Substantial decline of Northern European peatland bird populations: consequences of drainage

Sara Fraixedasa,b*, Andreas Lindénc, Kalle Mellera, Åke Lindströmd, Oskars Keišse, John Atle Kålåsf, Magne Husbyg, Agu Leivitsh, Meelis Leivitsi and Aleksi Lehikoinena

a The Helsinki Lab of Ornithology (HelLO), Finnish Museum of Natural History, University of Helsinki, FI-00014 Helsinki, Finlandb Observatoire des Zones Humides Méditerranéennes (OZHM), Institut de recherche de la Tour du Valat, Le Sambuc, FR-13200 Arles, Francec Novia University of Applied Sciences, FI-10600 Ekenäs, Finlandd Department of Biology, Biodiversity Unit, Lund University, Ecology Building, SE-22362 Lund, Swedene Lab of Ornithology, Institute of Biology, University of Latvia, Miera iela 3, LV-2169 Salaspils, Latviaf Norwegian Institute for Nature Research, P.O. Box 5685 Sluppen, NO-7485 Trondheim, Norwayg Department of Science, Nord University, NO-7600 Levanger, Norwayh Species Conservation Unit, Estonian Environmental Board, Nigula , EE-86107 Tali, Estoniai Department of Geography, Institute of Ecology and Earth Sciences, University of Tartu, Vanemuise 46, EE-51014 Tartu, Estonia

*Corresponding author’s e-mail and phone number: sara.fraixedas@helsinki.fi / (+33) 619240004

1

123456789

101112131415161718192021

Abstract

Northern European peatlands are important habitats for biological conservation because they support rich biodiversity and unique species compositions. However, historical management of peatland habitats has had negative consequences for biodiversity and their degradation remains a major conservation concern. Despite increasing awareness of the conservation value of peatlands, the statuses and ecological requirements of peatland species have remained largely understudied. Here, we first analysed temporal trends of Northern European peatland birds to document the status of their populations using bird data from five different countries. Second, we used Finnish monitoring data to assess habitat preferences of peatland bird species, hence helping to target conservation to the most relevant habitat types. There was a general decline of 40% in Northern European peatland bird population sizes in 1981–2014 (speed of decline 1.5% / year) largely driven by Finland, where populations declined almost 50% (2.0% annual decline). In Sweden and Norway, peatland bird populations declined by 20% during 1997–2014 (1.0% annual decline). In contrast, southern populations in Estonia and Latvia, where the majority of open peatlands are protected, showed a 40% increase during 1981–2014 (1.0% annual increase). The most important habitat characteristics preferred by common peatland species in Finland were openness and low tree height, while wetness proved to be an important feature for waders. Drainage of peatlands had clear negative effects on the densities of many species, with the only exception of rustic bunting, which specializes on edge habitats. Our findings call for more effective conservation actions in Northern European peatland habitats, especially in Finland where peatland drainage represents a major threat to biodiversity.

Keywords: boreal peatlands, bird biodiversity, habitat loss, ditching, protected areas.

2

222324252627282930313233343536373839404142434445

1. Introduction

Peatlands are wet habitat types characterized by peat accumulation and are typically dominated by Sphagnum moss vegetation in high latitudes (Pakarinen 1995, Parish et al. 2008). Globally, about 4.0 million km2 of the world’s land area is covered by peatlands. The vast majority are northern peatlands (~90%), although there are also 368 500 km2 of tropical peatlands and 45 000 km2 of southern peatlands (Patagonia) (Yu et al. 2010). Among all northern peatlands the majority occurs in Russia, Canada, the US, and Fennoscandia (Gorham 1991). In the European Union, almost one third of the peatland cover is located in Finland, which is considered to be the country with the highest proportion of peatlands worldwide (see Supplementary Table S1).

Many specialised species inhabit peatlands, making them critical habitats for biodiversity conservation (Pearce-Higgins and Grant 2006). Not least are they key habitat for many bird species, several of which are included in the Annex I of the European Commission Birds Directive (Littlewood et al. 2010, European Commission 2015). The high value of the peatland breeding bird assemblages at the European level has contributed to turning some peatland areas into Special Protection Areas (SPAs) (Rieley and Lubinaite 2014). Despite their importance as biodiversity reservoirs, peatlands are typically under-represented in protected area networks, both nationally and internationally (Parish et al. 2008, Čivić and Jones-Walters 2010). Yet, peatlands have been listed as some of the most threatened habitat types within the European Union (Janssen et al. 2016).

Natural peatlands have been globally drained for different purposes, including agriculture (responsible for 50% of peatland loss), forestry (30%) and peat extraction (10%) for energy production and/or gardening (Vasander et al. 2003). Although some authors have suggested that peatlands could be drained sustainably (e.g. Uda et al. 2017), research evidence shows that most of these studies have largely neglected the issue of peatland subsidence in the long-term (Evers et al. 2017, Wetlands International and Tropenbos International 2016). Indeed, a plethora of scientific papers in the last decade have documented the pervasive effects of peatland drainage on biodiversity (Carrol et al. 2011). In contrast to wetland drainage, draining peatlands irrevocably involves removing most water from the extraction area (Holden et al. 2004). This causes extensive peatland degradation, resulting in a complete loss of peatland ecosystem functions (Parish et al. 2008). Finally, peatland removal encompasses a release of vast quantities of carbon into the atmosphere, modifying the biogeochemical processes of their soils and decreasing their biological productivity (Limpens et al. 2008). These long-term irreversible impacts reduce biodiversity and accelerate climate change (Carrol et al. 2011). While several different peatland types have been identified (Čivić and Jones-Walters 2010), the term peatland in the present study includes various habitats of fens, bogs and mires (see Supplementary Table S1).

Climate is a major determinant of peatland function and species composition (Dieleman et al. 2015). Evidence of climate change driven range shifts has mounted for many species, regions and habitats (e.g. Chen et al. 2011), but little is known about range shifts in peatland species. Given the predicted increases in temperatures and changes in rainfall patterns, and the resulting precipitation-evaporation dynamics, climate change poses a threat to the longevity of peatland ecosystems and therefore to birds and other species dependent on them (Holden et al. 2007). With cool, wet northern peatlands becoming warmer many species are expected to lose suitable climatic conditions, therefore having implications for their population performance in the long-term (Carrol et al. 2015).

Peatland bird diversity and abundance are known to increase along a northern gradient in Europe (Järvinen and Sammalisto 1976). However, the ecological requirements of peatland birds remain

3

464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596

poorly investigated, except for a few well understood species in the UK uplands (Douglas et al. 2014, Newey et al. 2016). To our knowledge, only a few studies have examined habitat preferences of peatland birds and/or quantified the effects of drainage on bird populations (e.g. Poulin et al. 2006, Hancock et al. 2009), most likely because they are often scarce and thus difficult to study without additional effort. Attempts to increase the information on the importance of peatlands contribution to regional diversity are essential for peatland protection (Calmé et al. 2012). In this context, identifying species’ habitat preferences helps prioritizing in conservation and restoration (Noss et al. 2009, Fraixedas et al. 2015).

Here, it is our aim to increase the ecological knowledge about peatland birds by providing information on: a) the joint bird population trends from five Northern European countries, therefore producing the first pan-European peatland bird status indicator (see Gregory et al. 2005); b) the regional trends of Finnish, Scandinavian and Baltic peatland bird populations; and c) the species-specific habitat preferences and spatio-temporal trends of six common peatland birds and seven less common peatland wader species, based specifically on Finnish peatland bird counts. For the purposes of the present study, we use the term peatland to refer to fens, bogs and/or mires (see Supplementary Table S1).

In general, we hypothesize that peatland bird populations will show stronger declines in areas where peatlands have been drained and a smaller proportion of the habitat is protected (in this case Finland, where only 14% of the current peatlands are protected; Alanen and Aapala 2015) compared to areas with a high level of protection (e.g. Estonia, where 75% of the open mires are currently protected; Supplementary Table S1). Furthermore, if climate change acts as a driver of peatland bird populations, we expect poleward shifts in species distributions (Chen et al. 2011) and more pronounced declines in the southern part of the study area, i.e. Baltic countries compared to Finnish or Scandinavian populations (Virkkala and Rajasärkkä 2011). Concerning species-specific habitat preferences for common peatland birds in Finland, we assume that birds will be positively associated with habitat characteristics typical for open peatlands (i.e. large open areas with low tree heights). On the other hand, a negative relationship is expected between species densities and peatland drainage.

2. Materials and Methods

2.1. Study area, habitat description and habitat data

The study area comprises five North European countries: Finland, Sweden, Norway, Estonia and Latvia (Fig. 1, Table 1). Overall, the area represents well the Boreal Region in Europe. Given that most peatland types are found in Finland (Montanarella et al. 2006), we followed the Finnish zonal, eco-climatic peatland complex type classification by Eurola and Kaakinen (1979) and only considered raised bogs, aapa mires (also known as fens) and Arctic mires (all henceforth referred to as peatland). Both open (treeless) and forested peatlands were included within the raised bog and aapa mire categories. We excluded all other wetlands (e.g. lakes, freshwater and brackish-water marshes, reedbeds and wet meadows; Pakarinen 1995).

To analyse habitat preference we used data from the Finnish line transect bird censuses (hereafter Finnish line transects; see Supplementary Appendix S1), which also include some data from Sweden, Norway and Estonia (Supplementary Fig. S1). Censuses were done between late May and early July, when the vast majority of migratory birds have arrived to their breeding grounds (Koskimies and Väisänen 1991), well including the focal species in this study. Each route was surveyed only once per breeding season and most often by the same observer across years. Two distance belts are distinguished along the line transect routes: the main belt (25 + 25 m wide) and

4

979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147

the supplementary belt, including pairs observed > 25 m from the route and all individuals flying over (Koskimies and Väisänen 1991, Laaksonen and Lehikoinen 2013). Since 1986, habitat data have been recorded from the main belt and structured in twelve habitat categories (see Fraixedas et al. 2015 for additional information). In this study, only the habitat categories of forested and open peatland, and their corresponding sub-categories, were considered. In forested peatlands, the observer classifies the peatland as one of the following habitats: i) natural nutrient-poor pine peatland (a total of 970 km transect in our dataset), ii) natural nutrient-rich birch peatland (70 km), or iii) drained peatland with ditches (i.e. ditched; 930 km). In non-forested peatlands (i.e. open peatland), the habitats are: i) ‘intermediate’ open peatland (550 km), ii) wet (310 km), iii) dry (160 km), or iv) peat extraction area (peat is used largely for energy production; 30 km). Tree height of forested peatlands is estimated on a 5 m interval scale (min 2.5–7.5, max 27.5–32.5). The areas of open peatlands are classified into one of six different size classes: < 0.01 ha, 0.01–0.1 ha, 0.1–1 ha, 1–10 ha, 10–100 ha, or > 100 ha.

We did not explicitly model detection rate in any of the analyses, primarily because the data did not allow that (non-repeated censuses). However, the habitat preference analyses were done only based on the 50 m wide main belt, where the birds can be linked to the observed habitat and have in general high detectability. Most birds are actively displaying or warning at this time of year, and hence we expected no relevant differences in detection rates between habitats. The largest risk for bias, due to habitat-specific differences in detectability, concerns birds that are neither actively displaying nor warning.

2.2. Species selection

Numerous methods for choosing species for multi-species indicators have been applied over the last years, most of them founded on expert opinion (Sætersdal et al. 2005). In order to apply a more objective methodology (Wade et al. 2014), we based our species selection on empirical data using a species habitat preference ratio (Rolstad et al. 2002). We first assessed the preference for peatland habitats over non-peatland habitats based on Finnish line transect habitat data from the main belt (Fraixedas et al. 2015). We compared the species’ relative densities (pairs / walked km) in peatland habitat (forested and open peatlands) with the densities in non-peatland habitat (other habitat types), and selected species for which the peatland density was at least 3 times higher relative to non-peatland habitat types (i.e. the peatland preference ratio is ≥ 3). However, we excluded all ducks, gulls and terns (Anatidae and Laridae) because they are typically found in lakes or ponds close to peatlands, rather than in peatland habitats as such. Observations of these species may have been incorrectly assigned to the main belt of a certain peatland habitat type. We also excluded vole-eating avian predators because they have low sample sizes and their populations are characterized by large natural fluctuations, especially in northern latitudes (Korpela et al. 2014). Their inclusion in the trend analyses would break the principle of stability, one of the key attributes of an effective biodiversity indicator (Gregory et al. 2005). For each species, we required at least an average of 10 pairs and a minimum of two pairs observed annually in order to estimate population trends (altogether two cases of minimum value). On this basis, 15 common peatland bird species were chosen to represent the Finnish peatland bird assemblage (see Supplementary Table S2). The same selected species were also used for Sweden and Norway (no cases of minimum value), presuming that the habitat use of peatland birds is quite similar between Finland, Sweden and Norway (see Supplementary Table S3). As for the Baltic countries (two cases of minimum value), the species choice relied on expert opinion (species with large enough sample sizes preferring raised bogs; A. and M. Leivits), but mainly consisted of the same species as for the rest of the countries.

2.3. Trend analyses

5

148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198

2.3.1. National monitoring schemes and route selection

We used data from five national bird monitoring schemes, most of which are based on a web of systematically distributed routes to be surveyed once a year (mainly by volunteers) in late spring or early summer (Lindström et al. 2015). For the Finnish estimates of the annual abundances of breeding birds we mainly used data from the line transect census method explained above, but also from point count censuses (Koskimies and Väisänen 1991). Information on the monitoring schemes and route selection for all the countries can be found in Supplementary material Appendix S1 and in Table 1.

After the selection of routes including a considerable amount of peatland habitat (see Appendix S1) we obtained a total of 1398 peatland routes that were surveyed at least twice. Over 90% of these routes belonged to Finland and Sweden (Fig. 1, Table 1), where most of the peatlands are found (c. 60% of the peatland area in Europe; Montanarella et al. 2006). The area of peatland habitat per number of routes was very similar in Finland, Sweden and Estonia (c. 120–150 km2; Table 1). Peatland routes were underrepresented in Norway and Latvia (Fig. 1, Table 1); however, these countries have a lower representation of peatland habitat compared to the other three countries (Montanarella et al. 2006). The five countries altogether cover approximately 70% of the peatland area in Europe (Montanarella et al. 2006).

2.3.2. Overall and regional trends

Using data from all countries and a total of 15 species (see Table 2) we calculated a joint peatland bird status indicator for the period 1981–2014 (sensu Gregory et al. 2005). The species selection was the same as for the Finnish region alone. The indicator includes species breeding also in other habitats, but many of the species are actually found almost exclusively in peatlands (see subsection 2.2.). The habitat specific route selection also increases the probability that the observed birds are indeed breeding in peatlands. The base year for every species was set to 2006 (index = 1). From this year onwards all species and countries are represented in the data. Population sizes of species included in the joint indicator are shown in Supplementary Table S4.

Variable detection rates between species do not bias the indicators, as the species indices are scaled. Species with lower detection rates are somewhat noisier, similarly to rarer species. Among the species studied here and using Finnish line transect data, Lehikoinen (2013) found no temporal trends in the proportion of main belt observations relative to supplementary belt observations. This indicates no relevant concerns regarding temporal trends in detection rate e.g. due to changed phenology.

For the calculation of the regional trends, countries were grouped into three different regions: Finland, “Scandinavia” (Sweden and Norway), and “the Baltic” (Estonia and Latvia). This classification was based on geographical proximity and ecological similarity between habitat types. Trends for Finland and the Baltic were calculated for the period 1981–2014, and for Scandinavia for 1997–2014. The Finnish peatland bird assemblage was represented by 15 species, the Scandinavian by 12, and the Baltic by 9 species (Table 2).

We used log-linear Poisson regression to estimate species-specific annual abundance indices, controlling for overdispersion, using the software TRIM (TRends and Indices for Monitoring data; Pannekoek and van Strien 2005) version 3.53. We applied the model ‘Time effects’ (model 3, or ‘Effects for each time point’) where the expectation of the natural logarithm of the counts

6

199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248

(response variable) is explained with the fixed effect factor variables year and site (Pannekoek and van Strien 2005), the year effect being of primary interest.

We did not use any weighting in relation to national population size estimates when calculating multi-country trends for species, since the amount of peatland habitat per number of routes was similar in the three countries with a higher percentage of peatlands (Finland, Sweden and Estonia; see previous subsection). Any combined species index (indicator) was calculated as the geometric mean of the species-specific annual abundance indices (obtained from TRIM) for the species involved in each of the regions and in the whole study area. Standard errors for geometric means were computed from indices and standard errors of individual species (Gregory et al. 2005).

Since the population trends may not be linear throughout the study period, we analysed the indicators using piecewise linear regression (R package segmented; Muggeo 2008) and ordinary linear regression (always on the log-scaled index) with year as the explanatory variable. In the piecewise linear model, we allowed for a possible change in the trend with an estimated break-point for year (two additional parameters estimated; Fraixedas et al. 2015). We used a likelihood ratio test (LRT) to compare this approach with the null model, where the trend was uniform. The null model was applied whenever the test was statistically non-significant.

2.3.3. Species composition and indicator type

Although the species composition inevitably differed at the regional level, there were six species that were included in the regional indicators in all three areas (meadow pipit Anthus pratensis, common snipe Gallinago gallinago, common crane Grus grus, whimbrel Numenius phaeopus, Eurasian golden plover Pluvialis apricaria, and wood sandpiper Tringa glareola). As for the indicators, different length time series were considered for several species in almost all cases to make use of the best information available (Table 2). Because the state of the peatland bird communities is measured over different time periods, this could influence the magnitude of the trends, especially in the Baltic region which has the smallest sample sizes. However, it is unlikely that this would change the direction of the regional indicators.

2.4. Analysis of habitat preference and range shifts

Using habitat information and bird data from the Finnish line transects (years 1987–2014), we modelled the population densities of six common peatland bird species and a group comprising seven less common wader species (observations summed), all belonging to the Finnish peatland bird assemblage. We explained densities using peatland habitat characteristics and spatio-temporal trends. The model structure follows a separate generalized linear mixed model (GLMM) with a logarithmic link-function and Poisson error distribution. The full model can be written as:

ln λis = α + β1 × open + β2 × ditchedis + β3 × peatextis + β4 × wetnessis + β5 × areasizeis + β6 × treeheightis + β7 × lats + β8 × lons + β9 × yeari + β10 × lats × yeari + β11 × lons × yeari + ln(lengthis) + as + ɛis

Nis ~ Poisson(λis)as ~ Normal(0, σ²s)ɛis ~ Normal(0, σ²e)

Nis is the response variable (number of pairs observed in the main belt) for year i in site s, which is assumed to be Poisson distributed with expectation λis. The offset variable ln(lengthis), on the right hand size of the equation, scales the expected numbers to the km walked in a certain habitat;

7

249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299

hence, effectively, we modelled population densities (pairs / km). The parameter α is the overall intercept, while β1–β11 are estimated fixed effects of the explanatory habitat variables (open, ditched, peatext, wetness, areasize and treeheight) and spatio-temporal variables (lat, lon, year, lat × year and lon × year), which model patterns in population densities not explained by habitat. The term as represent random variation in the intercept among sites, and it is assumed to be normally distributed with mean 0 and variance σ²s. This effect also accommodates possible differences between observers. The observation-level random effect (OLRE) ɛis comprises residual extra-Poisson variation (accounting for overdispersion), and it is assumed to be normally distributed with mean 0 and variance σ²e.

To obtain the six explanatory variables describing the peatland habitat characteristics (i.e. open, ditched, peatext, wetness, areasize, and treeheight), we first calculated the peatland route and census specific lengths of different peatland habitat category combinations (for example, 300 meters of wet open peatland). Each of these data points was assigned a combination of variables describing the habitat characteristics along the relevant part of a line transect route. These were initially coded as numerical (dummy) variables:- open (1 for open peatlands) - ditched (1 for ditched forested peatlands)- peatext (1 for peat extraction areas, applies only to open peatlands)- wetness (1 wet peatland; 0 ‘intermediate’ peatland; –1 dry peatland). It applies only to natural open peatlands.- areasize (average log10 of peatland area size based on the six size classes, only for open peatlands)- treeheight (average tree height, only for forested peatlands)

After this, the mean of the variables ditched, peatext, wetness, areasize, treeheight, lat, lon and year was set to 0 by subtracting the mean, which was calculated only for the applicable observations. For the centered variables ditched, peatext, wetness, areasize and treeheight, we set to zero the values where the variable was not applicable (e.g. all open peatland categories for the variable ditched). There was no strong correlation between the explanatory variables for any of the individual species or set of species (maximum Pearson’s correlation coefficient was always below 0.5; Booth et al. 1994).

For these specific analyses, we included all peatland segments of at least 200 m length from the line transect data. In numerous cases, a single route contained several such peatland segments. We thus omitted all observations in segments with less than 200 m of peatland habitat (of any kind).

We considered all species whose minimum peatland preference ratio was at least 3 and had at least a total of 300 observations from the main belt. Only six species fulfilled this requirement: meadow pipit, reed bunting Emberiza schoeniclus, rustic bunting Emberiza rustica, yellow wagtail Motacilla flava, common snipe and wood sandpiper. In addition, we gathered the main belt observations for a set of seven less common waders (broad-billed sandpiper Calidris falcinellus, ruff Calidris pugnax, whimbrel, jack snipe Lymnocryptes minimus, red-necked phalarope Phalaropus lobatus, Eurasian golden plover and spotted redshank Tringa erythropus), all with sample sizes below 300 pairs, to model the general habitat preference for these species (see Supplementary Table S2).

For each species, we restricted the habitat preference analysis to a spatial range of main occurrence, ensuring that the species occur at least rarely and show variation (i.e. information to be analysed) throughout the focal spatial area. We restricted the analysed range according to the

8

300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350

Minimum Convex Polygon (MCP) method (Mohr 1947) based on all data points where the focal species had been observed at least one year. We excluded data points falling outside the polygon, therefore all being zeros for all years (see Supplementary Fig. S1).

The models were fitted using R package lme4 (Bates et al. 2015) using maximum likelihood estimation and a Laplace approximation of the likelihood. Further, we applied the BOBYQA optimizing algorithm and at most 100 000 function evaluations.

Whenever the among-site variance (σ²s) went to zero, we refitted the models without an observation level random effect (OLRE) to ensure σ²s was non-zero. This choice is justified because “route” constitutes the most relevant level of replication (unit) in the data in relation to the questions studied.

In order to assess the robustness of the obtained qualitative results, e.g. in relation to distributional assumptions, we performed a similar analysis, but using a binary response (presence/absence). In this case “loglength” was defined as a covariate instead of an offset, and all models excluded the OLRE (see justification above), which is often difficult to estimate in binary models. This alternative analysis is a natural choice given the character of the data – i.e. many zeros and ones.

We defined slightly different variants of the model structure whenever we found some habitat variable was close to being redundant (e.g. one or no observations within either one of two habitat categories defined by a binary variable; Supplementary Table S5).

Parameter estimates were presented along with their standard errors. For fixed effects we applied Z-tests to investigate whether they differed from zero. The effects were considered to be statistically significant when observed p-values were < 0.05, and showing a statistical tendency whenever p-values were < 0.1. Analyses were carried out in R version 3.2.2 (R Development Core Team 2013). MATLAB (version 8.1.0.604) was used for the calculation of the MCP.

3. Results

3.1. Population trends

Overall in Northern Europe, only two of the study species showed positive population trends (common crane, hereafter crane, and little bunting Emberiza pusilla), while a total of seven species declined during 1981–2014. The remaining six species did not show significant trends (i.e. their trends were uncertain or stable). The species overall and regional trends are given in Table 2. Based on the geometric means of species-specific annual abundance indices, the population sizes of peatland birds have declined on average 40% since 1981 (Linear regression, b = –0.015 ± 0.002 SE, F1,32 = 91.46, p < 0.001; LRT of piecewise trend, χ2 = 1.83, df = 2, p = 0.40) (Fig. 2A).

At the regional level, only one species (crane) increased in Finland, whereas seven species were declining in 1981–2014. The remaining species (n = 7) did not show any significant trend (Table 2). The Finnish indicator declined during the whole period by almost 50% (Linear regression, b = –0.020 ± 0.002 SE, F1,32 = 98.52, p < 0.001). However, there was a significant change in the slope during the study period (i.e. a piecewise trend was significantly better compared to a linear trend; LRT, χ2 = 11.72, df = 2, p = 0.003). The decline was significantly steeper after 2001 than before (Piecewise linear regression, period 1981–2001, b1 = –0.011 ± 0.004 SE; period 2001–2014, b2 = –0.038 ± 0.007 SE; Fig. 2B).

9

351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401

In Scandinavia, only one species showed positive population trends (crane), but five species declined from 1997 to 2014. The rest of the species (n = 6) showed no significant trends (Table 2). As such, the Scandinavian indicator declined c. 20% during 1997–2014 (Linear regression, b = –0.011 ± 0.005 SE, F1,16 = 4.71, p = 0.045; LRT, χ2 = 1.88, df = 2, p = 0.171; Fig. 2C).

In the Baltic, in 1981–2014, four species showed a positive trend (crane, wood sandpiper, common redshank Tringa totanus and northern lapwing Vanellus vanellus), whereas only one declined (meadow pipit). The rest of the species (n = 4) showed no significant population trends (Table 2). In contrast to the indicators in other areas, the Baltic indicator has on average experienced an overall 40% increase for the whole study period (Linear regression, b = 0.010 ± 0.004 SE, F1,32 = 5.27, p = 0.028; LRT, χ2 = 4.48, df = 2, p = 0.107; Fig. 2D).

3.2. Habitat preference and spatio-temporal trends in Finland

By applying the procedure described in section 2.4., all fitted models converged properly. The estimated parameters from the Poisson GLMMs on relative density as well as their estimates of uncertainty and statistical significance are reported in Table 3. The corresponding results from the binary analysis of occurrence are given as supplementary materials (Table S6). In general, the latter analysis strongly confirms the patterns of habitat effects observed in the Poisson model, with only a few exceptions.

In accordance with the aforementioned predictions, two out of four passerine species (meadow pipit and yellow wagtail), as well as common snipe, wood sandpiper and the rare wader set preferred open rather than forested peatlands. The densities of three passerine species (meadow pipit, reed bunting and yellow wagtail), as well as those of wood sandpiper, were negatively associated with ditching in forested peatlands, while rustic bunting showed higher densities in ditched habitats. We found no statistically significant relationships with peat extraction areas. There was no indication of a preference for the degree of wetness (open peatlands) among passerines, but wader densities were significantly or nearly significantly positively associated with wetness. Meadow pipit showed significantly higher densities in larger open peatlands, and the rare wader set showed a tendency for a positive effect of peatland size. All species, except for rustic bunting and common snipe, showed strong negative effects of tree height in forested peatlands.

Meadow pipit and yellow wagtail densities declined over time. Many species showed spatial trends inside the studied area, the most consistent pattern being higher densities towards the north in all species except for common snipe and the rare wader set. Rustic bunting, yellow wagtail and wood sandpiper showed significant latitudinal density shifts northwards during the period 1987–2014. In addition, rustic bunting also experienced a significant longitudinal density shift eastwards.

4. Discussion

4.1. Overall and regional trends

Peatland bird populations have undergone an overall decline of 40% in North Europe during 1981–2014 (1.5% annual decline), but there are regional differences. Finnish peatland bird populations experienced the strongest decline for the same period (almost 50%, or 2.0% annual decline). Although drainage has taken place in all the countries involved in this study, Finland has suffered the most from drainage (e.g. c. 60% of the original peatlands have been ditched for

10

402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452

forestry; Supplementary Table S1, Supplementary Fig. S2). According to Rassi et al. (2010), drainage continues to be the most important threat to peatland biodiversity in Finland, not least because of drainage associated with historical ditching, which slowly degrades the habitat (Vasander et al. 2003; Supplementary Table S1, Supplementary Fig. S2). The accelerated rate of bird population decline over the past 15 years is especially alarming. Since Finland holds the largest peatland bird population within the EU countries (Supplementary Table S4), it has the highest responsibility for sustaining the population in the EU area.

In line with the declining population trends in Finland, the Scandinavian indicator (represented by Sweden and Norway) showed a 20% decline during the period 1997–2014 (1.0% annual decline). In Scandinavia, more than half of the original peatlands are in natural or nearly natural state (Vasander et al. 2003; Supplementary Table S1). Despite that substantial drainage has occurred also in Scandinavia especially between the 1930s–1970s, it has apparently affected a lower proportion of peatland area as compared to e.g. Finland (less than 20%; Supplementary Table S1). Importantly, the Scandinavian declines are purely based on data obtained from systematic sampling, and the Finnish populations also show a declining tendency since 2006, when systematic sampling started, supporting the fact that long-term declines are unlikely driven by biased sampling.

In contrast with the declining trends in Fennoscandia, the Baltic peatland bird indicator (represented by Estonia and Latvia) increased by up to 40% during 1981–2014, however, showing considerable variation over time. Although peatlands in both countries have been drained extensively (Vasander et al. 2003; Supplementary Table S1), there has been a notable progress in terms of conservation and sustainable use of wetlands since the political changes in 1991 and the accession to the European Union in 2004 (Kimmel et al. 2010; Supplementary Table S1). The conservation situation is especially good in Estonia, where c. 75% of the open mires are currently protected and 90% are located within the Estonian Green Network, respectively (Supplementary Table S1). Because of the high protection status in Estonia, the majority of the censuses were carried out inside protected areas. In contrast, the proportion of protected peatlands is still very low in Latvia (Supplementary Table S1), and the monitoring network in peatland areas also requires improvements.

4.2. Habitat preference and range shift model

Passerines and waders showed preference for open peatlands – rather than forested ones – as well as for low tree height and, to some extent, large peatland area sizes. In addition, the degree of wetness proved to be highly relevant for waders. Although high densities in a certain habitat does not necessarily mean that it is a good quality habitat for the species (van Horde 1983), our results support earlier studies suggesting openness as an important factor for passerines inhabiting peatlands (Järvinen and Sammalisto 1976) and open wet peatlands being essential for several wader species (Väisänen et al. 1998). Importantly, species also showed a considerably lower preference for ditched forested peatlands (see also Väisänen and Rauhala 1983).

Drainage of peatlands gradually leads into afforestation, increased tree height and reduced wetness (Laine et al. 1995), all of which according to our results would decrease the densities of peatland species (Wilson et al. 2012). Peat extraction areas were very scarce in our dataset, and thus this category was not possible to include in the analysis of e.g. wood sandpiper. When included in the models, we were not able to identify any effects of peatland extraction. However, this was possibly due to low power, as this habitat was quite poorly represented in the data. Overall, our habitat preference results highlight and confirm the beneficial effects of typical peatland properties for the birds (Väisänen and Rauhala 1983). The fact that drainage tends to

11

453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503

remove these properties supports the view that this practise has been an important driver of the observed population declines.

We acknowledge that many of the habitat preference results are highly expected, or may seem almost self-evident – peatland species prefer habitat characteristics typical for open peatlands. However, it is important to provide scientific evidence that this is the case, and to quantify the effect in different species. As the whole analysis is done for a subset of data a priori defined as peatland, we do not use the data circularly. Rustic bunting is a good example of a species which clearly prefers peatland habitats (preference ratio 4.51; Supplementary Table S2), but has nevertheless very different preferences inside the range of peatland habitats compared to the other studied species. This species prefers forested peatlands and even shows evidence for higher densities in ditched peatlands. The explanation is simply that the species is specialized on drier forest peatlands and edges of natural peatlands (Väisänen and Rauhala 1983), while the rest of the selected species favour interior peatland habitat, which is clearly the habitat of conservation concern.

The observed density shifts towards northern latitudes (identified for rustic bunting, yellow wagtail and wood sandpiper) could have been caused both by climate change (Chen et al. 2011) and a higher proportion of habitat loss in the southern part of the country (Virtanen et al. 2003). These effects are difficult to separate, but since the wood sandpiper population has increased in Estonia but declined in Finland, habitat loss is the more likely driver of the northward density shifts in Finland. Although the current evidence for negative impacts of climate change on peatland species is weak, climate change may harm peatland populations in the future (Virkkala et al. 2008). The combination of climate change and a continuing anthropogenic degradation of peatland habitats (peat extraction, forest ditching and drainage associated with past management actions, e.g. drying of ditched peatlands or afforestation) may lead to worse declines than the two processes alone, as observed in other habitats (e.g. Burns et al. 2016).

In general, we expect high detectability in this analysis, which is based on the 50 m wide main belt, surveyed at a time of year when most birds are actively displaying or warning. However, whenever this is not the case and birds are cryptic, detection rates may be higher in more open, less densely vegetated habitats and e.g. with lower tree heights. While we believe that our results are indeed based on true habitat preference, the observed positive effects of mire openness and negative effects of tree height are indeed in the same direction as expected from a detection artefact. In contrast, the effects of ditching, wetness, area size and the spatio-temporal trends should by no means be biased by detection.

One topic that has not been examined in this study is the impact of predation on peatland bird populations. On the one hand, some studies have shown that several mammal and bird species can predate on peatland birds, therefore contributing to population declines (MacDonald and Bolton 2008). For instance, species like common crane are known to predate on eggs and chicks in peatland habitats (Cramp and Simmons 1980). However, despite population increases in common cranes, to our best knowledge there is no study examining the influence of crane predation on peatland bird numbers. On the other hand, reductions in mammal and avian predators have been shown to improve the breeding success of several moorland bird species (Fletcher et al. 2010). Further research on this front is called upon in order to make better conservation plans for peatland bird species, considering that several land management practices such as peat harvesting can increase nest predation (Haddad et al. 2000) and hence incur in bird population declines.

12

504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554

5. Conclusions

We show that peatland bird species are generally declining in their important North European stronghold, and provide substantial support for the view that current and historical peatland drainage has major negative effects on peatland biodiversity. To protect this biodiversity there is an urgent need to increase the level of protection and restoration of peatland ecosystems in order to mitigate the combined negative effects of human land-use and changing climate. However, our ability to coordinate efforts toward peatland conservation will inevitably rely on our success in gaining local community buy-in. The most biodiversity-rich peatlands (i.e. large open wet peatland areas) need to be actively restored to preserve the bird fauna (Similä et al. 2014). Not only restoration is recommended, but also the strict protection of the still pristine peatlands. These measures are most urgent in Finland, but also important in the Scandinavian countries. Without protection and habitat restoration declines of peatland species are expected to continue in the future. Furthermore, we encourage more research on poorly studied peatland species along the lines of the present work to obtain empirical results that support current conservation actions.

Acknowledgments

We thank all volunteers who participated in data collection and the institutions responsible for maintaining the databases. Special thanks to Joona Lehtomäki for providing information on Finnish peatlands, Andrea Santangeli for producing the map in Fig. 1, and Roald Vang for taking the main responsibility for the Norwegian TOV-E webpage and database. Martin Green and Richard Gregory gave valuable comments on an earlier version of the manuscript. The Finnish common bird monitoring has been supported by the Finnish Ministry of the Environment. The surveys of the fixed routes in Sweden were supported by grants from the Swedish Environmental Protection Agency, and carried out in collaboration with all 21 County Administrative Boards of Sweden. The bird surveys are carried out within the framework of the Centre for Animal Movement Research and the strategic research environment Biodiversity and Ecosystem Services in a Changing Climate (BECC). The Norwegian Environment Agency finances the Norwegian bird monitoring. Financial contributions for carrying out the monitoring project for mire birds in Estonia have been provided since 1994 by the Estonian State Monitoring Programme. Data compilation in Latvia (O. Keišs) in 2015 was supported by the European Social Fund (ESF) within the project no. 2014/0009/1DP/1.1.1.2.0/13/APIA/VIAA/044. S. Fraixedas received financial support from the Maj and Tor Nessling Foundation and the Finnish Cultural Foundation, and A. Lehikoinen from the Academy of Finland (grant 275606).

References

Alanen, A., Aapala, K. 2015. Peatland Conservation Working Group's proposal to improve mire conservation. Report no 26. Ministry of Environment, Helsinki. (In Finnish). Available at: https://helda.helsinki.fi/handle/10138/158285.

Bates, D., Maechler, M., Bolker, B., Walker, S. 2015. Fitting Linear Mixed-Effects Models Using lme4. J. Stat. Softw. 67, 1–48.

Booth, G., Niccolucci, M., Schuster, E. 1994. Identifying proxy sets in multiple linear regression: an aid to better coefficient interpretation. U.S. Department of Agriculture Forest Service, Washington, D.C.

Burns, F., Eaton, M.A., Barlow, K.E., Beckmann, B.C., Brereton, T., Brooks, D.R., Brown, P.M.J., Al Fulaij, N., Gent, T., Henderson, I., Noble, D.G., Parsons, M., Powney, G.D., Roy, H.E., Stroh, P., Walker, K., Wilkinson, J.W., Wotton, S.R., Gregory, R.D. 2016. Agricultural Management and Climatic Change Are the Major Drivers of Biodiversity Change in the UK. PLoS ONE 11(3), e0151595.

13

555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605

Calmé, S., Desrochers, A., Savard, J-P. L. 2002. Regional significance of peatlands for avifaunal diversity in southern Québec. Biol. Conserv. 107, 273–281.

Carrol, M.J., Dennis, P., Pearce-Higgins, J.W., Thomas, C.D. 2011. Maintaining northern peatland ecosystems in a changing climate: effects of soil moisture, drainage and drain blocking on craneflies. Glob. Change Biol. 17, 2991–3001.

Carrol, M.J., Heinemeyer, A., Pearce-Higgins, J.W., Dennis, P., West, C., Holden, J., Wallage, Z.E., Thomas, C. D. 2015. Hydrologically driven ecosystem processes determine the distribution and persistence of ecosystem-specialist predators under climate change. Nat. Comm. 6, 7851.

Chen, I.-C., Hill, J.K., Ohlemüller, R., Roy, D.B., Thomas, C.D. 2011. Rapid range shifts of spe-cies associated with high levels of climate warming. Science 333, 1024–1026.

Cramp, S., Simmons, K.E.L.1980. The birds of western Palearctic. Vol. II. Oxford University Press, Oxford.

Čivić, K., Jones-Walters, L. 2010. Peatlands in Ecological Networks in Europe. European Centre for Nature Conservation, Tilburg.

Dieleman, C.M., Branfireun, B.A., McLaughlin, J.W., Lindo, Z. 2015. Climate change drives a shift in peatland ecosystem plant community: Implications for ecosystem function and stability. Glob. Change Biol. 21, 388–395.

Douglas, D.J.T., Bellamy, P.E., Stephen, L.S., Pearce-Higgins, J.W., Wilson, J.D., Grant, M.C. 2014. Upland land use predicts population decline in a globally near-threatened wader. J. Appl. Ecol. 51, 194–203.

Evers, S., Yule, C.M., Padfield, R., O’Reilly, P., Varkkey, H. 2017. Keep wetlands wet: the myth of sustainable development of tropical peatlands – implications for policies and management. Glob. Change Biol. 23, 534–549.

Eurola, S., Kaakinen, E. 1979. Ecological criteria of peatland zonation and the Finnish mire site type system, in: Kivinen, E., Heikurainen, L., Pakarinen, P. (eds.), Classification of peat and peatlands. International Peat Society, Hyytiälä, pp. 29–32.

European Commission 2015. Wild Birds: Threatened bird species in Annex I. Available at: http://ec.europa-eu.

Fletcher, K., Aebischer, N.J., Baines, D., Foster, R., Hoodless, A.N. 2010. Changes in breeding success and abundance of ground-nesting moorland birds in relation to the experimental deployment of legal predator control. J. Appl. Ecol. 47, 263–272.

Fraixedas, S., Lindén, A., Lehikoinen, A. 2015. Population trends of common breeding forest birds in southern Finland are consistent with trends in forest management and climate change. Ornis Fenn. 92,187–203.

Gorham, E. 1991. Northern peatlands: role in the carbon cycle and probable responses to climate warming. Ecol. Appl. 1(2), 182–195.

Gregory, R.D., van Strien, A., Vořišek, P., Gmelig-Meyling, A.W., Noble, D.G., Foppen, R.P.B., Gibbons, D.W. 2005. Developing indicators for European birds. Phil. Trans. Royal Soc. B 360, 269–288.

Haddad, S., Desrochers, A. 2000. Artificial nest predation in bogs: Does peat harvest increase risk? Ecoscience 7(1), 32–37.

Hancock, M.H., Grant, M., Wilson, J.D. 2009. Associations between distance to forest and spatial and temporal variation in abundance of key peatland breeding bird species. Bird Study 56, 53–64.

Holden, J., Chapman, P.J., Labadz, J.C. 2004. Artificial drainage of peatlands: hydrological and hydrochemical process and wetland restoration. Prog. Phys. Geog. 28(1), 95–123.

Holden, J., Shotbolt, L., Bonn, A., Burt, T.P., Chapman, P.J., Dougill, A.J., Fraser, E.D.G., Hubacek, K., Irvine, B., Kirkby, M.J., Reed, M.S., Prell, C., Stagl, S., Stringer, L.C., Turner, A., Worrall, F. 2007. Environmental change in moorland landscapes. Earth-Sci. Rev. 82, 75–100.

14

606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656

van Horde, B. 1983. Density as a misleading indicator of habitat quality. J. Wildl. Manage. 47, 893–901.

Janssen, J.A.M., Rodwell, J.S., Garcia Criado, M. Gubbay, S., Haynes ,T., Nieto, A., Sanders, N., Landucci, F., Loidi, J., Ssymank, A., Tahvanainen, T., Valderrabano, M., Acosta, A., Aronsson, M., Arts, G., Attorre, F., Bergmeier, E., Bijlsma, R.-J., Bioret F., Biţă-Nicolae, C., Biurrun, I., Calix, M., Capelo, J., Čarni, A., Chytry, M., Dengler, J., Dimopoulos, P., Essl, F., Gardfjell, H., Gigante, D., Giusso del Galdo, G., Hajek, M., Jansen, F., Jansen, J., Kapfer, J., Mickolajczak, A., Molina, J.A., Molnar, Z., Paternoster, D., Piernik, A., Poulin, B., Renaux, B., Schaminee, J.H.J., Šumberova, K., Toivonen, H., Tonteri, T., Tsiripidis, I., Tzonev, R., Valachovič, M. 2016: European Red List of Habitats. Part 2: Terrestrial and freshwater habitats. Publications Office of the European Union, Luxembourg.

Järvinen, O., Sammalisto, L. 1976. Regional trends in the avifauna of Finnish peatland bogs. Ann. Zool. Fenn. 13, 31–43.

Kimmel, K., Kull, A., Salm, J.-O., Mander, Ü. 2010. The status, conservation and sustainable use of Estonian wetlands. Wetl. Ecol. Manag. 18, 375–395.

Korpela, K., Helle, P., Henttonen, H., Korpimäki, E., Koskela, E., Ovaskainen, O., Pietiäinen, H., Sundell, J., Valkama, J., Huitu, O. 2014. Predator-vole interactions in northern Europe: the role of small mustelids revised. Proc. Royal Soc. London B 281, 20142119.

Koskimies, P., Väisänen R.A. 1991. Monitoring bird populations: a manual of methods applied in Finland. Zoological Museum, Finnish Museum of Natural History, Helsinki.

Laaksonen, T., Lehikoinen, A. 2013. Population trends in boreal birds: Continuing declines in agricultural, northern, and long-distance migrant species. Biol. Conserv. 168, 99–107.

Laine, J., Vasander, H., Sallantaus, T. 1995. Ecological effects of peatland drainage for forestry. Environ. Rev. 3(3–4), 286–303.

Limpens, J., Berendse, F., Blodau, C., Canadell, J.G., Freeman, C., Holdens, J., Roulet, N., Rydin, H., Schaepman-Strub, G. 2008. Peatlands and the carbon cycle: from local processes to global implications – a synthesis. Biogeosciences 5, 1475–1491.

Lindström, Å., Green, M., Husby, M., Kålås, J.A., Lehikoinen, A. 2015. Large-scale monitoring of waders on their boreal and arctic breeding grounds in northern Europe. Ardea 103, 3–15.

Littlewood, N., Anderson, P., Artz, R., Bragg, O., Lunt, P., Marrs, R. 2010. Peatland Biodiversity. Report to IUCN UK Peatland Programme, Edinburgh.

MacDonald, M.A., Bolton, M. 2008. Predation on wader nests in Europe. Ibis 150, 54–73.Mohr, C.O. 1947. Table of equivalent populations of North American small mammals. The

American Midland Naturalist 37, 223–249.Montanarella, L., Jones, R.J.A., Hiederer, R. 2006. The distribution of peatlands in Europe. Mires

and Peat 1, 1–10.Muggeo, V.M.R. 2008. Segmented: an R Package to Fit Regression Models with Broken-Line

Relationships. R News 8, 20–25.Newey, S., Mustin, K., Bryce, R., Fielding, D., Redpath, S., Bunnefeld, N., Daniel, B., Irvine,

R.J. 2016. Impact of Management on Avian Communities in the Scottish Highlands. PLoS ONE 11(5), e0155473.

Noss, R., Nielsen, S., Vance-Borland, K. 2009. Prioritizing ecosystems, species, and sites for restoration, in: Moilanen, A., Wilson, K.A., Possingham, H.P. (eds.), Spatial Conservation Prioritization. Oxford University Press, Oxford, pp. 158–171.

Pakarinen, P. 1995. Classification of boreal mires in Finland and Scandinavia: A review. Vegetatio 118, 29–38.

Pannekoek, J., van Strien, A. 2005. TRIM 3 Manual (TRends and Indices for Monitoring data). Statistics Netherlands, Voorburg.

Parish, F., Sirin, A., Charman, D., Joosten, H., Minaeva, T., Silvius, M. 2008. Assessment on Peatlands, Biodiversity and Climate change: Main Report. Global Environment Centre, Kuala Lumpur, and Wetlands International, Wageningen.

15

657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707

Pearce-Higgins, J.W., Grant, M.C. 2006. Relationships between bird abundance and the composition and structure of moorland vegetation. Bird Study 53, 112–125.

Poulin, M., Bélisle, M., Cabeza, M. 2006. Within-site habitat configuration in reserve design: A case study with a peatland bird. Biol. Conserv. 128, 55–66.

R Development Core Team 2013. R: A language and environment for statistical computing. The R Foundation for Statistical Computing, Vienna.

Rassi, P., Hyvärinen, E., Juslén, A., Mannerkoski, I. 2010. The 2010 Red List of Finnish Species. Ministry of the Environment, Finnish Environment Institute, Helsinki.

Rieley, J.O., Lubinaite, S. 2014. International Conventions, Agencies, Agreements and Pro-grammes Implications for peat and peatland management. International Peat Society, Jyväs-kylä.

Rolstad, J., Gjerde, I., Gundersen, V.S., Sætersdal, M. 2002. Use of Indicator Species to Assess Forest Continuity: a Critique. Conserv. Biol. 16, 253–257.

Sætersdal, M., Gjerde, I., Blom, H.H. 2005. Indicator species and the problem of spatial inconsistency in nestedness patterns. Biolog. Conserv. 122, 305–316.

Similä, M., Aapala, K., Penttinen, J. 2014. Ecological restoration in drained peatlands – best practices from Finland. Metsähallitus, Natural Heritage Services, Finnish Environment Institute, Vantaa.

Uda, S.K., Hein, L., Sumarga, E. 2017. Towards sustainable management of Indonesian tropical peatlands. Wetl. Ecol. Manag. DOI 10.1007/s11273-017-9544-0.

Vasander, H., Tuittila, E.-S., Lode, E., Lundin, L., Ilomets, M., Sallantaus, T., Heikkilä, R., Pitkänen, M.-L., Laine, J. 2003. Status and restoration of peatlands in northern Europe. Wetl. Ecol. Manag. 11, 51–63.

Väisänen, R.A., Rauhala, P. 1983. Succession of land bird communities on large areas of peatland drained for forestry. Ann. Zool. Fenn. 20, 115–127.

Väisänen, R.A., Lammi, E., Koskimies, P. 1998. Distribution, numbers and population changes of Finnish breeding birds. Otava, Helsinki. (In Finnish with English summary).

Virkkala, R., Heikkinen, R.K., Leikola, N., Luoto, M. 2008. Projected large-scale range reductions of northern-boreal land bird species due to climate change. Biol. Conserv. 141, 1343–1353.

Virkkala, R., Rajasärkkä, A. 2011. Northward density shift of bird species in boreal protected areas due to climate change. Boreal Environ. Res. 16, 2–13.

Virtanen, K., Hänninen, P., Kallinen, R.-L., Vartiainen, S., Herranen, T., Jokisaari, R. 2003. Finnish peat resources in 2000. Report of Investigation 156, Geological Survey of Finland, Espoo. (In Finnish).

Wade, A.S.I., Barov, B., Burfield, I., Gregory, R.D., Norris, K., Vořišek, P., Wu, T., Butler, S. 2014. A Niche-Based Framework to Assess Current Monitoring of European Forest Birds and Guide Indicator Species’ Selection. PLoS ONE 9(5), e97217.

Wetlands International, Tropenbos International 2016. Can peatland landscapes in Indonesia be drained sustainably? An assessment of the ‘eko-hidro’ water management approach. Wetlands International Report.

Wilson, M.W., Gittings, T., Pithon, J., Kelly, T.C., Irwin, S., O’Halloran, J. 2012. Bird diversity of afforestation habitats in Ireland: current trends and likely impacts. Biol. Environ. 112B, 55–68.

Yu, Z., Loisel, J., Brosseau, D.P., Beilman, D.W., Hunt, S.J. 2010. Global peatland dynamics since the Last Glacial Maximum. Geophys. Res. Lett. 37, L13402.

16

708709710711712713714715716717718719720721722723724725726727

728729730

731732733734735736737738739740741742743744745746747748749750751752753754755

Tables

Table 1. Study period, number of years, number of peatland routes selected for each of the countries (FI = Finland, SE = Sweden, NO = Norway, EE = Estonia, and LV = Latvia), average number of routes surveyed annually (min–max), total peatland area according to Montaranella et al. (2006), and representatives of routes in peatland habitats (area of peatland divided by number of routes in the country; see also Table S1).

Country Study period No. years No. routes Routes Peatland Representative area Peatland area annually area (km2) per route (km2)FI 1981–2014 34 729 89 (11–176) 88 908 122SE 1997–2014 18 561 279 (59–455) 65 859 117NO 2006–2014 9 45 27 (5–40) 18 685 415EE 1981–2014 34 62 5 (1–15) 9353 151LV 2003–2014 12 1 1 (1–1) 7385 7385

17

756757758759760761762763764765766767768769770771

Table 2. The peatland bird assemblage representing the whole area (All) and each of the regions

(FI = Finland, SE–NO = Scandinavia, EE–LV = Baltic), the species-specific overall and regional trends (base year = 2006; see subsection 2.3.2.), and percentages of multiplicative net change. Significant increasing or decreasing growth rates (± SE) are in bold (otherwise the trend is estimated as either stable or uncertain). The dash ‘–’ means that the species was either not present in the country, not considered as a peatland specialist, or had too little monitoring data. The asterisk ‘*’ indicates species for which data started later than the period specified in the table. In those cases, even if the species where included in the indicators, sample sizes were too low and/or unbalanced to estimate a trend for the entire period.

Common name of species All (n = 15) Change FI (n = 15) Change SE–NO (n = 12) Change EE–LV (n = 9) Change1981–2014 % 1981–2014 % 1997–2014 % 1981–2014 %

Meadow pipit –0.022 ± 0.002 –53 –0.011 ± 0.003 –31 –0.028 ± 0.005 –40 –0.039 ± 0.004 –73Broad-billed sandpiper * –0.106 ± 0.031 –97 –0.106 ± 0.068 –97 – – – –Ruff –0.061 ± 0.009 –87 –0.106 ± 0.023 –97 –0.076 ± 0.024 –75 – –Little bunting * 0.101 ± 0.040 3000 0.053 ± 0.048 506 – – – –Rustic bunting –0.040 ± 0.007 –74 –0.047 ± 0.009 –80 –0.043 ± 0.011 –54 – –Reed bunting –0.018 ± 0.004 –46 –0.020 ± 0.004 –49 –0.016 ± 0.006 –25 – –Common snipe * –0.001 ± 0.003 –3 –0.005 ± 0.004 –16 0.010 ± 0.005 20 0.010 ± 0.017 40Common crane 0.039 ± 0.005 277 0.037 ± 0.006 252 0.049 ± 0.005 142 0.041 ± 0.007 303Jack snipe * –0.048 ± 0.040 –80 0.020 ± 0.049 97 – – – –Yellow wagtail –0.030 ± 0.003 –64 –0.033 ± 0.003 –67 –0.007 ± 0.005 12 – –Eurasian curlew * – – – – – – –0.028 ± 0.015 –61Whimbrel * –0.001 ± 0.005 –3 0.005 ± 0.006 19 0.014 ± 0.009 29 –0.009 ± 0.019 –26Eurasian golden plover –0.000 ± 0.002 0 –0.001 ± 0.004 3 0.002 ± 0.009 4 –0.001 ± 0.003 –3Spotted redshank * –0.052 ± 0.014 –83 –0.139 ± 0.053 –99 –0.039 ± 0.012 –50 – –Wood sandpiper 0.002 ± 0.002 7 –0.010 ± 0.003 –29 0.003 ± 0.004 6 0.013 ± 0.003 56Common greenshank 0.005 ± 0.005 19 –0.006 ± 0.005 –18 –0.001 ± 0.005 –2 – –Common redshank – – – – – – 0.025 ± 0.006 134Northern lapwing – – – – – – 0.013 ± 0.006 56

* Overall trends: broad-billed sandpiper (1997–), little bunting (2006–), jack snipe (2006–), spotted redshank (1997–); Regional trends FI: broad-billed sandpiper (2006–), little bunting (2006–), jack snipe (2006–), spotted redshank (2006–); Regional trends EE–LV: common snipe (1994–), Eurasian curlew (1994–), whimbrel (1994–).

18

772

773774775776777778779780781782783784785786787788789790791792793794795796797798799800801

802803804805

Table 3. Statistics for log-linear Poisson mixed models explaining variation in population densit-ies of Finnish peatland birds. The results are shown for six separate species and one composite group of waders. The estimated parameters are associated with standard errors (SE), Z-statistics and two-tailed p-values for the null hypothesis of zero effects. Statistically significant effects (p < 0.05) are in bold and tendencies (p < 0.1) are italicised. The estimated random-effect standard de-viations (SD) of both “site” (random intercept) and “observation ID” (within-site extra-Poisson variation) are presented. In three cases (reed bunting, rustic bunting and common snipe), where the site-level SD was estimated as zero, we refitted a model without the observation-level random effect. This was done because the analysis focuses on site-level variation. The effects of “lat” and “lon” are per 100 km, and the effect of year is per decade.

Parameters Estimate SE Z pMeadow pipitα intercept –3.438 0.166 –20.67 < 0.0001β1 open 2.396 0.169 14.16 < 0.0001β2 ditched –0.868 0.250 –3.48 0.0005β3 peatext –0.083 0.470 –0.18 0.8601β4 wetness 0.150 0.092 1.63 0.1037β5 areasize 0.178 0.079 2.26 0.0240β6 treeheight –2.691 0.541 –4.97 < 0.0001β7 lat 0.059 0.024 2.45 0.0144β8 lon –0.185 0.064 –2.91 0.0037β9 year –0.343 0.076 –4.53 < 0.0001β10 lat×year 0.026 0.026 0.98 0.3265β11 lon×year –0.068 0.057 –1.19 0.2352σs among-site SD 0.665 – – –σe observation-level SD 0.484 – – –Reed buntingα intercept –3.859 0.239 –16.14 < 0.0001β1 open 0.323 0.212 1.52 0.1274β2 ditched –0.637 0.275 –2.32 0.0206β3 peatext 0.748 0.896 0.84 0.4037β4 wetness 0.330 0.202 1.63 0.1028β5 areasize 0.017 0.177 0.10 0.9225β6 treeheight –1.173 0.598 –1.96 0.0499β7 lat 0.217 0.057 3.80 0.0001β8 lon –0.055 0.124 –0.44 0.6580β9 year 0.173 0.161 1.07 0.2849β10 lat×year 0.080 0.064 1.24 0.2159β11 lon×year –0.006 0.122 –0.05 0.9618σs among-site SD 1.464 – – –σe observation-level SD – – – –Rustic buntingα intercept –2.191 0.107 –20.49 < 0.0001β1 open –3.781 0.940 –4.02 < 0.0001β2 ditched 0.937 0.171 5.48 < 0.0001β3 peatext – – – –β4 wetness – – – –β5 areasize 0.516 0.921 0.56 0.5752β6 treeheight –0.174 0.382 –0.46 0.6481β7 lat 0.228 0.052 4.41 < 0.0001β8 lon 0.151 0.089 1.69 0.0912β9 year –0.092 0.103 –0.89 0.3717β10 lat×year 0.163 0.054 3.00 0.0027

19

806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859

β11 lon×year 0.189 0.087 2.18 0.0292σs among-site SD 0.083 – – –σe observation-level SD – – – –Yellow wagtailα intercept –2.611 0.126 –20.69 < 0.0001β1 open 1.341 0.128 10.45 < 0.0001β2 ditched –1.376 0.193 –7.12 < 0.0001β3 peatext –0.804 0.787 –1.02 0.3076β4 wetness –0.0004 0.098 –0.004 0.9967β5 areasize 0.082 0.084 0.97 0.3318β6 treeheight –2.566 0.382 –6.71 < 0.0001β7 lat 0.180 0.028 6.41 < 0.0001β8 lon –0.051 0.062 –0.82 0.4099β9 year –0.243 0.071 –3.40 0.0007β10 lat×year 0.089 0.030 3.02 0.0026β11 lon×year 0.004 0.055 0.08 0.9355σs among-site SD 0.697 – – –σe observation-level SD 0.646 – – –Common snipeα intercept –4.580 0.294 –15.58 < 0.0001β1 open 1.827 0.249 7.34 < 0.0001β2 ditched –0.653 0.417 –1.57 0.1172β3 peatext – – – –β4 wetness 0.474 0.210 2.26 0.0242β5 areasize –0.015 0.165 –0.90 0.3669β6 treeheight –0.882 0.822 –1.07 0.2834β7 lat –0.012 0.050 –0.24 0.8079β8 lon 0.037 0.140 0.26 0.7943β9 year –0.054 0.159 –0.34 0.7357β10 lat×year 0.022 0.063 0.34 0.7319β11 lon×year –0.010 0.137 –0.08 0.9394σs among-site SD 1.384 – – –σe observation-level SD – – – –Wood sandpiperα intercept –2.741 0.124 –22.06 < 0.0001β1 open 1.566 0.129 12.14 < 0.0001β2 ditched –1.197 0.218 –5.48 < 0.0001β3 peatext – – – –β4 wetness 0.378 0.105 3.59 0.0003β5 areasize –0.083 0.083 –1.00 0.3197β6 treeheight –1.073 0.388 –2.76 0.0057β7 lat 0.154 0.027 5.81 < 0.0001β8 lon –0.032 0.062 –0.52 0.6030β9 year –0.082 0.076 –1.09 0.2776β10 lat×year 0.090 0.030 3.00 0.0027β11 lon×year 0.007 0.059 0.12 0.9071σs among-site SD 0.601 – – –σe observation-level SD 0.640 – – –Rare wader setα intercept –5.033 0.302 –16.66 < 0.0001β1 open 2.669 0.268 9.97 < 0.0001β2 ditched –0.440 0.433 –1.02 0.3086β3 peatext –0.594 0.765 –0.78 0.4372β4 wetness 0.315 0.168 1.87 0.0618β5 areasize 0.246 0.134 1.83 0.0671

20

860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914

β6 treeheight –2.740 0.971 –2.82 0.0048β7 lat 0.012 0.040 0.29 0.7691β8 lon –0.079 0.097 –0.82 0.4131β9 year 0.066 0.138 0.48 0.6320β10 lat×year –0.036 0.049 –0.73 0.4668β11 lon×year 0.016 0.105 0.16 0.8755σs among-site SD 0.706 – – –σe observation level-SD 1.355 – – –

21

915916917918919920921922923

Figures

Fig. 1

22

924925926927

928

Fig. 2

23

929930

931

Fig. 1. Map of the study area comprising Finland, two Scandinavian countries (Sweden and Norway), and two Baltic countries (Estonia and Latvia). Country borders have been slightly smoothed for visualisation. Black dots show all census sites including peatland habitats (coordinates in WGS84).

Fig. 2. Indicators reflecting the state of Northern European peatland bird populations at both gen-eral (panel A) and regional (panels B–D) levels. They are based on the geometric means of an-nual species-specific abundance indices from 1981–2014 (15 species), 1981–2014 (15 species), 1997–2014 (12 species), and 1981–2014 (9 species), respectively. Annual indices for relative population density start from 1 in 2006 (see subsection 2.3.2.). The dashed grey lines represent either linear or segmented regressions (on the log-scaled index) against year. All the indices in-clude annual 95% confidence intervals defined as ± 1.96 SE of the geometric means (see Appen-dix A in Gregory et al. 2005).

24

932933934935936937938939940941942943944945946947

Supplementary material

Fraixedas, S., Lindén, A., Meller, K., Lindström, Å., Keišs, O., Kålås, J.A., Husby, M., Leivits, A., Leivits, M. and Lehikoinen, A.: Substantial decline of Northern European peatland bird populations: consequences of drainage.

Content Page Description

Appendix S1 2 Description of census methods in different countriesTable S1 5 Background information on peatlands in the study countriesTable S2 9 Information on the Finnish data analysed (indicators and habitat preferences)Table S3 10 Information on regional data used for the indicatorsTable S4 11 National estimates of population sizeTable S5 12 Number of observations in each habitat category (Finnish data) Table S6 13 Statistics for logistic binomial mixed modelsFigure S1 16 Illustration of the Minimum Convex Polygon (MCP) approachFigure S2 17 Map illustrating the degradation of peatland habitats in Fin-landReferences 18 Supplementary material references (excluding those of Table A1)

25

948

949

950951952

953

954

955

956

957958

959

960

961962

963

964

965966

967968

Appendix S1. Description of census methods in different countries.

In Finland, the point count census method for landbirds started in 1984. One point count route includes 20 points, each of which are located in uniform habitats (within a 50-m radius) and separated by 250 m in forests and 350 m in open areas (Koskimies and Väisänen 1991, Laaksonen and Lehikoinen 2013). Each point is surveyed for five minutes, and observed birds are classi-fied as inside or outside a 50-m radius from the point. Point count routes are performed once a year and they are annually censused from 20th May to 20th June in southern Finland and from 30th May to 30th June in northern Finland (Koskimies and Väisänen 1991). Since 1986 habitat data are also provided in point count routes, but its accuracy is lower compared with that gathered from line transects (e.g. average tree height is not specified; Fraixedas et al. 2015). As for the line transect census method, the old monitoring system started in 1975. Although the observers were free to select the routes, the Finnish Museum of Natural History encouraged a representative coverage of all habitat types in the census area. Since 2006 a new system based on 560 systematically chosen fixed line transect routes was set up. Each route is 6 km long and shaped like a 1 × 2 km rectangle (Lindström et al. 2015).

We excluded a total of 93 line transect sites from the Finnish dataset that presented deficiencies according to the standards of data quality estab-lished within the protocols of the monitoring scheme. Based on the habitat data from the line transects, we defined a peatland route as one with at least 200 m of main belt in an open peatland habitat (including peat extraction areas) or a natural forested peatland, or at least 1000 m in a ditched fores-ted peatland. In addition, in cases where habitat data were not available (routes surveyed before 1986 and also point counts), we selected only those routes in which either common greenshank Tringa nebularia or wood sand-piper Tringa glareola were observed (except for routes located in the mont-ane tundra region), given that these are the two most common species rep-resentative of Finnish peatland habitats (Väisänen et al. 1998). After apply-ing these two selection methods, a total of 729 peatland routes (641 line transects and 88 point counts) were included, covering the period from 1981 to 2014 (see Fig. 1 in the main text). Importantly, both main and supple-mentary belt observations were used in the trend analyses.

In Sweden, data originate from the so-called fixed routes (Lindström et al. 2013). There are in total 716 fixed routes of 8 km each distributed sys-tematically across the country, and the observer is allowed to deviate up to 200 m from the route (Lindström et al. 2013). We used Swedish Land Cover Data (Svensk Marktäckedata – SMD) 2000 to classify the peatland routes. A total of six habitat classes were selected to represent the peatland ecosys-tem: a) broad-leaved forest on peatlands; b) coniferous forest on peatlands; c) mixed forest on peatlands; d) wet peatlands; e) other peatlands; and f) peat extraction sites. The first three categories can be classified as peatland forests and the rest as open peatlands. Habitats were analysed within 200 m from the line transect, assuming that all birds recorded ‘belonged’ to the area surveyed. The total potential land area of such a buffer zone is 3.2 km2. Being so, 633 routes were chosen that contained any kind of peatland within

26

969

970

971972973974975976977978979980981982983984985986987

988989990991992993994995996997998999

1000100110021003

1004100510061007100810091010101110121013101410151016

200 m. Because in the CORINE system ‘peatland’ is understood as ‘any wet-land’, including rich eutrophic wetlands in southern Sweden, another selec-tion criterion was needed to better fit the ‘peatland’ definition. Therefore, the same criterion used for dividing Sweden into farmland production areas, where classes 1–4 are richer lowland areas and classes 5–8 represent poorer farmland areas, was applied in this case (the latter being within forested areas in northern Sweden and some parts of southern Sweden). All routes placed in production areas of classes 5–8 were selected, given that they cover quite well peatland areas according to our definition. Using both se-lection criteria, a total of 561 routes were finally obtained, with bird data covering the period from 1997 to 2014 (Fig. 1). In Sweden, unlike other countries, the study unit is the number of adult individuals (males and fe-males, but not young of the year), whereas in other countries a pair is re-garded as the census unit. However, the scale of abundance measurements does not affect the estimated multiplicative trends (Lehikoinen et al. 2014).

In Norway, common bird monitoring data are collected from almost 500 sites which are randomly selected among 1030 sites distributed systematic-ally over the country in an 18 km north-south and a 18 km east-west net-work (Husby and Kålås 2011, Lehikoinen et al. 2014, Lindström et al. 2015). Bird counting consists of point counts, each route containing 20 points situ-ated 300 apart and forming a 1.5 × 1.5 km square. In addition, nearly all ob-servations of non-passerine birds (and a few pre-selected passerine birds) observed while moving between the counting points are recorded (Le-hikoinen et al. 2014, Lindström et al. 2015). The number of observations for each sampling route is the sum of observed pair equivalents of birds at the counting points (5 min counting period at each point) and while moving between counting points (Kålås and Husby 2002). For some of the routes, the number of counting points can be less than 20 (but always > 12) be-cause of reduced availability (lakes, cliffs, rivers, etc.). Generally, counts are done between 23 May and 7 July (Lehikoinen et al. 2014, Lindström et al. 2015). Given that some of the peatland bird species are found in several types of habitats, data should be restricted only from routes including peat-land habitat. Similarly to Sweden, this selection is not that straightforward, because there is no access to the habitat around the censused routes. As such, route selection was based on the Norwegian Forest and Landscape In-stitute (Norsk Institutt for Skog og Landskap, Norsk Institutt for Bioøkonomi – NIBIO) land cover map which has been created and updated from the 2006 CORINE Land Cover database. Habitat types within 100 m from each count-ing point were included, the area of this buffer being 0.63 km2 per route. Routes considered as ‘suitable’ for analysing peatland bird species included the following percentages of coverage (for all sampling points in a route): > 0% of peatlands, < 10% of agricultural areas, and < 5% of open habitats (e.g. dry mountain and open coastal areas). Routes with significant coverage of open habitats as farmland and dry mountain areas were excluded because these are also regularly used by some of the peatland bird species in Nor-way. Results were expected to be more sensitive when incorporating also open dry habitats (as mountain areas) rather than only agricultural areas, because some of the peatland species (e.g. meadow pipit Anthus pratensis and Eurasian golden plover Pluvialis apricaria) are rare in agricultural habit-

27

101710181019102010211022102310241025102610271028102910301031

1032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065

ats compared to open dry mountain habitats. The Norwegian dataset con-sisted of 45 routes censused from 2006 to 2014 (Fig. 1), but the monitoring network is developing and more sites will be added in the near future.

In Estonia, the common bird monitoring scheme started in 1983 and it consists of point counts placed on permanent routes (20 points per route) that are freely selected by the observer (Kuresoo et al. 2011). Given that only few census routes are found in peatlands, the scheme cannot be con-sidered as representing this ecosystem. However, there are rather good baseline data on breeding bird numbers of Estonian open mires from 1948 to 1957 for large bogs (Kumari 1972) and from 1951 to 1967 for fens in western Estonia (Renno 1974) based on surveys using line transects. The in-terval between transects in mires varies between 200–500 m (in most cases 300 m) depending on the habitat, accessibility and available resources (Leiv-its 2014). Regular yearly censuses of mire birds begun in 1968 in Nigula bog (Irdt and Vilbaste 1974), where 2000 ha of bog area are censused using 9 to 10 transects per year with an average length of 5.5 km (Leivits 2014). Since 1986 the census sites started to be surveyed again in order to cover the ex-act same areas, the difference being the comparison units, which changed from transects (count of breeding pairs on a transect) to habitat patches (count of breeding pairs on a mire patch, usually mires or mire massifs; Lait-inen et al. 2007, Leivits et al. 2013). The interval between the baseline (1948–1957) and the first regular re-inventory varied from 16 to 30 years, and the interval between regular re-inventories varied between 1 and 20 years (Leivits et al. 2013). The number of line transects surveyed per year in each site also varies depending on the size of the mire patch. While large mire patches are surveyed with several transects to cover the total area of the mire patch (e.g. in Nigula), small patches are maybe surveyed only with one transect. For the data collection of a single-visit transect count, the ob-server has a special map of the transect and also base maps including a shaded orthophoto at a scale of 1:10 000 (Leivits 2014). The observer maps all the birds it encounters; overflights and flushes are specially marked. Singing or displaying, or individuals occurring in pairs (mostly waders) are interpreted as breeding pairs (Irdt and Vilbaste 1974, Leivits 2014). Estim-ates of the number of breeding pairs are based on the interpretation of the census maps (Leivits et al. 2013). Overall, data included altogether surveys of 62 different sites. Therefore, we assumed that censuses were random samples from the mires. After all good and large sites have more birds than poor and small sites, and thus they have more importance in the model. The study period comprised years from 1981 to 2014 (Fig. 1).

In Latvia, data were collected entirely in the Ķemeru Mire (6192 ha), one of the largest raised bogs in the country located in the Ķemeri National Park (Bambe et al. 2008). This site was censused yearly in mid-May (twice a year from 2003 to 2009 in mid-May and mid-June) and consisted of a total of 13 consecutive line transects across the terrain making a loop that covers typical open raised bog habitat. Each transect is 400 m length (except the last one, which is 110 m length). Data were collected by the same observer during the whole study period using a modified version of the Finnish line transect method (50-m main belt; Järvinen and Väisänen 1975). The counts started within one hour after sunrise and were continued for approximately

28

106610671068

106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104

1105110611071108110911101111111211131114

3–4 hours. There are also some transboundary peatland sites on the Esto-nian-Latvian border (Ziemelu Purvi – Sookuninga Ramsar areas) surveyed repeatedly on both sides of the border, but these data are included within the Estonian dataset. The study period covered the years 2003–2014 (Fig. 1).

In the trend analysis, the use of both point counts and line transects should not bias the results since we used the raw counts in the analyses to study multiplicative changes. Therefore, we expected e.g. that halving the number of birds in a point count route would correspond to halving the num-ber of birds in a line transect census during the same time period. The counting methods used do not reveal the absolute unbiased abundances of species (Calladine et al. 2009), or the exact boundaries between the habitats used by birds. Clearly, any bird assumed to be connected to a ‘peatland’ may have been seen outside the peatland habitat of that given route. However, the selection of peatland species based on Finnish main-belt data should not have large detectability differences between the habitats. Still, we cannot be entirely sure that some of the more cryptic species (e.g. common snipe) did not show somewhat lower detection rates in more densely vegetated habit-ats. Although a temporal change in species detectability should not be a problem in our case (no annual trend in the proportion of main-belt observa-tions; Lehikoinen 2013), we cannot discard the fact that potential spatial trends in detectability, as well as spatio-temporal trends in other countries, may have caused spurious trends.

29

11151116111711181119

112011211122112311241125112611271128112911301131113211331134113511361137

Table S1. Type of peatlands in the study countries, history of peatland management and habitat restoration techniques.

Country Types of natural peatland areas

History of peatland management Restoration techniques

Finland Limnogenic mire complex types, raised bogs (of various types), aapa mires: flark-string aapa mires, lawn (intermediate level) aapa mires, slope mires, palsa mires and Arctic mires.

Country with the highest proportion of peatlands in the world. Absolute area of peatlands in Finland 98 000 km2. Systematic drainage for forestry and farming started at the beginning of the 19th century. It developed as a nation-wide campaign in the 1960s. Total area drained especially for forestry 57 000 km2

(c. 60% of the original peatland area). Less than 25% of the original peatland area remains in southern and central Finland. Before modern forestry substantial area converted into arable land (7000 km2) affecting rich fens. Large-scale extraction of peat for fuel stared in the 1970s. Spatial distribution of the drained areas in Supplementary Fig. S2. Drainage mainly done using ditches, which slowly dry out peatlands and transform them into forests in a process that continues for decades.

Approximately 14% of the existing peatlands are protected. Peatland protection programme within the Action Plan (1997–2005); 4.1 million ha of pristine peatlands (one million protected by law). National requirements for the rehabilitation and restoration of cut-away peatlands (2001); 40 000–45 000 ha of cut-away peatland. Around 30 000 ha of peatlands restored in the last 30 years, but 18 000 ha need to be restored to safeguard ecological values of protected areas; 120 000 ha proposed for conservation (about half of them protected within Natura 2000).

Sweden Raised bogs, plane bogs, topogenous fens, sloping fens, mixed mires. Topogenous and sloping fens are the most common. Open mires are more abundant than forested mires.

Peat cutting for litter started in the 19th century. Since 1960 occurred on a greater scale for fuel, with increasing activity since 1980; 15% of peatlands have been influenced by forestry. Extensive survey of Swedish wetlands and their need for protection started in 1980. Subsidies for draining ceased in 1986.

47% of the peatland area in natural or near-natural state. Interest in peatland restoration as an alternative after-use option; 300 ha of cut-away areas could be left aside. Ways to obtain peat-forming vegetation are being investigated. Ongoing restoration projects in the southern part of central Sweden.

Norway Raised bogs, blanket bogs, flat fens, sloping fens, stringing fens, palsa mires. Sloping and string fens more abundant.

In total 7000 km2 of peatland strongly affected by direct human activity. Ditching of peatlands and swampy woodland to increase forestry production began at the end of the 19th century and lasted until nearly 2000; 4000 km2

were ditched to plant trees; 2000 km2 for agricultural exploitation, mainly affecting rich

More than half of the original peatland area left. Last proportion of most valuable remaining areas protected. Single projects aiming at restoring habitats for birds and amphibians.

30

1138

1139

fens; 300 km2 affected by production of peat moss litter and peat for fuel during 1919–1996, mainly affecting ombrotrophic bogs.

Estonia Minerotrophic mires (fens and transition mires) and ombrotrophic mires (raised bogs). Raised bogs in favourable conservation more common than fens. Large open raised bogs and large rich (calcareous) fens under the responsibility of Estonia in the whole boreal region.

Intensive agriculture, forestry, peat harvesting and forest drainage occurred in 1950–1980. There was a long tradition in using peat for heating. Milled peat harvesting dominating method till 1980; 70% of former mires drained (c. 50% for forestry). Drained minerotrophic mires used for agriculture (mainly grasslands) and peatland forestry. Minerotrophic mires have suffered most from drainage; 70% of bogs still in intact or near pristine conditions with high conservation status. Main problem for bogs: drainage systems in minerotrophic lagg areas and slow afforestation due to surrounding drainage also in protected areas.

240 000 ha in natural conditions (majority ombrotrophic); 17 Ramsar sites designated; 75% of open mire areas protected (90% located within the Estonian Green Network); 178 000 ha of mires preserved; 30 000 ha of peat-cutting and abandoned areas, but total area affected directly or indirectly from peat-cutting 50 000–60 000 ha. Peat harvesting requires restoration. No full-scale restorations done so far, but 2000 ha of peat abandoned extraction areas must be rehabilitated before 2020. Some 300 ha of cut-away sites restored in the 1970s using cranberry plantations to improve hydrological conditions. At least 10 000 ha of edge areas of large bog complexes and rich fens must be restored in Natura 2000 areas to stop degradation in the most valuable mire areas (e.g. Ramsar sites).

Latvia Minerotrophic mires (fens and transition mires) and ombrotrophic mires (raised bogs). Raised bogs more widespread but number of fens greater.

Extraction and utilization of peat for industrial purposes started at the beginning of the 20th century. Peatlands influenced mostly by drainage for peat extraction, agriculture and forestry. Altogether c. 70% of the historical peatland areas drained.

Only a small part (6000 peatland sites) protected as natural areas; 6 Ramsar sites designated. Half of the area left among open peatlands. Some restoration techniques initiated (e.g. Ķemeru Mire) resulting in increasing numbers of waders.

References for Table S1

FinlandAapala, K., Heikkilä, R., Lindholm, T. 1996. Protecting the diversity of Finnish mires, in: Vasander, H. (ed.), Peatlands in Finland.

Finnish Peatland Society, Vantaa, pp. 45–57.Alanen, A., Aapala, K. 2015. Peatland Conservation Working Group's proposal to improve mire conservation. Report no 26. Ministry of

Environment, Helsinki. (In Finnish). Available at: https://helda.helsinki.fi/handle/10138/158285.

31

1140

1141

1142

1143

1144114511461147

Auvinen, A-P., Hildén, M., Toivonen, H., Primmer, E., Niemelä, J., Aapala, K., Bäck, S., Härmä, P., Ikävalko, J., Järvenpää, E., Kaipiainen, H., Korhonen, K.T., Kumela, H., Kärkkäinen, L., Lankoski, J., Laukkanen, M., Mannerkoski, I., Nuutinen, T., Nöjd, A., Punttila, P., Salminen, O., Söderman, G., Törmä, M., Virkkala, R. 2007. Evaluation of the Finnish National Biodiversity Action Plan 1997–2005. Monographs of the Boreal Environment Research 29, Finnish Environment Institute, Helsinki.

Eurola, S., Kaakinen, E. 1979. Ecological criteria of peatland zonation and the Finnish mire site type system, in: Kivinen, E., Heikurainen, L., Pakarinen, P. (eds.), Classification of peat and peatlands. International Peat Society, Hyytiälä, pp. 29–32.

Heikkilä, R. 1995. Finnish unprotected nature valuable peatlands. Unpublished manuscript. Finnish Environmental Institute, Helsinki. (In Finnish).

Järvinen, O., Väisänen, R.A. 1976. Species diversity of Finnish birds, II: Biotopes at the transition between taiga and tundra. Acta Zoologica Fennica 145, 1–35.

Luoto, M., Heikkinen, R.K., Carter, T.R. 2004. Loss of palsa mires in Europe and biological consequences. Environ. Conserv. 31(1), 30–37.

Myllys, M. 1996. Agriculture on peatlands, in: Vasander, H. (ed.), Peatlands in Finland. Finnish Peatland Society, Vantaa, pp. 64–71.Peltola, A. 2004. Finnish Statistical Yearbook of Forestry 2004. Natural Resources Institute Finland, Helsinki. (In Finnish).Peltola A. 2014. Finnish Statistical Yearbook of Forestry 2014. Natural Resources Institute Finland, Helsinki. (In Finnish).Peltomaa, R. 2007. Drainage of forests in Finland. Irrig. Drain. 56, S151-S159.Picken, P.T. 2006. Land-use scenarios for Finnish cut-over peatlands – based on the mineral subsoil characteristics. Bull. Geol. Soc.

Finl. 78, 106–119.Punttila, P., Autio, O., Kotiaho, J.S., Kotze, D.J., Loukola, O.J., Noreika, N., Vuori, A., Vepsäläinen, K. 2016. The effects of drainage and

restoration of pine mires on habitat structure, vegetation and ants. Silva Fenn. 50(2), article id 1462. http://dx.doi.org/10.14214/ sf.1462.

Selin, P. 1999. Industrial use of peatlands and the re-use of cut-away areas in Finland. Jyväskylä Studies in Biological and Environmental Science 79, University of Jyväskylä, Jyväskylä. (In Finnish with English summary).

Similä, M., Aapala, K., Penttinen, J. 2014. Ecological restoration in drained peatlands – best practices from Finland. Metsähallitus, Natural Heritage Services, Finnish Environment Institute, Vantaa.

Vasander, H. 1996. Peatlands in Finland. Finnish Peatland Society, Vantaa.Vasander, H., Tuittila, E.-S., Lode, E., Lundin, L., Ilomets, M., Sallantaus, T., Heikkilä, R., Pitkänen, M.-L., Laine, J. 2003. Status and

restoration of peatlands in northern Europe. Wetl. Ecol. Manag. 11, 51–63.

SwedenHagen, D., Svavarsdottir, K., Nilsson, C., Tolvanen, A.K., Raulund-Rasmussen, K., Aradóttir, Á.L., Fosaa, A., Halldorsson, G. 2013.

Ecological and social dimensions of ecosystem restoration in the Nordic countries. Ecol. Soc. 18(4), 34. http://dx.doi.org/10.5751/ES-05891-180434.

32

1148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175

1176

1177

117811791180

Joosten, H. 2015. Peatlands, climate change mitigation and biodiversity conservation – An issue brief on the importance of peatlands for carbon and biodiversity conservation and the role of drained peatlands as greenhouse gas emission hotspots. Nordic Council of Ministers, Copenhagen.

Joosten, H., Clarke, D. 2002. Wise use of mires and peatlands – Background and principles including a framework for decision-making. International Mire Conservation Group, Greifswald, International Peat Society, Jyväskylä.

Swedish Environmental Protection Agency 2014. Environmental Protection Agency Annual Report 2014. Naturvårdsverket, Stockholm. (In Swedish).

Vasander, H., Tuittila, E.-S., Lode, E., Lundin, L., Ilomets, M., Sallantaus, T., Heikkilä, R., Pitkänen, M.-L., Laine, J. 2003. Status and restoration of peatlands in northern Europe. Wetl. Ecol. Manag. 11, 51–63.

NorwayHagen, D., Svavarsdottir, K., Nilsson, C., Tolvanen, A.K., Raulund-Rasmussen, K., Aradóttir, Á.L., Fosaa, A., Halldorsson, G. 2013.

Ecological and social dimensions of ecosystem restoration in the Nordic countries. Ecol. Soc. 18(4), 34. http://dx.doi.org/10.5751/ES-05891-180434.

Joosten, H. 2015. Peatlands, climate change mitigation and biodiversity conservation – An issue brief on the importance of peatlands for carbon and biodiversity conservation and the role of drained peatlands as greenhouse gas emission hotspots. Nordic Council of Ministers, Copenhagen.

Joosten, H., Clarke, D. 2002. Wise use of mires and peatlands – Background and principles including a framework for decision-making. International Mire Conservation Group, Greifswald, International Peat Society, Jyväskylä.

Moen, A., Dolmen, D., Hassel, K., Ødegaard. F. 2010. Mires, springs and flood plains, in: Kålås, J.A., Henriksen, S., Skjelseth, S., Viken, Å. (eds.), Environmental conditions and impacts for Red List species. Norwegian Biodiversity Information Centre, Trondheim, pp. 51–66.

EstoniaAnon 2013. A prioritised action framework (PAF) for Natura 2000 Estonia for the EU Multiannual Financing Period 2014–2020.

Ministry of Environment, Tallinn. Available at: https://www.envir.ee/sites/default/files/elfinder/article_files/paf_est.pdf.Anon 2015. Action plan for protected mires in Estonia. Ministry of Environment, Tallinn. (In Estonian). Čivić, K., Jones-Walters, L. 2010. Peatlands in Ecological Networks in Europe. European Centre for Nature Conservation, Tilburg. Kimmel, K., Kull, A., Salm, J.-O., Mander, Ü. 2010. The status, conservation and sustainable use of Estonian wetlands. Wetl. Ecol.

Manag. 18, 375–395.Leivits, M. 2016. Report of national biodiversity monitoring 2015: Breeding birds of fens and bogs. Estonian Environment Agency,

Tallinn. (In Estonian).

33

118111821183118411851186118711881189

1190

1191

11921193119411951196119711981199120012011202

1203

1204

12051206120712081209121012111212

Leivits, M., Leivits, A. 2009. Use of sequential aerial photography and LiDAR for mapping Scots Pine (Pinus sylvestris) encroachment and change detection in bird habitats from 1950 to 2008 in Nigula mire. Proceedings of 33rd International Symposium on Remote Sensing of Environment (ISRSE), May 4–8, Stresa.

Masing, V., Botch, M., Läänelaid, A. 2010. Mires of the former Soviet Union. Wetl. Ecol. Manag. 18, 397–433.Masing, V., Paal, J., Kuresoo, A. 2000. Biodiversity of Estonian wetlands, in: Gopal, B., Junk, W.J. and Davis, J.A. (eds.), Biodiversity in

wetlands: assessment, function and conservation. Vol. 1, Backhuys Publishers, Leiden, pp. 259–279.Paal, J. 2005. Estonian mires, in: Steiner, G.M. (ed.), Mires from Siberia to Tierra del Fuego. Staphia 85, Land Oberösterreich,

Biologiezentrum / Oberösterreichische Landesmuseen, Linz, pp. 117–146.Paal, J., Leibak, E. 2011. Estonian Mires: Inventory of Habitats. Regio, Tartu.Paal, T., Paal, J. 2002. Rehabilitation of milled peat areas by cranberry (Oxycoccus palustris L.) plantation, in: Schmilewski, G. and

Rochefort, L. (eds.), Proceedings of the International Peat Symposium, September 3–6, Pärnu. International Peatland Society, Jyväskylä, pp. 280–282.

Vasander, H., Tuittila, E.-S., Lode, E., Lundin, L., Ilomets, M., Sallantaus, T., Heikkilä, R., Pitkänen, M.-L., Laine, J. 2003. Status and restoration of peatlands in northern Europe. Wetl. Ecol. Manag. 11, 51–63.

LatviaBambe, B., Baroniņa, V., Indriksons, A., Kalniņa, L., Ķuze, J., Nusbaums, J., Pakalne, M., Petriņš, A., Pilāte, D., Pilāts, V., Priede, A.,

Rēriha, I., Salmiņa, L., Spuņģis, V., Suško, U. 2008. Mire Conservation and Management in Especially Protected Nature Areas in Latvia. Latvian Fund for Nature, Riga.

Joosten, H., Clarke, D. 2002. Wise use of mires and peatlands – Background and principles including a framework for decision-making. International Mire Conservation Group, Greifswald, International Peat Society, Jyväskylä.

Joosten, H. 2015. Peatlands, climate change mitigation and biodiversity conservation – An issue brief on the importance of peatlands for carbon and biodiversity conservation and the role of drained peatlands as greenhouse gas emission hotspots. Nordic Council of Ministers, Copenhagen.

34

12131214121512161217121812191220122112221223122412251226

1227

1228

12291230123112321233123412351236

1237

Table S2. Information about the peatland bird species included in the Finnish regional indicator (IND) and habitat analysis (HAB). Conservation status is described as inclusion in the Annex I of the EU Birds Directive, and the status in the 2015 Red List of Finnish Bird Species (Tiainen et al. 2016; categories: LC – least concern, NT – near threatened, VU – vulnerable and CR – critically endangered). Sample size is reported as the annual average number of pairs observed in the main belt based on Finnish line transect habitat data (1987–2014). Peatland preferences are given as the ratio of densities in peatland to non-peatland habitats (at least 3). Broad-billed sandpiper Calidris falcinellus was only present in peatland habitats.

Common and scientific name of species Analyses Annex I Red list AveragePreference

count / year ratioMeadow pipit (Anthus pratensis) IND, HAB no NT 1750 4.07Broad-billed sandpiper (Calidris falcinellus) IND, HAB* no NT 50

InfRuff (Calidris pugnax) IND, HAB* yes CR 46 27.42Little bunting (Emberiza pusilla) IND no LC 58 17.51Rustic bunting (Emberiza rustica) IND, HAB no NT 498 4.51Reed bunting (Emberiza schoeniclus) IND, HAB no VU 765 3.07Common snipe (Gallinago gallinago) IND, HAB no VU 352 3.72Common cane (Grus grus) IND yes LC 110 5.98Jack snipe (Lymnocryptes minimus) IND, HAB* no LC 16 46.68Yellow wagtail (Motacilla flava) IND, HAB no NT 1384 11.98Whimbrel (Numenius phaeopus) IND, HAB* no LC 73 8.54Red-necked phalarope (Phalaropus lobatus) HAB* yes VU 21

5.00Eurasian golden plover (Pluvialis apricaria) IND, HAB* yes LC 254

3.18Spotted redshank (Tringa erythropus) IND, HAB* no NT 26 36.68Wood sandpiper (Tringa glareola) IND, HAB yes NT 849 28.50Common Greenshank (Tringa nebularia) IND no LC 171

12.99

* Included in the set of rare waders analysed together in the habitat analyses.

35

1238123912401241124212431244124512461247

1248

12491250

1251

12521253125412551256125712581259126012611262126312641265126612671268126912701271

1272

1273

Table S3. The common peatland bird species selected for each of the regions: Finland, Scandinavia (Sweden and Norway), and the Baltic (Estonia and Latvia), whether the species is included in the Annex I of the EU Birds Directive, and the annual average number of pairs observed in peatland routes after combining data for each of the regions (adult individuals in Sweden; see Supplementary Appendix S1 and section 2 in the main text). The dash ‘–’ means that the species was either not present in the country, not considered as a peatland specialist, or had too little data to calculate a trend. Note that for some of the listed species, sample sizes are shown for a shorter time series (see Table 1 in the main text). The asterisk ‘*’ indicates a conservation status other than least concern.

Common and scientific name of speciesAnnex I Finland Scandinavia Baltic1981–2014 1997–2014 1981–2014

Meadow pipit (Anthus pratensis) ** no 315 (94–915)758 (131–1357)141 (64–403)Broad-billed sandpiper (Calidris falcinellus) no 19 (4–42) – –Ruff (Calidris pugnax) yes 25 (2–76) 18 (5–44) –Little bunting (Emberiza pusilla) no 23 (2–54) – –Rustic bunting (Emberiza rustica) * no 49 (5–130) 55 (15–104) –Reed bunting (Emberiza schoeniclus) no 134 (38–263) 263 (43–485) –Common snipe (Gallinago gallinago) no 131 (43–356) 254 (41–490) 45 (3–205)Common crane (Grus grus) yes 73 (9–227) 322 (62–604) 15 (2–50)Jack snipe (Lymnocryptes minimus) no 17 (4–33) – –Yellow wagtail (Motacilla flava) no 239 (37–826) 468 (90–761) –Eurasian curlew (Numenius arquata) ** no – – 16 (4–39)Whimbrel (Numenius phaeopus) no 77 (10–299) 146 (27–343) 39 (2–114)Eurasian golden plover (Pluvialis apricaria) yes 76 (20–366) 295 (13–532)

123 (31–344)Spotted redshank (Tringa erythropus) no 13 (3–28) 24 (10–49) –Wood sandpiper (Tringa glareola) yes 251 (71–692) 485 (118–816) 77 (16–221)Common greenshank (Tringa nebularia) no 84 (19–184) 249 (41–444) –Common redshank (Tringa totanus) no – – 33 (3–112)Northern lapwing (Vanellus vanellus) ** no – – 51 (5–150)

* E. rustica is globally listed as Vulnerable (VU; BirdLife International 2016). ** N. arquata, V. vanellus and E. rustica are listed as Vulnerable and A. pratensis as Near Threatened (NT) according to the European Red List of Birds (BirdLife International 2015).

36

12741275127612771278127912801281128212831284

1285

12861287

1288128912901291129212931294129512961297129812991300130113021303130413051306

130713081309

1310

1311

Table S4. National population size estimates (breeding pairs) of the 15 species included in the joint peatland bird indicator. Finnish, Swedish, Estonian and Latvian values are based on a recent EU bird reporting directive available at: http://bd.eionet.europa.eu/article12/summary. Norwegian values are based on a publication by Shimmings and Øien (2015).

Species Finland Sweden Norway Estonia LatviaMeadow pipit (Anthus pratensis) 500 000–750 000 544 000–1 105 0003 500 000–4 000 000150 000–200 00037 329–134 213Broad-billed sandpiper (Calidris falcinellus) * 25 000–35 000 4300–7600 1650–1800 – –Ruff (Calidris pugnax) 120 000–150 000 83 000–140 000 1030–1710 3000–4000 259–467Little bunting (Emberiza pusilla) * 13 000–40 000 50–430 15–50 – –Rustic bunting (Emberiza rustica) * 170 000–310 000 11 000–21 000 3–4 – –Reed bunting (Emberiza schoeniclus)210 000–330 000 192 000–448 000 200 000–500 000 60 000–120 000 10 617–36 926Common snipe (Gallinago gallinago) 92 000–180 000 72 000–197 000 50 000–75 000 40 000–60 000 38 329–72 808Common crane (Grus grus) 23 000–50 000 21 000–39 000 1500–2500 7000–8000 1513–2268Jack snipe (Lymnocryptes minimus) * 3500–11 000 6000–12 000 500–1000 30–60 0–1Yellow wagtail (Motacilla flava) 500 000–840 000 241 000–484 000 75 000–150 000 10 000–20 000 28 276–197 393Whimbrel (Numenius phaeopus) 31 000–54 000 7400–15 000 2000–10 000 400–700 43–70Eurasian golden plover (Pluvialis apricaria)10 000–15 00016 000–35 000 150 000–300 000 10–30 0–5Spotted redshank (Tringa erythropus) *9800–27 000 5200–11 000 6000–8650 – –Wood sandpiper (Tringa glareola) * 340 000–560 000 97 000–167 000 20 000–35 000 3000–4000 390–872Common greenshank (Tringa nebularia)46 000–70 000 19 000–390 000 7300–16 000 400–500 7–17

* Species marked with asterisks are almost exclusively found in peatland habitats (Cramp et al. 1977–1994).

37

1312131313141315

1316

1317

13181319

1320

1321

1322

1323

1324

1325

1326

1327

13281329

1330

1331

1332

1333

1334

1335

1336

1337

Table S5. Finnish bird observations for each of the peatland habitat variable combinations for both the individual species and the composite group: open and forested peatlands, ditched and non-ditched forested peatlands, peat extraction areas and non-peat extraction areas, and the three degrees of wetness (wet, ‘intermediate’, and dry). Peat extraction areas and the degree of wetness only refer to open peatlands. Peatland habitat combinations with either 0 or 1 observation (emboldened) were removed from the basic model structure of the habitat preference and range shift model. For the composite group (rare waders), observations have been summed up. More information regarding the composition of the wader group, the model structure and the different peatland habitat variables can be found in the main text (see section 2).

Species Open Forest Ditched Non-ditched Peatext Non-peatextWetness wetWetness intermWetness dry Variables removed

Meadow pipit 459 130 23 566 7 582 140 378 71 noneRustic bunting 2 172 101 73 0 174 0 174 0 peatext, wetnessReed bunting 78 109 21 166 2 185 24 154 9 noneYellow wagtail 431 374 39 766 2 803 122 596 87 noneCommon snipe 82 36 10 108 1 117 35 78 5 peatextWood sandpiper 381 225 29 577 1 605 137 440 29 peatextRare waders (n = 7) 189 43 9 223 3 229 77 134 21 none

38

13381339134013411342134313441345

1346

13471348

1349

13501351

1352

1353

1354

1355

1356

1357

Table S6. Statistics for logistic binomial mixed models explaining the occurrence (probability of observing at least one pair in a given peatland line transect segment) of Finnish peatland birds. The purpose of this analysis is to assess the robustness of the qualitative patterns in the Poisson-based population density analysis (see Table 3 in the main text), with different model assumptions (binomial distribution, “loglength” as a covariate, only “site” as a random effect ensuring a non-zero SD). The results are shown for a total of six separate species and one composite group of waders. The estimated parameters are associated with standard errors (SE), Z-statistics and two-tailed p-values for the null hypothesis of zero effects. Statistically significant effects (p < 0.05) are in bold and tendencies (p < 0.1) are italicised. The effects of “lat” and “lon” are per 100 km, and the effect of year is per decade.

Parameters Estimate SE Z pMeadow pipitα intercept –3.731 0.213 –17.48 < 0.0001β0 loglength 1.500 0.129 11.61 < 0.0001β1 open 3.160 0.234 13.48 < 0.0001β2 ditched –1.077 0.307 –3.51 0.0005β3 peatext –0.077 0.637 –0.12 0.9038β4 wetness 0.283 0.144 1.96 0.0499β5 areasize 0.228 0.131 1.74 0.0818β6 treeheight –2.992 0.686 –4.36 < 0.0001β7 lat 0.061 0.035 1.75 0.0808β8 lon –0.233 0.089 –2.62 0.0089β9 year –0.496 0.107 –4.65 < 0.0001β10 lat×year 0.034 0.038 0.88 0.3766β11 lon×year –0.068 0.079 –0.86 0.3882σs among-site SD 1.022 – – –Reed buntingα intercept –3.947 0.289 –13.65 < 0.0001β0 loglength 1.107 0.165 6.73 < 0.0001β1 open 0.539 0.261 2.07 0.0389β2 ditched –0.602 0.331 –1.82 0.0694β3 peatext 1.288 1.003 1.29 0.1990β4 wetness 0.350 0.240 1.46 0.1451β5 areasize –0.098 0.212 –0.46 0.6430β6 treeheight –1.709 0.743 –2.30 0.0215β7 lat 0.258 0.064 4.04 < 0.0001β8 lon –0.015 0.134 –0.11 0.9096β9 year 0.164 0.180 0.91 0.3639β10 lat×year 0.048 0.072 0.66 0.5067β11 lon×year 0.001 0.133 0.01 0.9913σs among-site SD 1.461 – – –Rustic buntingα intercept –2.153 0.099 –21.71 < 0.0001β0 loglength 1.149 0.139 8.29 < 0.0001β1 open –3.812 0.945 –4.04 < 0.0001β2 ditched 1.160 0.204 5.69 < 0.0001β3 peatext – – – –β4 wetness – – – –β5 areasize 0.443 0.928 0.48 0.6334

39

1358135913601361136213631364136513661367136813691370137113721373137413751376137713781379138013811382138313841385138613871388138913901391139213931394139513961397139813991400140114021403140414051406140714081409

β6 treeheight –0.010 0.428 –0.02 0.9817β7 lat 0.247 0.059 4.19 < 0.0001β8 lon 0.156 0.099 1.58 0.1148β9 year –0.120 0.120 –1.00 0.3154β10 lat×year 0.202 0.064 3.14 0.0017β11 lon×year 0.226 0.099 2.27 0.0230σs among-site SD 0.000 – – –Yellow wagtailα intercept –2.566 0.155 –16.59 < 0.0001β0 loglength 1.286 0.107 12.03 < 0.0001β1 open 1.666 0.176 9.44 < 0.0001β2 ditched –1.503 0.233 –6.46 < 0.0001β3 peatext –1.527 1.123 –1.36 0.1741β4 wetness 0.186 0.141 1.32 0.1856β5 areasize 0.153 0.125 1.22 0.2217β6 treeheight –3.150 0.509 –6.19 < 0.0001β7 lat 0.237 0.038 6.24 < 0.0001β8 lon –0.055 0.082 –0.67 0.5004β9 year –0.349 0.094 –3.71 0.0002β10 lat×year 0.087 0.040 2.20 0.0278β11 lon×year –0.014 0.073 –0.195 0.8450σs among-site SD 0.956 – – –Common snipeα intercept –4.812 0.425 –11.31 < 0.0001β0 loglength 1.097 0.206 5.33 < 0.0001β1 open 1.960 0.308 6.36 < 0.0001β2 ditched –0.707 0.446 –1.59 0.1126β3 peatext – – – –β4 wetness 0.754 0.254 2.98 0.0029β5 areasize –0.103 0.213 –0.48 0.6282β6 treeheight –0.975 0.878 –1.11 0.2667β7 lat –0.021 0.059 –0.35 0.7253β8 lon 0.026 0.160 0.16 0.8721β9 year –0.036 0.181 –0.20 0.8413β10 lat×year 0.014 0.073 0.20 0.8442β11 lon×year –0.045 0.156 –0.29 0.7714σs among-site SD 1.670 – – –Wood sandpiperα intercept –2.550 0.134 –19.01 < 0.0001β0 loglength 1.255 0.107 11.69 < 0.0001β1 open 1.813 0.165 10.96 < 0.0001β2 ditched –1.213 0.238 –5.11 < 0.0001β3 peatext – – – –β4 wetness 0.444 0.140 3.18 0.0015β5 areasize –0.071 0.119 –0.60 0.5464β6 treeheight –1.226 0.460 –2.67 0.0077β7 lat 0.192 0.034 5.72 < 0.0001β8 lon –0.074 0.076 –0.98 0.3282β9 year –0.123 0.092 –1.345 0.1787β10 lat×year 0.088 0.037 2.38 0.0175β11 lon×year –0.013 0.071 –0.19 0.8509σs among-site SD 0.782 – – –Rare wader set

40

14101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462

α intercept –4.648 0.296 –15.71 < 0.0001β0 loglength 1.384 0.171 8.11 < 0.0001β1 open 3.056 0.302 10.11 < 0.0001β2 ditched –0.438 0.436 –1.01 0.3146β3 peatext –0.902 0.884 –1.02 0.3079β4 wetness 0.395 0.182 2.17 0.0302β5 areasize 0.173 0.160 1.08 0.2795β6 treeheight –2.757 0.999 –2.76 0.0058β7 lat 0.016 0.046 0.36 0.7197β8 lon –0.122 0.112 –1.09 0.2769β9 year 0.057 0.150 0.38 0.7033β10 lat×year –0.006 0.055 –0.11 0.9141β11 lon×year 0.004 0.116 0.04 0.9696σs among-site SD 1.107 – – –

41

146314641465146614671468146914701471147214731474147514761477

Figure S1. Example of the Minimum Convex Polygon (MCP) constructed for whimbrel Numenius phaeopus and rustic bunting Emberiza rustica showing transects (data exclusively belonging to high-quality transects) that fall within the core distribution range of the focal species (in grey colour). Black filled circles are routes where the species are observed within the main belt, while open circles represent other routes. Note that some of the routes were sur-veyed in other countries than Finland (Sweden, Norway and Estonia), but we decided to include additional information outside the Finnish boundaries since habitat data were collected using the same protocol in countries which are part of this study, thus supporting the results from the habitat preference and range shift model at a broader scale.

42

14781479148014811482148314841485148614871488

1489

1490

1491

Figure S2. Map showing the percentage of degradation in peatland habitats along each Finnish fixed route due to ditching and peat extraction based on the main belt habitat type classifications done by the observers. This percentage is calculated as follows: (ditched peatlands + peat extraction areas) / (all peatland habitats) * 100. The size of dots denotes the amount of peatland habitat along the census route (from 0 to 6 km), and the different colours refer to the proportion of peatland habitat that has been degraded, with open dots with black edges representing those transects where less than 10% of peatland habitat has been degraded. Note that many peatlands have been historically turned into either farmlands or forests as a result of e.g. ecological succession after ditching, and so this illustration likely gives an underestimate of peatland habitat degradation (i.e. habitats classified in other categories than peatland).

43

1492149314941495149614971498149915001501150215031504

1505

1506

1507

1508

1509

1510

1511

1512

1513

1514

References

Bambe, B., Baroniņa, V., Indriksons, A., Kalniņa, L., Ķuze, J., Nusbaums, J., Pakalne, M., Petriņš, A., Pilāte, D., Pilāts, V., Priede, A., Rēriha, I., Salmiņa, L., Spuņģis, V., Suško, U. 2008. Mire Conservation and Manage-ment in Especially Protected Nature Areas in Latvia. Latvian Fund for Nature, Riga.

BirdLife International 2015. European Red List of Birds. Office for Official Publications of the European Communities, Luxembourg.

BirdLife International 2016. Emberiza rustica. The IUCN Red List of Threatened Species 2016: e.T22720960A89641304. Available at: http://www.iucnredlist.org.

Calladine, J., Garner, G., Wernham, C., Thiel, A. 2009. The influence of survey frequency on population estimates of moorland breeding birds. Bird Study 56(3), 381-388.

Cramp, S., Simmons, K.E.L., Perrins, C.M. 1977–1994. Handbook of the Birds of Europe the Middle East and North Africa: The Birds of the Western Pa-laearctic. Vol. I–IX. Oxford University Press, Oxford.

Fraixedas, S., Lindén, A., Lehikoinen, A. 2015. Population trends of common breeding forest birds in southern Finland are consistent with trends in forest management and climate change. Ornis Fennica 92, 187–203.

Husby, M., Kålås, J.A. 2011. Terrestrial bird species as indicators of sustain-able development in Norway. The state of the various habitats and the ef-fects of climate change. Utredning 128, 48. (In Norwegian).

Irdt, A., Vilbaste, H. 1974. Bird fauna of the Nigula peat bog, in: Kumari, E. (ed.), Estonian wetlands and their life. Estonian Contributions to the In-ternational Biological Programme 7, Valgus, Tallinn, pp. 214–229.

Järvinen, O., Väisänen, R. A. 1975. Estimating relative densities of breeding birds by the line transect method. Oikos 26, 316–322.

Koskimies, P., Väisänen R.A. 1991. Monitoring bird populations: a manual of methods applied in Finland. Zoological Museum, Finnish Museum of Nat-ural History, Helsinki.

Kumari, E. 1972. Changes in the bird fauna of Estonia peat bogs during the last decades. Aquilo, Series Zoology 13, 45–47.

Kuresoo, A. Pehlak, H., Nellis, R. 2011. Population trends of common birds in Estonia in 1983–2010. Estonian Journal of Ecology 60 (2), 88–110.

Kålås, J.A., Husby, M. 2002. Monitoring program for terrestrial ecosystems. Extensive monitoring of terrestrial birds in Norge. NINA Oppdragsmeld-ing 740, 25. (In Norwegian with English summary).

Laaksonen, T., Lehikoinen, A. 2013. Population trends in boreal birds: Con-tinuing declines in agricultural, northern, and long-distance migrant spe-cies. Biol. Conserv. 168, 99–107.

44

1515

1516

1517

1518

15191520152115221523

15241525

152615271528

152915301531

153215331534

153515361537

153815391540

154115421543

15441545

154615471548

15491550

15511552

155315541555

155615571558

Laitinen, J., Rehell, S., Huttunen, A., Tahvanainen, T., Heikkilä, R., Lindholm, T. 2007. Mire systems in Finland – special view to aapa mires and their water-flow pattern. Suo 58, 1–26.Lehikoinen, A. 2013. Climate change, phenology and species detectability in a monitoring scheme. Popul. Ecol. 55, 315–323.

Lehikoinen, A. 2013. Climate change, phenology and species detectability in a monitoring scheme. Popul. Ecol. 55, 315–323.

Lehikoinen, A., Green, M., Husby, M., Kålås, J.A., Lindström, Å. 2014. Com-mon montane birds are declining in northern Europe. J. Avian Biol. 45, 3–14.

Leivits, M. 2014. Report of national biodiversity monitoring 2014: Breeding birds of fens and bogs. Estonian Environment Agency, Tallinn. (In Esto-nian).

Leivits, A., Leivits, M., Pehlak, H. 2013. Reconstruction of population trends of waders using historical and repeated mire bird survey data in Estonia, in: "Every bird counts" Book of abstracts of the 19th Conference of the European Bird Census Council. Babeș-Bolyai University, Romanian Orni-thological Society (BirdLife Romania), Milvus Group, Cluj-Napoca.

Lindström, Å., Green, M., Paulson, G., Smith, H. G., Devictor, V. 2013. Rapid changes in bird community composition at multiple temporal and spatial scales in response to recent climate change. Ecography 36, 313–322.

Lindström, Å., Green, M., Husby, M., Kålås, J.A., Lehikoinen, A. 2015. Large-scale monitoring of waders on their boreal and arctic breeding grounds in northern Europe. Ardea 103(1), 59–71.

Renno, O. 1974. Bird fauna of the West-Estonian Fens and its variation during the reclamation of fens, in: Kumari, E. (ed.), Estonian wetlands and their life. Estonian Contributions to the International Biological Programme 7, Valgus, Tallinn, pp. 204–213.

Shimmings, P., Øien, I.J. 2015. Population estimates for Norwegian breeding birds. Report no 2015–2. Norwegian Ornithology Society, Trondheim. (In Norwegian).

Tiainen, J., Mikkola-Roos, M., Below, A., Jukarainen, A., Lehikoinen, A., Le-htiniemi, T., Pessa, J., Rajasärkkä, A., Rintala, J., Sirkiä, P., Valkama, J. 2016. The 2015 Red List of Finnish Bird Species. Ministry of the Environ-ment and Finnish Environment Institute, Helsinki.

Väisänen, R.A., Lammi, E., Koskimies, P. 1998. Distribution, numbers and pop-ulation changes of Finnish breeding birds. Otava, Helsinki. (In Finnish with English summary).

45

15591560156115621563

15641565

156615671568

156915701571

15721573157415751576

157715781579

158015811582

1583158415851586

158715881589

1590159115921593

159415951596

1597

1598

1599