Thursday, November 10, 2022

A Deep Dive into the Great Salt Lake Effect

This is a reproduction of a report prepared and presented to the Great Salt Lake Advisory Council at their 14 September 2022 meeting.  It has been modified for formatting, to remove personal information, and to correct a few typos.  I thank students and colleagues whose research or comments contributed to and improved the manuscript.

Contributions of Lake-Effect Periods to Precipitation and Streamflow in Northern Utah

W. James Steenburgh
Professor of Atmospheric Sciences
University of Utah

Executive Summary

Lake-effect periods are sometimes produced during cold-air outbreaks over the Great Salt Lake, contributing to snowfall and streamflow in the surrounding mountains. Prior research indicates that precipitation produced by lake-effect periods is greatest south and southeast of the Great Salt Lake and contributed 5.1-8.4% of the cool-season (16 September – 15 May) precipitation at observing sites in the Cottonwood Canyons and Oquirrh Mountains from 1998–2009. During this study period, the lake was at or below its average historical area. No studies have carefully examined the long-term influence of lake area on lake-effect precipitation or the contribution of lake effect to streamflow, underscoring the need for further research in these areas. Studies do indicate that salinity reduces the coverage and intensity of lake-effect storms and could exacerbate precipitation losses should lake levels and area continue to decline.

1. Introduction

The Great Salt Lake is the largest body of water in the contiguous western United States and the largest salt lake in the western hemisphere. It is a terminal lake with no outlet, so its elevation and area change because of variations and trends in climate, including snowfall, streamflow, and evaporation, as well as water diversions for human activities. Historical levels at Saltair on the south shore have been as high as 4211.6 feet in 1986 and as low as 4190.4 in 2021, although preliminary data indicate that the lake dropped below this level during the summer or 2022 (Lake surface elevations based on the USGS observing site at Saltair Boat Harbor (1001000) and based on National Geodetic Vertical Datum of 1929 (NGVD 1929). Elevations north of the Union Pacific rock-fill railroad causeway may differ after completion of its construction in 1959.).  Correspondingly, the lake area has been as large as 3,300 square miles and as small as 950 square miles (Figure 1).

Figure 1. Landsat satellite imagery of the Great Salt Lake based on data collected in September 1987 (left) and April/May 2021 (right) when the lake was near its high and low stands, respectively.  Images courtesy of the U.S. Geological Survey.  

Due to the accumulation of sodium chloride and other salts, the Great Salt Lake is hypersaline and much saltier than ocean water. There is a stark salinity contrast between the northern and southern halves of the lake, which are separated by a rock-fill railroad causeway that limits mixing between the two halves. The northern half receives very little freshwater inflow and typically has a salinity near 27%. The southern half receives most of the freshwater inflow and has a salinity that has been as low as 6% during higher lake stands, but since 2010 has varied between 10 and 18%. For comparison, ocean water has an average salinity of 3.5%. Lake salinity can be locally low near freshwater inlets.

The Great Salt Lake is also very shallow. At an elevation of 4200 feet, it has an average depth of 16 feet and a maximum depth of 33 feet. As a result, the lake-surface temperature responds relatively quickly to changes in air temperature associated with fronts and other weather systems. It also warms quickly in the spring, which contrasts with larger, deeper lakes such as the Great Lakes of eastern North America, which tend to warm or cool more slowly in response to weather systems and seasonal changes. During winter, because of its salinity, the Great Salt Lake develops little ice and can achieve low temperatures (near 28˚F). 

The Great Salt Lake has multifaceted influences on the climate and water resources of northern Utah. As directed by the Utah Department of Natural Resources and Great Salt Lake Advisory Council, this report summarizes current understanding of the influence lake-effect precipitation generated by the Great Salt Lake on precipitation and streamflow.

2. The Great Salt Lake effect

a. What is lake-effect precipitation? 

Lake-effect precipitation is precipitation that is produced or enhanced when cold air passes over a relatively warm body of water. This occurs due to the transfer of heat and moisture into the atmosphere, which destabilizes the atmosphere and leads to atmospheric circulations that initiate and organize clouds and precipitation systems. Lake-effect and related sea- or ocean-effect precipitation often falls as snow, although it can fall as rain, especially in the fall or spring at lower elevations in warmer climates. The most prolific lake-effect snowfall occurs in the snowbelts near the Great Lakes of eastern North America and in the heavy snow region of Japan near the Sea of Japan where frequent cold-air outbreaks occur over large water bodies. Less prolific but sometimes disruptive lake-effect storms can also be produced by smaller water bodies including the Great Salt Lake, Utah Lake, Bear Lake, Pyramid Lake, and Lake Tahoe. The water that falls in these storms can be traced back to both the upstream atmosphere and evaporation from the lake surface. For small lakes, upstream moisture is often critical for lake-effect development.

b. Characteristics of the Great Salt Lake effect

Great Salt Lake-effect precipitation occurs most frequently during cold-air outbreaks following the passage of a cold front when the flow is westerly, northwesterly, or northerly. The development, intensity, and coverage of lake-effect precipitation depends on several factors including the lake temperature and salinity, air temperature and humidity, wind direction and speed, and other factors. Additionally, low-level flow convergence frequently initiates and organizes lake-effect clouds and precipitation. Such convergence can be produced by flow interactions with the topography or land breezes from the lake shorelines (Figure 2).

Figure 2: Schematic depiction of the convergence of land breezes and the development of a lake-effect storm over the Great Salt Lake.  Source: Steenburgh (2014).  © University Press of Colorado.

Most Great Salt Lake-effect precipitation periods are disorganized and produce scattered or widespread precipitation (Figure 3a). Sometimes they organize into narrow bands that produce heavy, localized snowfall with rates that can approach 3 inches per hour (Figure 3b). It is also possible for lake effect to enhance or occur simultaneously with other precipitation features (Figure 3c). For this reason, it is not possible to completely disentangle lake-effect and non-lake-effect precipitation.

Figure 3. Radar imagery of a) widespread lake-effect precipitation, b) banded lake-effect precipitation, and c) lake-effect precipitation with other precipitation features.  Source: Alcott et al. (2013).  © American Meteorological Society. 

The ability to identify and monitor Great Salt Lake-effect precipitation increased significantly in 1994 with the installation of a NOAA/National Weather Service radar on Promontory Point. The characteristics described below are based lake-effect periods identified during the 1998 to 2010 cool seasons (16 September to 15 May). During this study period, the area of the Great Salt Lake declined from 1750 to 1200 square miles, which corresponds to near average (1700 square miles) and roughly halfway between average and the historical minimum (950 square miles).

During this study period, there were an average of 13 lake-effect periods per cool season with as few as 3 in 2005 and as many as 20 in 2010. Many of these periods were short lived, with less frequent but intense periods responsible for most of the precipitation accumulation (see section 2c). Lake-effect periods were most common from mid-October to mid-December and in early April. The early April peak is unusual compared to other bodies of water that are deeper and tend to be cold in the spring. In contrast, the shallowness of the Great Salt Lake allows it to warm rapidly, enabling it to generate lake-effect precipitation in the spring when a cold-air outbreak occurs after a warm period.

c. Contribution of lake-effect periods to cool-season precipitation

Estimating the contribution of lake-effect periods to cool-season precipitation and snowpack is not straightforward. For the estimates below, we identified lake-effect periods using radar imagery from the 1998–2009 cool seasons and determined how much precipitation they produced at National Resources Conservation Service (NRCS) Snowpack Telemetry (SNOTEL) stations. The estimates include some non-lake-effect precipitation since there are times when lake effect enhances or occurs with precipitation produced by other weather systems. On the other hand, it is possible that the lake influences or enhances some mountain snowstorms in ways that are not easily identified in radar.

During the study period, precipitation during lake-effect periods was greatest at SNOTEL stations in the Cottonwood Canyons and the Oquirrh Mountains (Figure 4a). In the Cottonwood Canyons, the average cool-season water equivalent precipitation during lake-effect periods was 2.06 and 2.38 inches at the Mill D North (Big Cottonwood Canyon) and Snowbird (Little Cottonwood Canyon) SNOTEL stations, respectively. In the Oquirrh Mountains it was 2.12 and 2.37 inches at the Rocky Basin-Settlement (Settlement Canyon) and Dry Fork (Butterfield Canyon) SNOTEL stations, respectively. In the Wasatch Range, average cool-season liquid equivalent precipitation during lake-effect periods generally decreased northward of the Cottonwood Canyons with 1.46, 1.61, 1.59, and 1.01 inches at the Parleys Summit, Lookout Peak, Farmington, and Ben Lomond Peak SNOTEL stations, respectively. Values at other stations ranged from 0.33 to 1.32 inches, the latter at the Payson Ranger Station southeast of Utah Lake. At some of these sites, especially those in the Bear River Range and western Uintas, much this precipitation was produced by non-lake-effect precipitation features that occurred simultaneously with lake-effect precipitation.

Figure 4. a) Water equivalent of precipitation during cool-season (16 September – 15 may) lake-effect periods during the 1998–2009 water years.  b) Fraction of cool-season precipitation produced during lake-effect periods during the 1998–2009 water years.  Lake shores based on high and low stands during the study period with Bear, Weber, and Jordan–Provo River Basins annotated.  Data from Yeager et al. (2013). 

The average fraction of total cool-season precipitation produced during lake-effect periods was 8.4% and 6.3% at Dry Fork and Rocky Basin-Settlement in the Oquirrh Mountains and 5.9% and 5.1% at Mill D North and Snowbird in the Cottonwood Canyons (Figure 4b). The Oquirrh Mountains are drier than the Cottonwoods during non-lake-effect periods, yielding the higher fractions. The lowest fractions are at the Ben Lomond Peak and Trail SNOTEL stations in the northern Wasatch (2.0% and 1.6% respectively).

Observations from Snowbird provide additional insights into the characteristics of lake-effect periods. During the 12 cool-season study period, just 13 lake-effect periods, or about 10% of the periods, produced 50% of the lake-effect precipitation at Snowbird. Thus, approximately one large storm per year was responsible for half of the lake-effect precipitation at Snowbird, equating to about 1.2 inches of water equivalent. From year-to-year, the amount of cool-season precipitation during lake-effect periods at Snowbird varied from as high as 5.04 inches in the 2002 water year (about 12% of the precipitation that cool season) to as low as 0.51 inches in the 2003 water year (about 1% of the precipitation that cool season). Such a wide swing in back-to-back years illustrates that year-to-year variations in lake effect cannot be explained solely due to changes in lake area. Meteorology, especially the frequency and characteristics of cold-frontal passages and associated cold-air outbreaks, also plays an important role. Lake area could have longer-term implications for precipitation, however, as discussed in section 2e.

d. Contribution of lake-effect periods to streamflow

For northern Utah and the major drainage basins of the Great Salt Lake, between 50% and 80% of the annual streamflow occurs during the four-month period of snowmelt (April through July), but even during low-flow periods stream water is predominantly composed of snowmelt that recharged groundwater and is released slowly during the year. The percentage varies from basin-to-basin and from year-to year depending on many factors including the altitude, aspect, and other geographic and ecological characteristics of the basin, antecedent groundwater storage, amount of precipitation that falls as snow and is retained in the end-of-season snowpack, and whether it is a high or low snow season.

To our knowledge, there are no peer-reviewed studies estimating the contribution of lake-effect precipitation to streamflow in any northern Utah hydrologic basin. The estimates in section 2c are the contribution of lake-effect periods to precipitation, which is not equivalent to the contribution to streamflow. With these caveats in mind, we provide some discussion below for general guidance but emphasize that further research is needed.

As discussed in section 2c, the amount of precipitation produced during lake-effect periods is greatest in the Cottonwood Canyons and the Oquirrh Mountains. However, it is likely that the percentage of lake-effect precipitation that is converted to streamflow is higher in the Cottonwoods where there is more high altitude, north-facing terrain and the climatology favors greater precipitation and a deeper snowpack. These factors favor a greater conversion of precipitation to streamflow. Therefore, we anticipate that lake effect contributes to a larger volume of streamflow in Little and Big Cottonwood Creeks than to the creeks issuing from the Oquirrh Mountains. It is possible, however, that lake effect contributes a larger fraction of streamflow to the creeks issuing from the Oquirrh Mountains where precipitation during non lake-
effect periods is lower than in the Cottonwoods.

As described in section 2c, lake-effect periods produced on average 5.1–5.9% of the precipitation at SNOTEL sites in the Cottonwood Canyons during the 1998–2009 cool seasons. It is not known if the contribution of lake effect to streamflow in Little and Big Cottonwood Creeks scales similarly, but we suggest that it may be close. It could even be slightly higher. This is because lake-effect storms tend to be colder, with lower snow levels, thus contributing to a greater fraction of the snowpack at low elevations. Additionally, the addition of lake-effect snow may increase the streamflow yield even the non-lake-effect by creating a deeper snowpack. On the other hand, these relatively low numbers illustrate that non-lake-effect precipitation is the primary driver of streamflow in the Cottonwood Canyons.

The difficulty in relating the small amounts of precipitation associated with lake-effect storms to streamflow arises primarily from the high site-to-site and year-to-year variability in runoff efficiency. In Little Cottonwood Canyon average annual streamflow is 63% of average annual precipitation, but that value ranges from 45% to 80% from year to year. In Big Cottonwood Canyon average annual streamflow is 49% of average annual precipitation but values range from 32% to 65%. Thus, the natural variability in streamflow generation is significantly larger than the lake effect. In addition, a warming climate is resulting in snowmelt beginning earlier and less efficient streamflow generation/ runoff efficiency. Lake-effect snow in the spring can also act to increase how much sunlight is reflected by snow surface, which may delay the melt following a storm. This is likely a small effect but would contribute to more efficient
streamflow generation.

e. Influence of lake size on precipitation

In any given year, lake-effect precipitation depends on lake characteristics and meteorology. During the 1998–2009 study period when the Great Salt Lake area declined from 1750 to 1200 square miles, lake area poorly explained year-to-year variations in lake-effect precipitation. The three biggest lake-effect seasons at Snowbird, for example were 2002 when the lake area was near the 1500 square mile average for the study period, 1998 when the lake area was near the 1750 square mile maximum for the study period, and 2009 when the lake area was near the 1200 square mile minimum for the period. This is because the characteristics of cold-air outbreaks that occur each cool season, especially those that contribute to intense lake-effect periods, also affect lake-effect precipitation. This obscures the signal of lake area if one examines a relatively short period of about a decade.

Nevertheless, it is likely that lake area does influence the characteristics of lake-effect storms and that this would be detectable over a longer record of multiple decades or potentially illustrated using regional climate modeling. We are unaware of any peer-reviewed studies that have attempted to do this. Further, we suggest that there may be tipping points at which a small change in lake elevation (and area) produce is a significant shift in lake-effect characteristics. Lake elevation affects not only area but also shape, which in turn influences the characteristics of lake-effect storms. At high-stand (4211 feet) and 4200 feet elevations, the Great Salt Lake occupies the Farmington and Bear River Bays (Figure 5). However, near and below the historical low stand (4291 feet), the lake is confined to an elongated region along the axis of the North and South Arms. At these levels, the lake is quite narrow, potentially reducing the range of flow directions that can effectively generate lake effect. For lower lake levels, declines in lake-effect precipitation might be greater, for example, in the Bountiful-area mountains than in the Oquirrh Mountains. Research is needed to explore these effects.

Figure 5. Great Salt Lake coverage at 4211 feet (high stand), 4200 feet (historical average), 4191 feet (low stand), 4180 feet, and 4170 feet.  Great Salt Lake bathymetry source: Tarboton, D. (2017). Great Salt Lake Bathymetry, HydroShare, http://www.hydroshare.org/resource/582060f00f6b443bb26e896426d9f62a. Hillshade sources: Esrii, USGS, FAO, NOAA. 

f. Influences of lake salinity on snowfall

Salinity reduces the transfer of water from the lake to the atmosphere through evaporation. The reduction increases with salinity, especially at high salinities like those found in the north half of the lake. Due to the contrast in salinity, the transfer of water from the lake to the atmosphere through evaporation tends to be greater over the south half of the lake than the north half.

Scientists have incorporated these effects into computer model simulations of lake-effect storms. For one event, the salinity produced a 17% reduction in snowfall compared to a simulation with fresh water. It is likely that increases in salinity accompanying a shrinking Great Salt Lake, which would mainly occur in the south half of the lake, would further reduce lake-effect precipitation.

3. Sources and bibliography 

Alcott, T. I., and W. J. Steenburgh, 2013: Orographic influences on a Great Salt Lake-effect snowstorm. Monthly Weather Review, 141, 2432–2450.

Alcott, T. I., W. J. Steenburgh, and N. F. Laird, 2012: Great Salt Lake-effect precipitation: Observed frequency, characteristics, and environmental factors. Weather and Forecasting, 27, 954–971.

Bardsley, T., A. Wood, M. Hobbins, T. Kirkham, L. Briefer, J. Niermeyer, and S. Burian, 2013: Planning for an uncertain future: Climate change sensitivity assessment toward adaptation planning for public water supply. Earth Interactions, 17, 1–26, https://journals.ametsoc.org/view/journals/eint/17/23/2012ei000501.1.xml.

Brooks, P. D., A. Gelderloos, M. A. Wolf, L. R. Jamison, C. Strong, D. K. Solomon, G. J. Bowan, S. Burian, X. Tai, S. Arens, L. Briefer, T. Kirkham, and J. Stewart, 2021: Groundwater-mediated memory of past climate controls water yield in snowmelt-dominated catchments. Water Resources Research, 57, e2021WR030605. https://doi.org/10.1029/2021WR030605.

Onton, D. J., and W. J. Steenburgh, 2001: Diagnostic and sensitivity studies of the 7 December 1998 Great Salt Lake-effect snowstorm. Monthly Weather Review, 129, 1318–1338.

Steenburgh, W. J., and D. J. Onton, 2001: Multiscale analysis of the 7 December 1998 Great Salt Lake-effect snowstorm. Monthly Weather Review, 129, 1296–1317.

Steenburgh, W. J., S. F. Halvorson, and D. J. Onton, 2000: Climatology of lake-effect snowstorms of the Great Salt Lake. Monthly Weather Review, 128, 709–727.

USGS, 2022: Great Salt Lake, Utah. https://pubs.usgs.gov/wri/wri994189/PDF/WRI99-4189.pdf. Downloaded 3 Aug 2022.

USGS, 2022: Water Quality Samples for the Nation, USGS 10010000 Great Salt Lake near Saline,


Yeager, K. N., W. J. Steenburgh, and T. I. Alcott, 2013: Contributions of lake-effect periods to the
cool-season hydroclimate of the Great Salt Lake Basin. Journal of Applied Meteorology and
Climatology, 52, 341–362.

9 comments:

  1. Can you have lake effect without low lever convergence or it is required to be officially be called lake effect?

    I wish it was more publicized that on average lake effect only contributes 5.1-8.4% of the cool season precipitation. So for places that average 500" of snow a year and assuming it all falls as snow the lake on average only contributes 25.5-42" a snow each year. A small amount for all of the hype it gets!

    ReplyDelete
    Replies
    1. It's pretty difficult to generate precipitation without ascent and you can't get ascent without low-level convergence. The two are intertwined. That said, the mechanisms driving the convergence might not be lake driven and could be related to mountain effects (we see this in lake-effect simulations) or transient weather features (sometimes lake-effect strengthens or initiated when surface boundaries move across the lake).

      Delete
    2. Thanks, I never consider it a lake effect event without some type of low level convergence. This seems to be a BIG point that is missed by so many broadcasters and even a few meteorologists, they talk non stop about lake effect without low level convergence. Like you said without ascent it is hard (impossible?) to get precipitation. The king of ascent in these parts, as you know, is orographics. Any idea of what percent of the annual snowfall in the Wasatch is due to orographics? A guess is fine.

      Delete
    3. The ratio of monthly precip at Alta to Salt Lake City is a guestimate. See https://wasatchweatherweenies.blogspot.com/2013/01/cold-fronts-and-orographic-ratio.html.

      Delete
    4. So looking at the article 40% seems to be a good guess?

      Delete
    5. Regarding precipitation, technically you can get precipitation without "net" ascent though this is likely rare. You can form a cloud via mixing of 2 subsaturated air parcels if the temperature and RH are sufficient since the saturation water vapor mixing ratio increases exponentially with temperature. To go from non-precipitating drops to precipitation with only liquid is difficult without ascent unless the airmass is exceedingly clean, but if ice is formed and the temperature is sufficiently cold, it can grow quickly at the expense of liquid droplets due to the difference in vapor pressure over liquid and ice. Note that this could happen in portions of a cloud with subsiding motion or in the cloud formed via mixing, though in reality, it is likely exceedingly rare that ascent isn't a primary contributor to precipitation formation.

      Delete
    6. Agreed. I was trying to avoid getting into the weeds :-).

      Delete
  2. Thank you for posting this. My understanding is that for the study period, radar was used to identify events where there was lake effect and the total amounts of precipitation observed at the SNOTEL sites during these events was used to calculate the percentage of precipitation from lake effect events. Therefore, there was no estimate of the amount of precipitation during lake effect events that fell due to large scale flow. Is that a correct statement? If so, I do understand how difficult it would be to estimate the amount of the total for each event due to lake effect enhancement of precipitation.

    ReplyDelete
    Replies
    1. Yes. We didn't try to separate lake-effect and non-lake-effect precipitation during lake-effect periods. That was a difficult ask. On the other hand, we also didn't try to account for more subtle lake influences that are less obvious.

      Delete