ozone are simulated on average throughout the basin below 1800 m
7
when driven by the FULL simulation compared to the NONE simulation.
8
Time-averaged east-west cross sections of ozone (Figs. 14c and 14d) were averaged along a
9
24 km wide swath approximately 20-30 km south of the red line in Fig. 2b for the FULL and
10
NONE simulations. Comparison between the two simulations demonstrates the higher ozone
11
concentrations generated in the FULL run. The CMAQ model appears to lift the ozone
12
slightly higher in the FULL simulation than observed with the drop-off to background levels
13
(less than 60 ppb) taking place at 2000 m in the FULL simulation versus that observed
14
roughly below 1900 m at Horsepool (Karion et al. 2014).
15
The time evolution of ozone concentrations at Roosevelt and Horsepool is shown in Fig. 15.
16
Data at Roosevelt (Fig. 15a) indicates that CMAQ struggles to simulate the buildup of ozone
17
during the CAP period in the western portion of the basin. This is likely due to the spatial
18
variation of primary precursor emissions in the emissions inventory. A relatively small
19
amount of emissions exists in the western section when compared to the southeastern
20
quadrant of the basin (Barickman et al. 2014). Both the FULL and NONE simulations keep
21
concentrations generally near 50–60 ppb, and below the NAAQS of 75 ppb. Closer to the
22
specified primary precursor emission sources in the southeastern section of the basin, ozone
23
concentrations are noticeably higher at Horsepool with substantively higher concentrations
24
when snow cover is specified (Fig. 15b). Within the primary precursor emission source
25
region (not shown), the time evolution of ozone in the FULL simulation is quite similar to
26
that observed, again with notably higher concentrations than when snow is absent. The
27
differences in ozone concentrations among these three locations illustrates that the CMAQ23
1
model with the emission inventory available at this time struggles to capture the broad areal
2
extent of high ozone concentrations that are observed in this case.
3
A Time-height plot in Fig. 15c of ozone and potential temperature at Horsepool for the FULL
4
simulation help to analyze the evolution of pollutant concentration as a function of CAP
5
depth. While the highest concentrations of ozone are confined within the CAP, elevated
6
concentrations in excess of 75 ppb extend too far upwards in the afternoons into the base of
7
the inversions. In addition, CMAQ fails to build ozone concentrations each day through the
8
CAP event, as was observed (Fig. 7). Instead, the highest concentrations appear to be related
9
to the depth of the CAP and inversion base. Concentrations are high on 1 and 2 February,
10
when the CAP is shallow, then they decrease on the 3rd and 4th as the CAP deepens and the
11
inversion base lifts to ~1800 m. As the inversion base lowers again on 5 and 6 February,
12
concentrations increase with a maximum during the afternoon on the 6th. A similar evolution
13
is noted in the NONE simulation (not shown), but the CAP is much deeper, concentrations are
14
lower, and the maximum occurs on the afternoon of the 5th. This inverse relationship between
15
CAP depth and ozone concentrations is supported by observations, signifying its potential
16
importance. Intuitively, when the inversion base lowers, it effectively decreases the volume
17
of the shallow mixed-layer in the CAP resulting in higher ozone concentrations.
18
Finally, WRF shortwave radiation values were compared for FULL and NONE simulations.
19
Downward shortwave radiation at the surface is generally similar between both simulations,
20
however, on days with extensive cloud cover (such as mid-day on 3 and 4 February), the
21
FULL simulation has ~150 W m-2 more downward shortwave radiation (not shown). This is
22
likely because the ice-dominant clouds in the FULL simulation allow for greater
23
transmittance of shortwave radiation than the liquid-dominant clouds in NONE. For all days,
24
the upward (or reflected) shortwave radiation is greater in the FULL case, due to the high
25
albedo of snow compared to the bare ground in the NONE case. Total shortwave radiation
26
near the surface (sum of upward and downward radiation) was used as a crude
27
approximation for actinic flux (omitting, for example, the contribution due to scattering).24
1
The total shortwave radiation in the FULL simulation is over ~1.5 times greater than that in
2
the NONE simulation during the day. The greater total shortwave radiation in FULL then
3
contributes to higher photolysis rates and greater simulated ozone concentrations compared to
4
NONE.
5
5 Conclusions and discussion
6
Numerical simulations of a persistent CAP in the Uintah Basin from 1–6 February 2013 were
7
conducted to investigate the sensitivity of cloud microphysics and snow cover on CAP
8
evolution and structure. Terrain-flow interactions and basin- and smaller-scale thermally-
9
driven circulations were identified as possible means to infuse clean air into the basin as well
10
as transport pollutants within the basin. Output from selected meteorological simulations was
11
input into the CMAQ model to investigate the sensitivity of ozone production to snow cover.
12
The key findings of this study can be summarized as follows:
13

Nested numerical simulations at high resolution (1.33 and 4 km) driven on the
14
outer domain by continually updating analyzed lateral boundary conditions are
15
capable of reproducing the large-scale features of the regional circulation over the
16
Uintah Basin that control the basic morphology of the CAP.
17

The intensity, vertical structure, and boundary-layer flows within the Uintah Basin
18
CAP below ~500 m AGL are heavily influenced by the numerical treatment of
19
cloud microphysics and snow cover.
20

The default settings in the Thompson microphysics scheme produce dense, liquid-
21
phase low clouds and fog that were not observed, whereas restricting ice
22
sedimentation and cloud ice conversion to snow in the lowest model layers
23
resulted in more realistic vertical profiles of temperature and low clouds in this
24
case.
25
1

Intrusions of clean air into the basin as a result of terrain-flow interactions, east-to-
2
west cross-basin transport above the surface, and shallow thermally-driven slope
3
and valley circulations are important factors for mixing pollutants throughout the
4
Uintah Basin.
5

CMAQ model-derived estimates of ozone concentrations that are forced by the
6
most realistic emission inventories available and the best specification of the snow
7
surface and meteorological conditions tend to be adequate near major precursor
8
emission source regions in the southeast quadrant of the basin but too low
9
throughout most of the basin and during overnight hours.
10

The presence of snow cover in simulating high ozone concentrations is important
11
for two reasons: (1) it strengthens the CAP and increases low-level stability, and
12
(2) the high albedo surface increases photolysis rates, contributing to rapid ozone
13
production.
14
For this study, the modifications to the microphysics scheme (autoconversion from cloud ice
15
to snow and sedimentation of cloud ice turned off) were potentially available to be triggered
16
everywhere within the entire model simulation period in the lowest 15 model levels. To apply
17
the changes more generally requires some additional code modifications to activate them
18
based on appropriate boundary layer thresholds of stability and temperature. In this manner,
19
the changes to the microphysics scheme would only be triggered to deviate from the default
20
settings when specific criteria appropriate only to CAPs were met.
21
sedimentation of ice particles in the lowest 15 model layers can lead to excessive ice
22
accumulation but cloud ice is still able to convert to snow once it reaches a large enough
23
diameter through depositional growth.
24
The work presented here has been limited to a single persistent CAP event in early February
25
2013. In order to obtain a more thorough understanding of how cloud microphysics and snow
26
cover affect the evolution of persistent CAPs, their wind flow patterns, and resulting
The restricted
26
1
impacts on air quality, further studies are needed. A natural extension of this research would
2
be to include more cases and examine other basins with similar air quality problems to help
3
determine the applicability of this study’s findings.
4
Another challenge to model CAPs is the weakness of PBL schemes to handle low clouds,
5
vertical temperature profiles, and mixing in stably stratified conditions.
6
generally allow for too much turbulent mixing, which results in boundary layers that are too
7
deep (Holtslag et al. 2013). Mesoscale forecast models also poorly forecast whether or not
8
low clouds will form within a persistent CAP and when they will break up. They also have
9
considerable uncertainty in representing surface variables, often leading to 2-m temperature
10
warm biases, and in some cases poorly simulate the inversion above the boundary layer
11
(Reeves et al. 2011; Shin and Hong 2011). Alternative PBL parameterizations, such as the
12
Mellor-Yamada Nakanishi & Niino and Quasi-Normal Scale Elimination schemes, were not
13
tested in this study, but may show some improvement for CAPs. Regardless, further work to
14
improve parameterization schemes for modeling very stable boundary layers and their impact
15
on CAP simulations is needed (Baklanov et al. 2011).
16
Finally, snow cover and albedo were shown to have a prominent impact on CAP evolution
17
and air quality. This study highlights the need for improvements in the representation of
18
snow variables in meteorological and air quality models and analysis initialization fields in
19
regions with shallow, persistent snow cover. Proper treatment of snow using a snow physics
20
model driven by local atmospheric and chemical properties may be needed to obtain a
21
sufficiently accurate evolution of the snowpack and surface albedo. Application of the Noah
22
Multi-Parameterization land surface model (Niu et al. 2011), which features a three-layer
23
snow model may prove useful. Additional research is also needed to understand the complex
24
cycling of water over the thin snowpacks in the Uintah Basin and its impact on surface
25
albedo, i.e., the interplay of very small sublimation rates, formation of ice fogs, and
26
deposition of ice crystals back onto the snow surface.
Most schemes
27
1
Acknowledgements
2
The authors thank their colleagues for continuing support and discussion around the coffee
3
breaks. The editor thanks X. Y. Furore and another referee for assisting in evaluating this
4
paper.
5
28
1
References
2
Alcott, Trevor I., W. James Steenburgh, 2013: Orographic Influences on a Great Salt Lake–
3
Effect Snowstorm. Mon. Wea. Rev., 141, 2432–2450.
4
Baklanov, Alexander A., and Coauthors, 2011: The Nature, Theory, and Modeling of
5
Atmospheric Planetary Boundary Layers. Bull. Amer. Meteor. Soc., 92, 123–128.
6
Barickman, Patrick, Seth Lyman, Marc Mansfield, Howard Shorthill, Randy Anderson,
7
Jordan Evans, Chad Mangum, and Tate Shorthill, 2014: Emission Inventory Development
8
Activities. In T. Stoeckenius and D. McNally (Eds.), Final Report: 2013 Uinta Basin Winter
9
Ozone Study, Chapter 9, 18 pp.
10
Bell, Michelle L., Aidan McDermott, Scott L. Zeger, Jonathan M. Samet, Francesca
11
Dominici, 2004: Ozone and Short-term Mortality in 95 US Urban Communities, 1987–2000.
12
J. Am. Med. Assoc., 292, 2372–2378.
13
Billings, Brian J., Vanda Grubišić, Randolph D. Borys, 2006: Maintenance of a Mountain
14
Valley Cold Pool: A Numerical Study. Mon. Wea. Rev., 134, 2266–2278.
15
Byun, D., and K. L. Schere, 2006: Review of the governing equations, computational
16
algorithms, and other components of the Models-3 Community Multi-Scale Air Quality
17
(CMAQ) modeling system. Appl. Mech. Rev., 59, 51–77.
18
Cai, X.-M, Luhar A. K., 2002: Fumigation of pollutants in and above the entrainment zone
19
into a growing convective boundary layer: a large-eddy simulation. Atmos. Environ., 36,
20
2997–3008.
21
Chen, Fei, Jimy Dudhia, 2001: Coupling an Advanced Land Surface–Hydrology Model with
22
the Penn State–NCAR MM5 Modeling System. Part I: Model Implementation and Sensitivity.
23
Mon. Wea. Rev., 129, 569–585.
29
1
Clements, Craig B., C. David Whiteman, John D. Horel, 2003: Cold-Air-Pool Structure and
2
Evolution in a Mountain Basin: Peter Sinks, Utah. J. Appl. Meteor., 42, 752–768
3
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and
4
Wickham, J., 2011:
5
Conterminous United States, Photogram. Eng. Remote Sensing, 77, 858–864.
6
Hart, Kenneth A., W. James Steenburgh, Daryl J. Onton, 2005: Model Forecast
7
Improvements with Decreased Horizontal Grid Spacing over Finescale Intermountain
8
Orography during the 2002 Olympic Winter Games. Wea. Forecasting, 20, 558–576.
9
Holtslag, A. A. M., and Coauthors, 2013: Stable Atmospheric Boundary Layers and Diurnal
10
Cycles: Challenges for Weather and Climate Models. Bull. Amer. Meteor. Soc., 94, 1691–
11
1706.
12
Horel, J., and Coauthors, 2002: Mesowest: Cooperative Mesonets in the Western United
13
States. Bull. Amer. Meteor. Soc., 83, 211–225.
14
Iacono, M. J., J. S. Delamere, E. J. Mlawer, M. W. Shephard, S. A. Clough, and W. D.
15
Collins, 2008: Radiative forcing by long-lived greenhouse gases: Calculations with the AER
16
radiative transfer models. J. Geophys. Res., 113, D13103.
17
Janjić, Zaviša I., 1994: The Step-Mountain Eta Coordinate Model: Further Developments of
18
the Convection, Viscous Sublayer, and Turbulence Closure Schemes. Mon. Wea. Rev., 122,
19
927–945.
20
Jeong, Jee-Hoon, Hans W. Linderholm, Sung-Ho Woo, Chris Folland, Baek-Min Kim,
21
Seong-Joong Kim, Deliang Chen, 2013: Impacts of Snow Initialization on Subseasonal
22
Forecasts of Surface Air Temperature for the Cold Season. J. Climate, 26, 1956–1972.
23
Jiménez, Pedro A., Jimy Dudhia, J. Fidel González-Rouco, Jorge Navarro, Juan P. Montávez,
24
Elena García-Bustamante, 2012: A Revised Scheme for the WRF Surface Layer Formulation.
25
Mon. Wea. Rev., 140, 898–918.
Completion of the 2006 National Land Cover Database for the
30
1
Kain, John S., 2004: The Kain–Fritsch Convective Parameterization: An Update. J. Appl.
2
Meteor., 43, 170–181.
3
Karion, Anna, Samuel Oltmans, Gabrielle Petron, Colm Sweeney, Russ Schnell, 2014:
4
Analysis of Aircraft Observations. In T. Stoeckenius and D. McNally (Eds.), Final Report:
5
2013 Uinta Basin Winter Ozone Study, Chapter 4, 26 pp.
6
Katurji, Marwan, Shiyuan Zhong, 2012: The Influence of Topography and Ambient Stability
7
on the Characteristics of Cold-Air Pools: A Numerical Investigation. J. Appl. Meteor.
8
Climatol., 51, 1740–1749.
9
Lareau, Neil P., Erik Crosman, C. David Whiteman, John D. Horel, Sebastian W. Hoch,
10
William O. J. Brown, Thomas W. Horst, 2013: The Persistent Cold-Air Pool Study. Bull.
11
Amer. Meteor. Soc., 94, 51–63.
12
Lareau, Neil P., 2014: The Dynamics of Persistent Cold-Air Pool Breakup. Ph.D. Thesis.,
13
University of Utah, 138pp.
14
Lareau, Neil P., J. Horel, 2014: Dynamically Induced Displacements of a Persistent Cold-Air
15
Pool. Bound.-Layer Meteor., accepted.
16
Lippmann M., 1993: Health effects of tropospheric ozone: review of recent research findings
17
and their implications to ambient air quality standards. J. Expo. Anal. Environ. Epidemiol., 3,
18
103–129.
19
Lyman, Seth, Howard Shorthill, 2013: Final Report: 2012 Uintah Basin Winter Ozone and
20
Air Quality Study, 285 pp.
21
Malek, E., T. Davis, R. Martin, and P. Silva, 2006: Meteorological and environmental aspects
22
of one of the worst national air pollution episodes (January, 2004) in Logan, Cache Valley,
23
Utah, USA. Atmos. Res., 79, 108–122.
31
1
Niu, Guo-Yue, et al., 2011: The community Noah land surface model with
2
multiparameterization options (Noah‐MP): 1. Model description and evaluation with local-
3
scale measurements. J. Geophys. Res., 116, D12109.
4
Pollack, I. B., Thomas B. Ryerson, Michael Trainer, J. A. Neuman, James M. Roberts, David
5
D. Parrish, 2013: Trends in ozone, its precursors, and related secondary oxidation products in
6
Los Angeles, California: A synthesis of measurements from 1960 to 2010. J. Geophys. Res.
7
Atmos., 118, 1–19.
8
Reeves, Heather Dawn, David J. Stensrud, 2009: Synoptic-Scale Flow and Valley Cold Pool
9
Evolution in the Western United States. Wea. Forecasting, 24, 1625–1643.
10
Reeves, Heather Dawn, Kimberly L. Elmore, Geoffrey S. Manikin, David J. Stensrud, 2011:
11
Assessment of Forecasts during Persistent Valley Cold Pools in the Bonneville Basin by the
12
North American Mesoscale Model. Wea. Forecasting, 26, 447–467.
13
Roberts, James M., and Coauthors, 2014: Intensive Chemical Measurements at Horsepool. In
14
T. Stoeckenius and D. McNally (Eds.), Final Report: 2013 Uinta Basin Winter Ozone Study,
15
Chapter 5, 96 pp.
16
Salmond, J. A., 2005: Wavelet analysis of intermittent turbulence in the very stable nocturnal
17
boundary layer: Implications for the vertical mixing of ozone. Bound.-Layer Meteor., 114,
18
463–488.
19
Schnell, Russ C., Samuel J. Oltmans, Ryan R. Neely, Maggie S. Endres, John V. Molenar,
20
Allen B. White, 2009: Rapid Photochemical Production of Ozone at High Concentrations in a
21
Rural Site during Winter. Nat. Geosci., 2, 120–122.
22
Schnell, Russ, Bryan Johnson, Patrick Cullis, Chance Sterling, Emrys Hall, Rob Albee, Allen
23
Jordan, Jim Wendell, Samuel Oltmans, Gabrielle Petron, Colm Sweeney, 2014: Tethered
24
Ozonesonde and Surface Ozone Measurements in the Uintah Basin, Winter 2013. In T.
32
1
Stoeckenius and D. McNally (Eds.), Final Report: 2013 Uinta Basin Winter Ozone Study,
2
Chapter 8, 48 pp.
3
Shin, Hyeyum H., Song-You Hong, 2011: Intercomparison of Planetary Boundary-Layer
4
Parameterizations in the WRF Model for a Single Day from CASES-99. Bound.-Layer
5
Meteor., 139, 261–281.
6
Shupe, Matthew D., Janet M. Intrieri, 2004: Cloud Radiative Forcing of the Arctic Surface:
7
The Influence of Cloud Properties, Surface Albedo, and Solar Zenith Angle. J. Climate, 17,
8
616–628.
9
Silcox, Geoffery D., Kerry E. Kelly, Erik T. Crosman, C. David Whiteman, and Bruce L.
10
Allen, 2012: Wintertime PM2.5 Concentrations in Utah’s Salt Lake Valley during Persistent
11
Multi-day Cold-Air Pools. Atmos. Environ., 46, 17–24.
12
Stoeckenius, Till, Dennis McNally, 2014: Final Report: 2013 Uinta Basin Winter Ozone
13
Study, Chapter 2, 55 pp.
14
Thompson, Gregory, Paul R. Field, Roy M. Rasmussen, William D. Hall, 2008: Explicit
15
Forecasts of Winter Precipitation Using an Improved Bulk Microphysics Scheme. Part II:
16
Implementation of a New Snow Parameterization. Mon. Wea. Rev., 136, 5095–5115.
17
U.S. EPA, 2006: Air Quality Criteria for Ozone Related Photochemical Oxidants, Volume 1,
18
821 pp.
19
U.S. EPA, 2013: National Ambient Air Quality Standards (NAAQS). [Available online at
20
http:// http://www.epa.gov/air/criteria.html.]
21
Verzijlbergh, R. A., H. J. J. Jonker, T. Heus, J. Vila-Guerau de Arellano, 2009: Turbulent
22
Dispersion in cloud-topped boundary layers. Atmos. Chem. Phys., 9, 1289–1302.
23
Whiteman, C. D., S. Zhong, W. J. Shaw, J. M. Hubbe, X. Bian, J. Mittelstadt, 2001: Cold
24
Pools in the Columbia Basin. Wea. Forecasting, 16, 432–447.
33
1
Zängl, Günther, 2005a: Formation of Extreme Cold-Air Pools in Elevated Sinkholes: An
2
Idealized Numerical Process Study. Mon. Wea. Rev., 133, 925–941.
3
Zängl, Günther, 2005b: Wintertime Cold-Air Pools in the Bavarian Danube Valley Basin:
4
Data Analysis and Idealized Numerical Simulations. J. Appl. Meteor., 44, 1950–1971.
34
1
Table 1. The History of America from Discovery to Present.
Date
Event
Ref.
1492
Discovery
Columbus (1492)
1776
Independence
James et al. (1776)
1954
Nothing much
Smith and Weston (1954)
1999
Present
Phillips (1999)
2
35
1
2
3
Figure 1. Sebastiano del Piombo painted this portrait thirteen years after Columbus’s death
4
(from the Columbus Navigation Homepage).
5
36
1
2
3
Figure 2. Columbus’s voyage to the New World (a rough approximation).
37