Applied Aviation Sciences - Prescott College of Aviation 11-2005 MMooddiifificcaattiioonn ooff PPrreecciippiittaattiioonn bbyy CCooaassttaall OOrrooggrraapphhyy iinn SSttoorrmmss CCrroossssiinngg NNoorrtthheerrnn CCaalliiffoorrnniiaa Curtis N. James Embry-Riddle Aeronautical University, [email protected] Robert A. Houze Jr. University of Washington - Seattle Campus Follow this and additional works at: https://commons.erau.edu/pr-meteorology Part of the Meteorology Commons SScchhoollaarrllyy CCoommmmoonnss CCiittaattiioonn James, C. N., & Houze, R. A. (2005). Modification of Precipitation by Coastal Orography in Storms Crossing Northern California. Monthly Weather Review, 133(11). Retrieved from https://commons.erau.edu/pr-meteorology/1 © Copyright 2005 American Meteorological Society (AMS). Permission to use figures, tables, and brief excerpts from this work in scientific and educational works is hereby granted provided that the source is acknowledged. Any use of material in this work that is determined to be “fair use” under Section 107 of the U.S. Copyright Act September 2010 Page 2 or that satisfies the conditions specified in Section 108 of the U.S. Copyright Act (17 USC §108, as revised by P.L. 94-553) does not require the AMS’s permission. Republication, systematic reproduction, posting in electronic form, such as on a web site or in a searchable database, or other uses of this material, except as exempted by the above statement, requires written permission or a license from the AMS. Additional details are provided in the AMS Copyright Policy, available on the AMS Web site located at (http://www.ametsoc.org/) or from the AMS at 617-227-2425 or [email protected]. This Article is brought to you for free and open access by the College of Aviation at Scholarly Commons. It has been accepted for inclusion in Applied Aviation Sciences - Prescott by an authorized administrator of Scholarly Commons. For more information, please contact [email protected]. 3110 MONTHLY WEATHER REVIEW VOLUME133 Modification of Precipitation by Coastal Orography in Storms Crossing Northern California CURTIS N. JAMES DepartmentofMeteorology,Embry-RiddleAeronauticalUniversity,Prescott,Arizona ROBERT A. HOUZE JR. DepartmentofAtmosphericSciences,UniversityofWashington,Seattle,Washington (Manuscriptreceived4October2004,infinalform14March2005) ABSTRACT Thisstudycompilesandinterpretsthree-dimensionalWeatherSurveillanceRadar-1988Doppler(WSR- 88D) data during a 2.5-yr period and examines the typical orographic effects on precipitation mainly associatedwithwinterstormspassingovercoastalnorthernCalifornia. Thethree-dimensionalmeanreflectivitypatternsshowechostructurethatwasgenerallystratiformfrom over the ocean to inland over the mountains. The flow above the 1-km level was strong enough to be unblockedbytheterrain,andthemeanechopatternoverlandhadcertaincharacteristicsnormallyasso- ciatedwithanunblockedcross-barrierflow,bothonthebroadscaleofthewindwardslopesofthecoastal mountainsandonthescaleofindividualpeaksoftheterrainonthewindwardside.Upward-slopingecho contoursonthescaleoftheoverallregionofcoastalmountainsindicatedbroadscaleupslopeorographic enhancement.Onasmallerscale,themeanstratiformechopatternoverthemountainscontainedastrong embeddedcoreofmaximumreflectivityoverthefirstmajorpeakofterrainencounteredbytheunblocked flowandasecondaryechocoreoverthesecondmajorriseofthecoastalmountainterrain. Offshore,upstreamofthecoastalmountains,thereflectivitypatternshowedaregionofenhancedmainly stratiformechowithin(cid:1)100kmofthecoast,withanembeddedechocore,similartothoseovertheinland mountainpeaks,alongitsleadingedge.Itissuggestedthattheoffshoreenhancementiscausedbyinten- sifiedfrontogenesisintheoffshorecoastalzoneand/orbytheonshoredirectedlow-levelflowrisingover athinlayerofcool,stableairdammedagainstthecoastalmountains. Theorographicallyenhancedprecipitationoffshoreandoverthecoastalmountainswaspresenttosome degreeinallthelandfallingstorms.However,thedegreetowhicheachfeaturewaspresentvaried.Allthe featuresweremorepronouncedwhenthe500–700-hPaflowwasstrong,themidlevelhumiditywashigh,and the low-level cross-barrier wind component was strong. When the low-level stability was greater, the offshoreenhancementofprecipitationwasproportionatelyincreased,andthegeneralbroadscaleenhance- mentinlandwasreduced. 1. Introduction (e.g., Nagle and Serebreny 1962; Elliott and Hovind 1964;Hobbsetal.1975,1980;Houzeetal.1976;Braun CoastalnorthernCaliforniaisanideallaboratoryfor etal.1997;Doyle1997;Colleetal.1999,2002;Yuand observingstabletoweaklyunstableorographicprecipi- Smull2000;Neimanetal.2004;Ralphetal.2004).The tation. Deep convection is rare along the West Coast, terrain in the region contains quasi-two-dimensional andheavyprecipitationisusuallyassociatedwithland- mountain ridges that are analogous to traditional ide- fallingbaroclinicsystemsthatdirectstrong,moistlow- alizedmodelstudiesofflowoverterrain(e.g.,Queney level flow against the terrain from the Pacific Ocean 1948).Theridgesareorientedfromnorth-northwestto south-southeast and are approximately orthogonal to theprevailinglow-levelflowduringheavy-precipitation Correspondingauthoraddress:Prof.C.N.James,Dept.ofMe- events(Fig.1).TheSouthForkMountainandtheKing teorology, Embry-Riddle Aeronautical University, 3700 Willow Range ridges will be prominent in the subsequent dis- CreekRoad,Prescott,AZ86301–3720. E-mail:[email protected] cussions of this paper. ©2005AmericanMeteorologicalSociety MWR3019 NOVEMBER2005 JAMES AND HOUZE 3111 quasi-perpendicular to the two-dimensional ridges of the coastal mountains of northern California and thus inanoptimalorientationfororographicmodificationof theprecipitationassociatedwiththepassingbaroclinic system. The modification of the precipitation that occurs as the LLJ impinges on a mountain range is a multiscale process. The flow pattern in the baroclinic system is modified on the mesoscale, depending on the stratifi- cation of the temperature, moisture, and wind stratifi- cation of the impinging flow. Over the Alps, Houze et al. (2001, hereafter HJM) found that the basic pattern ofprecipitationenhancementdependedonwhetherthe upstream flow was blocked or unblocked. In the blockedcases,theenhancementbeganupstreamofthe barrier (as suggested, e.g., by Grossman and Durran 1984). In the unblocked cases, the enhancement oc- curredalmostentirelyoverthelowerslopesofthebar- FIG.1.DigitalterrainmapofcoastalnorthernCaliforniawith rier with little activity upstream. Major Alpine floods terrain elevation shaded. The locations of important geographic occurprimarilywiththeunblockedcases,inwhichthe features and the Eureka WSR-88D are labeled. Black-filled circlesoverlandrepresentautomatedraingaugestations;black near-surface boundary layer air rises directly over the dotsoveroceandepictthelocationsofEtaModelgridpoints.The barrier, with little tendency to turn parallel to the bar- fourpointsenclosedbytheellipsewerehorizontallyaveragedto rier or flow around individual peaks. The unblocked produceasyntheticsoundingrepresentativeoftheupstreamflow. cases are marked by rapid growth and fallout of pre- cipitation over the lower windward slopes and espe- ciallyoverthefirststeepriseoftheterrain(Buzzietal. Whenstrong,moist,low-levelwindinteractswiththe 1998; Doswell et al. 1998; HJM; Yuter and Houze California coastal orography, heavy rainfall accumula- 2003). In the combined case where lower-level flow is tion (up to 250 mm in a day) can occur. The rain may blockedandanupperlayerisunblocked,itisalsoevi- combine with rapid snowmelt and produce extreme dent from the Mesoscale Alpine Program (MAP) that flooding of local rivers and streams. During one flood orographic precipitation enhancement can occur both inDecember1964,theEelRiver(Fig.1)roseapproxi- upstreamfromandoveramountainbarrier[theinten- mately 25 m above flood stage and reached a peak discharge rate of 21 300 m3s(cid:2)1 (Sommerfield et al. sive observation period 8 (IOP8) case studied by Me- dinaandHouze(2003,2005)andMedinaetal.(2005)]. 2002). These extreme events can occur when the low- Inthispaper,wesuggestananalogousbehavioroccurs level jet (LLJ) required by hydrostatic and semigeo- overcoastalnorthernCalifornia,wheretheflowabove strophicbalanceaheadofafrontconveysmoistureand 1kmisgenerallystrongandlargelyunblockedbutmay heat into the region of the mountains for a prolonged periodoftime,sometimesforseveraldays(Ralphetal. have a lower layer of blocked flow also affecting the 2004, 2005). Floods occur in the European Alps for precipitation processes. muchthesamereason,asthelow-levelmoistjetahead Small-scale processes also participate in orographic of a front impinges on the barrier (Buzzi et al. 1998; precipitation enhancement. Smith (1979) pointed out Doswell et al. 1998; Rotunno and Ferretti 2001). thelikelyimportanceofconvective-scalecellularityfor Browning (1986) emphasized the strong poleward enhancing orographic precipitation particle growth by sensibleandlatentheatfluxintheLLJtotheeastofan coalescence or riming. Elliott and Hovind (1964) sug- approaching cyclone by calling it a “warm conveyor gestedthatprecipitationassociatedwithfrontspassing belt.” LLJs approaching the California coastline origi- over California indeed manifested embedded small nateinthenorthernfringesoftheTropicsandarecor- convectioncells,whichtheythoughtenhancedthetotal respondinglyquitemoist.Ralphetal.(2004)referredto rainfall.AlsoexaminingCaliforniaprecipitation,White the narrow belt of moisture flux within a warm con- etal.(2003)suggestedthatsmallconvectivecellsmight veyor belt as an “atmospheric river.” When fronts ap- enhance the rainout. They also suggested, on the mi- proachthenorthernCaliforniacoast,thisnarrowzone crophysicalscale,thatthecoalescenceofliquiddropsin of moist flow is nearly always southwesterly, oriented thecloudlayerbelowthe0°Clevel(sometimesreferred 3112 MONTHLY WEATHER REVIEW VOLUME133 to as “warm rain”) is an important microphysical tivity)andairmotions(indicatedbyradialvelocity).By mechanism producing precipitation fallout in storms furtherrelatingtheradarreflectivityandvelocityfields passingoverthecoastalmountainsofCalifornia.White to the details of the topography of the terrain, we fur- etal.(2003)alsoshowed,aswillbeshowninthispaper, therinferhowtheorographymodifiesthemicrophysics thatmajorprecipitationsystemsaccountingformostof and dynamics. the rain in the northern California coastal region ex- In section 2 of this paper, we describe the data and tendwellabovethe0°Clevelandexhibitapronounced methodsusedinthisstudy.Section3presentsthelarge- bright band, signaling that the ice phase is also impor- scale setting of major rain events over the northern tant.Browningetal.(1975)suggestedthatbothliquid- Californiacoastalregion.Section4describestheover- water coalescence and ice-phase processes contribute all average three-dimensional radar echo climatology. to high precipitation efficiencies in orographic precipi- Section5explainshowwestratifythedatasetinto“ep- tation. Yuter and Houze (2003) found that small cells ochs” defined by combinations of environmental vari- embeddedinorographicprecipitationinbaroclinicsys- ables. Section 6 discusses the variability of the radar tems over the European Alps were important in en- echoclimatologywithrespecttomidlevelflow.Section hancing the growth of precipitation particles both 7 analyzes the orographic precipitation processes with aboveandbelowthe0°Clevel.Theratesofcoalescence respect to the wind velocity and stability of the low- belowandrimingabovethe0°Clevelwerecomparable level flow impinging on the mountains. Section 8 inte- within small-scale embedded updrafts. Medina and grates all the results. Houze (2005) have found that strong turbulent over- turning in the shear layer separating a lower layer of 2. Data and methods retarded or blocked flow from an upper layer of un- blocked flow may enhance the growth and fallout of The basic dataset for this study is a two-and-a-half- precipitationparticlesonthewindwardslopes.Larger, yeararchiveofmajorprecipitationeventsfromtheEu- deepconvectivecellscanoverwhelminglydominatethe reka,California,WSR-88D.Thiscoastalradar,located precipitationoverthewindwardslopeiftheflowovera nearCapeMendocino,hasarelativelyunimpededview mountainbarrierissufficientlyunstable(e.g.,Caracena ofprecipitationoverthemountainstoitseastandover etal.1979);however,thenorthernCaliforniaregionin theoceantoitswest.Thislocation(Fig.1)allowssam- winter is not a favored location for deep convection, pling of precipitation systems as they approach the and the weak instability leads primarily to embedded coast, make landfall, and move over the mountains. shallowconvectioninabasicallystratiformcloudlayer. Terrainclutterandshadowingarelessofaproblemfor In this study, we seek insight into the multiscale the Eureka WSR-88D than other West Coast WSR- physicalprocessesinvolvedinorographicprecipitation 88Dsites(Westricketal.1999).ArchivedlevelIIdata enhancement over the coastal mountains of northern of reflectivity and radial velocity (Crum and Alberty Californiaandhowthenatureoftheenhancementvar- 1993)wereobtainedformostoftheheavy-precipitation ies with the strength, stability, and layering of the up- days during 1 October 1995–31 March 1998, a period stream flow. Achieving this objective will help deter- that included the California Landfalling Jets Experi- mine whether processes occurring over this mountain- ment (CALJET; Ralph et al. 1999). ous region are similar to or different from processes Amajorprecipitationeventwasdefinedasadayon over the Alps and other major mountain ranges. We which at least 25% of the 73 automated rain gauges in use the Weather Surveillance Radar-1988 Doppler the region bounded by 39°N, 42°N, 122°W, and the (WSR-88D) operational radar data collected over a Californiacoastlinerecorded25mm(1in.)ormoreof two-and-a-half-year period at Eureka, California (Fig. precipitation.Theblack-filledcirclesinFig.1showthe 1).Thisradarcoversprecipitationbothovertheocean locations of the gauges. Radar archives were available and as it crosses the windward slopes of the mountain for 61 of the 67 heavy-precipitation days identified barrier. This is not a study of rain measurement by (Table 1). The basic unit of radar data was the three- radar. Rather, we aim to understand how the coastal dimensional volume scanned by the WSR-88D eleva- mountainsaffectthephysicalprocessesofprecipitation tion angle sequence. To reduce autocorrelation and growthandfallout.Toaccomplishthisaim,wecompile minimize data storage requirements, we reduced the time-mean three-dimensional patterns of radar reflec- timeresolutionoftheradardatabyusingonlythedata tivityandradialvelocity.Byexaminingsimultaneously volumes obtained at 1-h time intervals. the detailed mean spatial structure of both the reflec- Since the radar processor’s clutter suppression algo- tivityandradialvelocityfields,wededuceaspectsofthe rithmwasinsufficienttoremoveallterraincontamina- interplay of microphysics (indicated grossly by reflec- tion, we developed a digital terrain mask. The terrain NOVEMBER2005 JAMES AND HOUZE 3113 TABLE1.Heavy-raindaysidentifiedduring1Oct1995–31Mar1998whenWSR-88Darchiveswereavailable. 1995 1996 1997 1998 11,12,14,15Dec 15,16,18,27Jan 1Jan 2,3,11,12,14,16,18,25,26Jan 4,17,18,19,20Feb 16Mar 1,2,3,5,6,7,14,16,19,21Feb 4Mar 18Apr 12,21,22,23Mar 21May 3Jun 17,18,19Nov 8Oct 4,7,8,9,10,26,29,30,31Dec 15,16,26,29Nov 7,14Dec mask used an equivalent-earth-radius ray-propagation Meanreflectivityandradialvelocitywerecomputed modeltoapproximatethealtitudeoftheradar’slowest using the interpolated radar volumes. For reflectivity tilt (0.5°) at each radar gate (Doviak and Zrnic 1993, averaging,amissingvalueindicatesabsenceofprecipi- 14–23). The equivalent earth radius corresponding to tation. Therefore, the reflectivity factor (mm6 m(cid:2)3) at the strongest vertical refractivity gradient of all 61 the corresponding grid point was set to zero. Missing heavy-precipitation days was used to give a liberal es- radial velocity information, on the other hand, merely timate of the amount of beam refraction. Then, a ter- indicatestheabsenceofscatterers.Therefore,themean rain elevation dataset with 30-s spatial resolution was reflectivity factor Z and radial velocity V at each grid used to determine whether any terrain intersected the point were computed as bottomoftheradarbeamwithinahorizontallatitude– longitude element of dimensions 1(cid:3) (cid:4) 1(cid:3). If the main (cid:6)N Z lobe of the radar (width 0.94°) was intersected by ter- i rain,thenthatrangebinandallbinsatfartherrangein Z(cid:5)i(cid:5)1 (cid:7)1(cid:8) N thatradialwereremoved.Toreducesidelobecontami- nation,ifterrainwaslocatedwithin0.5°ofthebottom and ofthemainlobe,thenthecorrespondingradarbinwas deleted. This technique removed virtually all terrain (cid:6)n V clutterandshadowingfromthedatasetandallowedall i remaining radar reflectivity bins over terrain to be in- V(cid:5)i(cid:5)1 , (cid:7)2(cid:8) n terpreted as precipitation rather than clutter contami- nation. To remove noise, radial velocity data were removed if their corresponding reflectivity values were deleted orbelowathresholdof(cid:2)10.0dBZ.Theradialvelocity data were then dealiased using a University of Wash- ington algorithm similar in construction to the WSR- 88Dalgorithm(EiltsandSmith1990).Asmallfraction of the volumes that were not successfully dealiased by thealgorithmwererejected,leavingatotalof1176for analysis.Thevolumeswerebilinearlyinterpolatedtoa three-dimensional Cartesian grid with 2-km horizontal spacing and 0.5-km vertical spacing using the National Center for Atmospheric Research’s (NCAR) SPRINT software (Mohr and Vaughan 1979) and finally con- verted to Unidata’s Network Common Data Format (NetCDF)foranalysisusingMountainZebra(Jameset al.2000),whichisaversionofNCAR’sZebrasoftware (Corbetetal.1994)inwhichthedetailedterrainfieldis included.Theinterpolationgrid,superposedwithaver- FIG.2.HeightvsrangerepresentationoftheEurekaWSR-88D tilt sequence looking east from the radar site (indicated by the tical cross section of the most commonly used WSR- opencircleat767-maltitudeand0-kmrange).Eachradartiltis 88D volume coverage pattern (VCP 21), is shown in shaded, and interpolation grid points are indicated by “(cid:9).” Fig. 2. (AdaptedfromJamesetal.2000.) 3114 MONTHLY WEATHER REVIEW VOLUME133 whereNisthetotalnumberofvolumes((cid:5)1176),andn ((cid:1)N)isthenumberofvolumeswithnonmissingradial velocity values. From (1), it is evident that Z maxima indicate regions where the precipitation was either morefrequent,morereflective,orboth.Higherreflec- tivity is generally correlated with heavier rainfall. To reduce unwanted noise and radar artifacts, inverse range-squared horizontal smoothing was applied at eachinterpolationgridpointwithina16-kmhorizontal radius of influence for all horizontal maps (6-km hori- zontal radius for vertical cross sections). To investigate the sensitivity of rainfall to the up- FIG. 3. The interdependence of squared moist Brunt–Väisälä stream flow and stability, superposed epoch analyses frequencyandwinddirectioninthe900–800-hPalayerovervir- tuallyalloftheEta-derivedupstreamsoundings.NegativeBrunt– (e.g., Reed and Recker 1971; HJM), or “composites,” Väisäläfrequenciesdenoteconditionalinstabilityinthe900–800- were constructed by dividing the radar volumes into hPalayer. subsetsorepochsdefinedbysomespecificwindorther- modynamic condition. Then, the mean and standard deviationwerecomputedateachgridpointinthesub- Väisälä frequency was small in magnitude (i.e., less setofvolumes.Thestatisticalsignificanceofthesesub- than 10(cid:2)4 s(cid:2)2), indicating that deviation from moist set means were indicated by two-sided Student’s t dif- neutrality was minimal, with some days being slightly ference-of-means tests that were performed at each moist-unstable and others slightly moist-stable. Fur- grid point using an a priori confidence level of 95% thermore, the 900–800-hPa stability was generally un- (Spiegel 1972). As in HJM, the sample sizes in this correlated with wind direction, although northwest study and the low autocorrelation between successive flows were seldom absolutely stable. radarscansbothappeartobeadequatefordifference- of-means tests (e.g., Wilks 1995). 3. Large-scale flow and stability The superposed epoch analyses led to conclusions about the sensitivity of the precipitation in the vicinity Figures4aand4bshowthemeanlarge-scalesynoptic ofEurekatodynamicandthermodynamicvariablesof conditions indicated by 12-hourly NCEP global model the large-scale flow offshore. We estimate these up- reanalysisoutputat2.5°(cid:4)2.5°resolutionforthemajor streamvariablesusinganalysesand6-hforecastsfroma rainevents(Kalnayetal.1996).Thismapshowsthaton 90 km (cid:4) 90 km (cid:4) 50-hPa resolution National Centers average the 1000- and 500-hPa flows were generally forEnvironmentalPrediction(NCEP)EtaModelgrid. southwesterly,perpendiculartothemountainrangesof Themodeldataatthefourupstreammodelgridpoints coastalnorthernCalifornia,consistentwithRalphetal. bounded by the ellipse in Fig. 1 were horizontally av- (2004) and other studies mentioned earlier. A baro- eragedtoproducesmoothverticalsoundingsat6-hin- clinictroughwaslocatedoffshorewiththemaximumof tervals. If no model grid was available within 3 h of a large-scale upward motion located over the northern radar-volume time, the volume was not used for those Californiacoast.Figure4alsodepictsthemeansynop- calculationsthatrequiredsoundinginformation.Model ticmapsfortwosubsetsofcases.Forreasonsdiscussed gridswereavailablefor1116oftheradarvolumes.The in section 6, we focus on cases for which the 900–800- 700–500-hPa layer in the synthetic soundings was used hPa flow was west-southwesterly, which produces the torepresent“midlevel”characteristicsoftheflow.The greatest orographic effect on the precipitation. 900–800-hPa layer ((cid:1)1–2 km MSL) in the synthetic The west-southwesterly events are subdivided into soundings, corresponding to the LLJ altitude and the days in which the thermodynamic stratification of the strongest correlation with precipitation (Neiman et al. 900–800-hPalayerwasslightlyunstableorneutralver- 2002), was used to estimate the static stability, wind sus absolutely stable. The large-scale mean synoptic speed, wind direction, and dewpoint temperature up- patterns for the unstable/neutral cases (Figs. 4c,d) and wind. The static stability was represented by the moist stable cases (Figs. 4e,f) were qualitatively very similar, Brunt–Väisälä frequency (Durran and Klemp 1982) except that the stable events exhibited slightly weaker and was computed using finite differences. southwesterly gradient wind. The midlevel height pat- Figure 3 summarizes the static stability estimates in terns (Figs. 4d,f) were also very similar, with compa- the 900–800-hPa layer of onshore flow, as computed rable 700-hPa upward motion between unstable and fromtheEtaModeloutput.ThesquaredmoistBrunt– unstable events. The 700-hPa vertical motion patterns NOVEMBER2005 JAMES AND HOUZE 3115 FIG.4.MeanNCEPreanalysisfields,averagedover60heavy-precipitationeventsduring1Oct1995–31Mar1998.(a)MSLpressure, 1000–500-hPathickness,andisotachsofgradientwind;(b)500-hPaheightand(cid:10);(c)and(d)compositesynopticpatternsforthesubset ofheavy-raineventsthathadunstableorneutralwest-southwesterlyflowinthe900–800-hPalayer,withthe0°Clevelabove2.5km MSL;and(e)and(f)sameas(c)and(d),respectively,exceptthatonlystableeventsareshown. are both elongated in Figs. 4d and 4f, with apparent at 50-hPa vertical resolution derived from the Eta frontal orientation from south-southwest to north- Modelanalysisand6-hforecastfields(section2)forthe northeast.Inbothcases,theCaliforniacoastappeared 61 heavy-precipitation days. The average 0°C level is to be located in prefrontal flow. around 780 hPa (2.1 km MSL), and the (cid:2)15°C level is Figure5ashowstheoverallmeanupstreamsounding about560hPa((cid:1)4kmMSL).Thedewpointdepression 3116 MONTHLY WEATHER REVIEW VOLUME133 intheprofilegraduallyincreasesfromabout2°Catthe surface to 8°C above 4 km MSL, indicating that midlevel cloud was less frequent than at low levels. In addition, the temperature profile reveals the preva- lenceofweakconditionalinstabilitybelowthe850-hPa level, combined with abundant near-surface moisture. Within the layer of conditional instability below 850 hPa, the wind profile veered with height, indicative of warmadvectionand/orEkmanturning,withanaverage 900–800-hPa wind speed of 30–35 kt ((cid:1)15–18 ms(cid:2)1) fromthewest-southwest.ThemeanEta-derivedprofile inFig.5aisconsistentwiththepointdropsondeobser- vations of Ralph et al. (2005) within 17 different land- fallingLLJs,exceptthatthelargerareasamplinginour Eta Model profile depicts slight conditional instability rather than moist neutrality in the 900–800-hPa layer and dry air intrusion at midlevels. Figure 5b shows the average sounding for the un- stable/neutral events, while Fig. 5c shows the average sounding for the stable events. A small but notable difference between the stable and unstable events is that the stable events (Fig. 5c) had higher humidity (indicated by much smaller dewpoint depression val- ues)atmidlevels.Aswillbeshown,theintensityofthe orographicmodificationofthesynoptic-scaleprecipita- tion pattern was sensitive to these slight differences in stability and/or humidity. 4. Radar climatology The horizontal cross sections of echo patterns de- scribed in this section are mostly taken at the 2-km level,whichishighenoughtoavoidmuchofthenear- field blocking of the radar beam by the terrain while being in the rain layer below the 0°C level (Fig. 2). Vertical cross sections incorporate all available grid levels. Figure6ashowsthe2-km-altitudehorizontaldisplay of reflectivity over coastal northern California, aver- agedoverthe61heavy-precipitationdays.Partialbeam shadowingoccurredbehindSouthForkMountain(Fig. 1) and other terrain features, and corresponding radar gates beyond those obstacles to the beam were re- movedfromthedataset.Removaloftheseblockedand partially blocked beams results in the circle of radar observations being much smaller to the east of the ra- ← FIG. 5. Mean upstream soundings (location shown in Fig. 1) ofdayswithwest-southwesterlyflowat900–800-hPaflowand0°C derived from Eta Model output. (a) The average over all 60 levelabove2.5kmMSL.Thetemperatureprofileissolid,andthe heavy-precipitationdays.(b)and(c)Meansoundingsforsubsets dewpointtemperatureisdashed. NOVEMBER2005 JAMES AND HOUZE 3117 FIG. 6. Radar-derived precipitation climatology obtained for all heavy-precipitation events observed by the Eureka WSR-88D, comprisingatotalof1176hourlyradarvolumes.Constantaltitudeplotsof2kmdepict(a)meanreflectivity,(b)therainfallfrequency, orpercentageofradarvolumesinwhichthereflectivitywasatleast13dBZ,and(c)meanDopplerradialvelocity.Negative(positive) radialvelocityindicatesflowtoward(awayfrom)theradar.Rangeringspacingis20km,andazimuthlinesaredrawnevery45°.The whitecontourrepresentsthecoastline.(d)Averticalcross-sectionplotofmeanreflectivityfromsouthwesttonortheastalongthesolid redlinein(a),withtheunderlyingterrainshadedgreen.Thedashedredlinein(a)indicatesthepositionofverticalcrosssectionsin Fig.7. dar than to the west in Fig. 6 and similar figures reflectivityisroughly60kmfromthecoastline,whichis throughout the paper. roughly 150 km from the crest of the Coastal Range. TheoverallpatternofreflectivityinFig.6aindicates Estimates of the Rossby radius (L (cid:5) NH/f) using the both enhancement of echo directly over the coastal representative dry Brunt–Väisälä frequency (N (cid:5) 0.01 mountainsandupstreamofthecoastline.Theenhance- s(cid:2)1), characteristic terrain height (H (cid:5) 1.5 km MSL), mentovertheterrainisevidentfromtheechomaxima andCoriolisparameterf(cid:5)10(cid:2)4s(cid:2)1correspondtothis overtheKingRangeandoverthecrestandwindward offshoredistance,suggestingthatconvergentliftingdue slopes (southwest side) of South Fork Mountain and to geostrophic adjustment in subcloud air may be en- theimmediatelyadjacentterraintoitssouth(locations hancing precipitation upstream, before the southwest- in Fig. 1), while upstream enhancement of the precipi- erly airstream (Fig. 6c) directly encounters the coastal tation processes is evident from the echo pattern over terrain. The Rossby radius estimated using a moist the ocean. Brunt–Väisälä frequency (Durran and Klemp 1982) Theechocontoursoffshoreareorientedroughlypar- characteristic of absolutely stable events (N (cid:5) 0.002 alleltothecoast,withechointensitygenerallyincreas- s(cid:2)1)isonlyabout30km,andwithinthisdistancefrom ing toward shore. The strongest offshore gradient of the barrier the reflectivity was even higher, suggesting Fig 6live4/C 3118 MONTHLY WEATHER REVIEW VOLUME133 additional adjustment within the cloud layer itself, closer to shore. However, these are only rough esti- mates. The Doppler velocity data do not provide con- clusiveevidenceofthesuspectedupstreamflowadjust- ment,andthecomplexityoftheterrainandvariability in the static stability make scale analysis to obtain the appropriate expression of the Rossby radius problem- atic. Aquasi-circularmaximumofreflectivityisapparent offshoreataradarrangeofabout40kminFig.6a.This curved maximum is associated with the brightband ef- fect of melting ice particles. However, the ringed pat- tern also mimics the shape of the coastline, which bulges westward at Cape Mendocino, and the curved reflectivity pattern could therefore also indicate up- stream enhancement of the precipitation. Braun et al. (1997) analyzed aircraft data and found that the up- stream modification of echo patterns appeared as an echo maximum paralleling the Pacific coastline. Figure6ashowsthatmaximaofthemeanreflectivity occurred over specific areas of the terrain, especially overtheKingRangeandSouthForkMountain(Fig.1). Thequestionariseswhethertheprecipitationwasmore intense or more frequent in these locations. Figure 6b depicts the percentage of radar volumes in which the reflectivity equaled or exceeded 13 dBZ, which is roughly equivalent to a rainfall rate of 0.2 mm h(cid:2)1. FIG.7.Verticalcrosssectionsofmeanradialvelocityduring(a) Overall,thepatternsinFigs.6aand6barequalitatively unstable/neutraland(b)stableevents(analysisV)takenapproxi- similar, suggesting that orographic forcing generally matelyparalleltothewindalongthedashedredlineinFig.6a. Theterrainisshadedgreen,andnegative(positive)radialvelocity makestheprecipitationmorefrequentratherthanmore indicates flow toward (away from) the radar location shown at intense. (An exception to this rule will be discussed in 140-kmdistanceand767-maltitude. section 7b.) Calculations of mean conditional rainfall rate (not shown), which is related to mean echo inten- sity,confirmthisresult,withtheexceptionthatslightly ratedandunsaturatedair,indicatingthatflowblocking higherechointensityoccurredbothoverthehigherter- was not occurring at these levels. Figure 7 shows the rain and offshore within roughly 60 km of the coast. average Doppler radial velocity along a southwest– Figure6cshowstheprevailingDopplervelocityat2 northeast cross section passing through the radar site km MSL, averaged over all 1176 radar volumes. The forunstable/neutralandstableconditions(asdefinedin mean flow was nearly perpendicular to the Coastal section 3). Both sections show strong southwesterly Range from the southwest at speeds approaching 20 flow increasing gradually with height throughout the ms(cid:2)1 at 2 km MSL. Maps of the Doppler velocities at layer observed by the radar. The flow was slightly less otheraltitudesindicatedthatthewindwasveeringwith intenseinthestablecasesbutneverthelessrapidlymov- height,especiallybelow3-km-MSLaltitude,consistent ing over the coast and up over the mountains. These with frictional turning, flow blocking, and/or warm ad- sections are, again, consistent with unblocked flow vection. The strong horizontal low-level wind toward above the 1-km level. the coast in the storms examined in this study ((cid:1)20 The vertical cross section of average reflectivity in ms(cid:2)1) was at least twice the strength of the cross- Fig. 6d, taken parallel to the prevailing southwesterly barrier low-level wind component of (cid:1)8 ms(cid:2)1 ob- wind along the red line in Fig. 6a is consistent with served during heavy-rainfall events in the Alps during unblockedflowoverthecoastalterrain.Amaximumof MAP (see HJM). The mean soundings (Fig. 5) show echo intensity occurred over the first major peak of that the strong flow toward the barrier in the 1–3-km terrain(i.e.,theKingRange,locatedatabout70kmon (900–700 hPa) layer during heavy-precipitation events the horizontal scale in Fig. 6d). This maximum ex- had Froude numbers greater than unity for both satu- tendedtothehigherlevelsasanupwardprotrusionof Fig 7live4/C