UBC Faculty Research and Publications

Interannual variability of accumulated snow in the Columbia basin, British Columbia Hsieh, William W.; Tang, Benyang Dec 22, 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Hsieh_AGU_2001_2000WR900410.pdf [ 932.38kB ]
JSON: 52383-1.0041777.json
JSON-LD: 52383-1.0041777-ld.json
RDF/XML (Pretty): 52383-1.0041777-rdf.xml
RDF/JSON: 52383-1.0041777-rdf.json
Turtle: 52383-1.0041777-turtle.txt
N-Triples: 52383-1.0041777-rdf-ntriples.txt
Original Record: 52383-1.0041777-source.json
Full Text

Full Text

WATER RESOURCES RESEARCH, VOL. 37, NO. 6, PAGES 1753-1759, JUNE 2001  Interannual variability of accumulated snow in the Columbia basin, British Columbia William  W. Hsieh  Departmentof Earth and OceanSciences, Universityof BritishColumbia,Vancouver,BritishColumbia,Canada  BenyangTang Jet PropulsionLaboratory,Pasadena,California  Abstract. Snowwater equivalentanomalies(SWEA) measuredaroundApril 1 by stationsin the Columbia basin area in British Columbia, Canada, were studied for their  interannualvariabilityduringthe period 1950-1999,particularlyin relationto E1 Nifio/La Nifia eventsand to high and low Pacific-NorthAmerican(PNA) atmospheric circulation patterns.Composites of the SWEA showedthat SWEA were negativeduringE1 Nifio years,positiveduringLa Nifia years,negativeduringhighPNA years,andpositiveduring low PNA years.High PNA appearedto havethe mostimpacton the SWEA, followedby La Nifia, E1 Nifio, and low PNA. In the Columbiabasinarea,La Nifia effects(relativeto E1 Nifio effects)on SWEA decreasenorthwardand eastwardbut strengthenwith elevation.Composites of the Pacificseasurfacetemperatureanomalies(SSTA) duringthe 10 lowestSWEA yearsrevealedweak signals,with E1 Nifio warm SSTA presentonly duringspringand earlysummerin the precedingyear and the SSTA patternconsistent with a highPNA presentby fall andwinter.In contrast,composites of the SSTA during the 10 highestSWEA yearsshowedstrongLa Nifia coolSSTA startingaroundMay in the precedingyear and lastingonto winter. 1.  Introduction  In westernCanada,the provinceof British Columbiagenerates90% of its electricityby hydropower,of which45% came from the ColumbiaRiver. The interannualvariabilityof the  Nifio and La Nifia eventsand betweenhigh and low PNA conditions.The outline of this paper is as follows:section2 describesthe data used, section3 describesthe influence of E1 Nifio and La Nifia events, and section4 describesthe influence  of high and low PNA on accumulatedsnowin the Columbia basin.Section5 presentsthe sea surfacetemperature(SST) supplyandhencethe hydropowerfrom thisriver.The purpose composites duringhigh/lowsnowyears. of this paper is to examinethe interannualvariabilityof the accumulated  snow in the Columbia  basin affects the water  accumulated snow in the Columbia basin, in relation to E1  Nifio-SouthernOscillation(ENSO) eventsand to the PacificNorth American(PNA) atmosphericcirculationpattern. For the westernUnited States,relationsbetweentemperature, precipitation,and streamflowand ENSO and PNA have beenfound [Redmondand Koch,1991].Impactsof ENSO on the Canadiantemperature[Shabbar andKhandekar,1996]and precipitation[Shahbaret al., 1997] have also been studied. Significantcorrelationswere found betweenNorth American  2.  Data  The ENSO indexusedwas the Nifio 3.4 region (5øN-5øS, 170ø-120øW) SSTanomalies(SSTA),ascomputedby National Oceanic and AtmosphericAdministration(NOAA). From 1950to 1999,NOAA monthlySST fields[Reynolds and Smith, 1994; Smith et al., 1996] were obtainedfor the Pacific.The original2øby 2øresolutionSSTdatawere combinedinto 4øby 4øgriddeddata. The NOAA PNA indexwasthe standardized amplitudefrom a rotatedprincipalcomponentanalysisof the winter snow cover and the PNA and North Atlantic Oscillation and Livezey,1987]. (NAO) indices[Gutzlerand Rosen,1992].In a studyof the 700 mbar heightanomalies[Barnston In the Columbiabasin,20 stationswith longrecordsof snow Canadian snow cover the west coast winter snow cover was foundto be muchmorestronglyrelatedto PNA thanto ENSO water equivalent(SWE) datacollectedaroundApril 1 of each [Brownand Goodison,1996].Interannualvariabilityof snow- year were selected(Figure 1 and Table 1). Three other highpack in the westernUnited Stateswas alsofound to be low in elevation stationsin the vicinity (Yellowhead, McGillivray Pass,and Park Mountain) were addedto supplementthe relwinterswith high PNA index [Cayan,1996]. In manypreviousstudies,linearmeasures(e.g.,correlation) ativelyfew high-elevationstationswith long records.All stawereusedwhichdid not distinguish betweenENSO warm(El tions in Table 1 have recordsdating back to at least 1950, Nifio) and cold (La Nifia) eventsnor betweenhigh PNA exceptfor Mount Abbot, Koch Creek, WhiterocksMountain, (strongAleutianLow) and low PNA (weak AleutianLow) McGillivray Pass,and Park Mountain. The record for Revelconditionsin the atmosphere.In thisstudy,we alsousedcom- stoke stoppedafter 1995, while thosefor Mount Revelstoke posites,which tend to bring out the asymmetrybetweenE1 and Park Mountain alsostoppedafter snowpillowswere set up. Snowpillow data were used for Mount Revelstokefrom Copyright2001 by the AmericanGeophysical Union. 1998to 1999.For Park Mountainthe snowpillow data (scaled by 1.119for better matchwith the old stationdata) were used Paper number 2000WR900410. 0043-1397/01/2000WR900410509.00  for 1996-1999.  1753  1754  HSIEHAND TANG:INTERANNUALVARIABILITYOF ACCUMULATED SNOW  53'N  (a)  52'N  51'N  195•)'''' .... 1960 i .... , .... 1970 , .... j .... 1980 , .... • .... 1990 , .... , .... 2000 (b)  3 I• .--e--Nino3.4 SSTA I ..... 50'N  •  ll---.a--PNA '•'/'  •  •  .....  • o ?-  49'N 123 øW  , ....  '• -2  Figure 1. Map of the BritishColumbiasnowstationsusedin thisstudy.Theopensquares indicatethelow-elevation stations  . •-•  \i  , .........................  ß  •._:...... . .....  :" '"  •  •.  ........'-• .........................................................  -3  1950  -i960  1970  1980  1990  2000  Year  (elevation <1000m), andtheopencircles indicate themediColumbia basin snow waterequivum-elevation stations (1000m -< elevation < 1700m). The Figure2. (a)Theaverage solidtrianglesand diamonds denotehigh-elevation stations alentanomalies (SWEA)atApril1 of eachyearand(b) the (1700m < elevation), withthetriangles lyingwithintheCo- winter Nino3.4seasurface temperature anomaly (SSTA)(sollumbiabasin(bounded bysolidcurve)andthediamonds out- idcurve) andthewinterPacific-North American (PNA)index sidethebasin. Thenumber beside thesymbol gives thestation (dashedcurve),wherewinteris definedso that winter 1999  number, as listed in Table 1.  meansDecember1998to February1999.All variableshave  beennormalized bythestandard deviation before plotting. Therewereveryfewmissing data,whichwereinterpolated using values fromadjacent months (e.g.,March1 or May1). averageColumbia basinSWEA(Figure2a) for the period Thestations, withtheyearsof interpolated datagivenin pa- 1950-1999. Integrating theautocorrelation function yielded an renthesis, are Field (1995),MountRevelstoke (1955and integraltimescale of 1.62years;hencetheeffective numberof 1992),Kicking Horse(1995),Whiterocks Mountain (1954and degrees offreedom is31.Forcomparison, thewinter(Decem1957),Yellowhead (1952and1957),McGillivray Pass (1955), ber-February) Nifio 3.4 (SSTA)indexand the winterPNA andParkMountain(1981).  indexare shownin Figure2b. Boththe SWEA andthe PNA  Thestandardized SWEanomalies (SWEA)fromthe20sta- display noticeable decadal-interdecadal-scale variability inFigtionsintheColumbia basin wereaveraged together toyieldan ure 2. The generaldeclinein SWEA from 1950 to 1963coinTable 1.  List of Stations Station  Elevation, Latitude, Longitude,  Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23  Name  m  øN  øW  Canoe River Glacier Field Mt.Revelstoke Kicking Horse Mt.Abbot Revelstoke KochCreek Sinclair Pass Sullivan Mine Marble Canyon UpperElkRiver Ferguson Nelson GrayCreek (lower) Monashee Pass TroutCreek Summerland Reservoir McCulloch Whiterocks Mountain Yellowhead McGillivray Pass ParkMountain  910 1250 1280 1830 1650 1980 560 1860 1370 1550 1520 1340 880 930 1550 1370 1430 1280 1280 1830 1860 1800 1890  52047 ' 51o15 ' 51o23 ' 51o02 ' 51o26 ' 51o42 ' 50o59 ' 49o43 ' 50o40 ' 49o43 ' 51o12 ' 49o59 ' 50o41 ' 49o25 ' 49o37 ' 50o05 ' 49o44 ' 49o49 ' 49o47 ' 50o01 ' 52o54 ' 50o41 ' 50o27 '  119017 ' 117o30 ' 116o31 ' 118o09 ' 116o21 ' 117o30 ' 118o13 ' 117o59 ' 115o58 ' 116o01 ' 116o08 ' 114o55 ' 117o29 ' 117o14 ' 116o41 ' 118o30 ' 120o11 ' 120o01 ' 119012 ' 119o45 ' 118o33 ' 122o36 ' 118o37 '  Years  Mean a SDa Nifio a Nifia a Hi pa Lopa  1950-1999 120 1950-1999 730 1950-1999 149 1950-1999 1227 1950-1999 362 1959-1999 1248 1950-1995 235 1959-1999 759 1950-1999 137 1950-1999 344 1950-1999 358 1950-1999 124 1950-1999 590 1950-1999 400 1950-1999 477 1950-1999 345 1950-1999 181 1950-1999 228 1950-1999 162 1953-1999 571 1950-1999 509 1952-1999 639 1957-1999 912  63 -43 144 -73 46 -31 212 -100 70 -40 264 -122 110 -60 158 -40 53 -27 93 -50 89 -48 68 -30 122 -46 100 -31 102 -66 69 -38 68 -37 70 -29 50 - 19 164 -35 123 -51 170 -52 164 -83  48 -65 45 -96 8 -29 153 -145 22 -42 172 -176 58 -92 108 -107 14 -42 56 -68 54 -68 46 -61 48 -93 65 -95 66 -94 16 -60 53 -56 49 -58 36 -51 117 -99 44 -69 118 -53 132 - 114  42 46 10 57 22 100 88 67 16 60 26 45 52 72 52 30 35 32 27 83 -2 21 75  aTheApril1snow water equivalent (in10-3m),standard deviation, andcomposites during E1Nifio, LaNifia, high Pacific-North America  (PNA),andlowPNAyears,respectively.  HSIEH  AND  TANG:  INTERANNUAL  VARIABILITY  OF ACCUMULATED  SNOW  1755  cidedwith a general increasein the PNA during this period, the enhanced SWEA  from 1965 to 1976 coincided with rather  low PNA, and the low SWEA from 1977 to 1987 coincidedwith  high PNA. From the linearly detrendeddata, the Spearman rank correlation was -0.41 between SWEA and Nifio 3.4, -0.57 between SWEA and PNA, and 0.40 between Nifio 3.4  and PNA. With 31 effectivedegreesof freedom,the 1% significance level for the rank correlation is at 0.42, and the 5%  significance level is at 0.30 [van Starchand Zwiers,1999,AppendixK]. Hence thesethree rank correlationvaluesare all significantat the 5% level, and that betweenSWEA and PNA is significantat the 1% level. For comparison,the Pearson correlation(whichis lessrobustthan the Spearmanrank correlation) is -0.44 betweenSWEA and Nifio 3.4, -0.57 between SWEA and PNA, and 0.43 between Ni•o 3.4 and PNA.  For our compositestudieslater, we need to designatesome winters  as E1 Ni•o  winters  and others as La Ni•a  winters.  Using ñ0.75øC in the winter Nifio 3.4 index as the threshold for definingE1 Ni•o/La Nifia winters,we obtained13 E1 Ni•o winters(1958, 1964, 1966, 1969,1970,1973, 1977, 1983, 1987, 1988, 1992, 1995, and 1998) and 10 La Nifia winters (1950, 1955, 1956, 1971, 1974, 1976, 1985, 1989, 1996, and 1999) during the period of 1950-1999, where for brevity,we used 1958 to denote the 1957-1958 winter. Similarly, using ñ0.5 standarddeviationasthe thresholdfor defininghigh/lowPNA winters from the linearly detrended winter PNA index, we obtained13 high PNA winters(1958, 1960, 1961, 1963, 1970, 1977,1978,1981,1983,1986,1987,1992,and 1998)and 14 low PNA winters(1950, 1952,1956,1965,1966,1969,1971, 1972, 1979, 1982, 1988, 1989, 1997, and 1999). High PNA winters often occurredduringE1 Nifio winters,as the list of 13 high PNA  winters  and the list of 13 E1 Ni•o  winters  have seven  a  o  .z_ ..................................................... .-•; ...................................... a......... i/.................................................. o----, ......... •: 1 ............................................................................................... [3........... .•-............ o................................................. o  0  ,,,,,I,,,,,I,  123'W  ,I ....  121:•'  ,I  ,I ....  11•:•'  o  I,,,•,1,,,,  ;17'W  I .....  ;15'W  Longitude  membersin common.Similarly,5 of the 10 La Ni•a winters Fisure 3. (a) The ratio of the compositeSWEA durin• La also occurred in the list of 14 low PNA winters. Ni•a yearsrelativeto the (absolutevalue of the) composite durin• E1Ni•o yearsplottedas a functionof the latitude.(b) The sameratio plottedasa functionof the longitude.Symbols 3. Influence of El Nifio/La Nifia Events denotedifferentelevations,as in Figure 1, with the solidsymFrom the linearly detrendedSWE data at each stationthe bols indicatin• the hi,h-elevation stations. SWEA from the E1 Ni•o yearswere averagedtogether,yielding the compositeSWEA for E1 Nifio years.These and the composites for the La Ni•a yearsare shownin Table 1, where arriveseither as snowor as rain, whichwashesawaythe accufor everystationthe compositeSWEA wasnegativeduringE1 mulated snow. Of the 20 stationsin the Columbia basin, 6 have Ni•os andpositiveduringLa Ni•as; that is,the Columbiabasin a ratio of <1, and 13 have a ratio above 1, meaningthat in tendsto have lessaccumulatedsnowduring E1 Ni•o winters general,the La Ni•a effectisgreaterthanthe E1Nifio effectin the Columbia basin SWE. and more snowduring La Nifia winters. Hoerlinget al. [1997]showedthat La Ni•a effectsare not the This ratio is alsoplotted as a functionof the longitudein exactoppositeof E1Ni•o effects,especiallyoutsidethe tropics. Figure 3b, where the rank correlationbetweenthis ratio and The compositefor La Ni•as dividedby the (absolutevalueof the longitudeis -0.46 (significantat the 5% level, with the the) compositefor E1 Ni•os indicatesthe strengthof the La Pearsoncorrelationat -0.48), indicatingthat the La Ni•a Nifia effect relativeto the E1 Ni•o effect.This ratio is plotted effect declineseastward.Combiningwith the result from Figas a functionof the latitude for the variousstationsin Figure ure 3a, we conclude that for the SWE the La Ni•a effect 3a, where the rank correlation between this ratio and the (relativeto the E1Ni•o effect)decreases northeastward in the latitude is -0.49 (with the Pearsoncorrelationat -0.43); Columbia basin. hence the La Ni•a  effect declines northward.  With 23 stations  the 1% significance levelfor the rank correlationis at 0.49,and 4. Influence of the PNA the 5% levelis at 0.35 [vanStarchand Zwiers,1999,Appendix K]. The ratio is alsoconsiderably higherfor the high-elevation Composites for highandlowPNA yearswerealsocomputed stations than for the medium- and low-elevation stations, (Table1), withthe composite SWEA for highPNA yearsbeing meaningthat the La Nifia effect is enhancedat highereleva- negativefor all stationsandthosefor low PNAs beingpositive tions. While the cause of this effect is not known, one does for all stations(exceptfor Yellowhead,whereit wasessentially for low PNA relativeto the expectthe systemto be particularlysensitivewhen the tem- zero). The ratio of the composite perature is closeto freezing,as found in higher elevations:a (absolutevalueof the) composite for high PNA (Figure4) slightchangein the temperaturecouldmeanthe precipitation showedthat for all stationsthe ratio was below 1, indicating  1756  HSIEH  •  []  AND  TANG:  o  0oo  VARIABILITY  OF ACCUMULATED  SNOW  SSTA picturesby averagingthe monthly SSTA from the 10 yearsof lowestSWEA andfrom the 10yearsof highestSWEA. To testthe statisticalsignificance of the composites, a Monte Carlo experimentwith 1000randomshufflesof the SSTA fields  • 0.6............................. '•-.................. < ILl  INTERANNUAL  []  ß  []  was carried out: For each calendar month, 10 of the 50 SSTA  fields from the years 1950-1999 were randomly selectedand averagedto give a compositefield. From 1000 suchrandom • 0.4 ................................................ • ....... -A------•.............................................................................. composites, the fiftieth largestabsolutevalue (for each grid •: o Z point andfor eachcalendarmonth)waschosenfrom the setof ....O 0.2 ............................................................................................... 1000 compositesto representthe 5% significance level. Figure 5 showsthe evolutionof the compositeSSTA from 03 D ,. March in the precedingyear to February in the low SWEA year, with the areassignificantat the 5% level shaded.Even whenthere are no true signals,one would expect-5% of the .0.249 .... i .... 50, .... i.,,,,51i .... , .... 52I .... I .... 53 total area of the compositemaps to be shadedfrom pure Latitude(deg.) chance.The percentagesof total area shadedfor the various Figure 4. The ratio of the (absolutevalueof the) composite monthsare 4% (Januaryprecedingyear), 13% (February), SWEA duringhighPNA yearsrelativeto the compositeduring 10% (March), 11% (April), 16% (May), 14% (June), 2% low PNA yearsplotted as a functionof the latitude. Symbols denotedifferentelevations,asin Figure 1. The soliddiamond (July), 5% (August), 6% (September),7% (October), 12% (November),10% (December),9% (January),and 17% (Febat the lower right corner denotesYellowhead. •  0  0  o  ruary) (Figure 5 only displaysthe last 12 of thesemonths). Thus there is an absenceof significantsignalsprior to February, precedingyear, and during July to September.During that the low PNA effectwasmuchweakerthan the high PNA Februaryto June,precedingyear,there are significantSSTA in effect.Thisratio decreases with latitude(Figure4), with a rank the westerntropicalPacificand in the easterncentralequatocorrelationof -0.46 betweenthis ratio and the latitude (sig- rial Pacific(the classicalE1 Nifio warm region) (Figures5anificant atthe5%level, with thePearson correlation at-0.53). 5d). The significantSSTA patternsduringNovemberto FebIf Yellowheadis omitted,the rank correlationdropsto -0.39 ruary(Figure5i-51) are relativelystable.The pattern,with cool (significantat the 5% level). This positivecorrelationmay SSTA in the Aleutian Low regionandwarm SSTA off western simplybe dueto the fact that manyof the highPNA yearswere Canada, first arose in October, strengthenedin November, also E1 Nifio years and many low PNA yearswere La Nifia then weakenedgli•htlvin Inn]inn/hofnrointonci%,ing againin years(hencethe fact that La Nifia effects(relativeto E1Nifio February.This SSTA patternis consistent with the highPNA, effects)decreasedwith latitudewouldtranslateinto low PNA which has an intensifiedAleutian Low (generatingmore upeffects(relativeto highPNA effects)decreasing with latitude). welling of cooler subsurfacewaters) and a higher pressure Elevation did not separatethe stationsin Figure 4 as it did in region over westernCanada.(Of course,the PNA signalis Figure 3a. This ratio plotted as a functionof longitude(not manifestedmuch strongerin the 700 mbar height anomalies shown) did not have a significantcorrelationor show any than in the SSTA.) obviouspattern. Figure6 showsthe evolutionof the compositeSSTA for the To comparethe overall E1 Nifio, La Nifia, high PNA, and highSWEA years.The percentages of total areashadedfor the low PNA effectsin the Columbiabasin,we took the average variousmonthsare 7% (January,precedingyear), 11% (FebColumbiabasinSWEA data (Figure 2a) and computedcom- ruary), 9% (March), 19% (April), 37% (May), 39% (June), posites.Normalizedby the standarddeviation,the composites 42% (July),45% (August),52% (September),52% (October), for E1 Nifio, La Nifia, high PNA, and low PNA are -0.63, 52% (November),49% (December),50% (January),and52% +0.69, -1.01, and +0.59, respectively,indicatingthat the ef- (February)(with onlythe last 12 of thesedisplayedin Figure fectsare strongestfor the highPNA, followedby La Nifia, then 6). Thesepercentages are, in general,muchhigherthan those E1 Nifio, and finally low PNA. The asymmetryin the SWEA from the low SWEA years(Figure 5) and are well abovethe between high and low PNA winters was greater than that 5% expectedfrom randomnoise.The shadedareasin Figure between E1 Nifio and La Nifia winters. 6 started in the tropics,and as the monthsprogressed,the signalsbecamestrongerin the higher latitudes.By May, a La Nifia cool SSTA pattern is clearly visible in Figure 6c. In 5. Compositesof Pacific SSTA for High/Low contrastto the E1Nifio patternfoundin Figure5, whichfaded SWEA Years out by July, the La Nifia pattern in Figure 6 continuedto So far, we have formed compositesof the SWEA for E1 strengthenwell into winter. Nifio, La Nifia, high PNA, and low PNA conditions.Now, we turn the problem around and ask, What do the PacificSSTA look like duringyearsof highSWEA andlow SWEA? With the 6. Summary and Conclusion linearlydetrendedaverageColumbiabasinSWEA data, the 10 Compositesof the April 1 SWEA for stationsin the Columyearsof lowestSWEA (startingfrom the lowest)were 1963, bia basin showedthat SWEA were negativeduring E1 Nifio 1970, 1977, 1960, 1993, 1984, 1981, 1992, 1958, and 1987, while years,positiveduringLa Nifia years,negativeduringhighPNA the 10 yearsof highestSWEA (startingwith the highest)were years,andpositiveduringlowPNA years.High PNA appeared 1972, 1997, 1974, 1967, 1999, 1971, 1956, 1975, 1976, and 1982. to havethe mostimpacton the SWEA (-1.01 standarddeviFrom the SSTA (linearly detrended, with climatological ation),followedby La Nifia (+0.69), E1Nifio (-0.63), andlow monthly mean removed) we computedcompositemonthly PNA (+0.59). In the Columbiabasin area, La Nifia effects  HSIEHAND TANG:INTERANNUALVARIABILITYOF ACCUMULATEDSNOW  1757  Composites of SSTAforthe 10 lowestSWEAyears (b)Apr. '  (a) Mar.  (c) May 45N  45N  30N  15N  15N  150W  150E  150E  90W  (d)Jun.  150W  90W  150E  (e)Jul.  150W  90W  (f)Aug.  45N  30N  15N  0  150E  150W  90W  150E  150W  9OW  150E  I'  45N  90W  (i) Nov.  (h) Oct.  (g) Sep.  150W  ' '•, '• ß  30N  15N  150E  150E  -....•:;•,, ...::-,  45N'  45N .•. ••,•  .,  15N• [•'•' 150W  150E  ,, .... ',-/:•,•  '  90W  'F, ......... ß  45N  'i  30N  .• • 15N  15N [  90W  150W  (I) Feb.  / /xU'"'•'•'• •:"•  30N  150E  90W  (k) Jan.  (j) Dec. ,• • • •  150W  150E  150W  90W  150E  150W  90W  Figure 5. Composites ofSSTA forthe10lowest SWEA years withSSTA data fromMarch inthepreceding yeartoFebruary inthelow-SWEA year. Positive anomalies areshown bysolid contours, negative anomalies areshown bydashed contours, andzeroanomalies areshown bythick contours. Contour interval is0.1øC. Shaded regions indicate composite values above the5%significance levelfromtheMonteCarlotest.  eachother,exceptthat (relative toEl Nifioeffects) onSWEAdecrease northward and SWEAyearswouldhaveresembled positive anomalies inonewouldcorrespond tonegative anomwasfoundbeComposites ofthePacific SSTAduring the10lowest SWEA aliesin the other.Instead,majorasymmetry for low SWEA and thosefor high yearsrevealed relatively weaksignals. The El Nifiowarm tweenthe composites SWEA, implying that the relationbetween SSTAandSWEA SSTAwasonlypresent duringspring andearlysummer in the was nonlinear. As the SSTA in the composites for highSWEA preceding year.Byfallandwinter,onlytheSSTAconsistent thanthosefor lowSWEA witha highPNApatternat themiddle-high latitudes couldbe (Figure6) weremuchstronger more discerned. In contrast, composites of the SSTAduringthe 10 (Figure5), highSWEAwerelikelyto be significantly thanlowSWEAbySSTA,thanks toabettersignal highest SWEAyearsshowed strong La NifiacoolSSTAstart- predictable ingaround Mayin thepreceding yearandlasting ontowinter. to noise ratio. Thepresence of a strong La NifiacoolSSTAin thehigh If the relation between SSTA and SWEA were linear, then (Figure6) andtheabsence ofa correspondan SSTAcomposite for lowSWEAyearsandonefor high SWEAcomposites eastwardbut strengthen with elevation.  1758  HSIEH AND TANG: INTERANNUAL VARIABILITY OF ACCUMULATED SNOW  Composites of SSTAforthe 10 highestSWEAyears (a) Mar.  (b) Apr. I •.'(/;x...),•j  v•  •  (c) May '  •,?' ß-,;.',<, '---"' ß .'-"•-•-J' i• '•  45N  / •t  30N  15N  0  150E  150W  ..  90W  150E  (d) Jun.  150W  90w  150E  150W  (e) Jul.  (f) Aug. ß  45N  '  90W  V'l I'  •S•: 3  '  .  x•"'•' ':'•'•'•  30N •  v  30N  o 150E  150W  90W  .• -•:.'(•.• 150E  150W  90W  •5OE  15ow  (h) Oct.  ",.  ...... •:• .......... 9ow  (i) Nov.  "•-•'•!•¾-• '  '  ß  ß  .  • ,v'...•.•½•:%-: ,  30N  30N  • •':•[• '•'  15N  •L-'•-"½•i '½<'•;:-• '"-'-' •'"'½• F. 150E  (j) Dec. ).' ....•.•-'t'--:  150W  90W  (I) Feb.  .... , .  45N "  •:..-  15N  150E  (k) Jan.  :"' '""  /,  90W  ,  ...... '........... • •::• e.:•:.-  45N  150W  '5•.•?•  45N  .,  •....... '•......"• •k  •  o[.,•t• 150E  ;  u....-•'-•'•,-••-••-.•••:..••  •.••  •,• • :•,• .•..........••...._...••••:.•• .••• 1 ß • ...••:.':•;•--...•.--.•;e•-: ;•  150W  90W  150E  150W  90W  F•k /•  ...... '•''""'•"•'••••••••••  yt.....•. -..•2•-.:..•.. ß................... •:•,• •• 150E  150W  90W  Figure6. Composites ofSSTA forthe10highest SWEA years withSSTA datafromMarch inthepreceding yeartoFebruary inthehighSWEAyear.Same shading andcontouring convention asin Figure 5 isused.  ingE1NifiowarmSSTAfromJulyonward(Figures 5e-51)  Acknowledgments. We are gratefulto BritishColumbiaHydro  BrianFast,Stephanie Smith,andWubenLuo,for their suggest thatequatorial La Nifiaconditions mayexertfar more staff,especially inthisproject. W.W.HsiehandB.Tangweresupported by influenceon the Columbia basin SWEA than E1 Nifio condi- assistance  a strategic grantfromtheNaturalSciences andEngineering Research tions. Thissituation is'similar tothatfortheCanadian prairie Council of Canada. wheat yield, which wasalsofound tobemuch moreaffected by  La NifiathanbyE1Nifioevents[Hsiehetal., 1999]. Prese.ntly, reasonably longleadtimeforecast skillshavebeen References attained forE1Nifio/LaNifiaevents butnotforhigh/low PNA Barnston, A. G., andR. E. Livezey, Classification, seasonality and persistence of low,frequency atmospheric circulation patterns, Mon. patterns.Sincethe highPNA hasevenstronger impacton WeatherRev., 115, 1083-1126, 1987. Columbia basinSWEAthanLa Nifia,thismaypresent a limBrown,R. D., andB. E. Goodison, Interannual variability in reconitationto ourabilityto forecast lowSWEAconditions atlong structedCanadiansnowcover,1915-1992,J. Clim.,9, 1299-1318, lead time. 1996.  HSIEH  AND  TANG:  INTERANNUAL  VARIABILITY  Cayan, D. R., Interannualclimate variabilityand snowpackin the western United States,J. Clim., 9, 928-948, 1996.  Gutzler, D., and R. Rosen,Interannualvariabilityof wintertimesnow coveracrossthe Northern Hemisphere,J. Clim., 5, 1441-1447,1992. Hoerling, M.P., A. Kumar, and M. Zhong, E1 Nifio, La Nifia and the nonlinearityof their teleconnections, J. Clim., 10, 1769-1786, 1997. Hsieh, W. W., B. Tang, and E. R. Garnett, Teleconnections between Pacificseasurfacetemperaturesand Canadianprairie wheat yield, Agric.For. Meteorol.,96, 209-217, 1999. Redmond, K. T., and R. W. Koch, Surface climate and streamflow  variabilityin the westernUnited Statesand their relationshipto large-scalecirculationindices,WaterResour.Res., 27, 2381-2399, 1991.  Reynolds,R. W., and T. M. Smith,Improvedglobalseasurfacetemperatureanalyses usingoptimuminterpolation,J. Clim., 7, 929-948,  OF ACCUMULATED  SNOW  Shabbar,A., B. Bonsal,and M. Khandekar,Canadianprecipitation patternsassociated withthe SouthernOscillation, J. Clim.,10, 30163027, 1997.  Smith, T. M., R. W. Reynolds, R. E. Livezey, and D.C. Stokes, Reconstructionof historicalsea surfacetemperaturesusingempirical orthogonalfunctions,J. Clim., 9, 1403-1420, 1996.  von Storch,H., and F. W. Zwiers,Statistical Analysisin ClimateResearch,484 pp., CambridgeUniv. Press,New York, 1999. W. W. Hsieh, Department of Earth and Ocean Sciences,University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z4.  (william@eos.ubc.ca) B. Tang, Jet PropulsionLaboratory,M/S 300-323, Pasadena,CA 91109.  1994.  Shabbar,A., and M. Khandekar, The impact of E1 Nifio-Southern Oscillationon the temperaturefield overCanada,Atmos.Ocean,34, 401-416, 1996.  1759  (ReceivedAugust4, 1999;revisedDecember22, 2000; acceptedDecember22, 2000.)  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items