Is palaeoecology a ‘special branch’ of ecology?

Abstract

Palaeoecology and ecology have a lot in common. However, the two disciplines have evolved almost separately, leading to an ongoing debate on how to link them better, as both would undoubtedly benefit from a higher degree of cross-fertilisation of ideas, research insights and questions. In this paper, we explore similarities and differences in the two branches of ecology over the past 40 years by assessing the publication output from an unusually large cohort of ‘academic siblings’ – researchers in ecology and palaeoecology who all share one of the strongest career-shaping influences on researchers; their supervisor. This was made possible by John Birks’ long and active career within ecology and palaeoecology, and his supervision of a large number of students in both fields (n = 12 and 21, respectively). Among them, John Birks’ academic progeny has published 934 papers in the international peer-reviewed literature. We collected information on the year, titles, keywords and journals of all these publications and used these data to assess the extent of, and find potential explanations for, the historical and present-day separation of ecology and palaeoecology. Despite considerable thematic overlap, there is a real and possibly widening division between the two branches that is visible both at the scale of individual research careers, journals and papers. We argue that rather than being unique to the relationship between palaeoecology and ecology, this reflects trends for thematic and methodological specialisation evident across the research landscape. We propose that both individual researchers and journals would benefit from, and can contribute to, closing the gaps developing between various ‘special branches’ in science. A good title, an informative abstract, a careful and thought-through selection of keywords and a focus on readability and avoidance of jargon will likely improve readership and impact both within one’s own ‘special branch’ and beyond.

Keywords

academic cohort cooperation counter-intelligence journal keyword network ordination publication scientific communication time

Introduction

As defined, ecology is the study and understanding of the relationship between living organisms and their environment (Birks, 2008). Similarly, palaeoecology is the ecology of the past, and aims to study and understand the relationships between organisms and their environments in the past (Birks, 2008). Following the principle of uniformitarianism, the past, present and future are not discrete entities but points in a time continuum where species and communities change and evolve. Ecology and palaeoecology are thus different approaches towards a common objective: the ecological understanding of the biosphere (Rull, 2010).

It is therefore not surprising that ecologists and palaeoecologists are working towards many of the same research challenges, and often simultaneously. As an example, in the 1980s and 1990s, the causes and consequences of ‘acid rain’ received considerable attention both in the ecological (Likens et al., 1996; Stoddard et al., 1999) and palaeoecological (Battarbee and Charles, 1987; Birks et al., 1990) literature. More recently, the impacts of climate change, for example, in causing species range shifts, has been a pervasive research topic in both ecology (e.g. Felde et al., 2012; Grytnes et al., 2014; Odland et al., 2010) and palaeoecology (e.g. Giesecke et al., 2008; Ohlemuller et al., 2012). Still, research on these and other shared topics remains largely separate, with little cross-fertilisation in terms of publications or projects that combine palaeoecological and ecological data and methods (but see, for example, Birks, 1996; Kapfer et al., 2013; Moe and Bjune, 2009; Vandvik et al., 2014; Willis et al., 2007). The two fields have different strengths, whereas ecology offers opportunities for high spatial-resolution experiments that can establish cause and effect, and have a broad range of predictor and response variables, palaeoecology offers unique insights into long-term landscape and population dynamics and can contrast our current Anthropocene setting (Ellis et al., 2012) with more natural past landscapes. There is thus clearly a large untapped potential for cross-fertilisation between ecology and palaeoecology, as repeatedly pointed out in the recent literature (e.g. Birks, 2012; Lindbladh et al., 2013; Willis and Birks, 2006).

So why has the divide emerged, and why does it remain so strong? Because palaeoecology usually exploits totally different kinds of data than other branches of ecology, it has developed its own set of field, laboratory and analytical methods, partly independent of other ecological disciplines. Over the years, a number of specialised journals in palaeoecology have emerged and, as a consequence, palaeoecological papers are less often published in the general ecological journals. In some institutions, palaeoecologists and ecologists belong to different departments. But the division could also be caused by more deeply rooted mental and cultural differences. Rull (2010) raises four potential explanations of the historical and present-day separation of ecology and palaeoecology: (1) the past–present dissociation characteristic of the human mind, (2) the diversity of fields of provenance of palaeoecologists, (3) the contrasting nature of the evidence and associated methodological differences, and (4) misunderstandings caused by the use of prefix ‘palaeo-’. Older studies indicate that tea preferences also differ between the two groups (Cambridge School of Botany, 1977). Whether the underlying causes are methodological/practical or psychological/cultural, there is no doubt that a lack of possibilities for exchange, discussion and learning contributes to the existing barriers between the two subjects (Davies et al., 2014 and references therein).

To explore differences and similarities between ecology and palaeoecology over the past four decades and to assess evidence for scientific cross-fertilisation or the lack of such, we made use of a unique natural experiment (sensu McArthur and Wilson, 1967). Thanks to the unusually long and productive research career of John Birks (see Birks (2014) for an overview), we have access to a large cohort of researchers within both ecology (n = 12) and palaeoecology (n = 21) who all share one of the strongest shaping forces that exist in academia – a common supervisor. John Birks’ research interests cover broad spectra of ecology and palaeoecology, as well as combinations of the two. His first paper was published in 1964, and his CV (August 2014) counts 21 books and edited volumes, 351 scientific papers on palaeoecology and pollen analysis and 137 scientific papers on modern plant ecology, plant geography, and bryology. He has supervised a number of students in both fields, and we consider that the publication output from this group of students will provide a suitable test case for evaluating the hypothesis that palaeoecology is, indeed, a ‘special branch’ of ecology. We hypothesise that John’s students will have been exposed to a research and supervisory environment optimal for creating an atmosphere of cross-fertilisation and learning across the two disciplines. Awareness of the scientific contributions from the other field should thus be high, and cultural resistance and ignorance creep should be low in this group of students. Any remaining differences should therefore reflect inherent differences in scientific focus, direction, interests or ‘modus operandi’ between the two branches of ecology. This cohort of scientific ‘siblings’ has published at least 934 papers in the international peer-reviewed literature. The onset of their careers spans almost 40 years (first student defended in 1974), and we make use of their output to investigate differences and similarities in the development of the fields of ecology and palaeoecology over the past 40 years. Specifically, we focus on the scientific overlap and exchange between the two branches, which we assess at the level of researcher’s careers, publication outlets (journals) and the individual research papers.

Exploratory data analysis and data display

Our intelligence-gathering provided information on the titles, journals, publication year and keywords of papers published by PhD students supervised by John Birks, counting in total 33 (10 in Cambridge, 22 in Bergen and 1 in Aberdeen), from 1974 to 2014.

Armed with the names of the 33 students John has supervised, we searched ISI Web of Knowledge version 5.13.1 in December 2013 for papers that they have published. All hits were added to a marked list, thereby excluding duplicate papers in cases of papers co-authored by multiple ex-students. The search results were checked so that we excluded authors who share a name with an ex-student. All the papers on the marked list including their keywords, where available, were then extracted and made into a spreadsheet. Many of the earlier papers did not have keywords registered in the ISI database, and these were added manually after downloading the paper. A small subset of these did not provide keywords at all, and so these were subjectively created from the title and abstract. Papers not captured by ISI Web of Knowledge have not been included in our analysis.

We found 934 papers published in 174 different journals. Journals that have changed name during the time period of interest were combined, for example, Plant Ecology which used to be Vegetatio. Prior to the analyses, we made an effort to combine keywords that obviously had the same meaning, such as plurals of the same word (diatom and diatoms) and differences in spelling (palaeoecology and paleoecology, CO₂ and carbon dioxide). This was done consistently for words that occurred more than thrice, but because of the large number of words, we were not able to do this for all words occurring only once or twice. We subjectively classified each of the 174 journals in which the papers were published, based on their preponderance for the two topics, into primarily ecological (94), primarily palaeoecological (39) and general (41), and we classified each of the 934 papers, based on the methods applied, into ecological (446), palaeoecological (438), mixed (45) or neither (5).

Rank publication curves of the cumulative number of publications against time since PhD defence for each ecological and palaeoecological ex-student reveal a large variation between the authors in terms of total publication output, career duration and paper accumulation rates (Figure 1). The students with the longest post-graduate careers are all palaeoecologists, reflecting that John started out supervising students within this branch. Short careers and lower total output belong to recent graduates or students who subsequently quit academic research. Some students started publishing several years prior to their PhD defence, usually through pre-graduate scientific employment. Overall, ecologists tend to have higher rank publication rates (Figure 1), but this may be explained by generational differences as more recent graduates within both fields generally have higher rank publication rates (results not shown).

Figure 1.

Rank publication plot showing the cumulative number of publications against time (years) for the ecological (blue) and palaeoecological (red) authors in the dataset. Note that time is standardised to each author’s PhD defence (which is set as year 0) so that negative values on the x-axis indicate publications prior to obtaining the PhD degree. All curves are smoothed.

Degree of specialisation among John’s scientific offspring

We took the subset of John’s former students who had at least four papers registered in the ISI database and plotted the proportion of their papers that fell into the ecological, palaeoecological or mixed category (Figure 2). The bulk of the output is from 21 of the 33 ex-students (13 palaeoecologists, 8 ecologists). Only three of these former students have published papers exclusively within one category, while the rest show varying degrees of cross-disciplinarity. A few of the former palaeo students, a Cambridge three in particular, have even gone over to the other side in the course of their careers.

Figure 2.

Percentage of ecological (E; blue), mixed (M; yellow) and palaeoecological (P; red) papers produced by John’s former students. Only authors with four or more publications are included in the figure. E92-E11 are students with an ecological PhD with the number indicating their year of graduation and P74-P04 are students with a palaeoecological PhD.

Keyword abundance and distribution

We illustrate the relative abundance of different keywords in the two disciplines by means of word clouds (wordle.net) of the dataset of keywords summed across all the ecological and palaeoecological publications (Figure 3). There is considerable keyword diversity in both fields. Keywords such as vegetation, change and climate are dominant in both disciplines, and model, temperature, plant, forest and analysis are less dominant but still equally frequent in both ecology and palaeoecology. These keywords reflect, broadly, the interests of John Birks and his students over the past 40 years and show that there are areas where the disciplines can inform each other. Other keywords are clearly most dominant in the palaeoecological papers, notably history, pollen and Holocene, whereas species, richness, global and diversity dominate in ecology. The word clouds also indicate a tendency for stronger dominance of a few, abundant keywords in palaeoecology relative to ecology.

Figure 3.

Word cloud of the keywords used in the publications included in this study. The plot on the left side (a) displays the keywords used in the ecology papers (446 papers), the plot on the right side (b) shows the keywords used in the palaeoecology papers (438 papers).

To illustrate whether journals are exclusive in their choice of papers, we counted the number of times each keyword was used in the different journals and illustrate this with a network analysis. In this analysis, we used only keywords that appear more than 15 times and journals with more than 15 publications, and we colour-coded the journals according to their category (Figure 4). The network analysis distinguishes the ecological journals (blue; left side) from the palaeoecological journals (red; right side), with journals of a more mixed publication profile (yellow) towards the left and the middle of the diagram. The ecological journals are characterised by a number of relatively low-frequency keywords linked to biodiversity (species richness, diversity, biodiversity, conservation), ecological processes (succession, disturbance) and quantitative methodology (ordination, model). The predominantly palaeoecological keywords are more abundant, and represent organisms and proxies often used in palaeoecological research (diatoms, plant macrofossils, pollen), methodological approaches (palaeolimnology, palaeoecology, pollen analysis), ecological responses (vegetation history, environmental change) and specific time-periods (Holocene, last glacial maximum). As a reflection of topics that are common in the two disciplines, the keyword climate is the most abundant and widespread across the literature surveyed, as are climate change, temperature, forest and vegetation which span the whole spectrum of the publication landscape surveyed. The journals in the middle of the diagram, from Nature to New Phytologist, also span a wide range of the publication landscape and publish both clearly ecological and clearly palaeoecological work, as ascertained from their keywords. The network suggests that these journals are prepared to publish articles across both ecology and palaeoecology as well as cross-disciplinary, although the subjects may differ in prevalence and publication rate.

Figure 4.

Network displaying the main journals where the papers included in this study have been published. The colours are given to separate the journals appearing more than 15 times with ecological journals in blue on the left side to more palaeoecological journals on the right side in red.

Gradients across the publication landscape

To investigate the main gradients of thematic variation within and between the ecological and palaeoecological research published by John’s academic progeny, as well as to explore changes in thematic content through time, we used ordination techniques. Analyses were performed on a multivariate dataset where each publication is a ‘sample’ and the keywords used in the publication are an attribute or ‘species’ of that sample. As ordinations can be heavily influenced by rare attributes (Legendre and Legendre, 2012), we only included keywords that appeared five or more times, and publications that had more than three of these keywords. This left us with a dataset of 272 publications and 205 keywords. Based on these publications, we first performed a Detrended Correspondence Analysis (DCA; Hill and Gauch, 1980) to assess the gradient length and evaluate whether unimodal or linear methods are appropriate for this dataset. The gradient length was 7.7 standard deviations, and we therefore chose unimodal methods (Lepš and Šmilauer, 2003). As the aim of this analysis is primarily exploratory, we used indirect ordination methods. The eigenvalues of the Correspondence Analysis (CA) are 0.76 and 0.65 for the two first axes, and the total inertia is 27.7. The analysis reveals clear thematic differences between palaeoecological and ecological papers, as they occupy partly distinct regions of the CA diagram (Figure 5a) with ecological papers dominating the left half of axis 1, and palaeoecological papers dominating the upper half of axis 2. The ‘species’ plot (Figure 5b) illustrates how this divide is driven by differences in research themes; on the left-hand side of the diagram, we find topics or keywords that are rarely addressed in palaeoecological papers, such as elevation, species distributions, spatial scales, recruitment and biodiversity, but are more often found in papers classified as ecological, whereas palaeoecological papers are characterised by keywords found at the upper right-hand side of the diagram, such as pollen stratigraphy, ordination, biome, post-glacial, plant geography and ¹⁴C calibration (Figure 5b). In the middle of the diagram and towards the negative end of axis 2, we find a mix of ecological and palaeoecological papers, as well as the mixed-methods papers, indicating that both research fields address questions relating to alpine, canonical correspondence analysis, climate, ecosystem models, global warming and species richness. There is thus evidence both for considerable thematic overlap and for distinct ecological and palaeoecological regions in the research landscape.

Figure 5.

(a) Ordination (CA) plot separating the papers according to the three paper categories, palaeoecological (red), ecological (blue) and mixed (yellow). (b) Ordination plot (CA) with the keywords explaining most of the variation between the papers.

Publication trends in time

To investigate whether the diversity of the research has changed through time, we explored the range of keywords from all the 934 papers amalgamated into 5-year periods since 1977. The richness of keywords through time is dependent on the number of individual keywords in the same way as pollen richness is dependent on the number of pollen grains counted for a time period (Birks and Line, 1992). Birks and Line (1992) solved this by rarefaction, a technique that gives the expected number of species for a given number of individuals. To get enough keywords to have sufficient resolution, we started the counting in the first 5-year period from 1978 to 1982 (called 1980). We find that keyword richness is lowest in the 1985 period, and increases from there on, reaching a maximum value in 2005, and flattening out thereafter (Figure 6). We find no difference in the rarefied keyword richness between ecological and palaeoecological papers nor between ecological and palaeoecological journal categories.

Figure 6.

Rarefied keyword richness in 5-year periods between ecological and palaeoecological papers included in this study. Note that the number of keywords is too low to start with the earliest publications; the first time period included is 1978–1982.

We also investigated temporal trends in cross-topical publications, that is, whether palaeoecological papers are published more or less frequently in ecological journals over time, and, correspondingly, whether ecological papers are more or less common in palaeoecological journals today than before. This was tested by evaluating a temporal trend in the proportion of papers classified as either palaeoecological or ecological and published in ecological and palaeoecological journals respectively. Because of over-dispersion, a quasi-binomial distribution was assumed and an F-test was used to evaluate the statistical significance. We found no statistically significant trend in the proportion of ecological papers published in palaeoecological journals (Figure 6; F = 0.29, p = 0.59), although none of the ecology papers included in this study was published in a palaeoecological journal before 1990. After 1990, a low but relatively constant number of what we have classified as ecology papers are found in palaeoecological journals. For the palaeoecological papers published, there has been a marked decrease in papers in ecology journals (Figure 6; F = 8.2, p = 0.007). These two patterns can have similar explanations as most palaeoecological journals are relatively young and journals have become more and more specialised.

The finding that more palaeoecological papers were published in ecology journals before 1990 is a pattern we can also spy in the most commonly used keywords. Keywords such as pollen analysis and vegetation history appear among the three most common keywords in ecology journals before 1990 (data not shown). After 1990, climate change is the most used keyword in all 5-year periods for papers published in ecology journals, and this is a keyword that is also consistent among the top five for papers published in palaeoecological journals in the same period. Apart from this, there is very little overlap among the most used keywords throughout the period. It also appears that the most commonly used keywords for palaeoecological papers often describe the site, time period or the method used in the study, while for ecological papers, keywords more often describe the ecological topics they cover.

Discussion and future perspectives

Our analyses indicate that based on the keywords used in publications in the research output of John Birks’ academic offspring, there is considerable thematic overlap between the two branches of ecology. This holds at the scale of researchers’ careers, journals and individual papers. A majority of the researchers investigated here have had at least some experience with publication involving methods from the ‘other branch’, and some have even defected fully over to the ‘other branch’. Both ecological and palaeoecological journals publish papers from the ‘other side’, and papers with mixed methods are not uncommon. The meeting ground between the two branches appears to be both thematic, for example, related to climate change or landscape and land-use history (e.g. Telford et al., 2006; Vandvik et al., 2014; Vandvik and Goldberg, 2006), and methodological, relating to, for example, the development and application of quantitative methods and approaches (e.g. Bertrand et al., 2011). As also mentioned by Rull (2010), the diversity of background can lead to larger gaps between ecology and palaeoecology. The overlap between the two topics included in this study can be a result of the fact that many palaeoecologists are trained in Earth Sciences (Bennett, 1986; Bradshaw, 2013), while John Birks and many of his students started their careers in botanical or biological departments where palaeoecologists and ecologists work side-by-side and the research focus is usually not as descriptive as in Earth Science. This is not the case for all countries and groups as summarised by Bradshaw (2013).

Yet, we also find clear differences between the two branches of ecology, despite the common thematic foci and scientific upbringing of John’s students. Some of these differences are related to methodological development and testing within the two branches. There is, for example, a substantial literature on the development and testing of methods reconstructing past climates not often used outside palaeo-sciences that could be used and tested in the other branch (e.g. Bjune and Birks, 2008; Nesje et al., 2014; Velle et al., 2005).

However, not all differences in keywords between the two fields reflect real scientific differences. Some are better explained as cultural differences between the different branches of ecology. Our analysis shows that palaeo papers tend to highlight specifics (time, place, method) both in keywords and in titles (analyses not shown here), whereas ecological papers more often highlight theory, process or patterns. This may disguise palaeo research of potentially broad interest and give it the appearance of a narrower scope than it deserves by the omission of general titles or keywords. The papers may appear descriptive and specific to the outside eye, despite cloaking a general message that could be of relevance outside the ‘special branch’.

We found stronger keyword dominance in the palaeoecological literature we reviewed, relative to the ecological literature. This is probably an artefact of our delimitation of the two fields. Palaeoecology is the branch of ecology that studies the past using proxies (Rull, 2010), rather than an equally ranked subject. John’s palaeo students have largely continued their careers within palaeoecology, whereas the ecological students have spread out to different branches of ecology, from biogeography and macroecology via plant and community ecology to pollination ecology. A stronger coherence in keywords would probably be found if we had analysed these other ecological branches individually. However, the data did not allow further divisions.

The development of a separate set of journals within palaeoecology over the past decades – many palaeoecological journals have been launched since 1970 – is paralleled by similar trends in other branches of ecology, and of science in general. This results from an exponentially growing research literature, driven by the ‘publish or perish’ regime and associated changes in research funding (Fanelli, 2010; Garfield, 1996). We find that this rise of specialised journals corresponds with a reduction in the proportion of palaeoecological papers published in the general ecological literature. This may be explained by reduced submission rates to the general journals or by less service from the palaeoecological community in the general journals as editors or reviewers, potentially leading to a decreasing understanding of the relevance of palaeoecological research in these journals. Whatever the cause may be, this trend of decreasing cross-over publishing will potentially lead to less exposure to literature from other scientific branches. This specialisation may to some extent be mitigated by the easy access to papers published in out-of-topic journals provided by searchable electronic databases, but successful identification of relevant literature from other fields in such searches depends critically on the quality and cross-disciplinary relevance of paper titles, keywords and other search parameters.

Conclusion

So is there cross-fertilisation between ecology and palaeoecology – or is palaeoecology a ‘special branch’ that conducts investigations on the side of, or perhaps even underground and therefore hidden from, the mainstream ecological research society? Our analysis suggests there is cross-fertilisation within single research papers, within individual researchers’ careers and within the publication profile of both ecological and palaeoecological journals, but also that there is room for more, and worryingly, that there is a negative trend that should be counteracted.

Based on these analyses, and our experience as members of a research group that combines ecological and palaeoecological work, we believe that palaeoecology and ecology do, indeed, have a lot in common and that increased cross-fertilisation is both possible and scientifically stimulating. Both journals and authors can do their share to stimulate cross-disciplinary reading, learning and, ultimately, usage and citation of research results. We have the following advice to scientists (including ourselves) with ambitions to reach outside their own ‘special branch’: think broadly in terms of audience when you publish! This goes for where you publish, but also for how you advertise your paper by means of title, keywords and abstract, and how you present results and write the paper. Avoid writing in code and using subject-specific jargon: instead, make general points clear and messages transparent. Spell out the general and broader implications of your research. Envisage a mixed audience when preparing your manuscripts, rather than writing for the narrowest possible definition of your ‘peers’. Keywords are critical to determine which searches, and whose searches, will find your paper.

Footnotes

Acknowledgements

We dedicate this paper to John Birks on the occasion of his retirement from the University of Bergen. We appreciate all the support, advice, scientific cooperation and friendship over the years and wish him a very happy retirement, good health and a continuing publication accumulation curve for the future! The authors acknowledge the FBI (former Botanical Institute) for providing an arena for knowledge exchange and data sharing. We thank the two reviewers Richard Bradshaw and Keith Bennett for their helpful comments and suggestions to improve this manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References

Battarbee

Charles

(1987) The use of diatom assemblages in lake-sediments as a means of assessing the timing, trends, and causes of lake acidification. Progress in Physical Geography 11: 552–580.

Bennett

(1986) The place of Quaternary palaeoecology within ecology. British Ecological Society Bulletin 17: 57–58.

Bertrand

Lenoir

Piedallu

. (2011) Changes in plant community composition lag behind climate warming in lowland forests. Nature 479: 517–520.

Birks

HJB

(1996) Contributions of Quaternary palaeoecology to nature conservation. Journal of Vegetation Science 7: 89–98.

Birks

HJB

(2008) Palaeoecology. In: Jørgensen

(ed.) Encyclopedia of Ecology. Oxford: Elsevier, pp. 2623–2634.

Birks

HJB

(2012) Ecological palaeoecology and conservation biology: Controversies, challenges, and compromises. International Journal of Biodiversity Science, Ecosystem Services & Management 8: 292–304.

Birks

HJB

(2014) A diverse scientific life. Journal of Paleolimnology 51: 113–137.

Birks

HJB

Line

(1992) The use of rarefaction analysis for estimating palynological richness from Quaternary pollen-analytical data. The Holocene 2: 1–10.

Birks

HJB

Berge

Boyle

(1990) A test of the land-use hypothesis for recent lake acidification by using hill-top lakes in southwest Norway – An extended summary. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 327: 369–370.

10.

Bjune

Birks

HJB

(2008) Holocene vegetation dynamics and inferred climate changes at Svanåvatnet, Mo i Rana, northern Norway. Boreas 37: 146–156.

11.

Bradshaw

RHW

(2013) Overview of terrestrial pollen data. In: Elias

(ed.) The Encyclopedia of Quaternary Science. 2nd Edition. Amsterdam: Elsevier, vol. 2, pp. 699–704.

12.

Cambridge School of Botany (1977) Flora Europaea – sonderheft of The Tea Phytologist, edition anniversaire. Cambridge: Cambridge University Botany School Tea Club.

13.

Davies

Colombo

Hanley

(2014) Improving the application of long-term ecology in conservation and land management. Journal of Applied Ecology 51: 63–70.

14.

Ellis

Antill

Kreft

(2012) All is not loss: Plant biodiversity in the Anthropocene. PLoS ONE 7: e30535.

15.

Fanelli

(2010) Do pressures to publish increase scientists’ bias? An empirical support from US States data. PLoS ONE 5: e10271.

16.

Felde

Kapfer

Grytnes

(2012) Upward shift in elevational plant species ranges in Sikkilsdalen, central Norway. Ecography 35: 922–932.

17.

Garfield

(1996) What is the primordial reference for the phrase ‘publish or perish’? The Scientist 10: 11.

18.

Giesecke

Bjune

Chiverrell

. (2008) Exploring Holocene continentality changes in Fennoscandia using present and past tree distributions. Quaternary Science Reviews 27: 1296–1308.

19.

Grytnes

Kapfer

Jurasinksi

. (2014) Identifying the driving factors behind observed elevational range shifts on European mountains. Global Ecology and Biogeography 23: 876–884.

20.

Hill

Gauch

(1980) Detrended correspondence analysis: An improved ordination technique. Vegetatio 42: 47–58.

21.

Kapfer

Birks

HJB

Felde

. (2013) Long-term vegetation stability in northern Europe as assessed by changes in species co-occurrences. Plant Ecology & Diversity 6: 289–302.

22.

Legendre

(2012) Numerical Ecology. 3rd English Edition. Amsterdam: Elsevier Science BV.

23.

Lepš

Šmilauer

(2003) Multivariate Analysis of Ecological Data using CANOCO. Cambridge: Cambridge University Press.

24.

Likens

Driscoll

Buso

(1996) Long-term effects of acid rain: Response and recovery of a forest ecosystem. Science 272: 244–246.

25.

Lindbladh

Fraver

Edvardsson

. (2013) Past forest composition, structures and processes – How paleoecology can contribute to forest conservation. Biological Conservation 168: 116–127.

26.

McArthur

Wilson

(1967) The Theory of Island Biogeography. Princeton, NJ: Princeton University Press.

27.

Moe

Bjune

(2009) Attractive spring-food for willow grouse (Lagopus lagopus lagopus L) studied using plant macrofossils and pollen in faeces: A methodological discussion. Grana 4: 310–315.

28.

Nesje

Bakke

Brooks

. (2014) Late glacial and Holocene environmental changes inferred from sediments in Lake Myklevatnet, Nordfjord, western Norway. Vegetation History and Archaeobotany 23: 229–248.

29.

Odland

Høitomt

Olsen

(2010) Increasing vascular plant richness on 13 high mountain summits in southern Norway since the early 1970s. Arctic, Antarctic and Alpine Research 42: 458–470.

30.

Ohlemuller

Huntley

Normand

. (2012) Potential source and sink locations for climate-driven species range shifts in Europe since the Last Glacial Maximum. Global Ecology and Biogeography 21: 152–163.

31.

Rull

(2010) Ecology and palaeoecology: Two approaches, one objective. The Open Ecology Journal 3: 1–5.

32.

Stoddard

Jeffries

Lukewille

. (1999) Regional trends in aquatic recovery from acidification in North America and Europe. Nature 401: 575–578.

33.

Telford

Vandvik

Birks

HJB

(2006) Dispersal limitations matter for microbial morphospecies. Science 312: 1015.

34.

Vandvik

Goldberg

(2006) Sources of diversity in a grassland metacommunity: Quantifying the contribution of dispersal to species richness. American Naturalist 168: 157–167.

35.

Vandvik

Töpper

Cook

. (2014) Management-driven evolution in a domesticated ecosystem. Biology Letters 10: 20131082.

36.

Velle

Larsen

Eide

. (2005) Holocene environmental history and climate of Råtåsjøen, a low-alpine lake in south-central Norway. Journal of Paleolimnology 33: 129–153.

37.

Willis

Birks

HJB

(2006) What is natural? The need for a long-term perspective in biodiversity conservation. Science 314: 1261–1265.

38.

Willis

Araújo

Bennett

. (2007) How can a knowledge of the past help to conserve the future? Biodiversity conservation and the relevance of long-term ecological studies. Philosophical Transactions of the Royal Society B: Biological Series 362: 175–186.