Changes in alpine plant population sizes inresponse to climate changeEvelina Rostö

(1)

Changes in alpine plant population sizes in response to climate change

Evelina Rostö

Degree project inbiology, Master ofscience (2years), 2020 Examensarbete ibiologi 30 hp tillmasterexamen, 2020

Biology Education Centre and SLU Artdatabanken, Uppsala University Supervisors: Brita Svensson, Per Toräng and Mora Aronsson

External opponent: Håkan Rydin

(2)

1

Abstract

Alpine plants are assumed to be in particular danger as the climate changes rapidly worldwide. Specialist alpine species in Norrbotten County, northern Sweden have been surveyed over the last 20 years, providing insight to population dynamics and how the plants might respond to the changing climate. The main current threat to the species is habitat destruction as the climate changes. Variation in the number of plants among populations and years, and correlations with environmental variables were examined. Some species had increased while others had decreased over the years. No uniform relationship for all species and populations were discovered, but some of the species exhibited relationships between population size changes and temperature and precipitation. However, if the future climate in Norrbotten County changes according to the predictions, the habitats of the specialist alpine plants may be severely altered, leaving the species with no alternative places to establish and grow.

Introduction

Vascular plants and climate change

The climate has changed repeatedly since the Earth was formed, but today it is changing at a more rapid pace than before (SGU 2018). Vascular plants have endured naturally changing conditions many times during their evolutionary history, but previous changes have been slow and gradual (Körner 2003, Hunter & Gibbs 2016). Human actions the last couple of centuries have increased the concentrations of greenhouse gases in the atmosphere rapidly. The elevated concentrations of carbon dioxide (CO2) are increasing global temperatures, resulting in changing precipitation patterns, as well as frequency of extreme weather events. The emissions from fossil fuels and use of fertilizers have also increased the amount of soluble nitrogen (N) in the soil and air (Körner 2003, IPCC 2014).

Climate change will affect species richness, plant community composition and productivity of vascular plants. Global warming is predicted to affect species richness in different ways in different places. In habitats where temperature is a limiting factor for many vascular plant species, such as north-temperate and alpine habitats, an increase in species richness is predicted in response to increased mean temperatures (Sommer et al. 2010, Venevskaia et al. 2013, Harrison 2020). Near the equator, in arid areas where water-availability is low, the projected increase in precipitation is predicted to increase plant species richness as well (Harrison 2020).

In many other tropical and sub-tropical regions, where precipitation is projected to decrease as the temperature increases, a decrease in plant species richness is predicted (Venevskaia et al.

2013, Harrison 2020).

The elevated CO2 concentrations in the air caused by global, anthropogenic activity, has been documented to increase plant growth and photosynthesis rates, as well as decrease water use and concentrations of N and proteins in plant tissues in many plant species (Bowes 1993, Ainsworth & Long 2005, Taub 2010, Gray & Brady 2016). Species with C4-photosynthesis (see definition in Ricklefs & Relyea 2014) show less response to increased concentrations of CO2 (Bowes 1993, Ainsworth & Long 2005, Taub 2010, Gray & Brady 2016). Fertilization experiments by Clark et al. (2007) and Midolo et al. (2019) have shown that increased availability of N increases plant growth and biomass but also decreases species richness and diversity. The plants that are adapted to low N-availability are out-competed by the species that are able to utilize the additional nutrients for rapid growth (Körner 2003, Clark et al. 2007,

(4)

3

Midolo et al. 2019). The responses in plant communities to additional N are not uniform but depends on several environmental factors such as mean annual temperature, the annual amount of precipitation and cumulative effects of N addition (Midolo et al. 2019).

The poles, and particularly the Arctic, are more affected by global warming than the rest of the planet because of various feedback processes. Globally, the temperature has risen 1°C, in areas towards the poles the increase is several degrees centigrade (IPCC 2014). In places where there is less snow and ice than before, the exposed, darker open water and landmasses absorb more solar radiation, which increases the warming even more (ACIA 2005, Screen & Simmonds 2010, IPCC 2014). This already have a large influence on the permafrost in the Arctic, which is an important carbon sink, but as the temperature rises, the thawing of the frozen ground releases even more greenhouse-gases, further accelerating the global warming (Tarnocai et al.

2009, Schuur et al. 2015).

The Swedish Scandes are considered as part of the Oroarctic, at the end of the boreal zone and beginning of the arctic tundra, but are not part of the Arctic (CAFF 2013, Ehrich et al. 2016).

However, the alpine areas in Norrbotten County, the northernmost situated county in Sweden, is exhibiting similar effects of global warming as the Arctic regions. At Abisko Scientific Research Station, in the northwest of the county, scientists have measured numerous environmental variables since 1913 (Swedish Polar Research Secretariat 2018). The environment has changed dramatically over the last 100 years. For example, the large lake Torneträsk is ice-free 40 more days each year compared to 100 years ago, the tree-line has moved more than 20 m upslope in elevation, the growing season is about four weeks longer and the winter temperatures have increased (Jonasson et al. 2012, CIRC 2020a, CIRC 2020b). The depth of permafrost thawing by Torneträsk has increased by a meter from 1978 (0.48 m) to 2018 (1.46 m) (CIRC 2020a). The Swedish Metrological and Hydrological Institute (SMHI) modelled the future climate in Norrbotten County until 2100, based on the RCP4.5 and RCP8.5 models (IPCC 2014, SMHI 2015). The models predict dramatic changes in yearly mean temperature, annual amount of precipitation and growing season length, compared to the period 1961 to 1990 (Table 1). The county administrative board in Norrbotten County have predicted that 75% of the area that today are above the treeline will be covered with shrubs and trees by 2100 (Länsstyrelsen Norrbotten 2015).

Table 1. The predicted changes until 2100 in yearly mean temperature, yearly precipitation, and growing season length in Norrbotten County compared to mean values from the period 1961 to 1990 using the models RCP4.5 and RCP8.5 (IPCC 2014, SMHI 2015).

Environmental variable RCP4.5 RCP8.5

Increased mean temperature (°C) 3-4 6

Increased precipitation (%) ~20 ~40

Increased growing season (days) 30 50

Trends of alpine plants

The arctic and alpine flora is adapted to harsh conditions on the mountain slopes. The species can endure long periods under a thick snow layer, low temperatures, a short growing season and strong winds (Bliss 1962, Billings 1974, Carlsson et al. 1999, Körner 2003). These are stress-tolerant species, highly adapted to enduring hard conditions. Species adapted to increased stress are usually bad competitors, as there is a trade-off between being stress-tolerant and being competitive (Grime 1977). The low temperature slow decomposition rates, limiting the

(5)

4

availability of soil nutrients for the plants. This leads to slower plant growth both due to slower photosynthesis rate because of low temperatures as well as shortage of available nutrients.

Adaptations like low stature and growth forms such as cushions or rosettes, minimize risks of injury from wind and snow and creates a warm microclimate during the summer (Carlsson et al. 1999, Körner 2003). In fact, many alpine plants with low stature may have to cope with overheating as the wind barely stirs the air closest to the ground, and temperatures may exceed 40 °C (Körner 2003). Researchers have been surprised to find that many alpine plants have sexual reproduction despite the short growing season, although species growing in high elevations usually have a higher proportion of clonal reproduction than species found at lower elevations. Plants found in high elevations can be hundreds of years old, growing only a few mm per year and therefore takes a long time to spread to new sites (Körner 2003, Walther 2003).

This makes them vulnerable to disturbances like ski runs and hiking trails as just one step on a plant-cushion can break tissue that has been formed over decades or centuries (Körner 2003).

If the environmental conditions change according to the predictions from models of future climate in alpine areas, the habitats that alpine and arctic species are adapted to will change dramatically (Bliss 1962, Billings 1974, Körner 2003, Winkler et al. 2019). Upward migration on the slopes have already been recorded for alpine plants in many areas, most likely triggered by the increased temperatures, longer growing season and the higher availability of nutrients from emissions (predominantly N) (Grabherr et al. 1994, Körner 2003, Pauli et al. 2007, Frei et al. 2010, CAFF 2013, Öberg 2013, Kullman 2015, Winkler et al. 2019). The addition of atmospheric N have been documented to shift plant community composition, causing competitive species to increase while slow-growing species adapted to low amount of nutrients decrease (Crawford 2008, Alatalo et al. 2014, Winkler et al. 2019). A longer growing season, the addition of nutrients and increasing temperatures will thus allow lowland species to colonize upslope habitats that were previously unsuitable for them. This will cause an increase in species richness in the alpine habitat (CAFF 2013, Steinbauer et al. 2018, Niskanen et al. 2019, Harrison 2020). However, many alpine plant species lack competitive strength, which means that when colonizers migrate up the slopes, the alpine, stress-tolerant flora can be out-competed (Bliss 1962, Billings 1974, Grime 1977, Körner 2003, Crawford 2008). The most vulnerable species are the ones that are already growing at the top of summits, they have nowhere to migrate to when their habitat is no longer suitable for them or other species take over their current growing sites (Crawford 2008, Freeman et al. 2018).

The botanists T. C. E. Fries, G. E. Du Rietz, and G. Sandberg surveyed the alpine flora near Abisko a couple of years from 1917 onward, recording not only species, but also plant distribution and phenology (CIRC 2020b). To study the effects of climate change on alpine plant species, the study is now repeated by the Abisko Scientific Research Station and Umeå University. The data collected in the beginning of the 20^th century is compared to the findings of scientists today to investigate changes in plant populations and communities (CIRC 2020b).

The current project is on-going, and nothing is published so far (May 2020). However, Larson and colleges (at Abisko Scientific Research Station and Umeå university) have documented several new species for the region (Abisko National Park) and range shifts in the plant community. Some species have expanded their range upwards without changing their lower range limit while some have moved both their lower and higher elevation ranges upwards. Some species that are dependent on snow beds have shifted their range downwards as snow beds have disappeared at higher elevations but remained further down the slopes where there is more

(6)

5

shade and less direct solar radiation (Keith Larson, Abisko Scientific Research Station, personal communication, April 2020).

In this study, I surveyed alpine, arctic vascular plant species in Padjelanta National Park and at Mt Pältsa in Norrbotten County, Sweden. I also analysed data collected by others over the last 20 years. This data was collected in Muddus National Park, Torne Lappmark, Padjelanta National Park and at Mt Pältsa (Figure 1, Material and Methods and Appendix 1). All species surveyed are in the EU Habitats-directive (except for Potentilla hyparctica Malte) as well as on the Swedish Red List (Table 2) (1992/43/EEG, SLU Artdatabanken 2020a). As EU member states are obligated to monitor and preserve the species listed in the directive (Sundseth & Creed 2008), the Swedish Environmental Protection Agency (SEPA) has developed guidelines for the conservation of each of these species. They include information about the species autecology, their current conservation status, pressures and threats, and what conservation measures should be taken to preserve them (Naturvårdsverket 2020).

Aims

By comparing population sizes of alpine plant populations in Norrbotten County over time, I aim to examine whether the changing climate has affected population dynamics. My research questions are:

1. How has the population sizes of the studied alpine plant species varied over the last 20 years?

2. If the population sizes have changed over time, can the variation in population size be linked to climate change?

3. How will environmental change affect the studied populations in the future?

Hypotheses

Based on the knowledge of alpine plant ecology and physiology in Sweden and Scandinavia, the trends in environmental variables due to climate change and the trends of plant population dynamics in similar environments globally, I have the following hypothesis:

1. The alpine plant population sizes have increased with an increased mean annual temperature and longer growing seasons over the last 20 years.

2. In the future, I predict that the populations I monitored in Norrbotten County will decrease and eventually disappear as a response to competition from new plant species in the plant community, changes to their habitats and the unavailability of places to move to when the local environment changes.

Material and methods

Study species

Below follows a summary of the knowledge of the species discussed in this report, and what threats they face, particularly from climate change. All species presented are bound to calcareous soils, except for Arenaria humifusa Wahlenb. (Aronsson et al. 2013). Nomenclature follows Dyntaxa (SLU 2020).

(7)

6

Table 2. Vascular plant species studied in this paper. Red List status as assessed by SLU Artdatabanken (2020).

All species are listed in the Habitats’ directive except for Potentilla hyparctica (1992/43/EEG).

Species Family Red List status Additional information Arenaria humifusa Caryophyllaceae NT

Braya linearis Brassicaceae NT

Luzula nivalis Juncaceae NT

Papaver radicatum subsp. laestadianum

Papaveraceae VU Endemic to Sweden and

Norway Papaver radicatum

subsp. radicatum

Papaveraceae NT

Platanthera obtusata Orchidaceae EN Growing sites’ location secured from the public

Potentilla hyparctica Rosaceae VU

Primula scandinavica Primulaceae VU Endemic to Sweden and Norway

Silene involucrata Caryophyllaceae NT Viola rupestris subsp.

relicta

Violaceae NT Endemic to Sweden,

Norway and Finland

Arenaria humifusa grows on serpentine soils, a bed rock with a low content of major nutrient elements (e.g. Ca) and a relatively high content of toxic heavy metals such as Ni and Cr (Whittaker et al. 1954, Brady et al. 2005, Aronsson et al. 2013). The low availability of calcium as well as the amount of heavy metals limits plant growth on serpentine soils, but some plant species are adapted to this environment and consequently have few competitors (Whittaker et al. 1954, Kazakou et al. 2008). A. humifusa is known from 20 sites in Sweden, either on serpentine or olivine bedrock with gravel and high soil moisture (Aronsson et al. 2013). A drier, warmer climate could desiccate the sites it grows on and cause a decrease in the populations, but competition from other species will likely not affect the species because of the high stress from the toxic content of the serpentine soil (Naturvårdsverket 2011a).

South-facing screes and slopes are important in alpine environments as they provide a warm and sunny microclimate, creating living space for several species of plants (Naturvårdsverket 2014). Viola rupestris subsp. relicta F. W. Schmidt and Braya linearis Rouy are two of them, known from 14 and 20 sites in Sweden, respectively (Aronsson et al. 2013). Climate change threatens the state of the screes in two ways. First, if the climate becomes warmer and drier, screes may become stable and overgrown by other plant species that outcompete V. rupestris subsp. relicta and B. linearis. Second, extreme amounts of precipitation over a short period of time can make the screes unstable, destroying the growing sites for the two species (Mora Aronsson, pers. comm., Naturvårdsverket 2011b, 2011c). B. linearis also occur on sites with solifluction, a habitat that will change the same way as the screes due to climate change (Naturvårdsverket 2011c, Aronsson et al. 2013). V. rupestris subsp. relicta is endemic to Norway, Sweden and Finland, meaning Sweden has a global responsibility to preserve the species (Naturvårdsverket 2011b, Aronsson et al. 2013).

Silene involucrata (Cham. & Schltdl.) Bocquet is known from a few sites in Sweden. It is dependent on disturbances and do not grow constantly at the same site but moves every few years as new sites appear. It occurs on river and lake shores (both on gravel and sand), steep river banks gravel-shores in the low-alpine area and also on sites with solifluction

(8)

7

(Naturvårdsverket 2011d, Aronsson et al. 2013). A drier climate with less precipitation will decrease the availability of sites with solifluction and stabilize shorelines along streams, making them less suitable for S. involucrata. Heavy rainfall can increase the erosion of the banks and shores, destabilising the sites where the species is adapted to growing on (Naturvårdsverket 2011d).

Two subspecies of Papaver radicatum Rottb. were examined in this study: P. radicatum subsp.

laestadianum Nordh. and P. radicatum subsp. radicatum Rottb. (SLU Artdatabanken 2020a).

P. radicatum subsp. laestadianum is endemic to Mt Pältsa and a few neighbouring mountains in Norway (Aronsson et al. 2013). As with V. rupestris subsp. relicta, Sweden has a global responsibility to preserve this taxon. It grows on steep slopes, at solifluction sites and screes with rocks and gravel. Natural disturbances through soil movement and reindeer grazing and trampling decrease competition from other plants and provide germination sites for P.

radicatum subsp. laestadianum. The greatest threat is that the disturbances cease due to a warmer, drier climate (less soil movement) and less reindeer in the area, resulting in fewer germination sites and sites being overgrown with other species (Naturvårdsverket & Selin 2010, Naturvårdsverket 2011e).

Papaver radicatum subsp. radicatum occurs on 15 sites in Sweden as well as a few places in Norway, on Iceland and the Faroe Islands (Aronsson et al. 2013). It too grows in screes and on solifluction soils, but can also be found on sandy banks by streams and mountain lakes (Aronsson et al. 2013). Exposed soil is needed for the species to establish, it depends on soil movement and reindeer grazing and trampling, as its close relative. The major threat to the subspecies is that its growing sites change to the benefit of other species with less disturbance of the soil by solifluction and reindeer. Both subspecies are weak competitors and would therefore be disfavoured from having other species around (Naturvårdsverket 2011f).

Sweden also carry a global responsibility for Primula scandinavica Bruun, endemic to Sweden and Norway. It grows on moist, south-facing cliffs, alpine meadows, stream-shores and on gravel (Naturvårdsverket 2011g, Aronsson et al. 2013). The climbing treeline may become a threat to the species, as its habitats becomes birch forests. Ceased reindeer grazing and trampling increases the risk of other plant species out-competing P. scandinavica (Naturvårdsverket 2011g).

Luzula nivalis (Laest.) Spreng. grows on calcareous, peaty, gravelly or solifluction soils (Naturvårdsverket 2011h, Aronsson et al. 2013). It is a weak competitor and dependent on disturbances from soil movement and reindeer grazing. A warmer climate with less soil movement and less grazing can threaten the populations of L. nivalis in Sweden (Naturvårdsverket 2011h).

Potentilla hyparctica is known from two sites in Sweden. It grows in sun and wind exposed places such as summit ridges with early snowmelt, on 1350-1450 m a.s.l. The plants are usually found in crevices and depressions where snow gather and keeps the soil moist throughout the summer (Aronsson et al. 2013). In a warmer climate, the soil can become too dry for P.

hyparctica and threaten the preservation of the population in Sweden.

The orchid Platanthera obtusata (Banks ex Pursh) Lindl. is known from three sites close to Abisko, Sweden. It grows on soils on slopes with solifluction or moist cliffs in proximity to birch-forests (Naturvårdsverket 2011i, Aronsson et al. 2013). As for P. scandinavica, the

(9)

8

climbing treeline can become a threat as the sites are just above the birch-forest. However, the lake Torneträsk cools the air at these sites, creating a climate similar to that above the treeline.

Because of this phenomenon, these populations might be less threatened by the changing treeline than populations monitored in Norway (Skrede et al. 2018). As this is an orchid, the population is also threatened by people digging up the plants, which is why the sites are kept secret (Naturvårdsverket 2011i, SLU Artdatabanken 2020b).

Study sites and data collection

The plant population data used in this study were collected between 2000 and 2019. The county administrative board (CAB) in Norrbotten County, groups of consultants on behalf of the Swedish Environmental Protection Agency (SEPA) and SLU Artdatabanken at the Swedish University of Agricultural Sciences provided data from Padjelanta National Park, Muddus National Park, Pältsa mountain area and Torne Lappmark (Figure 1, Appendix 1). The plant species surveyed are listed in Table 3.

Figure 1. The locations of the study sites in Norrbotten County, Sweden. The red line shows the boundary of Norrbotten County. Top left: Torne Lappmark; top right: Mt Pältsa; bottom left: Padjelanta National Park and bottom right: Muddus National Park (Maps made in ArcGIS 10.8 by Stephanie Jonsson).

(10)

9

Table 3. The plant species surveyed in Padjelanta National Park, Muddus National Park, Torne Lappmark and at Mt Pältsa, which years the populations were surveyed and who did the survey. All data collected in 2019 (bold in the table) was done by the author. All population data used in this study, as well as a description of the surveys presented here, are listed in the Appendix 1. Nomenclature follows Dyntaxa (SLU 2020).

Species Site Survey years Data collectors

Arenaria humifusa Padjelanta National Park

2004, 2005, 2006, 2010, 2012, 2019

CAB Norrbotten, SLU Artdatabanken, Evelina Rostö

Braya linearis Padjelanta National Park

2000, 2004, 2005, 2006, 2008, 2010, 2012, 2019

Luzula nivalis Mt Pältsa 2011, 2019 Mora Aronsson & Niklas Lönnell, Evelina Rostö Papaver radicatum

subsp. laestadianum

Mt Pältsa 2011, 2019 Mora Aronsson & Niklas Lönnell, Evelina Rostö Papaver radicatum

subsp. radicatum

Torne Lappmark 2005, 2006, 2008, 2010, 2011, 2013,

2018

CAB Norrbotten, Enetjärn Natur AB, SLU

Artdatabanken, Naturcentrum, CAB Norrbotten

Platanthera obtusata Torne Lappmark 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2013,

2018

CAB Norrbotten

Potentilla hyparctica Padjelanta National Park

2004, 2008, 2010, 2012

CAB Norrbotten Primula

scandinavica

Padjelanta National Park

2000, 2004, 2005, 2006, 2008, 2010, 2012, 2019

Silene involucrata Muddus National Park

1971, 2017, 2018

Flora watchers, CAB Norrbotten

Viola rupestris subsp. relicta

2000, 2019 SLU Artdatabanken, Evelina Rostö

Plant surveys 2019

I surveyed populations of four plant species in Padjelanta National Park that had been visited in 2000 and 2005 by SLU Artdatabanken, and populations of two species at Mt Pältsa that previously had been visited in 2011 by Mora Aronsson and Niklas Lönnell. My inventories took place between the 25^th of July and the 17^th of August 2019. In Padjelanta, populations of B. linearis, V. rupestris subsp. relicta, and P. scandinavica were surveyed on Oarjep Slahpetjåhkka, as well as populations of A. humifusa on Vietjevaratj, Oarjep Slahpetjåhkka and Abmelvaratj. At Mt Pältsa, I surveyed P. radicatum subsp. laestadianum and L. nivalis.

Coordinates from previously visited sub-localities were downloaded from Artportalen.se (SLU Artdatabanken 2020c) and entered into a Garmin Etrex Summit HC that was used during the

(11)

10

data collection. Practically all sub-localities were surveyed (a few exceptions due to errors in entering the coordinates into the GPS and inability to get to some of the sub-localities), along with essentially all suitable habitats for the species surveyed. The area where the surveyed plant species grow is clearly defined from the surroundings (Mora Aronsson, personal communication). A. humifusa grow on and in cracks of serpentine bedrock that is easily recognized and where almost no other plants occur. B. linearis, V. rupestris subsp. relicta and P. scandinavica grow in screes, and these habitats were easily separated from non-suitable habitats. Papaver radicatum subsp. laestadianum and L. nivalis only grow in the upper slopes of Mt Pältsa, which are easily delimited from the surroundings (Mora Aronsson, pers. comm., Aronsson et al. 2013).

The number of individuals were counted on each sub-locality. The counting unit for each species are listed in Table 4, following the guidelines in Aronsson et al. (2013). The area for each sub-locality varied, when I found one specimen of a species, the coordinate was specific.

In other places, there was no clear beginning or end of the ramets. In such cases, I would take new coordinates every 20-25 metres to divide the population (if there were no precise coordinates for me to use). If a sub-locality had clear ”edges”, I tried to get a coordinate from the middle of the assemblage. Elevation was noted for the majority of the sub-localities.

Table 4. Units for counting population sizes for the species surveyed in Padjelanta National Park and on MT Pältsa in 2019. (Aronsson et al. 2013)

Species Unit for counts

Arenaria humifusa Tufts

Braya linearis Tufts/rosettes

Luzula nivalis Tufts

Papaver radicatum subsp. laestadianum Tufts

Primula scandinavica Flowering plants/vegetative rosettes Viola rupestris subsp. relicta Flowering plants/vegetative rosettes

All observations have been reported to Artportalen.se. A summary of the findings has been sent to Artdatabanken and SEPA, as this study was part of the biogeographical monitoring of species for the EU Habitats-directive.

Climate data

For Muddus National Park, Padjelanta National Park and Torne Lappmark, climate data (amount of precipitation per month and year, daily and annual mean temperature) was downloaded from Luftwebb from 1990 to 2018 to be used in the regression analyses (Table 5) (SMHI 2020a). The length of the growing season was calculated as the number of days with a mean temperature above 5°C. I used the first 30 days of spring with daily mean temperatures above 5°C, including five days below the threshold, as the start of the growing season. The end of the growing season was assumed to be the last 30 days of autumn with daily mean temperatures above 5°C, including five days below the threshold. Spring starts no earlier than 15^th of February and fall starts after the 1^st of August at the earliest. These assumptions are taken from the SMHI’s rules for seasons and climate data, with an adapted rule for growing season length in order to standardize the calculation (SMHI 2019, SMHI 2020b, SMHI 2020c).

(12)

11

Table 5. The climate data from the study site was downloaded from Luftwebb using the coordinates presented (SMHI 2020a).

Site Coordinates (SWEREF 99 TM)

Muddus National Park 727879, 7416460

Padjelanta National Park 571860, 7482481

Torne Lappmark 652640, 7584446

Statistical analyses

I tested for differences in population sizes between years using Wilcoxon signed-rank test and ANOVA in combination with Tukey tests (Zar 2010, Pace 2012, Beckerman et al. 2017). All populations surveyed two times were tested for statistical differences between years with the Wilcoxon signed-rank test. I used non-parametric tests as the data was not normally distributed.

Outliers were excluded in the tests of Viola rupestris subsp. relicta, Primula scandinavica and Papaver radicatum subsp. laestadianum (Appendix 1), based on boxplots and boxplot-outliers command in R (R Core Team 2019). Sub-localities 10 (outlier in year 2000), 11 and 16 (outliers in years 2000 and 2019) were excluded from the analysis of V. rupestris subsp. relicta, sub- locality 3 (outlier in year 2000) was excluded from the analysis of P. scandinavica and sub- localities 2 (outlier in year 2011), 4 (outliers in years 2011 and 2019), 14 (outlier in year 2011) and 19 (outlier in year 2019) were excluded from the analysis of P. radicatum subsp.

laestadianum. The analysis of V. rupestris subsp. relicta was non-significant with outliers included, and significant with outliers removed. The analysis of P. scandinavica was non- significant with outliers included as well as removed. The analysis of P. radicatum subsp.

laestadianum was significant with outliers included, but had higher significance with outliers removed.

All populations surveyed three or more times and with several sub-localities were tested for statistical differences in population sizes with ANOVA tests. The data were transformed to achieve normal distribution using log()-command. If zeros were produced in the process, +1 was added to the formula (log(x+1)). If statistical difference was found between years, Tukey- test was performed to see where the differences are. The population of Platanthera obtusata in Abisko was not analysed as there are only three sub-localities and the range between them is large (sub-localities 1 and 2 have 2-19 individuals while sub-locality 3 have 84-388 individuals, see Appendix 1 1). Arenaria humifusa on Bellopudat, Potentilla hyparctica on Gáhpesoajve and Primula scandinavica on Unna Titer were surveyed 2012, but this year was excluded in the ANOVA-analysis because the data consisted only of the total number of individuals and no counts for the sub-localities.

I tested whether different climatic variables could explain differences in total plant population size using regression analyses for all localities where plant population sizes had been estimated at three or more occasions (years). No regression analyses were done for the plants surveyed at Mt Pältsa as I only had two survey years with population counts for these populations. The tested climate variables included growing season length and mean temperature and amount of precipitation for the whole year, the summer (June-September) and the winter (December- March). The climatic variables used were from both the same year as the population was surveyed (t), and from the year prior to the survey (t-1). I tested the climate variables for the whole year for correlations before performing the regression analyses. In the regression analysis

(13)

12

for P. obtusata, only the data from sub-locality 3 was used as the two other sub-localities had so few plants compared to sub-locality 3.

All statistical analyses were done in RStudio version 1.2.5033 (R Core Team 2019).

Results

All species surveyed in Padjelanta National Park and at Mt Pältsa in 2019 were found and counted except for Braya linearis. The reason why B. linearis was not found is most likely human error, rather than its disappearance from the site. It is a rather small plant that can be hard to find if you have not seen it before, which I had not. The population of Luzula nivalis at Mt Pältsa is presumably underestimated as I found it difficult to find the species and to distinguish L. nivalis from other Luzula species in the area.

Changes in population sizes

Among populations that have been surveyed on two occasions there were statistically significant differences in mean subpopulation size between inventory years in Arenaria humifusa on Oarjep Slahpetjåhkka (decrease), Viola rupestris subsp. relicta (increase) and Papaver radicatum subsp. laestadianum (increase) (Table 6, Figure 2).

Table 6. Population sizes for populations surveyed two times in Padjelanta National Park and on Mt Pältsa. The V-value and the p-value given by the Wilcoxon signed-rank test are presented, as well as the direction of change.

Reference to the boxplots visualizing the tests are also presented (Figure 2). * denotes that outliers have been excluded in the dataset. Number of individuals each year can be found in Appendix 1.

Site V p Change Figure

Arenaria humifusa Vietjervaratj 451 1 Non-significant 2a Oarjep

Slahpetjåhkka

121 0.037 Decrease 2b

Abmelvaratj 167 0.91 Non-significant 2c Primula

scandinavica*

Oarjep

Slahpetjåhkka

9 0.44 Non-significant 2d Viola rupestris subsp.

relicta*

Oarjep

Slahpetjåhkka

21.5 0.031 Increase 2e

Mt Pältsa

Papaver radicatum subsp. laestadianum*

Mt Pältsa 14.5 0.0008 Increase 2f

(14)

13

a) b)

c) d)

e) f)

Figure 2. Sub-locality population sizes in the years 2000, 2005 or 2011 (left) and 2019 (right). The boxplots present the median (the bold line), the data between the 1^st and 3^rd quartile (the box), maximum and minimum value excluding outliers (whiskers) and outliers (dots).

Among populations that have been surveyed at three or more occasions, B. linearis had a statistically significant difference in mean population size between years (Table 7, Figure 3).

The population size B. linearis are the same in 2004 and 2010, and in 2005, 2006 and 2008, respectively (letters “a, b” and “c”, Figure 3b). Year 2004 and 2010 are not the same as years 2005, 2006 and 2008. The population size in 2012 differs from all other years (letter “a”, Figure 3b). The population of A. humifusa on Bellopudat show an increasing trend 2008, but it was not statistically significant (Figure 3a). The population size of Potentilla hyparctica on Gáhpesoajve have not changed over the years (Figure 3c) and neither has the population of Primula scandinavica on Unna Titer (Figure 3d). However, the population mean for each year show a slight decreasing trend for the latter.

(15)

14

Table 7. Analysis of variance test of the population sizes of plants surveyed three or more times in Padjelanta National Park. Degrees of freedom (df), F-value and p-values are presented.

Site df F p

Arenaria humifusa Bellopudat 4, 55 1.70 0.16

Braya linearis Unna Tukki 5, 54 5.53 0.0004

Potentilla hyparctica Gáhpesoajve 2, 18 0.17 0.84 Primula scandinavica Unna Titer 4, 20 0.46 0.77

a) b)

c) d)

Figure 3. Population sizes (mean ± 95% CI) each survey year for populations surveyed more than two times. In plot b), values with the same letter are not significantly different (Tukey). The results from the analysis of variance are presented in Table 7.

Relationships between environmental variables and population size changes

Using the data for each year as observations, annual mean temperature and growing season length were positively correlated in Padjelanta National Park, Torne Lappmark and Muddus National Park (Table 8). As growing season was estimated using the mean daily temperatures and the correlation between the two variables was significant, growing season length was not used in the regression analyses.

(16)

15

Table 8. Correlations between climate variables annual mean temperature, annual precipitation and length of the growing season for the three sites where plant population data has been collected more than three times. The correlation coefficient is presented for each correlation analyses. * denotes statistically significant results.

Climate variables Padjelanta National Park

Torne Lappmark Muddus National Park

Temperature x precipitation 0.04 0.34 0.07

Temperature x growing season 0.48* 0.48* 0.40*

Precipitation x growing season -0.25 0.15 -0.02

Relationships between environmental variables and total population size changes were found for some populations at all three sites (Table 9, 10 & 11).

Population size changes and temperatures

The population of A. humifusa on Bellopudat increases in size with higher annual mean temperatures the year prior to the surveys and to higher mean winter temperatures the same year as the surveys. On Unna Tukki, the population of B. linearis increases with higher mean annual and mean summer temperatures the same year as the surveys, and decreases in size with higher mean summer temperatures the previous year. Potentilla hyparctica on Gáhpesoajve increases while Silene involucrata in Muddus National Park decreases with higher mean annual temperature the year prior to the surveys. The population of Platanthera obusata in Abisko decreases with higher annual mean temperatures and mean summer temperatures the year prior to the survey, and increases with higher mean summer temperatures the same year as the surveys.

Population size changes and precipitation

The population of P. scandinavica on Unna Titer increase with increased winter precipitation the same year as the surveys, and with increased winter and annual precipitation the year prior to the surveys. P. radicatum subsp. radicatum increases in population size with increased annual precipitation the year of the surveys. In Muddus National Park, S. involucrata decrease with increased winter precipitation the year prior to the surveys.

Table 9. Significant relationships between total population sizes and mean temperature for winter, summer and the whole year the same year as the surveys (t) and the year prior to the surveys (t-1). Details of the analysis are presented in Table 11.

Species Site Annual

temp.

t-1

Annual temp. t

Summer temp. t- 1

Summer temp. t

Winter temp.

t-1

Winter temp. t

Arenaria humifusa

Bellopudat + +

Braya linearis Unna Tukki + - +

Potentilla hyparctica

Gáhpesoajve + Platanthera

obtusata

Abisko - - +

Silene involucrata

Muddus -

(17)

16

Table 10. Significant relationships between total population sizes and amount of precipitation for winter and the year the same year as the surveys (t) and the year prior to the surveys (t-1). As there were no significant results for precipitation during summer, results from those analyses are not presented. Details of the analysis are presented in Table 11.

Species Site Annual

precip. t-1

Annual precip. t

Winter precip. t-1

Winter precip. t

Primula scandinavica Unna Titer + + +

Papaver radicatum subsp.

radicatum

Nissoncorru +

Silene involucrata Muddus -

(18)

17

Table 11. Regression analyses of the plant populations in Padjelanta National Park, Torne Lappmark and Muddus National Park. The species, the environmental variables (mean temperature and amount of precipitation for the whole year (y), the winter (w) and the summer (s)), b (incline of graph), p and the R² are presented.

Results are presented for both the year prior to the surveys, and the same year as the surveys. As there were no significant results for precipitation during summer, results from those analyses are not presented. Statistically significant results are presented as plots in Figure 4. Species and sites are: Arenaria humifusa a) on Bellopudat and b) in the valley between Unna Tukki and Stuor Tukki; Braya linearis on Unna Tukki; Potentilla hyparctica on Gáhpesoajve; Primula scandinavica on Unna Titer; Papaver radicatum subsp. radicatum on Nissoncorru;

Platanthera obtusata in the Abisko area and Silene involucrata in Muddus National Park (Appendix 1).

Sites and species

Env.

variable

Previous year Same year

Padjelanta National Park Arenaria humifusa (a)

b p R² b p R²

Temp.

(y)

278 0.05 0.59 32.6 0.70 -0.20

Temp.

(s)

32.7 0.81 -0.23 15.7 0.86 -0.24 Temp.

(w)

-3.09 0.96 -0.25 75.6 0.05 0.58

Precip.

(y)

0.62 0.31 0.06 -0.24 0.68 -0.19 Precip.

(w)

0.68 0.67 -0.19 1.09 0.38 -0.004

Arenaria humifusa (b)

Temp.

(y)

2.5 0.90 -0.49 2.50 0.88 -0.48

Temp.

(s)

-13.1 0.20 -0.46 -2.21 0.83 0.46 Temp.

(w)

2.34 0.70 -0.37 5.01 0.09 0.73

Precip.

(y)

-0.07 0.17 0.53 -0.01 0.88 -0.48

Precip.

(w)

-0.06 0.85 -0.47 0.08 0.74 -0.40

Braya linearis

Temp.

(y)

-156 0.55 -0.13 203 0.03 0.66

Temp.

(s)

-332 0.04 0.64 226 0.01 0.78

Temp.

(w)

124 0.16 0.29 43.6 0.54 -0.13

Precip.

(y)

0.39 0.69 -0.20 1.00 0.20 0.21

Precip.

(w)

2.19 0.32 0.05 0.95 0.62 -0.16

(19)

18 Potentilla

hyparctica

Temp.

(y)

210 0.04 0.90 -30.1 0.75 -0.41

Temp.

(s)

159 0.26 0.32 -58.7 0.63 -0.29

Temp.

(w)

-60.5 0.26 0.33 27.9 0.49 -0.11

Precip.

(y)

0.54 0.36 0.12 0.26 0.81 -0.44

Precip.

(w)

-0.28 0.84 -0.46 0.66 0.55 -0.20

Primula scandinavica

Temp.

(y)

29.8 0.46 -0.07 24.4 0.16 0.29 Temp.

(s)

0.14 1 -0.25 20.6

0.28 0.11 Temp.

(w)

-1.60 0.92 -0.25 14.8 0.13 0.34

Precip.

(y)

0.24 0.04 0.62 0.09 0.49 -0.09

Precip.

(w)

0.61 0.02 0.70 0.51 0.01 0.78

Torne Lappmark Papaver radicatum subsp.

radicatum

Temp (y).

-16.2 0.76 -0.18 52.1 0.14 0.27 Temp

(s).

-27.3 0.61 -0.13 40.3 0.26 0.09 Temp.

(w)

-16.5 0.38 -0.02 13.2 0.42 -0.04

Precip.

(y)

0.16 0.70 -0.16 0.52 0.05 0.47 Precip.

(w)

0.93 0.24 0.12 -0.53 0.18 0.19

Platanthera obtusata

Temp.

(y)

-133 0.01 0.59 68.9 0.13 0.19

Temp.

(s)

-122 0.01 0.57 87.1 0.04 0.39

Temp.

(w)

-21.2 0.37 -0.01 -16.2 0.45 -0.05

(20)

19 Precip.

(y)

0.13 0.80 -0.13 0.62 0.06 0.32

Precip.

(w)

0.04 0.97 -0.14 -0.13 0.82 -0.13

Muddus National Park Silene

involucrata

Temp.

(y)

-7.48 0.02 1 -3.92 0.44 0.20

Temp.

(s)

3.02 0.71 -0.61 -0.60 0.92 -0.97 Temp.

(w)

-1.06 0.50 0.01 -1.17 0.56 -0.18

Precip.

(y)

-0.03 0.41 0.28 -0.02 0.36 0.43 Precip.

(w)

-0.23 0.01 1 -0.53 0.21 0.79

Discussion

Species monitored 2019

One of the aims of this study was to see if the plant population sizes had changed over the last 20 years. In the populations I surveyed 2019, the Arenaria humifusa population on Oarjep Slahpetjåhkka had decreased, while the Viola rupestris subsp. relicta population and the Papaver radicatum subsp. laestadianum both had increased since the last survey (Figures 2b, f

& g, Table 6). For the other populations I could not detect any change in population sizes since the last counting.

The site of A. humifusa where the population size had decreased is at lower elevation (610-674 m a.s.l.) than the other sites of the same species (Vietjervaratj 643-750 m a.s.l., Abmelvaratj 891-923 m a.s.l.). It is a rather small, isolated hill of serpentine rock close to the south-facing mountain Oarjep Slahpetjåhkka, and there is little room for distribution shifts. Vietjervaratj and Abmelvaratj offers more space to disperse as they cover a larger area. The summer of 2018 was very hot in Sweden, causing great forest fires and shortage of water (SMHI 2018). The hot and dry conditions might have had a greater impact on the small hill with less water supply from snow beds than the larger and higher elevated mountains. Scotland experienced low water levels and dry soils in the summer of 2018, but seemed to cope well with the heat wave according to Undorf et al. (2020). However, they argue that if heat waves becomes a more frequently occurring phenomena, this may greatly impact Scotland and measures have to be taken to withstand future climate change (Undorf et al. 2020).

The observed increase in population size of V. rupestris subsp. relicta compared to the survey in year 2000, is surprising. The taxon grows on moist soil and is expected to be negatively affected by drought. Drought is one of the threats to alpine plants as the temperature rises, because it decreases the availability of nutrients (Körner 2003, Winkler et al. 2019). Year 2018 was, as mentioned above, a hot and dry summer. Therefore, I would have expected the population to have decreased in year 2019 from the stress in 2018. There is little snow

(21)

20

accumulated in the summer on Oarjep Slahpetjåhkka as it is south facing, with an early snowmelt, which makes drought even more likely. Since the last survey of V. rupestris subsp.

relicta was 19 years earlier, events during other years may explain the increase.

The high temperatures in 2018 may have benefitted P. radicatum subsp. laestadianum at Mt Pältsa, as it is not limited by water-availability. Water is rarely in short supply due to the presence of large snow beds even in mid-August in the area (personal observation, Mora Aronsson, personal communication). The increased growing season length and higher temperatures are changing the limitations that alpine plants are adapted to, and this may change their reproductive success and productivity rates (Körner 2003). In Abisko, Sweden, an increase in number of flowers as well as earlier flowering have been observed due to higher temperatures in warming experiments of two common species; Andromeda polifolia L. and Rubus chamaemorus L. (Aerts et al. 2006). Warming experiments in cold habitats elsewhere have shown increased vegetative growth for plants, and after a few years also increased reproductive effort by earlier formation of flower buds (Aerts et al. 2006, Crawford 2008). This effect seem to wear off a couple of years into the study due to depletion of stored plant reserves of nutrients or exhaustion of soil nutrients (Mølgaard & Christensen 1997, Crawford 2008). In other words, the soil’s recycling of nutrients through the breakdown of litter did not match the plants increased growing rates due to increased temperature, leading to a shortage of nutrients (Crawford 2008). So, if the increased population size of P. radicatum subsp. laestadianum is caused by the high temperatures of year 2018, the increase may wear off as the nutrients are depleted. However, it is important to keep in mind that the only thing two repeated surveys can tell us is the difference between those two survey-years, it is not possible to find the mechanisms behind the change without more surveys. It can be normal fluctuations in the population dynamics, or a response to environmental cues, but with only two datapoints, no model can be made and no certain explanation to the changes can be given.

Population surveyed several years

The populations that have been surveyed more than two times can be discussed more extensively for potential trends and their explanation. For Braya linearis on Unna Tukki, the ANOVA-analysis showed a significant change in the population (Figure 3b, Table 7). The population increased until 2006, then it decreased and had its lowest number of plants in year 2012. The population has varied greatly in numbers, reaching a peak in 2006 with 670 tufts. In year 2012, only 133 tufts were found. This may be normal population dynamics for B. linearis as it has a large seedbank and can reappear at sites from which adult plants have vanished (Naturvårdsverket 2011c). A. humifusa on Bellopudat had a peak in number of tufts in year 2008 with 644 tufts found, however, the change was not large enough to be statistically significant (Figure 3a, Table 7). In contrast to the low number of B. linearis in year 2012, A.

humifusa had a high number of tufts, 611, this year (Appendix 1).

Population size changes and temperature

The population size A. humifusa on Bellopudat increased with higher annual mean temperatures the year prior to the surveys and with higher mean winter temperatures the same year as the surveys (Table 11). The high temperature the year before may increase production of seeds that then germinate the following year. Seeds of alpine plants usually germinate in spring, shortly after snowmelt, as germination is triggered by the increasing spring temperatures (Bliss 1962, Körner 2003). Since A. humifusa is perennial, the adult plants response to higher temperature

(22)

21

is also important to consider. The chances of survival, successful establishment and development of shoots may have increased for the adult plants as well. The formed shoots form new tufts that can be counted as a “new” plant the following survey, even though it is technically the same individual (clonal growth) (Naturhistoriska riksmuseet 1998). The increase in population size with increased mean winter temperatures could be due to earlier snowmelt and higher temperatures early in the growing season. A. humifusa may be able to utilize the additional time to grow efficiently, resulting in more individuals establishing that year.

However, several studies states that alpine plant seeds need low temperature periods to germinate (Cavieres & Arroyo 2000, Körner 2003, Cavieres & Sierra-Almeida 2018, Gremer et al. 2020). This has probably evolved to avoid seeds germinating in late summer, shortly after dispersal, to avoid seedlings being killed when the winter comes (Körner 2003). If that is the case with A. humifusa, the high winter temperatures may affect something else in the population than seed germination, for example a faster establishment and higher survival for new shoots formed the previous summer. If the plants become active too early, though, a sudden cold spell or snowstorm can damage or kill the young plants.

Potentilla hyparctica on Gáhpesoajve increased with higher annual mean temperatures the year prior to the surveys, showing the same response as A. humifusa. The species does not have vegetative reproduction (Svalbardflora.net 2019), but increased seed production and germination rates may explain the relationship between higher mean annual temperature and increased population size.

The population size of B. linearis on Unna Tukki increased with higher mean annual temperatures and with higher mean summer in the same year as the surveys were done (Table 11). This may indicate that the large seedbank of B. linearis can be activated directly when conditions are right, as the seeds are already in the ground (Naturvårdsverket 2011c). However, B. linearis also show a decrease in population size with higher mean summer temperatures the year prior to the surveys. This indicates that high mean temperature the previous summer decreases the number of individuals the following year. The trends of population size change and mean summer temperatures the previous year and the same year contradicts each other as one indicates a negative relationship with higher mean summer temperature and the other a positive relationship, respectively. Maybe the seeds that germinate the year of the high summer temperature increases the population size, but then the warm temperature takes a toll on them and they do not make it to the next year, meaning the population decreases the following year.

Why some of the species benefit from higher temperatures the previous year and some the present year can be due to different ecological adaptations. Population sizes of A. humifusa and P. hyparctica may increase the year after the high temperatures because beneficial conditions in the previous year make it possible for them to produce more seeds ready to germinate next year. This can explain why A. humifusa increased in 2012 and B. linearis decreased the same year. In 2011, a high mean annual temperature was recorded, and A. humifusa increased the following year. 2012 had a lower mean temperature than 2011, and B. linearis decreased that year, consistent with the trend of population size decrease with high mean summer temperatures the previous year.

Similar to the trends of B. linearis, the population of Platanthera obtusata in Torne Lappmark increases with higher mean summer temperatures the same year as the surveys are done (Table 11). However, it decreases with higher annual temperature and higher mean summer

(23)

22

temperature the year prior to the surveys. P. obtusata may, just as B. linearis, benefit from the high temperatures and grow more individuals one year, but then become affected by the warmth and form less individuals the next year. The growing sites of the population do not get as much rain as surrounding areas (Brita Svensson, pers.comm.), warm summer temperatures may therefore dry out the soil and stress the orchids which causes a decrease the following year. The present population have, however, shown a great variation in numbers over the years (Mora Aronsson, personal communication – no data available), some years there have been no observations at all, while thousands of specimens were found other years. Other orchids in Norrbotten County exhibit the same variation in population size (Persson 2006, 2014). In north- western Russia, near the Scandinavian countries, increased abundance of several orchid populations have been observed as a response to high temperatures, in contrast to the relationship found in this report (Blinova 2008). A negative relationship between population sizes and the number of days with frost the previous growing season was also noted for the orchids in Russia. Additionally, the number of new shoots were negatively correlated with snow depth (Blinova 2008). In Norway, Skrede et al. (2018) discovered new sites with P. obtusata but still redeems it as threatened, mostly by ecotourism, increased grazing by reindeer and sheep and the use of motor vehicles where the species grows. They also argue that due to global warming, the climbing alpine treeline might ruin growing sites of P. obtusata, as well as the predicted increased frequency of heavy rainfall can cause landslides that will also endanger the growing sites (Skrede et al. 2018). The treeline is very close to the population of P. obtusata in Abisko, but as mentioned in the introduction, the climate here is very similar to that above the treeline elsewhere, due to the cooling effect of the large lake Torneträsk. Maybe the local climate is more important than the occurrence of trees for the orchid.

The population of Silene involucrata in Muddus National Park showed a decrease in population size in response to higher annual mean temperatures the year prior to the surveys (Figure 4g, Table 11). This can be due to that the higher temperature stabilize the soils, ceasing the soil- movement that S. involucrata is adapted to. However, since the dataset only included three years, and two of them were 2017 and 2018 (Appendix 1), the relationship should be interpreted with caution.

An explanation to consider when discussing changes in plant population sizes are the preformation of flower and leaf buds. Plants can prepare the following years growth by forming buds that will start to develop and grow as soon as the conditions are right, usually when the snow disappears in spring. By doing so, they can produce flowers early in the season and finish their reproductive cycle the current year. Some species form several generations of buds and hence get a lot of flowers some years (Körner 2003). In the Tyrolean alps, Austria, Saxifraga oppositifolia L. forms buds in autumn that in spring develops into flowers (Larl & Wagner 2006). In a comparison of flowering and the reproductive cycle (from floral initiation to fruit maturation) of the species at sites at different elevation, Larl & Wagner (2006) observed different flowering dates as well as a six week difference in performing the reproductive cycle of flowering; the population at higher elevation finished the cycle in four months, the population at lower elevation in two and a half months. This difference highlights the variation in flowering and length of the seasonal cycle within a taxon, and presumably between taxons as well. S.

oppositifolia is a flexible generalist and can adjust flowering time and reproductive effort according to present conditions (Larl & Wagner 2006). The species in this paper are specialists, and not as flexible as the example above. Without knowledge of the different species’

(24)

23

adaptations and responses to changing conditions, it is difficult to say if preformation of buds could have caused the changes seen in the populations.

Population size changes and precipitation

The population size of Primula scandinavica on Unna Titer increased with increased annual precipitation the year prior to the surveys and increased winter precipitation both the year prior to the surveys and the same year as the surveys (Table 11). The species grows in habitats with moist conditions, and if there is a lot of precipitation during winter, this prolongs the snowmelt and keeps the growing sites moist during the growing season. A warmer and wetter climate in Norrbotten County may benefit P. scandinavica, as the seed production and seed germination rates may improve with higher temperature. Disturbance by herbivores create bare soil that the seeds need to germinate, it is therefore essential that grazing and trampling continues to preserve suitable habitats for P. scandinavica. In Norway, P. scandinavica occurs at more sites than in Sweden. Many of these sites have historically been used for traditional land-use, such as harvesting grass for winter and grazing by cattle. Since these activities have seized in many places, the previous open grasslands are overgrown by shrubs and trees (Wehn & Olsson 2015).

There is a risk of extinction in the future due to habitat degradation, but at the moment there are high densities of P. scandinavica left at the sites undergoing the transition (Wehn & Olsson 2015).

Papaver radicatum subsp. radicatum on Nissoncurro, Torne Lappmark, increased with increased amount of precipitation the same year as the survey (Table 11). As this species is adapted to soil movement, years with much rain improve conditions for establishment of new individuals. Less rain and snow may reduce soil movement. That will mean less suitable growing conditions for P. radicatum subsp. radicatum as well as increased competition from other species that may establish at the same sites. On Greenland, a relative to the P. radicatum subspecies described in this report, have been observed to increase its biomass as well as its number of flowers as a response to experimental warming. This increase was followed by a decrease in biomass the following years, probably due to nutrient deficiency (Mølgaard &

Christensen 1997). Maybe the P. radicatum subspecies in Sweden can benefit from higher temperatures, but it has to be in combination with enough precipitation for their growing sites to continue soil movement. As many alpine plants are adapted to low levels of nutrients, increased nutrient levels may actually harm them, and instead benefit more fast-growing species.

The population of S. involucrata decreased with increased winter precipitation the year prior to the surveys (Table 11). The species is dependent on soil disturbance, but too much precipitation can make them too unstable for the species to establish (Naturvårdsverket 2011d). As mentioned above, this population was only surveyed three times and therefore the results are to be interpreted with caution.

I only tested for linear relationships between plant population sizes and climate variables in this project. In ecology, non-linear relationships are however common (Per Toräng, personal communication). Growth and reproductive success rarely depend on a single variable, but several variables together. For example, high mean annual temperature may increase population numbers, if the same year also have a good amount of precipitation. If there is no precipitation, the high temperatures can mean a decrease in population numbers due to lack of water. These

(25)

24

relationships are tested in multivariate analyses, but the plant population data I had at my disposal had too few datapoints to do such an analysis.

Although I could not test for multivariate relationships, I visually observed a trend among several of the species: all populations surveyed year 2010 decreased in numbers compared to the closest survey year prior to 2010 (Appendix 1). In Padjelanta National Park, A. humifusa on Bellopudat, B. linearis on Unna Tukki, P. scandinavica on Unna Titer and P. hyparctica on Gáhpesoajve all decreased in population size. In Torne Lappmark, P. obtusata in Abisko and P. radicatum subsp. radicatum on Nissoncurro both decreased as well. Year 2010 had a low mean temperature, low amount of precipitation and a short growing season, and this seemed to affect all the species negatively. All populations increased after 2010, except for B. linearis at Unna Tukki. It would be interesting to see if the population has increased again since it was last surveyed in year 2012.

Another thing to consider is that regression analyses were only made for two environmental variables, and these were mean temperatures and amounts of precipitation for the whole year, the winter and the summer. Extreme weather events are possibly the most important environmental variable to take into account when predicting the future of alpine plants, and this variable was not considered in any of the analyses. Changed soil moisture conditions due to heat waves and heavy rainfall can destroy growing sites of species such as V. rupestris subsp.

relicta, P. obtusata, B. linearis, P. radicatum and S. involucrata. Most of the species in this study are dependent on disturbances, but these disturbances have to be on the right “level” for the species to benefit from them. Temperature extremes are also important to take into account, as a mean annual temperature does not tell you much more than a trend. A very cold winter followed by a very warm summer can give the same mean annual temperature as a mild winter and a cold summer. Summer temperatures are most interesting to look at, since the plants are usually protected under the snow during winter in alpine areas. Also, growing season length is one of the most important variables to consider when investigating the trends of alpine plants, since plants in alpine habitats are adapted for short growing seasons with relatively low summer temperatures (Körner 2003). If the growing season gets longer and warmer, this can change the productivity and reproductive success of species considerably. With more data available, growing season length should be considered in multivariate analyses with precipitation frequency and daily mean temperatures during the growing season.

Habitat factors and dispersal

Several studies state that species exhibit range shifts as a response to climate warming (Grabherr et al. 1994, Körner 2003, Pauli et al. 2007, Frei et al. 2010, Winkler et al. 2019). These upward shifts are not as likely to occur for the species discussed in this report since they have very specific habitat requirements and limitations (Mora Aronsson, personal communication). For a species to shift its range, there has to be suitable habitats close-by, and “corridors” for it to get there (Körner 2003). As many alpine plants have slow growth and dispersal rates and the changes to the climate is happening at a faster rate than ever before, many species will not be able to move fast enough to new sites (Körner 2003, Harrison 2020). The observations of elevational shifts of alpine plants are mainly for generalists, species that can find suitable habitats elsewhere and thereby migrate to new growing sites when the climatic conditions or competition from other species change (Mora Aronsson, pers. comm., Pauli et al. 2001, Pauli et al. 2007, Crawford 2008). Many specialist alpine species have been on their present growing sites for hundreds, if not thousands of years, since the last glaciation (Mora Aronsson, pers.

Changes in alpine plant population sizes inresponse to climate changeEvelina Rostö