Snow Cover and Climate Change in the Cairngorms National Park: Summary Assessment

cli­mate change Scotland’s centre of expert­ise con­nect­ing cli­mate change research and policy

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assess­ment Mike Riv­ing­ton, Mike Spen­cer, Aless­andro Gimona, Rebekka Artz, Douglas War­dell-John­son and Jonath­an Ball. James Hut­ton Insti­tute, Novem­ber 2019

Exec­ut­ive summary

Snow cov­er is a key aspect of what defines the char­ac­ter of the Cairngorms Nation­al Park (CNP). It under­pins the eco­logy, hydro­logy and eco­nomy, which are all depend­ent on how much snow falls, and where and how long it stays.

Mod­el­ling snow cov­er based on cli­mate pro­jec­tions is chal­len­ging. We com­pared his­tor­ic tem­per­at­ure and pre­cip­it­a­tion data (1918−2018) with observed snow cov­er days (1969−2005) to identi­fy how tem­per­at­ure affects snow days. We then mod­elled future snow cov­er days using the best avail­able data gen­er­ated by the UK Met Office to identi­fy some pos­sible trends for the Cairngorms Nation­al Park.

There is need for cau­tion in inter­pret­ing the future pro­jec­tion results:

1. The his­tor­ic­al snow cov­er data is incom­plete, and we have used single loc­a­tions for weath­er and snow elev­a­tion analyses
2. The mod­el­ling of future cli­mate con­sequences on snow cov­er is based on the pro­jec­tions gen­er­ated by the Met Office as part of UKCP 18. Scen­ari­os have been gen­er­ated for dif­fer­ent tem­per­at­ure increases, but only the high emis­sions scen­ario (pro­ject­ing warm­ing of 8.5 W m⁻², equi­val­ent to a glob­al tem­per­at­ure increase of 2.6 (2.0 to 3.2)°C by 2046 – 2065 and 4.3 (3.2 to 5.4)°C by 2081 – 2100 rel­at­ive to 1850 – 1900 tem­per­at­ures) is cur­rently avail­able for ana­lys­is of daily data. More detailed ana­lys­is is required as fur­ther data­sets are released. Rep­res­ent­at­ive Con­cen­tra­tion Path­way (RCP) scen­ari­os with lower cli­mate for­cing (RCP1.9. 2.6, 4.5 and 6.0), rep­res­ent lower rates of warm­ing and would be expec­ted to have lower impacts on snow cover.

Bear­ing this in mind, our ini­tial res­ults show a reduc­tion in snow cov­er as the observed warm­ing trend con­tin­ues and accel­er­ates. Suc­cess­ful glob­al efforts to reduce emis­sions may mod­er­ate this impact, whilst even high­er emis­sions rates (e.g. due to eco­sys­tem car­bon releases) may fur­ther increase impacts.

Key find­ings 1) There has been an over­all decline in observed snow cov­er in the Cairngorms Nation­al Park (1969−2005). This trend con­forms to those seen across oth­er moun­tain areas and the Arc­tic and is in keep­ing with the observed glob­al warm­ing trend. How­ever, some vari­ab­il­ity can also be seen with sig­ni­fic­ant snow events and a pos­sible increase in snow cov­er in the last dec­ade. The over­all declin­ing snow cov­er trend is pro­jec­ted to con­tin­ue and accel­er­ate in the future. 2) A warm­ing trend has been observed at met­eor­o­lo­gic­al sta­tions in the CNP since 1918 for both max­im­um and min­im­um tem­per­at­ure. There is vari­ation between months:

Cli­ma­teX­Change is Scotland’s Centre of Expert­ise on Cli­mate Change, provid­ing inde­pend­ent advice, research and ana­lys­is to sup­port the Scot­tish Gov­ern­ment as it devel­ops and imple­ments policies on adapt­ing to the chan­ging cli­mate and the trans­ition to a low car­bon soci­ety www​.cli​ma​tex​change​.org​.uk

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

a) Octo­ber and Novem­ber show approx­im­ately 1.6°C + max­im­um tem­per­at­ure and 0.8 °C min­im­um tem­per­at­ure rises. This may influ­ence the like­li­hood of when sea­son­al snow forms and cov­er becomes established.

b) March, April and May show a warm­ing trend indic­at­ing like­li­hood of earli­er onset of snow melting.

c) Pre­cip­it­a­tion (meas­ured as rain­fall and snow or ice) per month is vari­able between years with no strong trend observed.

3) There is a clear observed decrease in the num­ber of days of snow cov­er at all elev­a­tion levels over the 35 win­ters between ¹⁹⁶⁹⁄₇₀ and ²⁰⁰⁴⁄₀₅, with high­er elev­a­tions hav­ing a lar­ger pro­por­tion­al decrease.

4) In the near-term, our estim­ates indic­ate the poten­tial for a con­tinu­ation of snow cov­er at the cur­rent range of vari­ation, but with a sub­stan­tial decline from the 2040s. These find­ings are in line with res­ults from the UK Met­eor­o­lo­gic­al Office and Inter-gov­ern­ment­al Pan­el on Cli­mate Change (IPCC 2019). There will be some years in the future when the weath­er con­di­tions cre­ate snow and enable lying snow that may be com­par­able to the past, but such occa­sions will become few­er. This applies to all elev­a­tions, but with lar­ger pro­por­tion­al decreases at high­er levels. Res­ults indic­ate a like­li­hood of some years with very little or no snow by 2080.

5) Snow is com­plex to mod­el and pre­dict, espe­cially in tem­per­ate regions like Scot­land with its strong mari­time (Atlantic Ocean) cli­mat­ic influ­ence. Changes in sea­son­al vari­ab­il­ity will depend on how air flow over the UK (e.g. loc­a­tion of the jet stream) is affected by glob­al scale ocean-atmo­sphere cir­cu­la­tion pro­cesses. Our find­ings are a good indic­at­or of future trends, but there remain sub­stan­tial uncer­tain­ties and caveats that need to be con­sidered in mak­ing a more detailed assess­ment of future snow cover.

6) The pro­jec­ted decreases in the spa­tial cov­er­age and tem­por­al dur­a­tion of snow will have import­ant con­sequences, for example on the eco­logy and hydro­logy of the Cairngorms Nation­al Park and sur­round­ing areas. This may include changes to, for example, spe­cies com­pos­i­tion and dis­tri­bu­tion, and thus biod­iversity; the amount and tem­per­at­ure of ground­wa­ter, streams and rivers and flood risk due to rap­id melting.

www​.cli​ma​tex​change​.org​.uk Page 2

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Con­tents

Exec­ut­ive summary……………………………………………………………………………………………………………………….1

Key find­ings ………………………………………………………………………………………………………………………………….1

Con­tents ………………………………………………………………………………………………………………………………………3

Introduction…………………………………………………………………………………………………………………………………4

1.1 Ana­lys­ing snow cov­er ……………………………………………………………………………………………………..4 1.2 Pre­vi­ous stud­ies of snow cov­er ………………………………………………………………………………………..4

Ana­lys­is of past trends…………………………………………………………………………………………………………………….5

1.3 Tem­per­at­ure and precipitation…………………………………………………………………………………………5 1.4 Snow cov­er days …………………………………………………………………………………………………………….8 1.5 Apprais­al of his­tor­ic­al ana­lys­is ………………………………………………………………………………………..9

Future pro­jec­tions ………………………………………………………………………………………………………………………..10

1.6 Tem­per­at­ure ………………………………………………………………………………………………………………..10 1.7 Snow modelling…………………………………………………………………………………………………………….11

Caveats and uncer­tain­ties ………………………………………………………………………………………………………………15

Pos­sible implic­a­tions ……………………………………………………………………………………………………………………16

Conclusions………………………………………………………………………………………………………………………………..17

Next Steps………………………………………………………………………………………………………………………………….18

Acknow­ledge­ments ………………………………………………………………………………………………………………………19

Appendix A: Method…………………………………………………………………………………………………………………….19

Appendix B His­tor­ic­al cli­mat­ic trends and future pro­jec­tions – Bal­mor­al ………………………………………..21

Appendix C: His­tor­ic­al cli­mat­ic trends and White­hil­locks snow cov­er and weath­er data com­par­is­on …….28

Ref­er­ences …………………………………………………………………………………………………………………………………..30

www​.cli​ma​tex​change​.org​.uk Page 3

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Intro­duc­tion 1.1 Ana­lys­ing snow cover

The spa­tial extent of snow cov­er, here defined as its dur­a­tion (num­ber of days per year snow is on the ground) and its quant­ity (depth) is an essen­tial part of the eco­logy and hydro­logy in the Cairngorms Nation­al Park (CNP). It also influ­ences green­house gas emis­sions and sink poten­tial from peat­lands. Whilst there has been large inter-annu­al vari­ation in the past, there are sub­stan­tial con­cerns that, as a res­ult of cli­mate change, there may be sig­ni­fic­ant decreases in snow cov­er, quant­ity and spa­tial extent, in the future.

This ini­tial scop­ing study explores the like­li­hood of these decreases in snow cov­er in the future. It does not address snow depth. We first assessed past trends using observed weath­er for one site at Brae­mar and snow­line elev­a­tion data for a site in the east­ern CNP. The aim was to identi­fy pos­sible cor­rel­a­tions between observed weath­er and snow cov­er trends. We then used cli­mate mod­el pro­jec­tions from the Met Office in a snow mod­el to estim­ate future snow cov­er responses.

We used the UKCP18 daily data pro­jec­tions for the RCP8.5 (cur­rent emis­sions rate – See Appendix A meth­ods Text Box 2 for fur­ther details). These are the only daily data released (the snow mod­el used here needs daily data), hence this is just one pos­sible future scen­ario. Snow quant­ity is also import­ant, but it is impossible to mod­el it effect­ively at this scale, and it is not con­sidered here. There are many weath­er factors that determ­ine the cre­ation of snow, how long it snows for and what hap­pens to it once on the ground (e.g. move­ment by wind). These are bey­ond the scope of this sum­mary study and there are many uncer­tain­ties asso­ci­ated with pro­ject­ing future con­di­tions (see sec­tion Caveats and Uncertainties).

1.2 Pre­vi­ous stud­ies of snow cover

Snow­fall var­ies con­sid­er­ably in Scot­land and cor­rel­ates with alti­tude. Both nation­al and loc­al scale cli­mat­ic factors are involved in the observed spa­tial and tem­por­al pat­terns. Snow cov­er is highly sens­it­ive to cli­mat­ic vari­ations, glob­ally (IPCC, Box 1), region­ally (Brown 2019) and at spe­cif­ic loc­a­tions (Trivedi at al 2007).

Glob­al Con­text • Over the last dec­ades, glob­al warm­ing has led to wide­spread shrink­ing of the cryo­sphere, with mass loss from ice sheets and gla­ciers (very high con­fid­ence), reduc­tions in snow cov­er (high con­fid­ence) and Arc­tic sea ice extent and thick­ness (very high con­fid­ence), and increased per­ma­frost tem­per­at­ure (very high con­fid­ence). • Ice sheets and gla­ciers world­wide have lost mass (very high con­fid­ence). • Arc­tic June snow cov­er extent on land declined by 13.4 ± 5.4% per dec­ade from 1967 to 2018, a total loss of approx­im­ately 2.5 mil­lion km², pre­dom­in­antly due to sur­face air tem­per­at­ure increase (high con­fid­ence). • Per­ma­frost tem­per­at­ures have increased to record high levels (1980s-present) (very high con­fid­ence) includ­ing the recent increase by 0.29°C ± 0.12°C from 2007 to 2016 aver­aged across polar and high moun­tain regions glob­ally. • Between 1979 — 2018, Arc­tic sea ice extent has very likely decreased for all months of the year.

Source: Inter­gov­ern­ment­al Pan­el on Cli­mate Change Spe­cial Report: The Ocean and Cryo­sphere in a Chan­ging Cli­mate. 24^th Septem­ber 2019.

Box 1. Glob­al warm­ing influ­ences form­a­tion and longev­ity of glob­al snow and ice-con­tain­ing features.

www​.cli​ma​tex​change​.org​.uk Page 4

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

The UKCP18 Head­line Find­ings states: By the end of the 21 Cen­tury, lying snow decreases by almost 100% over much of the UK, although smal­ler decreases are seen over moun­tain­ous regions in the north and west (UKCP18 2019a).

In this study we use a well-estab­lished assump­tion that snow cov­er is more cor­rel­ated to tem­per­at­ure than pre­cip­it­a­tion, based on good evid­ence indic­at­ing tem­per­at­ure is a primary influ­en­cing factor as it influ­ences form­a­tion and con­trols depth and dur­a­tion (Har­ris­on et al 2001a, Ben­is­ton et al 2003, Trivedi et al 2007). А study at the Ben Law­ers Nation­al Nature Reserve found that snow cov­er dur­a­tion at mid to upper alti­tudes (600−900 m) responds most strongly to vari­ation in mean daily tem­per­at­ure: a 1 °C rise in tem­per­at­ure can cor­res­pond to a 15-day reduc­tion in snow cov­er at 130 m and a 33-day reduc­tion at 750 m (Trivedi et al 2007).

In the Arc­tic snow-cov­er extent has decreased by approx­im­ately 20% per dec­ade dur­ing 1979 – 2013 (Blun­den and Arndt, 2014) whilst the tim­ing of snow­melt onset has advanced 2 weeks on aver­age across the Arc­tic area since the start of the satel­lite era in 1979 (Tedesco et al., 2009).

Con­versely, some loc­a­tion spe­cif­ic evid­ence for the west­ern Cairngorms sug­gests an increased peri­od of snow dur­a­tion asso­ci­ated with a later melt­ing date, rather than onset of winter snow (Andrews et al 2016). Satel­lite data shows that Scot­land has areas with char­ac­ter­ist­ic com­bin­a­tions of snow­fall and melt­ing cycles (Pog­gio and Gimona, 2015). Some of these areas are char­ac­ter­ised by large vari­ab­il­ity in the num­ber of days of snow cov­er (e.g. repeated accu­mu­la­tion fol­lowed by melt­ing) espe­cially at lower alti­tudes, while at high­er alti­tudes the pat­tern is more stable.

Such vari­ab­il­ity makes it dif­fi­cult to inter­pret met­rics such as the aver­age snow­line. A bet­ter met­ric is there­fore the num­ber of days of snow cov­er dur­ing a giv­en peri­od (e.g. Octo­ber to May), which res­ult in a cor­rel­a­tion between snow depth and num­ber of snow lay­ing days. For this reas­on, in this study snow cov­er refers to the num­ber of days of lying snow. Note that the num­ber of days of snow cov­er does not indic­ate snow per­sist­ence, i.e. 50 days of snow cov­er in a year may not be 50 days of con­tinu­ous cover.

Ana­lys­is of past trends

The ana­lys­is was split into two sec­tions: firstly we assessed past trends examin­ing data from a rel­ev­ant met­eor­o­lo­gic­al sta­tion (Bal­mor­al); for future pro­jec­tions we then ran a mod­el pre­dict­ing snow cov­er across the whole Cairngorm Nation­al Park. Fur­ther details of our approach can be found in Appendix A.

1.3 Tem­per­at­ure and precipitation

We examined daily weath­er data from Bal­mor­al (as it has a long data record, 1918 – 2018), to assess evid­ence of the past trends for max­im­um and min­im­um air tem­per­at­ure and pre­cip­it­a­tion. These are shown as monthly aver­ages in Table 1. This site was chosen due to the length and qual­ity of its cli­mate record and loc­a­tion near the centre of the CNP.

We ana­lysed max­im­um and min­im­um air tem­per­at­ure to assess the change in range and poten­tial con­sequences of dif­fer­ences in rates of change and levels of vari­ation between them. Tem­per­at­ure is seen as a bet­ter indic­at­or of snow responses than precipitation.

Table 1 (and Fig­ures 7 – 12 in Appendix A) show there has been a sub­stan­tial warm­ing trend per month since

1. The rate of increase is great­er for max­im­um tem­per­at­ure (the winter aver­age is 1.30°C) than for min­im­um (aver­age of 0.71°C). How­ever, Decem­ber min­im­um tem­per­at­ure has decreased by c. 0.34°C.

Sub­stan­tial increases in max­im­um tem­per­at­ure has occurred in April (1.90°C), indic­at­ing an increase in melt­ing effects on snow. How­ever aver­age min­im­um tem­per­at­ures have not increase as much as the max­im­um tem­per­at­ures. In respect of snow form­a­tion and dur­a­tion on the ground, the low­est min­im­um tem­per­at­ures will have a key influence.

www​.cli​ma​tex​change​.org​.uk Page 5

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

www​.cli​ma​tex​change​.org​.uk Page 6

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

	Pre­cip­it­a­tion (mm)	Max­im­um Tem­per­at­ure (°C)	Min­im­um Tem­per­at­ure (°C)
Novem­ber	↑17	↑ 1.64	↑ 0.85
Decem­ber	↓-111	↑ 0.64	↓-0.34
Janu­ary	↓-24	↑ 1.34	↑ 1.20
Feb­ru­ary	↑45	↑ 0.98	↑ 0.37
March	↑47	↑ 1.23	↑ 1.28
April	↑33	↑ 1.90	↑ 0.82
May	↓-117	↑ 1.34	↑ 0.81

Table 1: Changes in monthly pre­cip­it­a­tion and tem­per­at­ure across all years between 1918 and 2018 for Balmoral

The daily tem­per­at­ure range between max­im­um and min­im­um will also have an influ­ence on snow cre­ation con­di­tions and dur­a­tion once on the ground. Tem­per­at­ures below freez­ing will help pro­long snow dur­a­tion but increas­ing num­ber of days above will shorten it.

Pre­cip­it­a­tion has seen decreases in Decem­ber, Janu­ary and May, but increases in the oth­er winter months; yet it remains highly vari­able over time and between years. These changes are in line with glob­al trends but may be slightly lower than those seen for high­er lat­it­udes (which is prob­ably due to the strong cli­mat­ic effect of the Gulf stream on Scotland’s climate).

Fig­ure 3 shows that at Bal­mor­al (Novem­ber example), there has also been an increased warm­ing of the highest max­im­um and low­est min­im­um tem­per­at­ures observed with­in the month. This tells us that for Novem­ber: • The sever­ity of cold below freez­ing is decreas­ing: the low­est min­im­um tem­per­at­ures have reduced (warmed to be closer to 0) by c. 2°C. Whilst these are still sub­stan­tially below freez­ing, the trend indic­ates that the degree of freez­ing has reduced. • The warmest min­im­um tem­per­at­ure (Max Tmin) has increased by c. 0.6°C. • The low­est max­im­um tem­per­at­ure has not increased as much, c. 0.5°C. • The warmest max­im­um tem­per­at­ure has increased by c. 2.0°C

These changes imply that there was less low tem­per­at­ure to cool the ground and help con­sol­id­ate any exist­ing fallen snow.

www​.cli​ma​tex​change​.org​.uk Page 7

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Tem­pa­ra­u­ture (°C) 45 ‑10 20 15 10 5 0 Novem­ber W 1918 1921 1924 1927 1930 1933 1936 1939 1942 1945 1948 1951 1954 1957 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 2014 2017 Year Max Tmax Min Tmin Min Tmax ….. Lin­ear (Max Tmax) Mean Tmax Max Tmin Mean Tmin ….. Lin­ear (Min Tmax) Lin­ear (Mean Tmax) Lin­ear (Max Tmin) Lin­ear (Mean Tmin) ….. Lin­ear (Mean Tmin) ….. Lin­ear (Min Tmin) Fig­ure 1: Changes in Balmoral’s highest and low­est val­ues of daily max­im­um and min­im­um air tem­per­at­ure and estim­ated trends for November.

1.4 Snow cov­er days

For snow cov­er dur­a­tion, the Snow Sur­vey of Great Bri­tain (SSGB) site of White­hil­locks was used, as it has the longest con­tinu­ous tem­por­al cov­er­age for the CNP. The dis­tance to Bal­mor­al is 24km. The obser­va­tion point of White­hil­locks is a sim­il­ar elev­a­tion (c. 250m) but the hills assessed for snow cov­er exceed 900m.

There has been sub­stan­tial year-to-year vari­ab­il­ity in snow cov­er dur­a­tion but a clear declin­ing trend in the num­ber of days per year with snow cov­er at a range of elev­a­tions between the win­ters of ¹⁹⁶⁹⁄₇₀ to ²⁰⁰⁴⁄₅ at White­hil­locks (Fig­ure 2). The mean decrease across all elev­a­tions was 52.8 days. The decrease in snow cov­er days per year was steep­er at high­er altitudes.

www​.cli​ma​tex​change​.org​.uk Page 8

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

250

200

150

100

50

0 69 – 70 71 – 72 73 – 74 75 – 76 77 – 78 79 – 80 81 – 82 83 – 84 85 – 86 87 – 88 89 – 90 91 – 92 93 – 94 95 – 96 <900m y = ‑1.7985x + 152.94 <750m y = ‑1.7529x + 129.37 < 600m y = ‑1.4211x + 104.07 <450m y = ‑1.4305x + 88.714 <300m y = ‑1.1477x + 71.9

6 – 88 97 – 98 99 – 00 01 – 02 03 – 04 Winter <150m <750m Lin­ear (<600m) <300m <900 Lin­ear (<750m) <450m Lin­ear (<300m) Lin­ear (<900) <600m Lin­ear (<450m) Fig­ure 2: Num­ber of days per year of snow cov­er at or below spe­cif­ic elev­a­tions and estim­ated trends for White­hil­locks between 1969 and 2005.

1.5 Apprais­al of his­tor­ic­al analysis

Estab­lish­ing a cor­rel­a­tion between observed weath­er and snow elev­a­tion data is prob­lem­at­ic. Match­ing the two data types for one loc­a­tion over a suf­fi­cient length of time cov­er­age was not pos­sible. Here we have ana­lysed weath­er data from Bal­mor­al (1918−2018) and snow elev­a­tion data from White­hil­locks (1969−2005), which are 24km apart. Oth­er loc­a­tions had observed weath­er and snow elev­a­tion data, but for much short­er time peri­ods. Our assess­ment of data util­ity was that Bal­mor­al provided good evid­ence of long-term cli­mate trends and over­lapped well with the tem­por­al records of snow elev­a­tion from Whitehillocks.

The indi­vidu­al loc­a­tion ana­lyses are inform­at­ive but there is a large assump­tion that the increased warm­ing at Bal­mor­al is cor­rel­ated to the change in snow­line elev­a­tion at White­hil­locks. The moun­tains assessed for snow elev­a­tion as part of the SSGB from the White­hil­locks obser­va­tion point those closer to Bal­mor­al. Our inter­pret­a­tion is that whilst not ideal, the res­ults gained from the com­par­is­on of the two loc­a­tions are indic­at­ive of over­all trends.

www​.cli​ma​tex​change​.org​.uk Page 9

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Future pro­jec­tions 1.6 Temperature

The tem­per­at­ure increases per month seen in Table 1 are pro­jec­ted to con­tin­ue per month through to 2050 accord­ing to the UKCP18 cli­mate pro­jec­tion data for the RCP8.5 scen­ario. As such the res­ults presen­ted here are for one pos­sible future, based on our cur­rent glob­al rate which puts us on a tra­ject­ory towards the high­er tem­per­at­ure rise range (c. 3 to 4°C). Fig­ure 3 shows warm­ing trends at Bal­mor­al since 1918 for both monthly mean max­im­um and min­im­um tem­per­at­ure and how these are pro­jec­ted to con­tin­ue through to 2050 (see Fig­ures 7 – 12 in Appendix A for oth­er winter months). Note: future tem­per­at­ure pro­jec­tions are shown for three cli­mate mod­el sim­u­la­tions from the 12 avail­able from the UKC18 Fig­ure 3 also shows the observed and monthly pre­cip­it­a­tion total (mm) and the future mean from all 12 UKCP18 cli­mate mod­el projections.

Pre­cip­it­a­tion (mm) Novem­ber Bal­mor­al Observed 300 Tmax y = 0.0164x + 6.3282 250 Tmin y = 0.0085x + 0.0514 Pre­cip­it­a­tion y = 0.0175x + 87.235 200 150 100 50 0 Bal­mor­al Area (5km) future pro­jec­tion 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018 2022 2026 2030 2034 2038 2042 2046 2050 Obs Pre­cip­it­a­tion Obs Tmax EM5 Tmax EM5 Tmin Lin­ear (Obs Tmin) Ensemble Mean Pre­cip­it­a­tion Obs Tmin EM1 Tmin Lin­ear (Obs Pre­cip­it­a­tion) EM1

Tmax

EM4 Tmax EM4 Tmin … Lin­ear (Obs Tmax) 0 5 ‑10 Fig­ure 3: Weath­er data trends for Novem­ber, mean monthly max­im­um and min­im­um tem­per­at­ure and total pre­cip­it­a­tion for Bal­mor­al 1918 – 2018 and estim­ated future 2020 – 2050 pro­jec­tion (RCP8.5)

Data presen­ted in Fig­ure 3 shows that there is an observed warm­ing trend for all months, except min­im­um tem­per­at­ure in Decem­ber (see Appendix B for all oth­er winter months). These are likely to con­tin­ue in the future. Pre­cip­it­a­tion how­ever has remained highly vari­able but with no clear observed trend across all winter months.

The estim­ated lin­ear trend lines (dot­ted lines in Fig­ure 3) have been exten­ded to 2050 to show how the observed trends relate to the future pro­jec­tions. In the Novem­ber example above the max­im­um tem­per­at­ure trend matches well to the cli­mate mod­elled estim­ates. Here only three cli­mate mod­el estim­ates are shown from the avail­able 12, but these are rep­res­ent­at­ive of the range.

The pro­jec­ted future min­im­um tem­per­at­ure is great­er than indic­ated by the observed trend. This may be due to some cli­mate mod­el error, but research indic­ates that daily min­im­um tem­per­at­ure has and will con­tin­ue to increase more rap­idly than max­im­um tem­per­at­ure in many parts of the world (e.g. IPCC 2014). For the oth­er winter months, the cli­mate mod­el estim­ates fit well to the exten­ded observed trend lines.

www​.cli​ma​tex​change​.org​.uk Page | 10 5 15 10 Tem­per­at­ure (°C)

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

These res­ults imply an increas­ing prob­ab­il­ity of less favour­able snow form­ing con­di­tions and more rap­id rates of snow melting.

1.7 Snow modelling

To estim­ate future snow cov­er we ran a snow mod­el over all the 5 km grid cells cov­er­ing the Nation­al Park (226 cells in total, Fig­ure 4). Input fore­cast data to the mod­el were the UKCP18 cli­mate pro­jec­tions (see Appendix A). The mod­el estim­ates snow cov­er based on daily tem­per­at­ure and pre­cip­it­a­tion. When tem­per­at­ure is below a threshold, pre­cip­it­a­tion accu­mu­lates as snow and when tem­per­at­ure rises above the threshold the snow melts. For more inform­a­tion on the mod­el and cal­ib­ra­tion see Spen­cer 2016a.

Elev­a­tion (m) 200 400 600 800 1,000 1,200 Mean Elev­a­tion (m) 0 to 200 200 to 400 400 to 600 600 to 800 800 to 1,000 Fig­ure 4: Map of Cairngorm Nation­al Park and mod­el grid cells. Elev­a­tion is shown on a 50 m (left) and 5 km (right) grid. Con­tains Ord­nance Sur­vey data © Crown copy­right and data­base right 2019

Out­put from each 5 km grid cell were col­lated based on the mean elev­a­tion of the cell. Elev­a­tions were grouped into four bands, allow­ing us to com­pare what may hap­pen to snow cov­er at dif­fer­ent heights in the nation­al park. The res­ults of this exer­cise are in Fig­ure 5, and can be com­pared to the ski centre elev­a­tion ranges: • Cairngorm: 630 to 1150m • Glen­shee: 650 to 920m • Lecht: 580 to 780m

Note the bulk of ski activ­ity lies in the 600 — 800 m elev­a­tion range. The trend for the num­ber of days of snow cov­er below 400m and 400 – 600m elev­a­tion ranges sim­u­lated (Fig­ure 5¹) all approach zero by 2080, but with large vari­ations between cli­mate mod­el ensemble mem­bers and years. At elev­a­tion ranges 600 – 800m and over 800m, the trend indic­ates a reduc­tion by more than a half of the cur­rent num­ber of days with snow cov­er, with some cli­mate pro­jec­tions indic­at­ing poten­tial for very few days with snow cov­er even at high­er elevations.

There is a great deal of uncer­tainty in snow mod­el­ling, but our res­ults pro­ject a dra­mat­ic decline in the dur­a­tion of annu­al snow cov­er. Trends appear to remain con­sist­ent until ~ 2030, after which we estim­ate a steep­er decline in snow cov­er dur­a­tion. This change is likely due to a threshold tem­per­at­ure being exceeded, caus­ing less snow to accu­mu­late or per­sist. These declines are most notice­able for the highest elev­a­tions in the nation­al park, with elev­a­tions above 800 m estim­ated to have 30 – 40 days of snow cov­er on aver­age each year by 2080.

¹ Estim­ated using a degree-day snow mod­el with observed data from UKCP09 (1960−2010) and 12 UKCP18 cli­mate pro­jec­tions (2020−2080).

www​.cli​ma​tex​change​.org​.uk Page | 11

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Snow cov­er (days per year) 80 60 Annu­al snow cov­er Cairngorms nation­al park, 1960 — 2080 below 400 m 400 – 600 m 100 75 50 25 0 600 — 800 m over 800 m 40 20 0 W wwwhw 100 50 0 2000 2050 150 100 50 0- 2000 Year 2050 Fig­ure 5: Annu­al snow cov­er (days per year) at four elev­a­tion ranges for the Cairngorms Nation­al Park. Light grey lines show indi­vidu­al mod­el runs with the heavy black line indic­at­ing the smoothed aver­age of these.

www​.cli​ma​tex​change​.org​.uk Page | 12

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Snow cov­er (days per year) Snow cov­er dur­a­tion 1960 – 2080 Cairngorm Nation­al Park 50

www​.cli​ma​tex​change​.org​.uk Snow cov­er above 550 m Dal­whin­nie Tomin­toul Aviemore Spit­tal of Glen­shee Brae­mar Kin­gussie Scot­tish Gov­ern­ment Riaghaltas na h‑Alba White­hil­locks Clova YEARS Fig­ure 6: Sum­mary of the mod­elled spa­tial changes in snow cov­er 1960 – 2080. Loc­a­tions above 550 m (green) may have more severe decline in days with snow cov­er than areas at lower alti­tude (purple). The future pro­jec­tions in Fig­ure 5 are mul­tiple snow mod­el sim­u­la­tion res­ults gained by using the 12 cli­mate mod­el sim­u­la­tions avail­able from the UKCP18 for the RCP8.5 emis­sions. This means the snow mod­el was run SRUC CRE­ATED BY MIKE SPEN­CER DOI: 10.5281/ZENODO.3518297 Con­tains Ord­nance Sur­vey data Crown copy­right and data­base right 2019. Con­tains Met Office UKCP09 and UKCP18 data licensed under the Open Gov­ern­ment Licence v3.0. Down­scal­ing and Cor­rec­tion copy­right 2019 The James Hut­ton Insti­tute. Grant­own-On-Spey YEARS cli­mate change Bal­later Blair Atholl Snow cov­er below 550 m LEAD­ING IDEAS FOR BET­TER LIVES Pàirc Nàiseanta a’ Mhon­aidh Ruaidh Cairngorms NATION­AL PARK The James Hut­ton Insti­tute gov.scot Achn­agoi­chan Page | 13

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

with 12 dif­fer­ent data sets pro­duced from 12 dif­fer­ent runs of the Region­al Cli­mate Mod­el used by the UK Met Office. This helps cap­ture some of the uncer­tainty in the cli­mate mod­el­ling and enables us to present the vari­ab­il­ity in likely scen­ari­os. The aver­age across this range (the heavy black line in Fig­ure 5) indic­ates the over­all pro­jec­ted trend.

The res­ults show that snow cov­er (days per year) may con­tin­ue to be sim­il­ar to the past for the next 1 – 2 dec­ades but will decline sub­stan­tially after­wards. This applies to all elev­a­tions, but with lar­ger pro­por­tion­al decreases at high­er levels. These declines may be asso­ci­ated with passing a tem­per­at­ure threshold where pre­cip­it­a­tion no longer falls as snow and any lying snow melts soon­er. An over­view timeline is approx­im­ately: • 2020 – 2030: sim­il­ar amounts and level of annu­al vari­ation of snow cov­er to the past at all elev­a­tions. Some years likely to be sim­il­ar or even pos­sibly great­er snow cov­er than in the past. • 2030 – 2040: declin­ing snow cov­er but with sim­il­ar levels of annu­al vari­ation to the past at all elev­a­tions. Some years likely to be sim­il­ar to the past but not achiev­ing the lar­ger quant­it­ies or spa­tial cov­er­age of snow cov­er, espe­cially at the low- to mid-range elev­a­tions. • 2040 – 2050: rate of decline increases at all elev­a­tions to approx­im­ately half of his­tor­ic long-term aver­age snow cov­er. Aver­age amounts of snow cov­er sim­il­ar to the low­est levels seen in the past. • 2050 – 2080: con­tin­ued increas­ing rate of decline par­tic­u­larly at high­er elev­a­tions, approach­ing <25 days above 600m on aver­age, but with some years where the largest amount of snow cov­er is sim­il­ar to the his­tor­ic low amounts. There is poten­tial for some years to have no snow even at the highest elevations.

These res­ults are in line with site-spe­cif­ic stud­ies of obser­va­tions. e.g. Trivedi et al (2007) found an observed rela­tion­ship of a 1°C tem­per­at­ure rise at a met­eor­o­lo­gic­al sta­tion at Ben Law­ers cor­res­pond­ing to a 15-day reduc­tion in snow cov­er at 130m elev­a­tion and a 33-day reduc­tion at 750m.

Cur­rently some evid­ence indic­ates an increase in snow cov­er in the last dec­ade (Andrews et al 2016 and anec­dot­al), with sub­stan­tial snow events in 2018, 2013, 2010 and 2009 (against a back­ground of an over­all drop since 1960) (UKCP18 2019b). Near-term cli­mate pro­jec­tions indic­ate a poten­tial for con­tinu­ation (Fig­ure 5 2020 to c. 2030), with large inter-annu­al vari­ab­il­ity and hence poten­tial for some years with long snow cov­er duration.

www​.cli​ma​tex​change​.org​.uk Page | 14

Snow Cov­er and Cli­mate Change in the Cairngorms Nation­al Park: Sum­mary Assessment

Our res­ults are con­sist­ent with oth­er stud­ies. Using a low emis­sions scen­ario, Tre­vidi et al (2007) mod­elled pro­jec­tions of a 93% reduc­tion in snow cov­er at 130m, 43% at 600m and 21% at 1060m. For a high­er emis­sions scen­ario they pro­jec­ted 100%, 68% and 32% for these elev­a­tions, respect­ively. The UKCP18 report a decrease in both fall­ing and lying snow across the whole UK for the peri­od 2061 – 2080 (using the same cli­mate mod­el data as our study). This decrease is smal­ler for the Scot­tish moun­tains, but still in the order of 20 – 60% (snow­fall) and 60 – 100%