Greetings from Bodega Head!
What was supposed to be an unremarkable series of evening rain showers turned into nearly an inch of rain here yesterday (Monday, March 3, 2014)!
The official forecasts were slow to pick up on this rain. in fact, as late as Sunday afternoon, the national weather service was predicting 50% PoP and was not assigning a quantitative precipitation forecast. There was much about this storm that was unexpected. Let’s investigate a bit to see why it was so hard to predict in advance.
We are thoroughly soaked here! It was a damp chilly morning, but winds are light out of the SSE and things are slowly warming up. Surface weather timeseries:
Timeseries of surface met conditions at BBY for the 48 hour period ending at 0800 PST on March 4, 2014. Credit: NOAA-HMT.
Notice two things in the above plot: 1) rain started falling just after 2 pm yesterday and by 2 am an inch had fallen at BML. 2) AR conditions existed for 24 hours over BML, from roughly midnight Monday to midnight Tuesday. This second point can be validated by the integrated water vapor (IWV) in the bottom panel. IWV greater than 2 cm corresponds to AR conditions. Interestingly, the NE Pacific did contain a long filamentary feature of enhanced moisture over the weekend and leading into Monday:
MIMIC PW analysis for the NE Pacific valid 2000 UTC on March 3, 2014.
We had an inch of rain, over 2 cm of IWV at BML for more than 24 hours, and a long filament of enhanced water vapor which connected to the tropical Pacific and made “landfall” on the Northern California coast.
Yet, as I mentioned above, the rainfall was somewhat a surprise. The reason is that there was not an obvious dynamic forcing mechanism to provide lift – and convert water vapor into water condensate. This can be seen in the GCM forecasts for Monday afternoon.
The upper level flow over the Northeastern Pacific Ocean has been strictly zonal for several days. At the surface, the nearest large-scale (thousands of kilometers or greater) storm has been slowly making it’s way east from the Central Pacific. As of 00 UTC Sunday, the ECMWF deterministic forecast for 4 pm PST Monday looked like:
ECMWF 48 hour forecast MSLP (left) and 500 hPa geopotential (right) valid at 00 UTC on March 4, 2014. Credit: ECMWF.
There isnt much in the synoptic scale surface (left panel) or upper level (right panel) fields that suggest rain on the west coast. If we look at the drift in this solution for the following 2 days, nothing changes:
Same as above, except 24 hour forecast.
Same as above, except analysis.
The surface and 500 hPa analysis from ECMWF above make it clear that yesterdays AR conditions were not typical, in that the AR did not arrive in the warm sector of a pacific extratropical cyclone.
Nature of the March 3 storm:
So how did so much rain fall at BML? We are not on a steep mountain ridge, so orographic lift will not provide rising motion – and condensation, clouds, autoconversion and rain – unless there is a coastal jet. There was no coastal jet yesterday, as can be seen in the 449 MHz wind profiles:
Timeseries of wind vectors measured by the 449 MHz wind profiling radar at BML and upslope IWV flux for the 48 hours ending at 16 UTC on March 4, 2014. Credit: NOAA-HMT.
The plot above does elucidate two features, however. The first is a mid-level trough which propagates above the profiling radar. Winds at 3 km turn cyclonically with time – from southwesterly to westerly to northwesterly – during the time period starting at noon yesterday and ending at 4 am.
The second feature is the delay in BML rain compared to the rain at Cazadero. Note that the coastal rain rate peaks while the mid-level trough passes above, while the CZC rain rate peaks when the wind projected along 230 degrees (SW) is strongest.
Coastal rain at BML was primarily due to lift and pressure dropping due to low-level thermal vorticity advection. Wind at the surface was southerly to south-southeasterly thoughout the past 48 hours. The wind speed trace on the surface meteorology timeseries was remarkably constant (see first figure, this post). When the trough approached above 3 km, positive vorticity advection increased with height. From quasi-geostrophic theory, lift should be forced in the lowest levels as a result. Even though there was not a strong upper trough approaching the area, there was local dynamically forced ascent. Its signature can be seen in the NAM analysis valid at 12 UTC on March 3, 2014:
Analysis wind shear in the layer bounded by the surface and 850 hPa, valid at 12 UTC on March 3, 2014. Credit: NCAR (model), Weathernerds (visualization).
The low (surface to 850 hPa) shear vectors clearly show the trough approaching the Northern California coast. Because this mid-level feature was not very widespread, significant precipitation was confined to the Northern California coastal ranges:
24-hr rain estimate from composite of WSR-88D radar network for March 3, 2014. Credit: NOAA-NWS.
Lift and rain at CZC were forced both by thermal vorticity advection and by orographic uplift. In the IWV plot above, rain at CZC (green bars) becomes heavy as soon as moderate southwesterly winds are established in the layer from 500 m to 2 km AMSL. Because these southwesterly winds flow directly uphill from the coast to the Cazadero site, there is forced ascent. Additionally, the timing of the strong low-level southwesterlies coincided with the arrival of extended AR conditions at BBY. In fact, the Southwesterly winds can be thought of as an extension of the AR itself:
GFS Analysis water vapor mixing ratio (g/kg) at 850 hPa valid at 00 UTC on March 4, 2014. Credit: NCAR (model), Weathernerds (figure).
The GFS analysis of water vapor mixing ratio at 850 hPa from 00 UTC yeseterday picked up on a channel of moist Southwesterlies intersecting the CA coast. This is the same flow that can be seen in the 449 MHz profiling radar plot, and likely extends SW well into the tropical Pacific. This flow is the AR.
The information presented above confirms that yesterday’s rain storm was an AR without an extratropical cyclone. Or, an AR without upper-level dynamic forcing. The final piece of interesting data I will share is the S-Band vertical profiling radar timeseries from Cazadero. Rain was heavy throughout the afternoon, and clouds were deep – reaching to 7 km at echo top height. However, there is a clear lack of seeder-feeder development in this storm:
S-Band signal-to-noise ratio (with brightband) timeseries from CZC ending at 16 UTC on March 4, 2014. Credit: NOAA-HMT.
Long-range transported aerosol:
A precursor seeder (upper) cloud, along with satellite and aircraft aerosol measurements, is something we use to indicate that upper-level long-range transported ice nuclei were participating in the precipitation process above CZC. In this case, the upper level cloud was, and the upper level ice nuclei may have been, missing.
In this post, I showed a GEOS-5 forecast which indicated dust may arrive just before and during Monday’s rain event. We took time-resolved precipitation samples during the rainy periods yesterday. It will be interesting to see if the rain samples contain significant dust from this storm; and whether that dust is present from time periods when clouds were likely shallow, or whether it appears only when echo tops are well above the freezing level.