The Sierra Nevada of eastern California receives heavy snowfall each year. However, it is the snowstorms that deposit heavy snowfall in a relatively short period of time that can cause major inconveniences and even life-threatening situations for the residents and visitors to the region. Some of these snowstorms are so extreme as to become legendary, and with increased population in this region a synoptic climatology of these extreme snowstorms can be a useful tool for assessing snowfall potential by operational forecasters. Additionally, the hydrological and climatological implications of extreme Sierra Nevada snowfalls are important for state and local resource- and emergency-planning purposes.
A climatology of these snowstorms will be presented. The period of study will include the snowfall seasons (October–May) 1949/50 through 2004/05. A total of 542 snowstorms occurred during these 56 snowfall seasons. These snowstorms were analyzed to determine any common synoptic features. The most intense snowstorms in the highest decile of snowfall totals were analyzed in more detail in order to determine the parameters associated with these strongest snowstorms.
Upper-level synoptic and thermodynamic characteristics associated with each snowstorm were then diagnosed to determine what common synoptic hydrodynamic and thermodynamic parameters the snowstorms share. Synoptic patterns were studied using the National Centers for Environmental Prediction (NCEP) model reanalysis data. Wind speeds at 200 hPa, and height anomalies at 500 hPa, were analyzed for each snowstorm from 3 days prior to the start of snowfall and continuing through the end of the storm. Anomalies and the transport of precipitable water were studied in order to determine the relative amount of moisture that was available to each snowstorm.
A conceptual model for forecasting the strongest snowstorms was developed. Key findings include the following: 1) the importance of a fetch of moisture from the subtropics with relatively large positive moisture anomalies, 2) the importance of the atmospheric moisture stream being normal to the Sierra, 3) the low static stability accompanying these snowstorms, and 4) the importance of relatively strong upper-level dynamics, which helped to intensify the systems as they approached the Sierra.
Translated as “snow-covered mountain range” in Spanish, the Sierra Nevada (especially the elevations of the High Sierra around Mount Whitney) can be snow covered for most of the year. The range is known for heavy and even legendary snowfalls including the infamous Donner Pass snowstorm of January 1952 (Ludlum 1952). Since snowfalls in the Sierra Nevada can be extreme, relative to most of North America, they require considerable analyses to understand their trends and unique organizational factors.
Research in the western United States has revealed upper-level circulation patterns that are associated with heavy snowfall. Younkin (1968) studied 22 snowstorms west of 100°W during four snowfall seasons (1963/64 through 1966/67). Younkin found that heavy snowfall occurred between the 5340-geopotential-meter (gpm) and the 5460-gpm 1000–500-hPa contours. The 5400-gpm contour is still used as an initial indicator of the approximate location of the rain–snow line for locations both west and east of the Continental Divide.
Klein et al. (1968) studied upper lows at various mandatory pressure levels and their relationships to winter precipitation over the Intermountain West. They found that low pressure areas at the lower levels (e.g., 850 hPa) had the most influence on the occurrence of precipitation. Topographic elevation also was an important factor. Since the 850-hPa level is very close to the surface in the intermountain United States, they stated that “precipitation, at least in winter, appears to be more closely related to the circulation at the surface than at any upper level.”
Reitan (1974) studied cyclogenesis for the period 1951–70, and his analysis showed a maximum in winter cyclone frequency over the Gulf of Alaska during this 20-yr period. Results from a study by Gyakum et al. (1989) also found cyclone formation and intensification areas east of Japan and in “another important region situated in the area extending southward from the Gulf of Alaska.” The cyclone maximum in the Gulf of Alaska matches well the genesis region of many of the cyclones in this study of Sierra Nevada snowstorms.
With the exception of these studies, little has been written about the forecasting of heavy snowfall in the Sierra from the perspective of a multiday lead time. As part of the science necessary to address this deficiency, a study was performed of the most significant snowstorms to affect the Sierra Nevada during the last half-century in order to update our knowledge base of contemporary trends in these phenomena. This manuscript presents a synoptic climatology of these snowstorms. This contemporary climatology will be important in affirming, for hydrological purposes, contemporary extrema in snowfall amounts, distribution, and duration.
Heavy snowfall can have a major effect on travel in the Sierra Nevada. Although the upper elevations are sparsely populated, many highways (including a major interstate highway) traverse the range and serve a growing winter tourism industry, in addition to interstate travel and transport. If one or more of these roads are blocked, it can have a major economic impact on the region including the important agricultural goods transportation industry in California. Grumm and Hart (2001) have stated that “extreme weather events have the greatest economic and human impact due to either their intensity or areal coverage.”
This study will also consider key synoptic parameters, based on an analysis of the modest current state of the literature, that are likely to be the most important in forecasting heavy snowfall in the Sierra. In the next section we will describe the data and methodology employed. Results of the climatological analyses and comparison to synoptic features will be presented in section 3. In this section the kinematics associated with upper-level jet maxima will be considered along with the dynamics associated with 500-hPa systems. Moisture transport will also be considered as a major component of snowstorms impacting the Sierra. Finally, section 4 will summarize the results.
2. Data and methodology
Daily data for this study were taken from the U.S. Department of Commerce publications (available online at http://www7.ncdc.noaa.gov/IPS/cd/cd.html) for the snowfall seasons (October–May) of 1949/50 through 2004/05. All data were converted to metric units.
Snowfall data were used for twenty-three locations along the entire length of the Sierra Nevada (Fig. 1). Fifteen locations are on the west side of the Sierra crest, with eight east of the crest. The data for the locations in this study (divided among the northern, central, and southern Sierra) are listed in Table 1.
None of the twenty-three sites was active continuously during the 56 yr of this study. A few locations were inactive for a decade or more at a time. However, all of these sites were active for a majority of the 56 yr and together form a robust record of snowfall data for the Sierra.
Daily snowfall data were initially analyzed in order to determine how often snowfall of at least 30 cm (12 in.) occurred during one calendar day and/or 60 cm (24 in.) occurred during the duration of a snowstorm at one or more study locations. For the purposes of this study, a snowstorm (or “event”) is defined as starting when a majority of the study locations reported at least 5 cm (2 in.) of snowfall on a calendar day before the 30-cm (or 60 cm) criterion was met, and ended when a majority of the sites recorded their last 5-cm daily snowfalls. An additional requirement was that each snowstorm, or event, was characterized by being associated with a distinct 500-hPa low pressure area or trough.
The analysis showed that, for the 56 snowfall seasons studied, a total of 542 snowstorms occurred. This is an average of 9.7 snowstorms per snowfall season (October through the following May). No snowstorms meeting these criteria occurred in the Sierra from June through September of any year. During two-thirds of the 542 snowstorms (357 events), at least 16 of the 23 locations were active. During only 81 of the events (15% of the total) were 14 or fewer active observation sites reporting snowfall.
The overwhelming majority of these snowstorms averaged less than 50 cm of snowfall per event across the study area. The average snowfall for each event was determined by averaging the total snowfall reported at each active location during that event. However, during a majority of these events only one or two locations might have barely managed to reach the 30-cm daily snowfall amount required for the event to be defined as a snowstorm in this study.
Many of the snowstorms among the top decile had impressive average snowfall totals. However, some of these events did not deposit overly large snowfall amounts that would hamper travel across the Sierra. It typically takes snowfall amounts of 100 cm or more to adversely affect a region such as the Sierra, which receives, and is known for, such large snowfall amounts each winter. The authors decided to restrict their analysis to those snowstorms that deposited what might be considered by residents to be heavy snowfall that would impact the area adversely.
For this study, a “major” snowstorm in the Sierra is defined as being among the top decile of snowstorms of the 542 derived from the 56 yr of this study. However, since many of the snowstorms in the top decile did not deposit extremely large amounts, a second criterion was created. For a storm to be considered significant, it also had to produce snowfall of 75 cm (during one calendar day) and/or 150 cm (storm total) at one or more of the study locations. This more stringent criterion would help to ensure that a snowstorm affected not only a relatively large part of the Sierra but that it also produced enough snowfall to adversely affect at least one location in the Sierra.
By using the above criteria a total of 34 snowstorms were classified as major snowstorms from the winter of 1949/50 through the winter of 2004/05 (Table 2). This averages to approximately two major events every 3 yr.
Standardized anomalies of 500-hPa heights and precipitable water values for each synoptic time were studied from the NCEP–NCAR reanalysis data. Thus, at each synoptic time, it was possible to determine how many standard deviations each parameter was from the climatological average for that date. An advantage of displaying data in this way is that meteorological parameters that are extreme stand out.
A sounding for Oakland, California (KOAK) was constructed from the raw radiosonde data. The data for the composite were derived from the 1200 UTC upper-air soundings that were taken on the day that each storm began to deposit snowfall in the Sierra (defined as the first day of the event). A moisture flux at KOAK for the same dates was derived by averaging the products of the mixing ratio and wind speed at pressure levels of 1000, 925, 850, 700, 500, and 400 hPa.
3. Results and discussion
A spatial and temporal overview of Sierra Nevada snowstorms that satisfy the above criteria will be discussed. A composite average snowfall total for each of the study sites will then be computed from these snowstorms. These averages will be compared in order to see how the snowstorms have affected different parts of the Sierra Nevada (northern, central, and southern). Differences resulting from elevation and storm duration will also be studied.
a. Snowstorms in the Sierra Nevada
Snowstorms that deposit at least 30 cm of snow at a location during one calendar day and/or 60 cm of snow during the duration of an event are common in the Sierra. This occurs, on average, nearly 10 times per winter. However, these events vary in intensity. When the snowfall totals that are recorded during a storm at the reporting stations in this study are averaged, the vast majority of the storms have an average snowfall of 50 cm or less (Fig. 2).
The snowstorms that record larger average snowfall totals tend to be associated with more intense synoptic parameters. As can be seen in Table 3, the average strength of the 200-hPa jet maxima, 500-hPa height anomalies, and precipitable water anomalies associated with the 542 snowstorms in this study increases as you move from the lowest to the highest decile.
Sierra Nevada snowstorms are typically associated with 200-hPa jet maxima that reach at least 90 kt in strength. Jet maxima with the stronger storms usually reach at least 110 kt. Of the 542 snowstorms occurring in the Sierra since the winter of 1949/50, only 25 had 200-hPa jet maxima of only 70 kt in strength. The average jet maximum associated with snowstorms in the top decile was almost 13 kt stronger than the average jet maximum of the lowest decile.
The vast majority of Sierra snowstorms (84%) have 500-hPa heights falling to at least two standard deviations below normal. This becomes even more pronounced with the strongest snowstorms. The snowstorms among the top decile have average 500-hPa height anomalies of −2.6. In fact, all 54 snowstorms in this decile had an anomaly of −2 or stronger, with most of them having an anomaly of −3. In each of the other nine deciles the most common 500-hPa anomaly was −2.
A similar distribution can be seen with precipitable water anomalies. Nearly all of the snowstorms in this study had moisture streams which were pulled into the storms’ circulations and pulled to the east toward the Sierra. However, the average snowfall for each event greatly depended on how much moisture reached the Sierra. Many of the storms with the lowest average snowfall totals (of 10.0 cm or less) had precipitable water streams that moved south of the Sierra. Larger average snowfall totals were associated with moisture that was aimed more directly at the Sierra, with the largest totals occurring as a result of moisture impacting the Sierra at an angle that was normal to the range.
b. Major snowstorms in the Sierra Nevada
For the October–May snowfall seasons that were examined, we found that major snowstorms affect the Sierra mainly from December through March, with only two major snowstorms beginning in November. This temporal pattern is consistent with the climatological location of the storm track in this region. The storm track passes to the south across the region during the winter and then back to the north in spring.
Another feature of note is the lack of temporal clustering of the events. Only five winters had two snowstorms that met the study criteria while most winters did not have any. This clustering is likely due to a predominant synoptic pattern that steers systems into the region over a period of a couple of months before changing.
There is a positive correlation between elevation and average event snowfall. The correlation coefficient between these two parameters is +0.709. This modest correlation is not surprising considering how orographic effects could help to increase snowfall at higher elevations.
It is important to note that the elevations of the stations in this study range from 1000 to 2500 m mean sea level (MSL). Locations at higher elevations see even larger snowfall totals as a result of winter storms. The mean freezing level is typically near the elevation of these stations during the winter. As a result, the ratio between snowfall and water equivalent increases with higher elevations in the colder air above the freezing level. Knowing the elevation of the freezing level is an important consideration in forecasting the changeover from rainfall to higher-elevation snowfall. Most of the Sierra passes, especially in the central and southern Sierra (including those utilized by U.S. Interstate 80 and U.S. Highway 50), are above 2000 m MSL.
Unlike the effect of elevation, there is virtually no correlation between the average event snowfall for a location and its latitude. The correlation coefficient for this relationship is only +0.103. However, it is interesting to note that the largest snowfall totals tend to occur between 38.5° and 39.5°N. This may be due to the fact that winter storms move across this area as the storm track moves south and then back north as the snowfall season progresses.
The duration of an event can also be important. There is a greater chance that heavy snowfall will affect an area the longer a snowstorm lasts. Twenty-five of the 34 events lasted at least 4 days. In fact, 17 of the top 20 events lasted at least 4 days. As might be expected, there is a positive correlation (r = +0.700) between snowstorm duration and the resulting average event snowfall.
The duration of an event did not have as much influence on the spatial pattern of snowfall as it did on the amount of snowfall received (as measured by the average snowfall for each event). Of the two longest-duration events (one of which lasted 8 days and the other 7 days), each of the three sections of the Sierra Nevada (north, central, and south) had at least one location that met the criteria to be classified as a major snowstorm. Of the 14 events that lasted five or more consecutive days, 9 of them affected at least two of the sections (meaning that at least one location in each section met the criteria to be listed as a major snowstorm). To decrease the bias resulting from the various durations of the events, the average snowfall per day for each event was calculated (Table 4).
Another way of comparing events was by normalizing the duration of each snowstorm based on synoptic–dynamic parameters or circulation features. The normalization algorithm was similar to that used by Hart and Grumm (2001). Averages for the greatest negative 500-hPa height anomalies associated with each snowstorm from each daily synoptic time (0000 and 1200 UTC) were determined (in the range 0 to −4). Data were analyzed from 3 days before the start of the event through the duration of the snowstorm. The same was done for the greatest positive precipitable water anomalies located within 500 km of each 500-hPa low (ranging from 0 to +4).
In the NCEP–NCAR analyses, winds at 200 hPa were displayed in 20-kt increments from 50 through 130 kt. To have a range similar to those of the other two parameters, 50 kt was given the value 0. Higher wind speeds give successively higher values: 70 kt was given a value of 1, 90 kt a 2, 110 kt a 3, and 130 kt a 4.
To determine the total intensity magnitude (MTotal) of each snowstorm, an average for each of the three parameters associated with the storm was derived. The parameters consist of the average of the maximum 200-hPa wind speeds from each synoptic time for the snowstorm (MWind), the average of the lowest 500-hPa height anomalies for each synoptic time (the absolute value of the average is used since each of the lowest 500-hPa height anomalies is negative) (|MHeight|), and the average of the highest precipitable water anomalies for each synoptic time (MPW). The sum of the three parameters is then divided by 3 to come up with the value for that snowstorm [see Eq. (1) below]. As a result, kinematic, dynamic, and moisture parameters are combined into this one parameter. Table 5 ranks the 34 snowstorms by total anomaly:
c. Tracks of systems at 500 hPa
By using the NCEP–NCAR reanalyses data, it was possible to track systems at 500 hPa as long as 3 days prior to the beginning of snowfall in the Sierra Nevada. Of the 34 systems, 23 had 500-hPa synoptic patterns that developed into closed lows during their evolution. With the other 11 snowstorms the 500-hPa pattern only developed into an open trough (although the base of these deep Pacific troughs often extended as far south as 40°N).
As Table 6 shows, the heaviest snowfall was associated with 500-hPa lows. Of the top 20 events 15 had 500-hPa patterns developing into closed lows. Of the remaining 14 major systems, only 8 developed into lows while the other 6 had upper features that remained as troughs at 500 hPa. In addition, the three 500-hPa patterns that intensified to four standard deviations below normal were all associated with closed lows. A long-duration event, such as the snowstorm of 10–17 January 1952, may have one parent low over the northeastern Pacific Ocean just off the coast of the Pacific Northwest, but short waves passing through the upper flow south of the low would produce periods of intensified snowfall over the Sierra.
The 34 synoptic systems associated with the major snowstorms approached the west coast of the United States from various directions. Each of the 500-hPa lows and troughs developed either over Alaska or the Yukon, or over the northern Pacific or Bering Sea (west of 150°W). Twenty-three systems remained east of 150°W during their entire path [defined as “short overwater” lows (SOLs) or troughs (SOTs)]. These systems would form over Alaska and then enter the Gulf of Alaska as they moved south. These systems tended to begin deepening within 12 h of reaching and moving through the Gulf of Alaska.
An additional eight systems formed over the western Pacific or the Bering Sea and then moved to the southeast. These systems spent part of their trajectory west of 150°W before approaching the Sierra. These were defined as “long overwater” lows (LOLs) or troughs (LOTs).
The remaining three systems formed over Alaska or the Yukon and then moved south along the Rocky Mountains inland of the British Columbia, Washington, and Oregon coasts before impacting the Sierra [defined as “inland” lows (L) or troughs (T)] (Fig. 3 and Table 6).
d. Upper-level dynamics
The kinematics associated with jet maxima in the upper atmosphere can have a major influence on the development and ultimate strength of snowstorms (e.g., Uccellini and Kocin 1987; Mote et al. 1997). This was the case with the major snowstorms that have affected the Sierra Nevada during the last half-century. Table 6 shows the parameters associated with each of the snowstorms in this study. The various parameters will be discussed in the text below.
During the winter, the polar upper-level jet maxima are typically located at or just above 300 hPa. However, the NCEP–NCAR reanalysis data only displayed upper-level winds at the 200-hPa level so the upper-level kinematics had to be inferred from these data. The polar jet stream will often show some influence as high as 200 hPa during the winter, so data restricted to this level did not pose a significant problem in the analyses.
The snowstorms in this study were associated with relatively strong jet maxima at 200 hPa. Eighteen of the 34 systems were associated with 200-hPa jet maxima that reached 130 kt at some point in their evolution (Table 6). The remaining 16 had jet maxima reaching 110 kt for at least 12 h. The vast majority of the systems had their jet maxima south of the central areas of low pressure, thus contributing speed shear and resulting in positive vorticity advection. In addition, Uccellini and Kocin (1987) have pointed out that there is divergent ageostrophic flow in both the right-entrance region and the left-exit region of an upper-level jet maximum. Nearly one-quarter of the major snowstorms in this study exhibited this jet structure.
Each of the 500-hPa low pressure areas associated with the major snowstorms in this study was more intense than normal at some point in its evolution (i.e., usually during the 3 days prior to the beginning of snowfall in the Sierra Nevada as the upper low or trough was deepening). Over half of the 500-hPa lows (20 of 34) had standard anomalies of either −3 or −4 at some point in their evolution during the 3 days prior to and including the date of landfall (Table 6). Three systems had 500-hPa heights dropping to −4 standard deviations. The cold air associated with these lower 500-hPa heights would correlate with lower freezing levels over the study region and would produce snowfall starting at correspondingly lower elevations.
This is similar to results found in a study of extreme rainfall events in the Sierra Nevada (Junker et al. 2008). Junker et al. determined that “the frequency of events with standard deviations of greater than three is quite small.” The remaining 14 major snowstorms in the present study had 500-hPa anomalies reaching two standard deviations below normal. None had a lowest anomaly of −1, 0, or was positive.
e. Transport of precipitable water
Even though the vast majority of the upper-level lows and troughs in this study had an overwater trajectory, this does not prove that this is a sufficient condition for heavy snowfall in the Sierra. The vast majority of lesser systems that affect the Sierra also come from off the northeastern Pacific Ocean.
Significant moisture is needed if heavy precipitation is to occur in the Sierra (Junker et al. 2008; Kaplan et al. 2009). All of the major snowstorms in this study had moisture amounts associated with positive standard anomalies—precipitable water amounts that were at least one standard deviation above normal for at least one 12-h period during the lead-up to and duration of the snowstorm (Table 6).
The moisture transport for each of the major snowstorms had some sort of westerly component associated with it. The precipitable water transport was divided into three groups depending on which direction the moisture came from (i.e., the time-dependent motion of upstream moisture maxima or the motion of positive precipitable water anomalies).
Transport was considered as coming from a northwesterly direction if the areas of positive precipitable water anomalies moved from the northeastern Pacific north of 40°N. Transport was from the west if the positive precipitable water anomalies came toward the Sierra from an area between 30° and 40°N. Finally, the precipitable water transport was considered to be from the southwest if the areas of positive precipitable water anomalies came from an area of the northeastern Pacific south of 30°N.
Of the 542 snowstorms in this study, 228 of them (42%) had moisture plumes that approached the Sierra from a northwesterly direction (Table 7). This is understandable when it is remembered that all of the upper lows or troughs formed in the Bering Sea or northern Pacific, or inland over Alaska or the Yukon, and then moved to the southeast. Moisture over the northern Pacific was easily entrained into the circulation around the upper low or trough. This moisture was then pulled southeastward as the system made its way toward the Sierra. The rest of the systems were nearly equally divided between moisture advection from a westerly or southwesterly direction.
This however is true only for 90% of the events in this study. The snowstorms among the top decile did not show this preference for a northwesterly flow but were almost equally divided among trajectories from all three directions. Eleven of the top 20 major snowstorms had moisture transport from the southwest. In fact, 7 of the top 10 snowstorms (in terms of average snowfall) had precipitable water transport from the southwest.
Much of this moisture would have been transported northward ahead of the low pressure system. The subtropical air entrained into this trajectory would have also corresponded to what Martin (1998, 1999) calls a “trowel airstream.” This trajectory is also consistent with the concept of “atmospheric rivers” (Ralph et al. 2004). Atmospheric plumes of moisture with precipitable water anomalies of at least three standard deviations above normal were associated with 19 of the 34 events in this study. Eleven systems had precipitable water anomalies rising to +2, while the remaining four snowstorms had anomalies of +1. The incredible snowstorm of 10–17 January 1952 had a stream of subtropical moisture with an anomaly of +5 for a 12-h period on 13 January.
Systems that transport moisture from the west, although not having the large precipitable water values that air masses from the southwest might have, would impact the Sierra Nevada with a more zonal trajectory, thus enhancing the upslope component of motion. The orographic effects realized would help to increase the snowfall that would ultimately result (Dettinger et al. 2004). The remaining thirteen major systems transported moisture from a westerly or northwesterly direction. Five of these moisture streams eventually reached the Sierra at an angle nearly normal to the range. The other eight systems had moisture that brushed the Sierra from a northwest to a southeast direction. However, the upper-level dynamics were strong enough to intensify the snowfall rates.
f. Moisture flux
It would be wrong to infer that a southwesterly trajectory of moisture is the only variable that separates the major Sierra snowstorms from lesser storms. Moisture flux into the Sierra is even more important. As can be seen in Table 7, the angle at which the moisture plume impacts the Sierra is more important than the source region of the moisture.
The 542 snowstorms in this study were divided according to how their moisture plume interacted with the Sierra Nevada. The five categories are
the bulk of the moisture plume moves north of the Sierra,
the moisture plume first affects the northern Sierra and then continues to move south along the range,
the moisture plume impacts the Sierra at an angle normal to the range,
the moisture plume first affects the southern Sierra and then continues to move north along the range, and
the bulk of the moisture plume moves south of the Sierra.
Table 7 shows that only 20 snowstorms had moisture plumes that moved north of the Sierra. Snowfall from these events, as might be expected, was concentrated in the northern Sierra. In addition, because the bulk of the moisture missed the Sierra, the average snowfall for these snowstorms was relatively low.
Another category that was not well represented was moisture plumes that impacted the Sierra and then were pulled north along the range. It may be rather intuitive why this occurrence was rare. The low pressure systems associated with these snowstorms generally moved from the north or northwest to the southeast. Any moisture associated with these systems would be pulled south along with the parent low or trough. Thus, moisture would rarely affect the southern part of the Sierra and then move to the north along the range. Only 35 snowstorms showed this pattern, although they were more equally represented among the 10 deciles than were the systems whose moisture plumes stayed north of the Sierra.
The other three categories greatly differ from each other. Systems that had moisture plumes that affected the northern Sierra first and then moved south along the range were fairly well balanced among all deciles. This is also the most common type of Sierra Nevada snowstorm (38%). The evolution of these snowstorms is not surprising. A low pressure system moves to the southeast toward the west coast of North America. The moisture that it entrains is directed at the northern Sierra first and then it is brought south along the range as the area of low pressure continues moving to the southeast, eventually making landfall along the California coast and then moving into the southwestern United States.
The distributions in the other two categories are near mirror images of each other. Looking first at the category containing those events that had their moisture plumes moving south of the Sierra, it is clear that these snowstorms are concentrated in the lower half of the data distribution. This might be understandable when it is realized that moisture that misses the southern part of the Sierra would not tend to deposit much snowfall during its passage. However, unlike the snowstorms whose moisture plumes passed north of the Sierra, these southern systems tended to advect subtropical moisture into their circulations. These relatively moist plumes, combined with the cyclonic circulation around the low center (which would direct moisture northward as the system moved east), produced larger snowfall totals than those with moisture plumes that moved north of the Sierra. But as Table 7 shows, relatively few of the larger snowstorms developed with this scenario.
This leaves the fifth category, which contains those snowstorms that are disproportionately represented in the top half of the data distribution. These snowstorms had moisture plumes that impacted the Sierra (anywhere along its extent) at an angle normal to the range. The angle at which a moisture plume interacts with the Sierra has a greater effect on the resulting snowfall than does the source region where the moisture comes from. Two-thirds of the snowstorms in this category are in the top five deciles. Among the top decile, 30 of the 54 snowstorms had moisture plumes that impacted the Sierra directly at an angle normal to the range.
The combination of larger amounts of moisture in the stream, along with the stream contacting the Sierra at an angle nearly normal to the range, will increase the probability of heavy snowfall. Fourteen of the top 20 major snowstorms (and 21 of the total of 34) had precipitable water streams impacting the Sierra at an angle that was normal to the range. The largest snowfall totals are also associated with moisture streams that remain aimed at a location in the Sierra. This was often a factor in one location receiving twice as much snowfall as other study sites during an event.
Figure 4 shows a composite of the moisture flux into the region associated with the major snowstorms in this study. The composite depicts the approaching stream of precipitable water (PWAT) as reflected in mixing ratio values recorded at KOAK on the dates when snowfall began in the Sierra. Moisture flux associated with the major snowstorms was maximized near 850 hPa. This concentration of moisture may be due to the presence of the low-level jet being located at this elevation. Satellite imagery of total precipitable water (TPW) over the Pacific can be monitored by forecasters when trying to determine where the stream of moisture may impact the Sierra.
g. Static stability of the lower atmosphere
In forecasting any type of weather phenomenon, it is important to have an idea of the vertical composition of the air mass. In forecasting snowfall in the Sierra Nevada, atmospheric soundings from upwind locations can provide extremely valuable information.
The composite sounding from KOAK (Fig. 5) depicts air masses that were relatively unstable at low levels (i.e., nearly moist neutral) with significant moisture from the surface up to 850 hPa (with a maximum near 950 hPa). The significant low-level lift that the air mass would receive from the flow up the Sierra Nevada would likely cause the lapse rate to approach moist adiabatic below 750 hPa.
As the upper-level systems and their low-level jets got closer to the Sierra Nevada, the moisture in the air mass over California would likely increase. In a study of atmospheric river characteristics, Ralph et al. (2005) found that the air in these narrow regions was nearly saturated below 800 hPa and that the moisture was “flowing in a direction approximately normal to the northwest–southeast-oriented California coastline.”
The air mass does not have to be highly unstable in order to contribute to heavy snowfall in the Sierra. In studying two snowstorms that affected the Sierra Nevada, Marwitz (1987) found that one storm “was very stable, and the second storm was slightly stable.” Even with the lack of background instability, the winds (which were normal to the Sierra) would have helped to maximize any orographic effects and to overcome any low-level stability.
This is what was seen with snowstorms in this study. Even with air masses that were not extremely unstable, impressive snowfall totals were realized due to the moisture that was advected into each storm and the orographic effects that were produced by the Sierra Nevada.
h. Conceptual model for major Sierra Nevada snowstorms
To aid meteorologists in the forecasting of heavy snowfall in the Sierra a conceptual model was developed (Table 8). The various parameters associated with snowstorms in the Sierra have been ranked from most important to least important in the forecasting of heavy snowfall. Although some of the parameters are more important than others, none is sufficient in determining whether heavy snowfall will occur. However, the presence of some of the more important parameters can help to increase a forecaster’s confidence that a heavy snowfall event is possible or even likely.
Our research shows that the orientation of the moisture stream to the Sierra was the most important variable in determining the possibility of heavy snowfall occurring. The probability of heavy snowfall increases as the moisture plume becomes more normal to the Sierra. This relationship was not seen with any other orientation. There was no correlation between heavy snowfall and the moisture plume moving from the north to the south along the range of the Sierra (this is the most common interaction of the moisture plume with the Sierra). There was a negative correlation between the moisture plume moving south of the Sierra and heavy snowfall occurring. Under this scenario, as might be expected, any heavy snowfall that occurs is typically restricted to locations in the southern Sierra. Only 10% of the 542 snowstorms in this study resulted from the other two scenarios (the moisture plume passing north of the Sierra and the moisture plume moving south to north along the range) and resulted in a negative correlation and no correlation, respectively (Table 7).
The next most important variable associated with heavy snowfall in the Sierra was determined to be the relative amount of precipitable water in the moisture stream. Among all of the snowstorms studied, there was an increasing likelihood that heavy snowfall would occur the greater the precipitable water anomaly associated with the moisture stream (Tables 3 and 6). In addition, of the 34 major snowstorms studied, 30 were associated with anomalies of +2 or greater (19 had anomalies of +3 or greater). The other four snowstorms had anomalies of +1. Precipitable water anomalies associated with snowstorms in each of the lowest seven deciles averaged less than +2, while those in the top three deciles averaged greater than +2. The top decile showed the largest increase from the next lower decile in the distribution, with an average anomaly of +2.5 (0.3 greater than that of the ninth decile).
The static stability of the approaching air mass is next in importance. Although large low-level instability is not a prerequisite for heavy snowfall in the Sierra (as found by Marwitz 1987), an unstable air mass certainly contributes to the upward vertical motion necessary for snowfall rates to increase. The composite sounding derived from KOAK data shows nearly moist neutral stability from the surface up through 850 hPa (Fig. 5). This is also the level of the greatest moisture flux from off the Pacific (Fig. 4). As the air mass is lifted by the Sierra Nevada, low-level instability will increase. This increased instability, along with the increasing moisture as the system approaches the region, will contribute to the heavier snowfall rates that were seen with the stronger snowstorms.
The strength of the 500-hPa low pressure area or trough associated with a snowstorm was deemed the fourth most important factor in forecasting heavy snowfall in the Sierra. The relationship of this parameter among deciles is similar to that seen with the precipitable water anomalies described above. The strength of the average 500-hPa low or trough associated with each decile increases from the lowest to the highest decile (this is reflected by the average anomalies becoming more and more negative). Since each decile is related to the average snowfall from snowstorms, there is thus a good correlation between the strength of the 500-hPa system and the resulting snowfall. Again, as with the precipitable water anomalies, the greatest increase (decrease) in the strength of the 500-hPa low or trough (average anomaly) occurs between the 9th and 10th deciles (Table 3). All 34 major snowstorms studied had 500-hPa lows or troughs intensifying to at least two standard deviations below normal (Table 6).
The trajectory of the moisture plume is not as important a consideration as the parameters listed above; however, it does have an influence. The upper-level lows and troughs associated with Sierra snowstorms invariably move from the northern Pacific to the Sierra. Thus, it is probably not surprising that moisture is advected south with these systems. More snowstorms in the Sierra (42%) result from moisture approaching the region from the northwest than any other direction (Table 7). However, this moisture may have originally been advected from the subtropics (as can be seen when monitoring TPW satellite imagery). Moisture approaching from the west and moisture being advected from the southwest are nearly equally distributed among the rest of the 314 snowstorms. The top decile (containing all of the major snowstorms in this study) is the only one with more snowstorms receiving moisture coming from the southwest more often than from the west or northwest. All other deciles show snowstorms receiving moisture from the northwest more often than from any other direction. When other factors are in place (e.g., large positive precipitable water anomalies associated with the moisture stream or large negative 500-hPa height anomalies), a southwesterly trajectory of moisture will increase the probability of heavy snowfall in the Sierra.
A strong 200-hPa jet maximum was seen with the vast majority of snowstorms in the Sierra. This is probably not surprising since upper-level jet maxima tend to be stronger during the winter than during the warmer half of the year. And relatively strong jet maxima are typically associated with deepening upper-level low pressure systems. As a result, the strength of the upper-level jet maximum is not a major concern in forecasting heavy snowfall since a strong jet maximum will typically be in place regardless. However, it was seen that the more intense snowstorms were associated with relatively stronger jet maxima (Tables 3 and 6). Within the lowest five deciles, the most common strength of the 200-hPa jet maximum was 110 kt. Among the top five deciles, the most common jet maximum that was recorded with snowstorms was nearly equally divided between 110 and 130 kt. In addition, all but one of the deciles recorded more than 20 snowstorms with 200-hPa jet maxima reaching 110 kt. Among the other jet maxima speeds, there was a positive correlation between increasing the jet maximum wind speed and heavier snowfall. The proportion of snowstorms associated with jet maxima reaching 130 kt increased from the lowest to the highest decile (Table 3). With weaker jet maxima (70 and 90 kt), there was little or no correlation with heavier snowfall.
The track of the associated 500-hPa low or trough showed the least correlation with heavy snowfall (Tables 3 and 6). An almost equal number of upper lows or troughs approached the Sierra by a long overwater trajectory as along a short overwater trajectory (254 and 243, respectively). Only 45 systems followed an inland trajectory. The 34 major snowstorms showed a preference for short overwater trajectories as opposed to long overwater trajectories (23 versus 8) but these were distributed throughout the rankings (Table 6).
4. Summary and conclusions
This study looked at snowstorms that have deposited relatively large amounts of snowfall throughout the Sierra Nevada. In the 56 winters studied, it was found that 34 snowstorms within the top decile met our criteria by depositing at least 75 cm of snow in a calendar day, and/or at least 150 cm of snowfall from an entire event, at one or more of the study sites.
There was a positive correlation between the average event snowfall for a particular study site’s location and its elevation in the Sierra; however, no correlation was found between the average event snowfall for a study site and its latitude.
The duration of an event had a positive effect on the amount of snowfall that was eventually realized from each snowstorm. The event duration however did not have a discernable effect on the spatial pattern of the snowfall across the Sierra.
This study found that significant snowstorms (those that deposit at least 75 cm of snow in one calendar day and/or 150 cm of snow over the duration of a snowstorm) have many meteorological parameters in common. These range from upper-level dynamics to low-level moisture transport. Table 8 summarizes a conceptual model that can be applied to forecasting heavy snowfall in the Sierra Nevada. Similarities among the snowstorms in this study include the following:
Eighteen of the 34 snowstorms in this study were associated with jet maxima reaching at least 130 kt during their evolution. These jet maxima were frequently south of the associated upper low, which helped to intensify the cyclones. Interacting jets were too infrequent to have a significant effect on the snowstorms through the period of study.
All of the major snowstorms included in this study had 500-hPa lows or troughs that intensified enough to have negative height anomalies. In 20 of the snowstorms, the associated 500-hPa heights dropped to at least three standard deviations below normal. Three of the systems had 500-hPa heights falling to four standard deviations below normal.
Thirty-one of the 34 systems were associated with upper-level lows or troughs that either formed in the Gulf of Alaska or intensified as they moved through it. This is similar to what Reitan (1974) found in his study of cyclogenesis for the period 1951–70. His analysis showed a maximum of winter cyclone frequency over the Gulf of Alaska during this 20-yr period.
All but three of the snowstorms had tracks that remained over the Pacific Ocean for up to 3 days before snowfall began in the Sierra. The overwater trajectories of these systems allowed moisture to become entrained and thus contribute to the large snowfall totals seen.
A large amount of moisture is necessary for heavy snowfall to occur in the Sierra. Over half (19 of 34) of the major snowstorms were associated with precipitable water anomalies of +3 or greater. All but four had anomalies of +2 or greater. This above normal amount of moisture undoubtedly contributed to the large snowfall totals observed.
How a snowstorm’s moisture plume impacted the Sierra had a major influence on how much snowfall would eventually be produced. The highest correlation between the moisture plume orientation and the resulting snowfall occurred with moisture streams that impacted the Sierra at an angle normal to the range. This angle allowed for the greatest amount of moisture to be involved in the production of snowfall. It also provided the maximum lift to the air mass as it ascended the western slopes of the range.
Finally, the air masses that were associated with these snowstorms were close to being moist unstable in their lower levels. Orographic effects as the air masses impacted the Sierra Nevada may have been important in the additional increase of low-level instability inland from coastal locations such as KOAK. Low-level moisture was also abundant in these maritime air masses.
We thank Chris Smallcomb (science and operations officer at the National Weather Service Forecast Office in Reno, Nevada) for offering support and encouragement throughout the process of writing this article. The authors would also like to thank Laura Edwards (assistant research climatologist at the Western Regional Climate Center) for reviewing an early version of the manuscript. Three anonymous reviewers also provided very helpful suggestions that made this a much improved article. Shane Cleary of the Department of Geography, University of Nevada—Reno, created the map of the study area used in Fig. 1.
Corresponding author address: Brian F. O’Hara, National Weather Service Forecast Office, 2350 Raggio Pkwy., Reno, NV 89512. Email: firstname.lastname@example.org