National Weather Service United States Department of Commerce

This weather event was unusual in many ways. The overall synoptic pattern that led to this event was more typical of January rather than mid-August. A ridge of high pressure had built over the Gulf of Alaska and amplified northward into Alaska. This brought record warm temperatures of upper 70s and lower 80s (normals are in the 60s) to the 49th state. In response, the jet stream over western Canada dug to the south, bringing a cool air mass into the northern Rockies and Northwest U.S. The hot temperatures of early August began to slowly moderate on the 10th as the Canadian jet gradually sagged into the area. The initial surge of a cooler low-level air mass arrived into the Inland Northwest on Thursday evening, August 11th as winds shifted to the northeast.

What transpired on the afternoon and evening of August 12th intially appears to have originated as a outflow gust front from thunderstorms over northeast Washington. The loop below shows the gust front dropping southwestward as the thunderstorms continue to move off to the southeast.

While this is not an unusual occurance, what was atypical was the longevity and persistent strength of this boundary. By 03Z (800 pm PDT), the boundary had progressed all the way to Walla Walla to the south, and Ellensburg to the west, more than 4 hours after it's inception. Additionally, the wind gusts experienced at Hanford and Pasco were equal to the gusts as it moved through Spokane.

Typically a thunderstorm-generated gust front will slowly decay over time. This is due to the fact that the initial surge of cold air which originated from the thunderstorm is limited in supply. As the gust front spreads out, this supply of cold outflow air becomes increasingly more shallow. Additionally, the outflow air will also undergo a mixing process with the environment as well as frictional affects.

The boundary on August 12th did have a couple of things working in its favor. The pre-existing flow was already from the northeast. So the boundary moving to the south didn't have a head wind to "fight". Also, the topography from Spokane to the Tri Cities is a gradual downslope, changing in elevation from about 2500' to around 400' above sea level. Thus, gravity would have been assisting the boundary as well.

The problem with assigning the thunderstorms as the cause of this boundary, is that the numerical models also predicted a cold front well before any thunderstorms has developed. The loop below shows the 500mb heights and vorticity during this event. Note the strong short wave that moves over the area during the afternoon and evening hours.

At the 850mb level, the model predicted a front to back into the area from the northeast. Note the strong temperature gradient with this front. Also note the near-perfect placement of this front in the model forecast compared to the actual boundary on the radar. The end of this loop is 03Z (800 pm PDT), which is the same time as the radar mosaic above.

As the 500mb trough moved over the area, the cooling temperatures aloft helped to destabilize the atmosphere. The dynamic lift from this 500mb short wave also aided in the thunderstorm development. Below is a loop of the hourly CAPE and CIN as analyzed by the LAPS model. CAPE values of near 2000 J/kg developed during the afternoon. But note how the atmosphere rapidly stabilizes over northeast Washington as the boundary moves south. More importantly, note how this stabilization occurs north and east (in southern BC and western Montana) of where the boundary was first spotted on radar at about 2145Z.

When the boundary moved through the Spokane area, the VAD Wind Profile from the radar observed the deepening and strengthing winds.

Again, note the pre-existing northeast flow before the passage of the boundary (around 2235Z). After the boundary passage, northeast flow up to about 7000 ft MSL was present under a prevailing northwest flow aloft. As the boundary continued to progress southward, it's depth was suprisingly consistent, remaining at about 8000 ft MSL over an hour later.

The Eta model actually did a good job of predicting this, again, before any convection had developed. It clearly shows a windshift just before 00Z (500 pm PDT) with the depth of the northeasterlies up to about 8000 ft. The white trace line is model omega, and shows a marked change from upward (negative) to downward (positive) motion as the front moved through.

Since the Eta model cannot predict outflow boundaries from convection, and the scale of the 850mb wind and thermal field is much larger than any convective complex, the data appear to indicate that this boundary was more of a synoptic scale front rather than from thunderstorm outflow.

The MSAS analysis of pressure (green lines) and 3-hour pressure change (brown lines) below, clearly shows a "bubble" of pressure rises that pinch off over western Montana and move into northeast Washington. This pressure rise center developed well behind any convection But more important are the observations at Cranbrook (CYXC) and Creston (CWJR), BC, just north of the Idaho Panhandle. At the start of the loop, Cranbrook has just experienced a wind shift to the northeast. Two hours later, Creston has a similar wind shift. Using the time of these two wind shifts and extrapolating ahead in time, note how well the cursor (purple circle with a dot inside) lines up with the wind shifts in eastern Washington at Spokane and Hanford as the front moves south.

Lastly, look at the loop below. This time, the extrapolation line extends backwards in time. The "boundary" that appears south of the thunderstorms at the end of the loop was extrapolated backwards. Note how convection is enhanced ahead of this line, but then dies behind it. Thus, the question becomes, did the boundary exist up in BC, but just wasn't observable on radar until it got close enough to Spokane? Or did the thunderstorms actually produce outflows that became the front.

Taking another look at the initial radar loop, the boundary does appear to be thunderstorm outflow in origin. But a closer look shows that this boundary is much too continuous to have been generated all at once by thunderstorms in different stages of development and decay.

Inspection of some of the radar images does show a discontinuous nature to the boundary, including some instances where it appears that semi-circles of outflow intersect the pre-existing boundary. Additionally, there is another boundary that appears in southwest Shoshone county and moves into Benewah county that does not appear to have originated from any thunderstorms. But our pre-conceived ideas of smooth continuous synoptic fronts is likely unjustified in complex terrain. The problem is that we aren't able to actually "see" the surface front.

In conclusion, it appears from observed and modeled data that the boundary on Aug 12th, 2005 was primarily a synoptic scale front which moved into the the area from southeast BC and northwest Montana. Convection developed ahead of this front due to surface-based instability and upper-level dynamics. As the front progressed southwestward under the convection, it stabilized the atmosphere. Fairly steep lapse rates in the mid-levels coupled with the dynamics aloft were able to sustain some of the convection in elevated form in the northwesterly flow aloft. The front proceeded southwestward in a somewhat discontinuous manner, likely due to the topography. Some modifcation and enhancement of the front due to thunderstorm outflow cannot be ruled out. But as the front moved over the smoother terrain of the Columbia Basin, it became better organized and more continuous in nature.


By Ron Miller