Figure 1.1 above Peak gusts (mph and km/h) for the December 11, 2014 windstorm. Wind speeds are largely from long-term surface airways weather observation sites, data buoys, lighthouses and C-MAN stations, with limited data from other networks (e.g. RAWS). Stations with long histories are preferred because of the research focus on intercomparison of historic storms. Numbers preceded by a tilde (~) represents the highest gust report in a dataset that has been interrupted at the height of the storm--usually data loss is from power outages. Values in italics are gust values estimated from peak wind, typically 2-minute or 5-minute, using a 1.3 gust factor. Stations with high-wind criteria gusts (≥58 mph or 93 km/h) are denoted with white-filled circles. Isotachs depicting ≥60 mph (~100 km/h) gusts are included to highlight the regions that had concentrations of the indicated magnitudes. The track of the extratropical cyclone center is shown (yellow arrow). Click on the map to see a larger version. Here is a map listing the station names.

Figure 2.1 above Storm track estimation largely based on surface maps provided by the US. NOAA Weather Prediction Center. Date and time in PST and central pressure in hPa (mb).

Figure 2.2 above The December 11, 2014 track (heavy black line) compared to other classic-path windstorms (various colors).

The extratropical cyclone intensified far to the south off of the California coast (Figure 2.1), moved northeast to east-northeast inside 130ºW at about 40ºN and then tracked north-northeast to Tatoosh and Vancouver Islands, with the low center gradually closing in on the coast as storm tracked northward. This is a great example of the classic path. The midlatitude cyclone reached a peak depth of approximately 974 hPa around 1600 PST on December 11, 2014, while off the mouth of the Columbia River. Other classic windstorms have also showed this tendency--reaching max depth while approximately off the Oregon coast and weakening before landfall on either the northwest tip of Washington or Vancouver Island. Thus, the low had peak depth while closest to Oregon, but in fact moved closer to Washington even as it weakened, roughly evening out the potential difference in impact.

The center of the December 11, 2014 extratropical cyclone tracked very close to the path of the 1962 Columbus Day Storm (Figure 2.2), in fact much closer than the major November 14, 1981 and December 12, 1995 windstorms.

The low-pressure center traveled approximately 690 miles (1,100 km) during the 24 hours from 0100 PST on Dec 11th to 0100 on the 12th. Thus, the system moved with an average forward speed of 29 mph (46 km/h), not as fast as some other classic windstorms. For example, the Columbus Day Storm tracked 465 miles (765 km) from 1000 to 2200 on October 12, 1962 for an average forward speed of 39 mph (62 km/h). The pace of the 2014 storm is closer to but still below the November 14, 1981 extratropical cyclone, which covered 645 miles (1040 km) in 21 hours as it moved up the coast, for an average speed of 31 mph (50 km/h). The January 16, 2000 windstorm had the relatively slow pace of 24 mph (38 km/h) as it moved from the waters off of southwest Oregon to west of Quillayute (KUIL) from 16:30 on January 15th to 13:00 on the 16th, a distance of 484 miles (780 km). Note that these are averages over fairly long time periods of 12- 24 hr, and that the forward speed of extratropical cyclones can vary over smaller time scales. For example, the Columbus Day Storm had a rate of about 51 mph (82 km/h) over 3 hr as it tracked off the Oregon Coast, and the December 11, 2014 storm had a 3 hr speed of 44 mph (72 km/h) off nearly the same stretch of coast.

Figure 2.3 above Synoptic chart for 0000 UTC December 12, 2014 (1600 PST December 11, 2014). Orange shading depicts the 300 hPa jet stream, with brown italicized numbers labeling isotachs in m/s. Some radiosonde wind observations are included (in black) mainly to show conditions in the study region (2.5 m/s per half barb and 25 m/s per pennant). For the 500 hPa level black lines denote heights in dm. Upper lows and central heights are also marked in black. Isotherms in ºC are indicated with white dashed lines. Surface lows and central pressures in hPa are indicated with dark blue, with tracks in light blue. Key surface anticyclones are indicated in red.

Figure 2.4 above Synoptic chart for 1200 UTC (0400 PST) December 12, 2014. See Figure 2.3 caption for the key.

Figure 2.5 above Meteogram for Portland (KPDX, orange), Seattle (KSEA, gray) and Vancouver (CYVR, blue) showing 2-min wind speed (mph), temperature (T, ºF) and dew point (Td, ºF) for the time period 00:00 PST to 10:00 PST on December 11-12, 2014.

Figure 2.4 above Satellite photo composite of: a) four km resolution visible; b) four km water vapor; c) four km enhanced infrared; and d) one km visible images for 1800 UTC (1000 PST) on December 11, 2014 (d).

Figure 2.5 above Satellite photo composite showing four km resolution water vapor images for: a) 1500 UTC (0700 PST) on December 11, 2014; b) 1800 UTC; c) 2100 UTC; and d) 0200 UTC on the 12th (d). The images depict the upper-level low center moving from the waters offshore of Coos Bay, OR, to just west of Willapa Bay, WA, with the associated fronts sweeping far inland.

Table 3.1 above Minimum sea-level pressure for 26 stations in the study region, timing and averages. The 11-station average uses ACV, OTH, AST, UIL, MFR, EUG, SLE, PDX, OLM, SEA and BLI.

Minimum pressures for the December 11, 2014 windstorm (Table 3.1) are roughly comparable to the major and much deeper November 14, 1981 extratropical cyclone. This is in part because the 2014 weather system tracked closer to the coast than the intense storm of '81.

Figure 3.1 above Coastal sea-level pressure trends (hPa) during the December 11, 2014 windstorm. Southern stations are shaded in orange, with northern in blue.

Figure 3.2 above Interior sea-level pressure trends (hPa) during the December 11, 2014 windstorm. Southern stations are shaded in orange, with northern in blue.

Sea-level pressure trends (Figures 3.1 and 3.2) show a classic progression with minimums being reached at southern stations first and northern stations last. A pronounced double-dip structure is present especially for Oregon locations, the first dip being associated with the passage of the leading cold/occluded front. Rapid pressure rises occurred at those sites that had close passage of the low center, most evident at the coastal stations Hoquiam (KHQM), Quillayute (KUIL) and Tatoosh Island (TTIW1). The broad, bowl-like form of the traces indicate an extratropical cyclone that is not progressing particularly fast, and also one that began weakening well before reaching its landfall location in the vicinity of Tatoosh Island.

Figure 3.3 above Coastal sea-level pressure cross-sections (hPa) during the December 11, 2014 windstorm. Stations are arranged by latitude. Earlier times (PST) have lighter shades.

Figure 3.4 above Interior sea-level pressure cross-sections (hPa) during the December 11, 2014 windstorm. Stations are arranged by latitude. Earlier times (PST) have lighter shades.

Pressure cross-sections (Figures 5 and 6) reveal an extratropical cyclone tracking from north to south that has, in typical fashion, the strongest pressure gradients to the south of the center. Keep in mind that these cross-sections do not cut through the low center until it is landing--the 2100 PST trendline is closest to landfall. Also, there is some noise because stations are not necessarily aligned perfectly south-to-north. Due to the path that the storm followed, most of the cross-sections depict a slice of the low to the east of the center. The coast, nearest the low center, had the steepest gradients. At 1500 PST on the 11th, the storm center neared Astoria and by 1800 PST, closed in on Tatoosh Island, carrying the strongest gradient northward just behind the low.

Table 3.2 above Maximum one hour pressure tendencies for 26 stations in the study region. Separation is the number of hours between the maximum rate of fall and maximum rate of rise.

Figure 3.5 above Hourly pressure tendencies at coastal stations for the December 11, 2014 windstorm. Southern stations are shaded in orange, with northern in blue.

Figure 3.6 above Hourly pressure tendencies at interior stations for the December 11, 2014 windstorm. Southern stations are shaded in orange, with northern in blue.

Overall, pressure tendencies were in a typical range for windstorms (Table 3.2). Storm averages were a little faster than the January 16, 2000 classic windstorm and on par to slower than the major November 14, 1981 event. Tendencies were generally the strongest on the coast where the low-pressure center tracked the closest. The rapid rate of fall at Newport (KONP) is in a rarefied category--the cold front likely contributed to this intense pressure drop. Tatoosh Island (TTIW1) experienced a strong pressure surge as the low tracked nearly right over the station. Values above +7.5 hPa/hr are quite rare and are almost always associated with the passage of a strong bent-back front, as was the case at Tatoosh during the December 11, 2014 windstorm. The pressure tendency trendline for Tatoosh shows a spike that stands out above all the other locations (Figure 3.5), the clear signal of a bent-back front. The interior stations generally had similar pressure tendencies, with no stations showing an intense surge like Tatoosh (Figure 3.6). The bent-back front weakened as the extratropical cyclone moved inland, which likely lessened the impact at inland locations. Such weakening is a common outcome, but sometimes landfalling lows bring inland strong bent-back fronts, such as February 7, 2002 and the 2006 Hanukkah Eve Storm.

Intense and fast-moving extratropical cyclones, two ingredients for high winds, tend to produce narrow separations between the peak rates of pressure fall and rise. The December 11, 2014 windstorm had some wide values even at locations close to the landfall point, largely a reflection of a slow moving storm, but also one that underwent its most rapid deepening well offshore.

Table 3.3 above For a selection of the many possible station pair combinations in the study region, standard (or one-dimensional) pressure gradients. Station-pair latitude (º), distance (km) and bearing (º) are provided. The raw peak gradient magnitude in hPa is shown along with the same value scaled to hPa/100 km. Scaling the gradients to a standard measure provides more consistent intercomparison between regions. Times and dates are in PST.

Figure 3.6 above Coastal scaled pressure gradients (hPa/100 km) for medium-distance station separations. Southern station pairs are shaded in orange, with northern in blue.

Figure 3.7 above Interior scaled pressure gradients (hPa/100 km) for medium-distance station separations. Southern station pairs are shaded in orange, with northern in blue.

Figure 3.8 above Intercomparison of coastal and interior scaled pressure gradients (hPa/100 km) for medium-distance station separations, with locations arranged by latitude. Depicted are the peak absolute values, therefore negative gradients are shown as positive (not to be confused with the absolute, or 2-D, pressure gradients discussed in section 3.4 below). The coast is shown in black, interior gray.

In general, the standard 1-D pressure gradients for the December 11, 2014 windstorm were strong but not near record territory (Table 3.3). The highest values in standard fashion tended to be along the coast for those regions that ended up in the southeast quadrant of the extratropical cyclone. Among the south-north measures and especially for the interior locations, negative values, indicating gradients with a northerly component to their pressure slope (likely northeast), tended to be stronger at a larger number of locations than is typical. This suggests a northeast quadrant with gradients somewhat more intense than the southeast quadrant, and may be indicative of the storm running into relatively strong high pressure over the intermountain region to the east.

There is a nice progression of ever strengthening pressure gradients from south to north along the coast (Figure 3.6). The Washington coast experienced the strongest gradients, not surprising given the track of the low almost over Tatoosh Island. The interior stations show a similar progression, but more weakly, with the highest values over the Northwest Interior of Washington (Figure 3.7). Broadly speaking, coastal locations had higher values than the interior, in part a result of the closer proximity of the low to the coast (Figure 3.8), though pressure slope interacting with the orientation of the station-pairs confounds interpretation.

More detailed discussion of pressure gradients using 2D measures is in section 3.4, below.

Table 3.4 above For 18 pressure-wind triangles located in the study region, peak absolute pressure (or 2-D) gradient (hPa/100 km), associated pressure slope (º), geostrophic potential wind (Mg in mph) and estimated peak surface wind and gust (mph) for overland regions based on the gradient magnitude and the presence of turbulent drag from surface roughness.

Figure 3.9 above Absolute (2-D) pressure gradients (hPa/100 km) for coastal regions. Trendlines from southern pressure-wind triangles are shaded in orange, with northern in blue.

Figure 3.10 above Absolute (2-D) pressure gradients (hPa/100 km) for interior regions. Trendlines from southern pressure-wind triangles are shaded in orange, with northern in blue.

Figure 3.11 above Intercomparison of the peak absolute pressure (or 2-D) gradient (hPa/100 km) for coastal (black) and interior (gray) locations arranged by mean latitude.

Table 3.5 above For 24 key stations in the study region, peak wind and gust (knots, mph and km/h), with direction (º) and time of occurrence (PST). Regional and overall maximums, minimums and averages are provided, along with the legacy 11-station average. Average gust ratios (average wind/average gust), or AGR, are also included. Wind direction average is based on vector components.

Figure 3.12 above Peak wind (mph), peak gust and peak wind direction (º) for coastal stations arranged by latitude.

Figure 3.13 above Peak wind (mph), peak gust and peak wind direction (º) for interior stations arranged by latitude.

Figure 3.14 above Peak wind (mph) and peak wind direction (º) for coastal and interior stations arranged by latitude.

Peak wind and gust for the December 11, 2014 shows notable similarity between the coast and interior (Table 3.5, Figures 3.12-3.14). Indeed for some locations at similar latitudes, the interior had higher speeds (e.g. Portland vs. Astoria). Such an outcome is not unprecedented (e.g. April 14, 1957), and in fact a pattern where coastal and interior peak wind speeds are roughly comparable--or interior speeds are higher--is evident in a number of past classic-path windstorm events. This is interesting given that coastal locations more typically, being in close proximity to a large body of open water, tend to have higher wind speeds than the interior during routine storms and other types of windstorm scenarios (e.g. extratropical cyclones with northeast to east tracks). For a like pressure gradient--and during the 2014 storm pressure gradients were similar between the coast and interior save for in parts of Washington (Figure 3.11)--one might anticipate higher surface winds at coastal locations due to lower turbulent drag over the water resulting in faster speeds. The overwater momentum can be carried inland, though it is reduced fairly rapidly. One example of this is at Vancouver, BC, where a westerly wind blowing inland off of the Georgia Strait can have its speed cut in half over just a few km as the airflow encounters higher drag forces overland.

During southerly windstorms on the Oregon and Washington coast, things are complicated as the wind is roughly parallel to the shoreline and the potential overwater speed boost is likely reduced save for at exposed headlands and places very close to shore. However, given the tendency for a southeasterly pressure slope as the low passes to the north of a given region, and the tendency for wind direction over the water to be nearly parallel with the isobars--turned toward the low by a few degrees--there is the tendency for a modest westerly component to the wind direction during classic windstorms along the coast. For example, the peak wind directions at nearly all the coastal stations during the December 11, 2014 windstorm were from the SSW to SW. These wind directions are favorable for bringing some over-water momentum inland to coastal communities.

Other factors that favor higher speeds at the coast, especially during classic windstorms, include a closer proximity to the low-pressure center that usually results in higher pressure gradients. In the case of 2014, being nearer the low did not necessarily translate into steeper gradients, pointing to part of the reason for the relatively slow winds on parts of the Pacific shore. A coastal low-level jet can also develop under the right circumstances, which may boost local wind speeds depending on a number of factors. It appears that this phenomenon did not take on a strong form during the 2014 windstorm--but this is outside the scope of the current analysis.

One important reason for the pattern of roughly comparable winds between the coast and interior is the already discussed good upper support for high winds over the Willamette Valley (Section 2.2). By the time the low-pressure center reached a position to deliver strong southerly winds to Oregon's north coast, the best upper support had shifted inland, over the Willamette Valley. This probably boosted wind speeds in the Valley, especially in the Portland Metro area, relative to the coast where upper winds were slower.

The average peak wind and gust speeds for the December 11, 2014 windstorm (Table 3.5) were below the 33.4 mph and 52.5 mph averages for the January 16, 2000 classic windstorm. Given a track closer to the coast, and an apparently deeper minimum central pressure (974 hPa in 2014 vs. 980 hPa in 2000), the slower peak winds for the 2014 storm are interesting. Consider also the fact that the 2000 windstorm had the relatively slow pace of 24 mph (38 km/h), even slower than the 29 mph (46 km/h) speed of the December 11, 2014 extratropical cyclone (section 2.1 above). Though the two events appear to have had relatively similar upper-level wind setups, the 2014 windstorm almost certainly had an edge over the 2000 storm, with close to an ideal arrangement over parts of western Oregon. By all accounts, the 2014 windstorm should have outperformed the event from 2000. Then, add in the fact that many of the wind measurements in 2014 were on new very sensitive sonic anemometers that do not have inertial response considerations and gusts are determined with a 3-second moving average. In 2000, reported gusts were a 5-second block average on cup-based anemometers. These differences in instrumentation and methodology make the lower peak winds in 2014 even more remarkable. It appears that there are still some mysteries to be solved with regard to Pacific Northwest windstorms, though it is important to recognize that the two windstorms being compared here are close in magnitude which means that noise is potentially a confounder to interpretation. The broad patterns described here (e.g. the influence of central pressure and track location on peak wind speeds), when applied over a wider range of events including the major storms, stand out more clearly.

A few select regions did experience faster winds during the December 11, 2014 windstorm. The Portland Metro area appears to have had the strongest winds since the December 12, 1995 windstorm. Given the differences in the peak gust measurement methodology between storms, it is perhaps better to consider the maximum 2-min wind speed when comparing events. For Portland (KPDX), the 2014 storm produced a faster peak wind than the 2000 event, but not as strong as the peak in 1995.

Table 3.6 above Peak 2-minute wind at Portland (KPDX) for the three most recent classic-path windstorms. The maximums are from the daily data, which captures the highest wind speed during the day. These speeds may not match those listed in the standard peak wind table (e.g. Table 3.5), since the numbers in the standard table are obtained from the hourly and special observations.

The Portland Metro area is one of the few regions where the December 11, 2014 extratropical cyclone outperformed the January 16, 2000 storm. Parts of the southwest Oregon coast also appear to have had stronger winds (Figure 1.1). But for the coastlines of central and north Oregon also much of Washington, the storm in 2000 generally produced stronger gusts, and this with 5-second averages employed. The Washington interior also had generally stronger gusts in 2000, though around Whidbey Island the winds may have been a little slower.

Figure 4.1 above Hourly 2-minute wind compared to hourly precipitation at Portland (KPDX), Seattle (KSEA) and Vancouver (CYVR). CWMM, or Pitt Meadows, is used in place of CYVR for precipitation because CYVR appears to report only 6-hourly totals.

Temperature and dew point in relation to the timing of peak winds and the idea of vertical mixing causing dew point depression are discussed above (Section 2.2, Figure 2.5). In this section, other weather variables will be examined for the key population centers of the region.

The main precipitation with this storm occurred well ahead of the peak winds (Figure 4.1), and is clearly associated with the leading front. Totals were not very high, ranging from 0.15" (3.8 mm) at Portland to 0.37" (9.4 mm) at Pitt Meadows. According to the weather lore of the Pacific Northwest, windstorms tend to be relatively "dry", and the December 11, 2014 event fits the mold. However, some recent windstorms have brought with them intense precipitation, including the 2006 Hanukkah Eve Storm and the 2007 Great Coastal Gale, and thus one should not expect every windstorm to produce little precipitation.

Also, the December 11, 2014 windstorm brought a deluge to northern California. The extratropical cyclone left a long trailing front, in essence an atmospheric river, draped across much of the state. For example, San Francisco (KSFO) had heavy rain with 0.37" (9.4 mm) in the hour ending 08:56 on December 11th. Given the strong meridional jet stream, the front slowly progressed southeast, followed by the development of a secondary wave that caused the baroclinic band to lift northward again as a pseudo-warm front (NCDC 2014), bringing another round of intense rain to northern California. The protracted rain in the Bay Area resulted in a storm total of 3.57" (90.7 mm) at San Francisco. This is a lowland station--higher elevation locations, and places north received two to three times as much (Figure 4.2). Indeed some locations reported hourly rain totals of two to three inches (50-75 mm). The storm ameliorated a historic drought underway in the region, but only temporarily as this was essentially a one-day rain event right in the middle of a remarkable multi-year dry period--reminiscent of desert thunderstorms that drop inches of rain between months of arid conditions.

Figure 4.2 above Northern California 36-h precipitation (inches) for the December 11, 2014 windstorm. Map courtesy of the National Weather Service. Hat-tip to Charlie Phillips for finding this map.

Heavy rains also occurred in California during great 1962 Columbus Day Storm. Indeed, this appears to be a general pattern among classic-path windstorms: California tends to get the heaviest rainfall and locations north tend to get the highest wind. The strong meridional pattern during classic windstorms appears to favor heavy rain in California because the coastline, and associated mountains, jut southeast to northwest. South to southwest upper air flow runs nearly perpendicular to the terrain axis, forcing copious rainfall. Also, the already mentioned tendency for frontal systems associated with deep upper troughs to have a slow progression eastward likely contributes.

In San Francisco on December 11, 2014, wind gusts to 48 mph (78 km/h) lashed the heavy rain and the combination caused widespread disruption of transportation and electrical systems. This earned the tempest the moniker #Hellastorm, apparently the first classic-path windstorm in history to get a hash tag for a name.

Returning to the Pacific Northwest, for the north-south trending valleys, the onset of strong winds with classic windstorms are often very closely tied with passage of the low-pressure center just north. In essence, as the pressure slope shifts from easterly to southeasterly, i.e. going from unsupported to supportive of southerly ageostrophic winds, the surface wind shifts from the northeast quadrant (E to ESE at KPDX due to the Columbia Gorge gap winds) to the southerly quadrant and accelerates fairly rapidly--sometimes it is like a switch has been thrown. Portland (KPDX) most clearly shows this pattern (Figure 4.3), with wind speeds accelerating right after the second barometric pressure minimum, which marks the time that the low has moved north of the station latitude.

Figure 4.3 above Hourly 2-minute wind and 3-second gust (mph) at Portland (KPDX), Seattle (KSEA) and Vancouver (CYVR) compared to sea-level pressure (hPa). Gust is not reported during most observations as there is an algorithm that prevents gust records if the separation between wind and gust is below a specific value. In these cases, gust is estimated using a sliding gust factor based on reported wind speed. The gust values during the main wind surge at each station were recorded--they met the criteria--and therefore these are not estimated.

At Seattle (KSEA) and Vancouver (CYVR), the pattern is not so clear, mainly because there is an interesting delay between low pressure and the onset of the main wind surge (Figure 4.3). The delay may be due in part to the low weakening as it moved up the coast, resulting in a slowly rising pressure before the cyclone center reached the latitude of Seattle.

Regional pressure slopes provide a stronger picture of the relationship between low position and the onset of the main wind surge (Figure 4.4). During weak pressure gradients (e.g. <1.0 hPa/100 km), small difference in pressure due to local effects or measurement error can lead to large swings in pressure slope. Also, incoming frontal systems can also cause big short-period changes in the gradient orientation. These two factors explain the sharp jumps in pressure slope for some regions early in the storm period.

For Portland, the main winds did not arrive until the pressure slope has exceeded approximately 135º, or southeast. If the Willamette Valley is assumed to have a 180º orientation (for the sake of simplicity), then about 71% (cos 45º) of the pressure gradient's remaining potential could be realized after surface boundary-layer factors such as turbulent drag are taken into consideration.

Figure 4.4 above Hourly 2-minute wind for Portland (KPDX), Seattle (KSEA) and Vancouver (CYVR) compared to the hourly pressure slope (º) as determined using the Willamette Valley, Puget Lowlands and South Georgia Strait pressure-wind triangles.

At Seattle, the main winds started when the pressure slope reached approximately 110º, not as supportive of south winds as the 135º orientation at Portland, but still above 90º (Figure 4.4). When considering conditions in the Seattle area, it is important to keep in mind that troughing in the lee of the Olympics likely affected the pressure field around Port Angeles (KCLM). A mesoscale lee low is indicated in the numerical weather prediction model forecast just before the storm (Figure 4.5). Such lee troughing results in a tendency for more easterly pressure slopes when using the Puget Lowlands pressure-wind triangle, which uses Port Angeles pressure readings for one of its legs. Due to using Victoria (CYJJ) pressure data, the South Georgia Strait pressure gradient orientations are also likely influenced by the Olympic lee low in a similar manner, but generally not as strongly.

Figure 4.5 above WRF-GFS 4-km domain 12:00 UTC December 11, 2014 initialization. This is a forecast for sea-level pressure, wind and 925 hPa temperature (ºC) valid for 22:00 PST on December 11th (18 hr in the future). Position of the extratropical cyclone center and an associated Olympic Mountain lee low are indicated. NWP output courtesy of the University of Washington Department of Atmospheric Science.

During southeasterly windstorms, Vancouver (CYVR) tends to have a more gradual escalation of winds toward peak values than places like Portland or Seattle (Figure 4.4). This is largely due to Vancouver being situated in a geography that favors southeast winds, including its close proximity to the Georgia Strait, a large body of water relative to narrow channels like the Puget Sound. Pressure slopes from the east can support SE winds at Vancouver. Usually, the winds start out E (with a northeast pressure slope) and gradually migrate to SE as the storm approaches, becoming ever faster until the low center makes landfall, at which time peak winds tend to occur followed by a gradual decline. Looking at the data for the December 11, 2014 windstorm, the peak wind phase of the storm started after the pressure slope shifted from northeast (~45º) to nearly due east (~90º), with the wind continuing to blow at elevated speed as the gradient orientation quickly aligned nearly due southeast (~135º) over the next hour.

In summary, at all three stations, the pressure slope reached 90º or above before the onset of the peak wind phase, indicative of the low-pressure center having reached, or passed the station latitude.

Peak winds generally occurred at the time of maximum absolute pressure gradient at the three locations (Figure 4.6). This is what I call a "well-behaved" storm, with an apparent relationship between surface pressure gradient and wind speed. In the complex terrain of the Pacific Northwest, this does not always occur. Indeed, the relationship between pressure gradient and wind speed may be quite weak, depending on the situation.

Figure 4.6 above Hourly 2-minute wind at Portland (KPDX), Seattle (KSEA) and Vancouver (CYVR) compared to the hourly absolute pressure gradient (hPa/100 km) as determined using the Willamette Valley, Puget Lowlands and South Georgia Strait pressure-wind triangles.

The much stronger pressure gradient over the Puget Lowlands seen in Figure 4.6, nearly 9 hPa/100 km, is likely due to the influence of the abovementioned Olympic lee low. Very tight isobaric packing is evident south of the meso-low in Figure 4.5, covering part of the Puget Lowlands. Interestingly, lee troughing is thought to have the potential to prolong the strong wind phase over the Puget Sound (Steenburgh and Mass 1996), but in the case of the December 11, 2014 windstorm a protracted wind event did not occur. This despite six consecutive observations with a pressure gradient higher than what occurred over the Willamette Valley or South Georgia Strait. The tendency for a more easterly pressure slope may have mitigated the winds in Seattle somewhat. However, as noted earlier, the orientation shifted nearly due south during peak gradient, a very supportive alignment for south winds. There is a bit of a mystery here.

The South Georgia Strait shows a double peak in pressure gradient. As discussed earlier, the bent-back front appears to have had a noticeable influence on the weather conditions in the Vancouver Metro area. The second spike in pressure gradient is also probably associated with the bent-back front. Note how there is a modest escalation in wind speed, off of near-maximum values, associated with the second increase in pressure gradient.

The leading front sweeping inland ahead of the low caused an initial dip in sea-level pressure at the three stations Portland, Seattle and Vancouver (Figure 4.7). These pressure changes influenced local pressure slopes (Figure 4.4), which in turn caused the 2-minute wind direction to change dramatically over a short period of time (Figure 4.7). After the front moved through, and the pressure slopes realigned in response to the cyclonic curvature of the incoming extratropical cyclone center, winds resumed, roughly an initial offshore configuration, E to NE. The direction abruptly shifted to S and then SSW at Portland as the low passed the station latitude. The persistence of the southerly direction at Portland is likely a reflection of the close to ideal wind-tunnel like configuration of the Willamette Valley. Seattle experienced a more gradual alteration of direction, shifting through E to SE to S over a six hour period. The winds at Vancouver also underwent a more gradual change, following the fairly common pattern as discussed above, moving from E to SE and eventually SSE as the low center passed to the west on its north-northeast track.

Figure 4.7 above Hourly 2-minute wind direction (º) and sea-level pressure (hPa) at Portland (KPDX), Seattle (KSEA) and Vancouver (CYVR).

The December 11, 2014 windstorm had a classic path, i.e. strongly meridional, with the low tracking north-northeast just off the Pacific coast. In fact, the storm had a trajectory quite similar to the infamous Columbus Day Storm--the closest of any classic windstorm since 1962. The extratropical cyclone went through what might be considered a fairly typical progression of development for this type of storm, reaching peak intensity (Shapiro and Keyser Stage-III) while off the southwest Oregon coast and reaching maximum depth (while in Stage-IV) as the low moved over northwest Oregon waters.

Due to following a classic path, the storm brought strong winds from northern California into southwest British Columbia. However, overall peak wind speeds generally did not approach those of the major windstorms (e.g. the 1962 Columbus Day Storm, November 13-14, 1981 and December 12, 1995), being comparable to what might be considered middle-of-the-road classic windstorms such as January 16, 2000. The minimum central pressure, around 974 hPa, was above the rare 947-960 hPa range observed during key historical events. And the 2014 windstorm appears to have progressed more slowly than the major windstorms. These two characteristics probably are the main reasons for lower wind speeds relative to the major storms, though other factors likely also contributed.

Interestingly, compared to the January 16, 2000 windstorm, the December 11, 2014 storm had somewhat weaker overall peak winds. This is despite many characteristics in favor of higher winds for the 2014 storm, including: 1) track closer to the coast; 2) deeper central pressure; 3) faster northward migration; 4) better upper-level wind support including favorable jet-core positioning right over the region; and 5) wind speeds being measured on sonic anemometers, as opposed to cup-based sensors that have inertial response considerations, with a shorter averaging period for gust. These facts suggest that there is more waiting to be learned about Pacific Northwest Windstorms.

Figure 6.1 above Southeasterly wind gusts on in the vicinity of Whidbey Island, WA, climbed to 60-70 mph (95-115 km/h) during the December 11, 2014 windstorm, causing damage to trees and property. In this scene, a large red alder succumbed to the gale, bringing down lines as it crashed across a byway. Photo by Charlie Phillips.

Figure 6.2 above A routine sight throughout the storm strike zone, fences in Sandy Hook on Whidbey Island, WA, were disrupted by the December 11, 2014 gale. Photo by Charlie Phillips.

Figure 6.3 above A Sandy Hook, Whidbey Island, boat shed lost a portion of its roof during the high winds. Photo by Charlie Phillips.

Figure 6.4 above Known as "salad" in some circles, conifer trees shed numerous small branches during the December 11, 2014 windstorm, nearly burying this Whidbey Island road. Photo by Charlie Phillips.

Figure 6.5 above As the December 11, 2014 extratropical cyclone landed on Vancouver Island, southeast wind gusts climbed to 50-55 mph (80-90 km/h) throughout much of the inland areas of Metro Vancouver, BC, with speeds reaching 60-65 mph (95-105 km/h) along the shores of the Georgia Strait. Enough to topple trees such as this plum in the Oakridge area. Power outages, mainly due to tree and branch failures, were widespread, with 78,000 customers affected (5.5% of the customer base) at peak according to BC Hydro.