Weather and Climate (1991) 11: 43-47 4 3

MODELLING PRESENT AND PAST SNOWLINE ALTITUDEAND SNOWFALLS ON THE REMARKABLES

Introduct ion

James R. F. Barr ingerDivision o f Land and Soil Sciences, DSIR

A compu te r m o d e l s imu la t ing snowl ines a n d snowfa l l s f r o m r e a d i l y ava i lab ledaily meteorological data has been used t o obtain estimates o f snowlines ove r thelast 6 0 y e a r s o n t h e Remarkables (Bar r inger, 1 9 8 6 ) . T h e m o d e l , based o nextrapolation o f temperature and precipitation f r o m a nearby l o w alt i tude s i te ( i e ,Queenstown A i r p o r t o r Queenstown), uses da i l y calculations o f snow accumulationand ablat ion t o estimate t h e mountain snowpack o v e r a range o f altitudes. Snowablation i s calculated us ing a s imple degree-day m e l t equation, a n d accumulationis modelled us ing a complex rain/snow threshold calculation t o g i ve sensit ivity t oestimates o f snowline altitude. T h e model re l ies upon t he use o f f ou r temperaturelapse r a t e s f o r ex t rapo la t ing temperatures t o h i g h e r a l t i tudes . Se lec t i on o ftemperature l a p s e r a t e t y p e i s m a d e u s i n g a d i s c r i m i n a n t a n a l y s i s o fmeteorological d a t a f r o m t h e l o w a l t i tude base s i t e . T h i s o f f e r s a substantialimprovement o v e r temperature ext rapolat ion techniques i n o t h e r s tud ies ( e g .Moore & Owens, 1984a) and i s crucial t o the models abi l i ty t o accurately simulatesnowline a l t i tude. De ta i l s o f t h i s model l ing approach a r e out l ined i n Barr inger(1986 and i n press).

The model w a s calibrated us ing photographs o f the da i l y w in te r snowline i n t h estudy area during 1984 and 1985. The model was optimised t o a best f i t o f the dai lywinter snowline da ta f o r each year, a n d t h i s cal ibrat ion tested against snowl inedata f o r the other year. Goodness o f f i t o f modelled t o observed data was variable,but i n a l l cases acceptable.

Scenarios

Because t h e mode l c a n b e used t o estimate snowl ine a n d snowfal ls f r o m l o waltitude d a i l y meteorological records, i t i s possible t o use i t i n conjunction w i t hscenario data ( ie . r ea l data modif ied t o approximate a di fferent cl imate) t o assessthe sensitivity o f snowline altitude and snowfalls t o changes i n climate. The use o fscenarios t o assess t h e impacts o f a cl imate d i f ferent f r o m t h e present does n o tconstitute a predict ion. Instead, scenarios "represent a systematic process t h a tuses ava i l ab le t h e o r y, f a c t s a n d judgments t o e x p l o r e t h e imp l i ca t i ons o fhypothesised conditions" (Lave & Epp le, 1985). Clear ly t he scenarios used i n th i sstudy a re considered t o b e plausible since there i s considerable evidence tha t o u rclimate i s n o t static, and may b e warming (Salinger, 1988), b u t they a re entirelyartificial a n d a re subject t o considerable uncertainty. T h i s i s part icularly t h e casebecause mos t scenarios f o r changing c l imate a r e der ived f r o m Globa l Circulat ionModels (GCMs) which are di ff icul t t o apply a t a regional o r local scale. None o f thescenarios used can be said t o be any more probable than the others.

In t h i s case t he implications f o r snowline alt i tude o f a cl imate scenario (Salinger,pers. comm.) o f cooler temperatures i n the 1850s ( n 1°C cooler than a t present) areconsidered. Cooler temperatures i n the 1930s a re a lso hypothesised, being assumedto b e 0 .5 °C coo le r than present, a n d snowl ine a l t i tude under these intermediateconditions est imated. T h e est imated snowl ines f o r t h e s e scenar ios a r e t h e ncompared w i t h snowlines f o r warm c l imate scenarios.

44 S n o w l i n e Altitude

Snowlines i n t h e 1850s a n d To d a y

By us ing meteorological records f r o m Queenstown A i r p o r t f o r the per iod 1972 t o1981 a n d altering the da i l y temperature data t o match the 1 ° C cooler scenario f o rthe 1 8 5 0 s , t h e m o d e l i nd i ca tes t h a t t h e m e a n w i n t e r s n o w l i n e w o u l d b eapproximately 1330 m . Values ranging f r om 1160 m t o 1420 m are estimated f o rindividual years i n t h i s scenario "decade". These f igures can b e compared w i t h avalue f o r mean w in te r snowline o f 1524 m calculated f o r t h e unaltered 1972 t o1981 period, indicating that snowline altitude may have been lower b y as much as190 m i f the scenario f o r t he 1850s i s correct. W i t h respect t o temperature th i ssuggests that snowline altitude may change b y as much as 190 m per ' C change i nmean w i n t e r temperature.

These f igures c a n b e compared w i t h those obtained f r o m s im i l a r analyses us ingother scenarios f o r c l imate d i f f e r i ng f r o m t h a t be ing experienced today (F igure1).

These estimates o f mean w i n t e r snowl ine a l t i tude unde r a range o f condit ionsprovide u s w i t h t h e fo l low ing values f o r change i n snowl ine al t i tude relat ive t ochange i n m e a n temperature.

1. Te m p e r a t u r e and rainfal l changing a t same rate as f o r the periodsince 1930 (ie. +0.6°C by 2050 wi th +10% precipitation) s n o w l i n e altituderises by E 100 in / "C r ise i n mean winter temperature. Calculated usingmodified 1972-1981 data.

2. + 1 . 5 ° C by 2050, no change i n rainfall (Salinger & Hicks Scenario 1) -snowline alt i tude rises b y 1 1 3 W C rise i n mean winter temperature,calculated using modif ied 1972-1981 data.

3. + 3 . 0 ° C by 2050, no change i n rainfall (Salinger & Hicks Scenario 2 ) -snowline alt i tude rises b y -a 8 6 m/"C rise i n mean winter temperature,calculated using modif ied 1972-1981 data. Th i s f igure i s an underestimatebecause o f the e ffec t o f undefined snowline where t h e snowline alt i tude i sabove the highest po int o n the range and hence defaults t o the maximumaltitude o f 2300 m.

2000

1800 -

1600 -

1400 -

1200 -

10001800 1900

Years

2000

Fig. 1. A comparison o f rates o fmean w i n t e r snow l i neretreat f o r the 1850

3 s c e n a r i o ( D ), the 1930A/ s c e n a r i o ( • ) , a

/ •-• 2 2 0 5 0 scenario( a ) ,/ • • • • a n d a n u m b e rof o ther scenariosfor changed

co c l i m a t e ( ) . Alsoshown a re snowlines

estimated b y dai lyrainfall a n dtemperature d a t a f r o mQueenstown 1930-1985(4).

2100

Snowline Altitude (metres)

Snowline Altitude 4 5

4. A s for 1. but calculated using 1930-85 (ie. 55 years) data to estimate altitudegives the same estimate o f rate o f change but higher snowline altitudes.The analysis in this case is o f poorer quality being based on temperature

and rainfall data only.

These analyses suggest a range o f values from a minimum o f 86 m to a maximumof 1 9 0 m r ise i n snowline altitude p e r degree centigrade r ise i n mean wintertemperature. I t i s notable however, that a l l t h e analyses involving estimates f o rsnowline altitude f o r conditions warmer than a t present suggest snowline altitudewould r ise b y about 1 0 0 m p e r degree centigrade, whi le t h e analysis o f pastclimate scenarios gives t h e higher figure o f 1 9 0 m p e r degree centigrade. Th ismay we l l b e a function o f the temperature lapse rate profiles a t this site which,because temperature inversions a r e common i n winter, leads t o l o w e r averagetemperature lapse rates a t altitudes below 1 2 0 0 m . W i t h lower temperature lapserates a t lower altitudes a small change i n temperature w i l l lead t o a large changein snowline altitude.

Summer a n d W i n t e r S n o w line Al t i tudesThe rates o f snowline retreat relative t o temperature increase indicated b y t h escenario based analysis above can be compared wi th figures derived b y modellingpresent snowlines o n a monthly basis f r o m summer t o w in ter. T h i s a l lows u s t oconsider snowl ine al t i tudes o v e r a range o f temperatures f a r greater t han thoseanticipated b y c l imate change scenarios f o r t he n e x t 5 0 years. T h i s method a lsoincludes a n a t tempt t o account f o r changes i n temperature lapse r a t e t y p e a n dfrequency i n wa rmer condit ions, s ince t h e scenario approach assumes s im i la r i t yof temperature lapse ra te under warmer conditions w i t h those o f present winters,while f o r t he seasonal analysis, lapse prof i les a re calculated o n a seasonal basis,and d isp lay s ign i f i cant differences between seasons.

This analysis d o e s assume a s im i la r i t y o f mean w i n t e r snowlines f o r possiblewarmer ( o r coo le r ) w in te rs w i t h mean mon th l y snowl ines w i t h o u t a l l ow ing f o rthe antecedent e f fec ts o f snowl ine a l t i tude i n prev ious months. T h i s e f f e c t i sclearly seen i n F igure 2 where a hysteresis e f fec t i s evident i n t h e plot ted data,with s p r i n g s n o w l i n e s ( i e . w i n t e r antecedent cond i t i ons ) b e i n g l o w e r t h a nautumn snow l i nes ( i e . s u m m e r antecedent cond i t ions) . F i g u r e 2 s h o w s d a t acalculated f o r the 1972 t o 1981 decade w i th a maximum snowline altitude o f 3500 m(cf. 2300 m f o r the top o f the Remarkables).

The analysis has been repeated us ing a longer t imes series ( ie . 5 0 years 1 9 3 0 -1984), b u t o f precipitation and temperature da ta o n l y f o r Queenstown (F igure 3 ) .The results o f these t w o analyses a re remarkably s imi la r. W i t h 1 0 years o f f u l lmeteorological record f r o m Frankton t h e ra te o f snowline change w i t h respect t omean mon th l y temperature i s calculated t o b e 1 0 1 m r C , a n d f o r t h e 5 0 y e a rQueenstown record 1 0 6 m P C . Both these results are also remarkably s imi lar t o theestimates obta ined f r o m t h e scenar io analyses i n t h e prev ious sec t ion , w h i c hapart f r o m t h e coo le r scenario, gave rates o f snowl ine change ranging f r o m 9 0mi0C to 113 m r C .

Temperature a n d Precipitation - S n o w line a n d SnowfallsWhen analys ing t h e resu l ts o f t h e mode l , i t i s interest ing t o corre late m o d e loutputs o f mean w i n t e r snowl ine a l t i tude a n d m e a n w i n t e r s n o w accumulat ionagainst m e a n w i n t e r temperatures a n d p rec ip i ta t ion , t o s e e w h e t h e r s i m p l erelationships ex i s t between these parameters a n d mode l output . The re i s a g rea tdeal o f var iabi l i ty present i n t h e i npu t a n d ou tpu t data,which means correlationsare n o t a lways good . Nonetheless, t he re i s genera l ly a s ign i f icant re lat ionshipbetween t h e amount o f snow accumulation, par t icu lar ly a t h i ghe r alt i tudes, a n d

46 S n o w l i n e Altitude

(1)

›,E

4000 -

3000‘*ma0 =

2000

1000

Fig. 2. A scatter p l o t o f mean monthly snowline al t i tude and mean monthlytemperature f o r the decade 1972 t o 1981 ( i e . us ing f u l l meteorologicalrecords as model input data) indicates a good correlation betweentemperature and mean monthly snowline al t i tude (r2=0.696). T h eregression l i n e suggests tha t snowline al t i tude changes b y 101 m r C .

(,)

E 3 0 0 0tu

o 1:1

.ot2000

. -3

CA

4000

1000

Fig. 3 A scatter p lo t o f mean monthly snowline altitude and mean monthlytemperature based on model results f o r the period 1930 t o 1985 ( le.using maximum a n d min imum temperature, a n d precipitat ion o n l yas model inputs) indicates a good correlation between temperatureand snowline alt i tude ( r 2 = 0.717). The slope o f the regression l inesuggests tha t snowline alt i tude changes b y 106 m r C .

-10

-

-

0

o S u m m e r

• A u t u m n

▪ W i n t e r

o S p r i n g

0Mean Monthly Temperature (degrees C)

Queenstown Airport

Ea a

10

• 0 DC l •

••

• 0 0 ••

• 0 3 )0 •

0

0 0 4Po 4)• , a 0

01:0 o

10 2 0

Mean Monthly Temperature (degrees C)Queenstown

20

Snowline Altitude 4 7

precipitation. T h i s r e l a t i onsh ip i s o b v i o u s g i v e n t h a t a t h i g h e r a l t i t u d e stemperatures d o n o t exceed freezing very of ten, s o tha t most precipitation fa l l s assnow a n d temperature i s a secondary effect . However, nearer t h e snowl ine t h i srelationship i s less c lea r a n d temperature more important.

Just a s w e m i g h t e x p e c t prec ip i ta t ion t o b e t h e p r i m a r y c o n t r o l o f s n o waccumulation, mean w i n t e r temperature shou ld b e t h e p r imary con t ro l o n meanwinter snowl ine a l t i tude, w i t h precipi tat ion a secondary fac to r. However, w h e nthese data are correlated the result i s no t significant. Th is i s a t least i n part due t othe complex relationship between temperature a n d precipitation t h a t leads t o a n ygiven snowline alt i tude, b u t i t i s a lso due t o t he relat ively smal l range o f mean'winter temperatures ( i e . -1-1.5°C). When w e consider a w ide r range o f temperatureconditions, l i k e those shown f o r mean month ly temperatures i n Figures 2 a n d 3(ie. -1-7.5°C), t h e temperature s igna l cont ro l l ing snowl ine a l t i tude shows c lea r l ythrough t h e no ise created b y natural var iab i l i ty.

Conclusions

1. S n o w line altitude i s complex t o model, and displays a h igh degree o f naturalva r i ab i l i t y.

2. Te m p e r a t u r e i s t he f i r s t order control o n snowline altitude, b u t w i th in therange o f mean win ter temperatures experienced over the last 3 0 t o 5 0 years(ie. 4-1.5°C), variabil i ty o f snowline caused b y the interaction o ftemperature, precipitation and other factors has masked any t rend o frising s n o w l i n e .

3. R a t e s o f change o f snowline alt i tude w i t h respect t o temperature a re about100 m r C change i n mean winter temperature, b u t may b e greater undercooler conditions than t he present, when cooler va l ley temperatures lead t omore f requent a n d more intense va l l ey inversions.

References

Barringer, J A R , 1986: So i l Erosion i n relation t o Snowline i n the Remarkables, Central Otago.Unpub. MSc thesis, Geography Department, University o f Otago, New Zealand.

Barringer, J.R.F., i n press: A Var iab le Lapse Rate Snow line Mode l f o r t h e Remarkables,Central Otago, New Zealand. N.Z. J. Hydro'.

Lave, B.L. & Epple, D. , 1 9 8 5 : Scenario Analysis I n Kates, R M . , Ausubel, J.H. & Berberian, M .(Eds), Climate Impact Assessment, John Wiley & Sons Ltd., 511 - 528.

Moore, R .D . & Owens, L E , 1984a: Modell ing A lp ine Snow Accumulation and Ablat ion UsingDaily Climate Observations. N . Z . J. Hyd ro l . , 23(2), 73 - 83.

Salinger, M.J. , 1988: N e w Zealand Climate Change: Past and Present. I n C l imate Change: T h eNew Zea land Response, Proceedings o f a workshop he ld i n Wellington, March 29-30, 1988,Ministry f o r the Environment, 17 - 24.

Salinger, M.J . & Hicks, DAC, 1989: Regional Climate Change Scenarios. Unpubl ished workingscenarios prepared f o r members o f the Impacts Working Group, New Zealand Climate ChangeProgramme, Min is t ry f o r the Environment.