Southern California pattern

The composite mean rainfall for a number of heavy rainfall events in southern California is presented below.  The composite mean 500-hPa anomaly pattern for these events is also shown.  The negative anomaly center is farther to the south than for the cases when the plume is impacting northern California, almost directly west of San Francisco.  Such a position leads provides a a fairly strong southwesterly to south-southwesterly gradient across the mountains of southern California. normal to the mountains. 

Composite of 24-hr rainfall for heavy rainfall events in the mountains near Los Angeles.  From Bell et al.,  2005

Composite mean 500-hPa height anomaly pattern for the beginning of the period of heavy rainfall on southern California.  From Bell et al.,  2005

Composite mean 250-hPa winds pattern for the beginning of the period of heavy rainfall.  From Bell et al.,  2005.

Composite mean 850-hPa winds pattern for the beginning of the period of heavy rainfall.  Outer magenta area is where the anomaly of the v-component (southerly) of the 850 winds have standard deviation (S.D.) of a greater than 1.8 and the inner oval is where the departure is greater than 2.1 S.D. 

The Bell et al study suggested that the left exit region of an upper level jet streak might be enhancing the upper level divergence across southern California while also contributing to stronger than normal v-component to the 850-winds due to the ageostrophic response to the upper branch of the jet streak’s transverse circulation. 


The most extreme events flood cases often result from multiple storms with one or two being of them producing extreme rainfall.  February of 1980 in southern California and Arizona is a case in point.  The steady onslaught of storms brought many streams record volumes of run-off in southern California.  The 500-hPa analysis (bottom right) shows one

An example of a southern California event. 

The NCEP operational models often do a good job forecasting the basic synoptic pattern associated with many of the heavy rainfall events along the west coast especially in the shorter time ranges. 

12hr GFS 500-hPa and vorticity v.t. 0000 UTC 28 Dec. 2004.

Analysis of 500-hPa heights and standardized anomalies v.t. 1200 UTC 27 Dec. 2004.

30hr GFS 500-hPa and vorticity v.t. 18 UTC 28 Dec. 2004.

Legend (feet)

The GFS forecast of the 500-hPa pattern indicated that the intense 500 low with a greater than 3 S.D. height departure would be slow moving during the period.  The 500-hPa forecast fits the composite for 4 inch or greater rainfall,  What about other fields?  Note how similar the forecast 250-hPa isotach (below) field is to the one in the composite for the southern California heavy events. 

30-hr GFS 250-hPa heights,  isotachs, ageostropic winds (arrows) and divergence v.t. 1800 UTC 28 Dec. 2004

30-hr GFS 850-hPa winds and PW v.t.   1800 UTC 28 Dec. 2004.

The predicted 850 winds in the vicinity of southern California is actually stronger than on the composite mean for the southern California events and the PW field shown above is very similar to the composite mean suggesting that above normal moisture flux was also present at 850-hPa.   The pattern predicted by the models is very conducive to getting 5 inches of rain in a 24-hr period along the coast or over the east-west oriented mountains of southern California. 

10-km resolution depiction of terrain. 

30hr NAM boundary layer moisture transport (arrows), Orographic omega (color filled are positive values), and 80-km depiction of model terrain (brown lines, in meters. 

Now look at the 10-km find the areas where the changes in elevation are greatest.  The areas with the strongest gradients on the 10-km depiction are considerably different than shown on the 80-km eta depiction.  Now try to visualize where the strong orographic lifting will be based on the boundary layer moisture flux arrows predicted by the ETA and gradients using the higher resolution terrain.  The yellow arrows point to the location of the tightest gradient and the location where the terrain induced lifting is greatest.   The PRISM precipitation climatology has rainfall maximum over both regions where the gradient is tight. Neither rainfall maximum is right on the coast. Admittedly,  the PRISM relies on gradients of terrain to infer  

10-km PRISM climatology 

Rainfall maximum

precipitation in data scarce regions. Western region uses PRISM to help downscale their precipitation analyses so it is usually wise to model precipitation maximum fairly close to the PRISM maxima if the flow is typical for significant rainfall event.   

Neiman et al. (2002) have developed a conceptual representation of the distribution of rainfall in California’s coastal mountains for blocked and unblocked flow.  Blocked flow is essentially flow in which there is considerably veering of the winds with height at the lowest levels and situations when there also is a marked terrain parallel barrier jet. 


Conceptual representation of orographic rainfall distribution in California’s coastal mountains, and the impact of terrain-blocked flow on this distribution:  (a) plan view of unblocked and blocked flow, and (b) and © show the cross-section perspective of unblocked and blocked flow respectively, with representative coastal profiles of wind velocity (flags and barbs) and the correlation coefficient (based on the magnitude of the upslope flow at the coast vs. the rain rate in the coastal mountains) shown on the left.  The variable h in (b) and © is the scale height of the mountain barrier.  The spacing between the rain streaks in (b) and © is proportional to rain intensity.  The small circle in © represents the location of terrain-parallel barrier jet.  From Neiman et al. 2002. 

The research by Neiman and his colleagues suggests when the airmass is stable in the low layers with veering winds with height that the precipitation is less concentrated moderate to heavy rainfall extends upstream farther from the mountains then cases without veering winds with heights or the stable layer. Their research was conducted for the coastal mountains but there is reason to suspect that similar blocking can also occur along the Cascade and Sierra ranges (Doyle 1997, Steenburgh and Mass 1996). 


When might the modeling the precipitation to the PRISM data distribution be misleading and lead you to underpredict the rainfall in the valleys of California.

A)  With the post frontal rainfall

B)  a system with strong veering with heights and warm advection.

C) With strong frontogenesis with some instability.

D)  with strong and deep southwesterly flow when an atmospheric river is impinging into the coast. 

For the answer, see next page

Polar diagrams of floods on the Sauk, Snoqualmie, and Chehalia Rivers during a) Oct-Dec., b) Jan-Mar. for 1979-2000.  The magnitudes of the floods in terms of streamflow are proportional to the radius (the outer ring represents the all time record peak on each river): the timing of these floods with respect to the phase of the MJO is shown by the azimuth (see labels on the perimeter of each diagram).  From Bond and Vecchi, 2003.

The Bond and Vechii research shown above suggests that the MJO may be an important player in flood events during Oct.-Dec. period in western Washington as 81% of the floods during that period occurred in MJO phases from very late in phase 5-through early phase 8 with the bulk of the events falling in phase 7.  From the Figure,  Bond and Vechii avowed that there was evidence of MJO modulation of the Oct-Dec floods with a statistical significance of 99%. 


The relationship to floods on the western Washington rivers to the MJO is less clear during Jan-March.  The phases associated with flooding in Jan-Mar. period also differed from the phases associated with the fall and early winter floods.  It is unclear how much help the MJO can be at anticipating the onset of extreme rainfall along the west coast in the medium range.