WESTERN REGION TECHNICAL ATTACHMENT
NO. 01-12
SEPTEMBER 18, 2001

FORECASTING APPLICATIONS FOR ELEVATED THUNDERSTORMS
PART 2: NOCTURNAL THUNDERSTORMS OVER CENTRAL
CALIFORNIA IN AUGUST 1999

Alexander Tardy- Weather Forecast Office, Sacramento, CA

[Note: All figures and tables appear on the web page only.]

INTRODUCTION

Thunderstorms developed on the afternoon of 9 August 1999 across the Sierra Nevada foothills upward to along the crest. Thunderstorm formation over the Sierra Nevada is common in the afternoon and early evening hours during the summer months. This is primarily the result of subtropical moisture being advected northward across Mexico into the Intermountain West and California on the westside of the seasonably strong high pressure system over the southwest United States (monsoon flow). This moisture interacts with diurnal instability over the higher terrain and often results in intense deep moist convection. In this study, the thunderstorms were not the more typical diurnal activity, but rather, became more numerous during the night hours. The thunderstorms on 9 August dissipated during the evening hours, however, moist convection redeveloped during the early morning of 10 August. None of the thunderstorms in this paper showed severe radar characteristics on Weather Surveillance Radar-1988 Doppler (WSR-88D) located in KDAX (Davis). However, several thunderstorms showed reflectivity values exceeding 60 dBZ, and radar echo tops (ET) between 30,000 and 40,000 ft mean sea level (MSL). These values are similar to the surface-based diurnal thunderstorms that occur over the higher terrain during the monsoon flow. In addition, the thunderstorms during the early morning of 10 August were almost continuously forming in the same general area and tracked northwest from the northern Sierra Nevada to the southern Cascades. A direct result of the thunderstorms was several grass and forest fires ignited by the lightning strikes.

The Sacramento Valley (the northern portion of the Central Valley in California) is usually much less favorable for summer thunderstorms because of, but not limited to, downsloping drying effects, a dry boundary layer, and the lack of surface moisture convergence. On 23 August 1999, a similar event to the one previously discussed occurred, except these thunderstorms developed in the central Sacramento Valley. The thunderstorms showed similar radar characteristics to the event on 10 August, and occurred during the late night hours. Numerous grass fires were also caused by these thunderstorms despite measurable rainfall (Oroville 0.18 in; Paradise 0.15 in).

These two cases show how important it is to accurately forecast thunderstorms in northern California because of the extreme fire danger during the summer months. This study will explore these cases, which displayed similar magnitude and duration, but occurred over greatly different topography. It will be shown in this study that synoptic scale processes lead to explosive moist convection when subtropical moisture interacts with mid-tropospheric weather disturbances.

On 10 and 23 August 1999, an upper level trough of lower pressure was situated just offshore along the west-central coast and was slowly drifting inland (Figs. 1 and 2). This pattern was similar to the thunderstorm event which occurred on 8-9 September 1999 (Tardy 2001). This feature provided the upper air support for the development of deep moist convection. This includes a well-defined upper level divergent region positioned over the area where thunderstorms occurred, and a cyclonically curved jet maximum entering the base of the upper low with the left front region placed over northern California. On 9 August, a strong vorticity maximum rounded the base of the trough and eventually lifted into western Nevada on the morning of 10 August (Fig. 1). Weaker vorticity maximums were present on 23 August (Fig. 2). The main moisture source in these events was supplied from a subtropical plume which contained seasonable precipitable water values ranging between 0.65 and 1.15 in. This moisture was drawn northward in the southerly flow created by the mid-tropospheric counterclockwise circulation over the southwest United States and the approaching upper low.

OROGRAPHIC EFFECTS

Topography may have been a factor leading to the initiation of thunderstorms, but the effects were likely minimal. Both events had a southeasterly flow above the 800-mb level which is not a favorable upslope flow for the west slopes of the Sierra Nevada. The thunderstorms that were initiated near the crest of the Sierra Nevada (Fig. 3) show that the southeasterly flow could have provided upslope conditions on the east side and crest of this mountain range. However, most of the thunderstorms developed along the west slope between elevations of 2,000 and 5,000 ft MSL (Motherlode foothills and lower Sierra Nevada), and once they were carried to the northwest along the mountain slopes, they encountered a more easterly steering wind component which forced them to descend into the Sacramento Valley. Elevated terrain could also aid in a parcel's ascent to the level of free convection (LFC) since the surface would then be closer to this level. The final step toward reaching the LFC can come from random atmospheric mixing of air parcels that is always present to some degree. Once air parcels are buoyant, the additional mass evacuation of air associated with the upper jet streak and divergent wind field will sustain and enhance moist convection.

RADAR AND SATELLITE DATA

Case 1

Figure 3 shows the early stages of thunderstorm development on 10 August. The thunderstorms persistently developed over the Motherlode and Sierra Nevada, as well as parts of western Plumas County (Figs. 4 and 5) between 0500 and 1200 UTC. During this time period, radar echo tops for the thunderstorms were as high as 40,000 ft MSL and maximum reflectivity reached 60 dBZ (Fig. 4). Figure 5 shows the diminishing reflectivity returns over the Sacramento Valley indicative of the weakening thunderstorms in this area. The mountain and foothill thunderstorm activity reached its peak intensity around 1000 UTC when numerous thunderstorms were depicted over the Motherlode and Sierra Nevada (Fig. 6). Corresponding GOES-10 infrared (10.7 m) satellite imagery showed significant cloud top cooling (-40 to -45 C) associated with the thunderstorms (Fig. 7). This satellite imagery also showed deep moist convection occurring over the southwest United States where more abundant subtropical moisture was available. By 1152 UTC the last area of thunderstorms moved over the I-80 corridor and diminished to light showers shortly afterwards (Fig. 8).

Case 2

On 22 August, prior to the thunderstorm outbreak, KDAX composite reflectivity detected an outflow boundary associated with thunderstorms that occurred during the late afternoon across the Sierra Nevada (Fig. 9). A radar time lapse showed that this mid-level boundary moved into the far west side of the southern Sacramento Valley and dissipated, while manual observations by the author showed altocumlus castellanus clouds were produced along it. This type of cloud is indicative of mid-tropospheric moisture and instability, and sometimes precedes thunderstorm development. Late in the evening on 23 August, satellite images showed that skies became mainly clear (not shown). At 0702 UTC 23 August, a narrow line of thunderstorms rapidly developed in the central Sacramento Valley (Fig. 10). Radar reflectivity and areal coverage continued to increase in this region, but showed little organization, (Fig. 11) resulting in numerous cloud-to-ground lightning strikes through 1100 UTC (not shown). Figure 12 shows intense thunderstorms that became concentrated at 1015 UTC over the valley floor of Butte County and produced locally heavy rain. All of the activity was carried to the north in the southeasterly flow.

OBSERVED INSTABILITY, MOISTURE, AND WIND

In these cases, the boundary layer stabilized after the loss of solar insolation, but this did not preclude deep moist convection. Conditions present above the boundary layer were supportive for thunderstorm development independent of the boundary layer stability. This stable boundary layer is clearly visible in Figure 13, which displays a sounding near Blue Canyon from Rapid Update Cycle (RUC) initialization at 1200 UTC 10 August 1999. It is very important to understand what type of atmosphere this sounding is depicting, specifically the elevated instability and moisture above the boundary layer. The sounding shows high static stability and dry air in the boundary layer. The sounding in Figure 13 shows that the moisture source for potential deep moist convection was in a layer between 650 and 400 mb. This classic so-called inverted-V sounding has been documented as the Beebe "Type IV" (Bluestein 1993), and characterized as producing high-based thunderstorms with little rainfall. The Convective Available Potential Energy (CAPE) at 1200 UTC, using the most unstable air parcel in this sounding, was measured at 1660 Jkg-1, and the 500-mb lifted index (LI) was an impressive -6.1. Thunderstorms had developed in this area during the afternoon hours on 9 August, but diminished in the evening. A time lapse of the Local Analysis and Prediction System (LAPS) analysis of surface-based CAPE showed a decreasing trend through the night (not shown). An example of surface-based stability indicators to be misleading for cases similar to those in this study. The addition of subtropical moisture to an initially dry air mass would increase the instability (steeper lapse rates) in that layer when it is vertically lifted since the dry air will cool faster (adiabatic rate) than the nearly saturated layer. Less stable conditions could also result from warm air advection at the lower levels (above the boundary layer). All of these factors indicate that the thunderstorm development was produced by synoptic upward vertical motion, and the divergence associated with the east side of an upper level circulation interacting with the available elevated instability and subtropical moisture. Even though synoptic scale forcing is not strong enough in magnitude to directly produce moist convection, this atmospheric lift will effectively lower the LFC through air mass layer lifting.

On 23 August, the wind profiles showed a well-defined boundary layer while on 10 August, this was also observed, but it was more shallow (Figs. 13 and 14). The KDAX wind profile on 23 August depicts northerly winds (Fig. 14) below 7,000 ft MSL, similar to the RUC sounding (Fig. 13). A north wind would not be conducive to surface moisture convergence, and tends to produce drying effects in northern California. Above 7,000 ft MSL, within the southeasterly wind field, is where the important subtropical moisture and instability existed. After studying the RUC sounding and the wind profile, it becomes apparent that the thunderstorms in these cases were not surface-based. During the period of thunderstorm activity on 10 August, the observed surface winds at Blue Canyon were primarily light easterly, indicative of downsloping nocturnal drainage wind conditions. This wind direction is also not conducive to surface moisture convergence.

MODEL SOUNDINGS

Numerical weather predictions have always had difficulty in accurately forecasting thunderstorms because of the complexity of the parameters involved, and the mesoscale features not resolved by synoptic scale models. Examining model-generated quantitative precipitation forecasts (QPF) for these events would not have helped predict the convective activity because the models did not produce precipitation. However, there are clues that are available and once recognized can significantly aid in the forecast process. One such tool is the model forecast sounding which is available using BUFKIT (Mahoney and Niziol 1997). Figure 15 displays a BUFKIT forecast sounding at Beale Air Force base (KBAB) from the 0000 UTC 10 August 1999 Eta run. This sounding depicts adequate instability with a CAPE (lowest 500-m layer average) value of 281 Jkg^-1 and a precipitable water value of 1.12 in. The sounding also denotes a large area of convective inhibition (CIN) below the LFC. This region of stable air (nocturnal radiational inversion) was not a limiting factor since the thunderstorm development occurred in the mid-troposphere. The model sounding accurately depicted steep environmental lapse rates above 10,000 ft MSL (700 mb) and correctly indicated a significant increase in moisture above 15,000 ft MSL (600 mb), denoted by the smaller dew point depression (see fig. 15).

FORECASTING APPLICATIONS

The use of the forecast or observed soundings alone is not sufficient to recognize the elevated thunderstorm potential. Using basic model depictions of 500-mb geopotential height and vorticity fields similar to those in Figures 1 and 2, and 500-mb relative humidity (or preferably a 600 to 400 mb mean layer), layer precipitable water trends, and omega contours, can give the forecaster an idea of the most favorable area where subtropical moisture and atmospheric lift will be sufficient for elevated thunderstorm formation. Water vapor satellite image trends and characteristics will help in determining whether the regions of best upward motion (associated with short waves and upper divergence) and instability will interact with the moisture necessary for elevated thunderstorms (Tardy 2001). Near real-time instability analyses on AWIPS can be helpful, but may sometimes severely underestimate the convective potential in the mid-levels of the atmosphere. Synoptic scale cold and warm air advection associated with the upper low, could also result in steeper mid-level lapse rates similar to what was observed (see fig. 13). Most important, sufficient destabilization will result from the synoptic scale layer lifting which effectively lowers the LFC.

CONCLUSION

This study showed a type of deep moist convection that can occur over northern California, which can result in lightning initiated forest fires. Two cases were presented in an attempt to show that elevated thunderstorms can and will occur over valleys or higher terrain given certain synoptic conditions. Despite the extreme difficulty in forecasting thunderstorms, there are available resources, such as model soundings which, when combined with pattern recognition and satellite trends, can yield the best forecast of the more infrequent weather events.

Most important, these thunderstorms are not rooted in the boundary layer, and are not dependent on terrain or boundary layer conditions, but can be influenced by them. This paper showed how mid-level convective processes need to be recognized in order to capture the potential for elevated thunderstorms. Given a sufficient amount of subtropical moisture, air mass lift and instability, the atmosphere can often yield deep moist convection when it might otherwise not be expected above a stable boundary layer. Forecasting this type of event is very important across northern California because of the extreme fire hazards present during the hot and dry summer months. In addition, since precipitation is infrequent across the Central Valley during the summer, any rainfall can have an impact on people and agricultural.

ACKNOWLEDGMENT

Thanks to Scott Cunningham, Scientific Operations Officer at WFO Sacramento, for several reviews of this paper.

REFERENCES

Bluestein, H. B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes Vol. II Observations and Theory of Weather Systems. Oxford University Press, 2, 444-455.

Mahoney, E. A., cited 2000: BUFKIT Documentation [Available on-line from http://www.nws.noaa.gov/er/buf/bufkit/bufkitdocs.html.]

Mahoney, E. A., and T. A. Niziol, 1997: BUFKIT: A software application toolkit for predicting lake-effect snow. Preprints 13th Intl. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Long Beach, CA, Amer. Meteor. Soc., 388-391.

Tardy, A. O., 2001: Forecasting Applications for Elevated Thunderstorm in California, Part 1: The 8 September 1999 Outbreak. Western Region Technical Attachment No. 01-09.