Journal of Ornithology

, Volume 159, Issue 2, pp 413–424 | Cite as

Climate change, predictive modelling and grassland specialists: assessing impacts of changing climate on the long-term conservation of Lesser Grey Shrikes (Lanius minor) in Romania

Original Article

Abstract

Climate change is considered one of the greatest challenges that current biodiversity is facing. Successful adaptation of different species to climate-related changes depends on their ability to follow the habitat shift by modifying their range. We assessed the projected future range changes for a grassland specialist bird using two available climate scenarios. The model subject, the Lesser Grey Shrike (Lanius minor), is a vulnerable grassland specialist, distributed in southeastern Europe, with its European population concentrated in Romania. We created a distribution model for the species using data provided by the national Common Bird Monitoring Programme for the years 2002–2005. Several different statistical models, based on the generalised linear model and multivariate adaptive regression splines, were tested with use of the available habitat and climate data. The final working model was selected by means of the lowest root mean square error from the cross-validation process. The model was tested against two climate scenarios—A1 (integrated world, rapid economic growth) and B2 (regional development, environmentally friendly scenario)—on a long-term (2050) scale. To assess the efficiency of site-based conservation (Natura 2000 sites) as the only tool currently in place for the preservation of Lesser Grey Shrike populations in Romania, we evaluated the mean changes in suitable habitats inside the national protected area network. The projected changes show large-scale reduction of suitable habitats, both inside protected areas and at the national level, due to the forecasted shifts in grassland belts. Our results show that under both scenarios, two thirds of the seminatural grasslands will be out of the recent protected area system by 2050. Small protected areas will lose more habitats than larger ones, irrespective of the Lesser Grey Shrike populations breeding therein. These results suggest that current site-based protection measures will become largely insufficient for the conservation of seminatural grasslands and the associated flora and fauna in the long term in Romania.

Keywords

Grassland Climate change Lanius minor Distribution model Protected area Range 

Zusammenfassung

Klimawandel, Modellvoraussagen und Weidelandspezialisten: Einschätzung der Bedeutung des sich ändernden Klimas auf den langfristigen Schutz von Schwarzstirnwürgern ( Lanius minor ) in Rumänien Der Klimawandel gilt als eine der größten Herausforderungen, mit denen die derzeitige Artenvielfalt konfrontiert ist. Eine erfolgreiche Anpassung verschiedener Arten an klimatisch bedingte Veränderungen hängt von ihrer Fähigkeit ab, der Verschiebung des Lebensraumes durch eine Änderung ihres Verbreitungsgebietes zu begegnen. Wir schätzten die zu erwartenden zukünftigen Änderungen in der Verbreitung einer auf Weideland spezialisierten Vogelart basierend auf zwei zur Verfügung stehenden Klimaszenarien. Die Modellart, der in Südosteuropa verbreitete Schwarzstirnwürger (Lanius minor), ist ein gefährdeter Weidelandspezialist, dessen europäische Population ihren Schwerpunkt in Rumänien hat. Auf der Grundlage von Daten aus dem nationalen Monitoring häufiger Brutvogelarten aus den Jahren 2002-2005 erstellten wir ein Verbreitungsmodell für die Art. Wir testeten eine Reihe verschiedener statistischer Modelle auf der Grundlage Generalisierter Linearer Modelle (GLM) und Multivariater Regressiver Regressions-Splinen (MARS), unter Verwendung der verfügbaren Habitat- und Klimadaten. Das endgültige Arbeitsmodell wurde anhand des niedrigsten RMSE-Wertes aus dem Kreuzvalidierungsprozess ausgewählt. Dieses Modell wurde für zwei Klimaszenarien A1 (Globalisierung, schnelles Wirtschaftswachstum) und B2 (regionale Entwicklung, umweltfreundliches Szenario) im langfristigen Maßstab (bis 2050) getestet. Um die Wirksamkeit von flächenbezogenen Schutzmaßnahmen (Natura 2000-Gebiete) zu beurteilen, welche derzeit das einzige Mittel zum Erhalt der rumänischen Schwarzstirnwürger-Populationen darstellen, bewerteten wir die mittleren Veränderungen an geeigneten Lebensräumen innerhalb des Netzwerkes nationaler Schutzgebiete. Die vermuteten Veränderungen zeigen großflächiges Verschwinden geeigneter Habitate, sowohl innerhalb der Schutzgebiete als auch auf Landesebene, aufgrund der zu erwartenden Verschiebungen der Weidelandgürtel. Unsere Ergebnisse zeigen, dass bei beiden Szenarien bis 2050 zwei Drittel des halbnatürlichen Weidelandes außerhalb des derzeitigen Netzwerkes von Schutzgebieten liegen wird. Kleinere Schutzgebiete werden höhere Habitatverluste erleiden als größere, unabhängig davon, ob dort Schwarzstirnwürger-Populationen brüten. Diese Befunde legen nahe, dass die derzeitigen flächenbezogenen Schutzmaßnahmen in Rumänien langfristig zum Erhalt des halbnatürlichen Weidelandes und der dort lebenden Flora und Fauna nicht mehr ausreichen werden.

Introduction

Climate change, along with habitat alterations at landscape level, are considered by most as major driving forces of the current and future biodiversity loss all over the world (Thomas et al. 2004; Harrison et al. 2006; Pearce-Higgins et al. 2011; Regos et al. 2015). A number of theoretical and empirical studies showed the impact of climate change (and the associated changes in environmental characteristics) on the survival or distribution of different species (La Sorte and Jetz 2010; Schwartz 2012; Saunders et al. 2013). In the case of birds these changes are primarily caused by changes in food resources (e.g. shifts in the temporal or spatial distribution of prey species) (Pearce-Higgins et al. 2010) or predation risk (Hamer 2010). Moreover, there are a number of indirect effects on the life of birds through the land-use changes implemented by humans to counteract the influences of climate change especially on agriculture (Jetz et al. 2007; Kleijn et al. 2010), but also other sectors such as forestry (Ruiz-Labourdette et al. 2012) and fishery (Cheung et al. 2010). Several studies highlighted the potential conservation problems related to climate change, such as those caused by shifts in breeding range elevation and/or latitude (Tryjanowski et al. 2005; LaSorte and Jetz 2010; Chamberlain et al. 2013), in migration timing (Sauter et al. 2010) or in breeding (Tryjanowski et al. 2004; Smallegange et al. 2011; Saunders et al. 2013). Climate change thus may accelerate the risk of extinction by pushing species range limits out of the available habitat cover or decreasing densities below recovery levels (Thomas et al. 2004; Conroy et al. 2011). It has been stressed that future conservation scenarios should build climate aspects into the planning process in a proactive way (Conroy et al. 2011; Strange et al. 2011; Wiens et al. 2011).

The cornerstone of biodiversity conservation in the European Union is the protected area network (Natura 2000 network) established under the Habitats Directive (92/43/EEC) and the Birds Directive (79/409/EEC) (BirdLife International 2004b). This network promotes species conservation primarily through the protection of sites which contain important populations of these species (the ‘‘species of Community interest’’, species listed in annexes of these two directives). These sites were selected with use of historical data on the presence of focal species or habitats. Thus, they represent the actual state of knowledge of fauna and flora at the moment of designation (Gaston et al. 2008). According to Article 6(1) of the Habitats Directive (92/43/EEC), each European Union member state should ensure that the conservation of its national populations of any species of Community interest achieves and/or maintains the focal species in a ‘favourable conservation state’. In most readings ‘favourable conservation status’ for a species means a stable population figure, or geographical range, to prevent major population declines (López–López et al. 2007). Moreover, this network has a continental-wide approach, with a recent assessment showing its suitability and coverage being adequate (Albuquerque et al. 2013) and also effective in conserving selected species (Gamero et al. 2017). Romania launched its Special Areas of Protection (SAP) network in 2007 (Papp and Sándor 2007), with further improvement and enlargement in 2011 (Sándor and Domsa 2012). Although the designation of these sites was based on the ‘best available data’, there was no assessment of the efficacy of these sites in maintaining, in the long term, the populations of threatened species in spite of future climate change events.

Modelling species distribution has a long history, and a large number of different methods and approaches are available (La Sorte and Jetz 2010). The basic principle applied is that environmental conditions at a certain occurrence and location represent the ecological niche of the species in question, and thus may be used to develop distribution models (Guisan and Zimmermann 2000). Although this may be generally appropriate, the performance of such models may differ depending on the amount and quality of data input (del Barrio et al. 2006; Elith and Graham 2009). Factors creating bias may include the number of records (Wisz et al. 2008), the ratio of presence to absence (Brotons et al. 2004), heterogeneity of the study area (Brotons et al. 2012), species’ habitat tolerance (Hernandez et al. 2006), type of bioclimatic envelope (Elith et al. 2006) and the resolution of environmental data layers (Araújo and Guisan 2006). Usually, the distribution of species with a limited range and specialised and homogenous habitat needs can be predicted with high accuracy with the existing methods even in spite of reduced location records, while creating distribution maps for generalist species with large distribution areas requires more data (Leathwick et al. 2006). The use of bioclimatic models is generally accepted for the assessment of potential climate-induced range shifts (Schwartz 2012), to estimate extinction rates (Leathwick et al. 2006), to examine the efficacy of existing reserve systems (Araujo et al. 2011), or to identify priority areas for conservation (Conroy et al. 2011). Recent studies examined such relationships at the level of animal groups at national levels (Triviño et al. 2013) or for species range (Huntley et al. 2007); however, to our knowledge, there is no study which uses a bioclimatic envelope in assessing the future distribution of a certain species at the national level in relation to protected areas designated for the conservation of the species.

We selected a species with limited distribution and highly specialised habitat needs, the Lesser Grey Shrike (Lanius minor), to model the future distribution of a conservation-dependent species in Romania using habitat as a surrogate as a conservation goal. Our aim is to estimate the probable future distribution of the species in Romania to assess the usefulness of the site-based conservation approach for maintaining populations of Lesser Grey Shrikes in the long term in Romania.

Methods

Model species

The Lesser Grey Shrike is a grassland-dependent species, living primarily in the steppe belt of eastern Europe and the Middle East, with a continuously shrinking range (Lefranc 1995; Lefranc and Worfolk 1997; Giralt et al. 2008). Although the species is not an exclusive grassland specialist (requires the presence of lonely or scattered trees as nest support), its distribution is linked to the presence of extensive grasslands used primarily as pastures, avoiding forests or Mediterranean maquis/garrigue, while the remnant central European populations breed in traditional orchards (Lefranc 1995; Krištín et al. 2007). The European strongholds of the species are located in Romania, Ukraine and southern Russia, while important populations are concentrated in Turkey and the Caucasus (BirdLife International 2004a). In this vast region, the species breeds nearly exclusively in the former forest steppe (these are dry grasslands, in terms of mesic and xeric grasslands), commonly breeding in treelines along roads and the margins of plantations (Lefranc 1995). It is a common breeding species in most seminatural lowland grasslands of Romania, occurring especially in the central, southwestern, eastern and southeastern parts of the country. The Romanian population is very important at the European level, accounting for 55–60% of the total European population (BirdLife International 2004b), with large-scale declines noted recently (A. Sándor, unpublished data). Being a long-term migrant, it spends only 4–5 months in the breeding areas, with the main wintering range being located in sub-Saharan Africa (Herremans 1998). It is a declining species all over its range, with habitat alteration and decrease being the most important causes identified up to now (Lefranc 1995; Krištín 2008). In the European Union (and Romania), the species is protected under the European Union Birds Directive, listed under Annex I (species requiring the establishment of special protected areas for the maintenance of their respective national populations).

Distribution data

The distribution data used for the present study were collected during the Romanian Common Bird Monitoring Programme (coordinated by the Romanian Ornithological Society) in the period from 2001 to 2011. The programme uses a 5-min point count survey technique on randomly selected, 2.5 × 2.5 km squares (2001–2005) in the breeding period (April to June). While the Common Bird Monitoring Programme uses two counts (early and late season), in our analysis we used only the data collected in the second count (after 15 May) to avoid possible bias caused by passage migrants. Each square has 15 survey points, and the data collected at the point level are treated as one sample (Fig. 1). The survey is performed each year in a standardised manner by volunteers. Although the survey uses volunteers with general bird-identification skills, they were especially trained for the identification of shrikes—Lesser Grey Shrike and Red-backed Shrike (Lanius collurio)—in the years 2001–2004. In the period from 2003 to 2005 all volunteers were required to fill in a form to assess their identification skills on a yearly basis, and data from only those volunteers who were familiar with Lesser Grey Shrikes were used. Given the fact that our aim is to have a general picture of the distribution of populations, the monitoring data were considered as a unit; therefore, we may have a “snapshot” picture of the birds’ distribution for this period. For most monitoring points, data from more than one year are available. In those cases, we applied the following approach: for a given year we used the maximum count and between years we used means. So, our data represent the means of maximum counts. We considered a “zero count” to mean the species was not recorded at a particular point. This assumption is based on the fact that our target is a widely recognisable common species, breeding in loose colonies, highly territorial and vocal in the period of the monitoring (May to June). The climatic limits of the species’ global distribution according to Cramp and Perrins (1993) were used. For modelling purposes we used only the higher (southern) thermal tolerance limit, as the lower thermal tolerance limits (and species’ distribution borders) lay to the north of Romania. As a consequence we used only the southern (higher) thermal limits of the Lesser Grey Shrike distribution for our modelling exercise. We ran the models using two different limits, as there are differences in thermal tolerance between populations breeding in Europe (13.5 °C multiannual isotherm corresponding to the actual southern range limits of the species in continental Europe) and the Middle East (an upper thermal tolerance at 14.5 °C multiannual isotherm, which borders the species range in the Middle East; Huntley et al. 2007; Lefranc and Worfolk 1997).
Fig. 1

The distribution of randomly selected point count locations at the national level

Ecological data sets (predictors) used

Several ecological data sets were used to model the birds’ distribution at the national level. These include the spatial location (national projection system, Stereografic 1970), altitude (WorldClim), habitat/land use (general distribution of land use based on CORINE LandCover 2006, CLC) and climatic variables (temperature, precipitation and evapotranspiration, available from the WorldClim database; Hijmans et al. 2005). Given the low representation of some habitats, several similar habitats were grouped together. In total, 18 different land-cover categories were assumed to cover grassland habitats and were used in this study (Table 1). The CLC data set has, in general, a low spatial resolution, with a minimum mapping unit of 25 ha, and thus for statistical analysis we used the relative numerical values. In this analysis, we used the percentage of each habitat category type divided by the total possible amount (200-m observation circle, 12.56 ha). For national scale modelling of the Lesser Grey Shrike’s distribution we used the percentage of each habitat category type from the total 2 km × 2 km grid cell. As a premise, throughout the modelling process we assumed that there will be no changes in the spatial distribution of grasslands with time. This presumption is based on the current legal requirements of land-use changes, which forbid any alteration of grasslands and forested areas in Romania.
Table 1

CORINE LandCover (CLC) categories used in modelling the geographical distribution of Lesser-Grey Shrike (Lanius minor) populations in Romania

CLC level 3

Land-use variable no.

Category name

111

1

Inhabited areas

112

141

142

121

2

Industrial units

122

123

124

132

133

131

3

Mineral extraction sites

211

4

Agricultural areas

212

213

244

221

5

Vineyards

222

6

Orchards

231

7

Pastures

242

8

Complex cultivated areas

243

9

Agricultural areas with natural vegetation

311

10

Deciduous forests

312

11

Coniferous forests

313

12

Mixed forests

321

13

Natural vegetation areas

322

324

14

Transitional forest—scrub areas

331

15

Sparsely vegetated areas

332

333

334

411

16

Wetlands

412

421

511

17

Rivers

512

18

Lakes and sea

521

523

The climatic data used are available from the WorldClim database, obtained for the 1960–2000 period. Climate raster data cells have a minimum size of 630 m × 630 m. The data were used in statistical modelling in two ways: as annual means and as April–June (monitoring period) means. In spatial analysis, each grid cell got its value by calculation of the mean among the overlapping climatic grid cells. For different future climate scenarios we used the available data from the CGIAR Research Program on Climate Change, Agriculture and Food Security. Different data sets were used to assess the predicted climatic changes that may occur in the future. All the data sets used belong to the HadCM3 general atmosphere circulation model (Ramirez and Jarvis 2008). Two different climate scenarios were used—A1 (integrated world, rapid economic growth) and B2 (regional development, environmentally friendly scenario)–with the time limit set to 2050 (Araujo et al. 2011; Seavy et al. 2008).

For statistical and spatial modelling, two general statistical methods were used: the generalised linear model (GLM) and multivariate adaptive regression splines (MARS) (Hastie and Pregibon 1992; Leathwick et al. 2006). Among the large number of statistical models tested, these two had the best performance for the data sets used. Combined with ecological predictors, six different models were built and used for the analysis (Table 2).
Table 2

The statistical models tested

Name

Statistics used

Predictors

RMSE

GLM1

GLM

Coordinates, altitude, habitats

0.3647

GLM2

GLM

Coordinates, altitude, habitats, annual climatic data

0.3614

GLM3

GLM

Coordinates, altitude, habitats, monthly climatic dataa

0.3625

MARS1

MARS

Coordinates, altitude, habitats

0.3651

MARS2

MARS

Coordinates, altitude, habitats, annual climatic data

0.3663

MARS3

MARS

Coordinates, altitude, habitats, monthly climatic dataa

0.3649

GLM generalised linear model, MARS multivariate adaptive regression splines, RMSE root mean square error

a“Monthly” means for April to June.

The statistical modelling was implemented in the R software environment (R Development Core Team 2009) with use of the modules available on the software’s website. We used TriMmaps (developed by Sovon, Nijmegen, Netherlands), for it possesses the capacity of combining the statistical and spatial modules available for R. The advantage is obvious, since data processing can be done in one step. First, the software builds the statistical model (based on regression analysis), then, by extrapolation, using the fitted regression equation, it calculates the predicted values for each grid cell. The selection of the most appropriate model was done by a tenfold cross-validation process. For each model, this process outputs the calculated values, such as root mean square error (RMSE). The model having the lowest RMSE was selected (Hastie and Pregibon 1992). The final model selected was considered the best available image (based on the best available data) of the Lesser Grey Shrike’s distribution in Romania, and hence was used in evaluating how future predicted climate change will shape the range of the species. All the analyses for this article (including species’ habitat distribution in the protected area network) are based on the selected model. To test the suitability of the current protected area network in maintaining Lesser Grey Shrike populations in the long term, we assessed the changes in habitat cover inside protected areas in each climate scenario and for each time scale. To test whether the mean habitat cover inside a particular protected area increased or decreased, we performed one-sample t tests. Each test tested the null hypothesis that the mean change in occupancy inside a given protected area is zero. We tested differentially the model for specially protected areas (SPAs; sites of Community interest declared for the long-term conservation of bird species under Annex I of the Bird Directive) and the whole protected area network in Romania (including all protected natural sites, irrespective of the designation). Performance of a high number of repeated tests using the same dataset would result in an elevated risk of a type 1 error (Zar 1996). To account for this factor, we applied the Bonferroni correction, adjusting the 0.05 level of significance (α) to 0.00065 (n = 76). To assess the climate-buffer capacity of differently sized protected areas, we divided the current SPA network with an arbitrary set size of 10,000 ha (which is close to the median of the size distribution of the SPAs, 9650 ha) into ‘small’ (below 10,000 ha) and ‘large’ (above 10,000 ha) SPAs and repeated the one-sample t tests, adjusting the Bonferroni correction according to the sample size.

Results

Distribution range

For the purpose of this article, we used data collected in the Common Bird Monitoring Programme between 2002 and 2005. These data represent a ‘snapshot image¨ of the Lesser Grey Shrike distribution in this time frame. In total, we used point-count data from 3227 points. The species was recorded at 197 points, representing only 6.1% of the total. The total number of recorded pairs was 332 (x = 1.69, standard deviation 1.16). Information was used from all the monitoring points (both zero and positive counts) to build the statistical and spatial distribution models.

Six different statistical models were fitted to calculate the spatial distribution of the abundance for the Lesser Grey Shrike, based on field data. All six models performed well, with similar results. The model with the lowest RMSE (0.3614) was GLM2 (based on coordinates, altitude, habitats and annual climatic data), and this was chosen as the best fit for the available data and was subsequently used in all the analyses for this article (Table 2). The resulting map using the best model is presented in Fig. 2. For modelling the future distribution, we used the available habitat for the species, based on CLC data. As a proxy for suitable habitats for Lesser Grey Shrikes we used land-use categories which cover natural and seminatural grasslands (CLC categories 231, 243, 321, and 322). We considered all suitable habitats where the species was recorded as of equal value for conservation of the species. We tested the correspondence between the modelled distribution and the available habitat. There is a good correspondence, with 90% of the modelled species distribution (in terms of breeding pairs) falling inside the land-cover categories considered (Fig. 3). Hence, we considered the selected habitats a good proxy in modelling the presence of the species, and used those values for future scenarios. According to the distribution model, there are two areas with high concentration at the national level: the southeast and the northwest of the country.
Fig. 2

The modelled distribution of Lesser Grey Shrikes (Lanius minor) in Romania

Fig. 3

The distribution of grasslands suitable for Lesser Grey Shrikes in Romania (based on CORINE LandCover 2006 categories 231, 243, and 244)

From the total of 149 SPAs, 76 sites were designated also for the Lesser Grey Shrike protection. These represent 51.4% of the total number and 39.9% in terms of total SPA area.

Climate-driven changes in distribution

The results project a sharp decline in available habitat for both climate scenarios for both thermal limits considered, with maximum loss in the case of climate scenario A1 and 13.5 °C limit (Figs. 4, 5). The change is similar if we consider only the habitats inside the current SPA network, with nearly 60% habitat loss in the case of the higher temperature tolerance limit (14.5 °C), and 84% habitat loss in the case of the lower thermal tolerance limit. The synthetic results are presented in Table 3. The minimum values of suitable habitat loss refer to the 14.5 °C temperature limit and the maximum values of suitable habitat loss refer to the 13.5 °C temperature limit.
Fig. 4

Projected range change for Lesser Grey Shrike habitats at the national level (2050, scenario A1, both thermal limits)

Fig. 5

Projected range change for Lesser Grey Shrike habitats at the national level (2050, scenario B2, both thermal limits)

Table 3

Projected changes in suitable habitat (per cent changes of habitat cover inside protected areas) suitable for Lesser-Grey Shrikes (Lanius minor) in specially protected areas (SPAs) designated for the species in Romania

Thermal minimum (°C)

Climate scenario

Total SPA (n = 76)

‘Small’ SPA (n = 39)

‘Large’ SPA (n = 37)

Change in habitat surface

t

P

Change in habitat surface

t

P

Change in habitat surface

t

P

13.5

A1

− 0.62 (±0.47)

− 11.59

**

− 0.68 (± 0.46)

− 9.21

**

− 0.56 (± 0.47)

− 7.21

**

B2

− 0.52 (±0.48)

− 9.48

**

− 0.57 (± 0.49)

− 7.19

**

− 0.48 (± 0.47)

− 6.16

**

14.5

A1

− 0.33 (±0.45)

− 6.58

**

− 0.40 (± 0.48)

− 5.24

**

− 0.27 (± 0.41)

− 4.04

**

B2

− 0.11 (±0.28)

− 3.45

*

− 0.13 (± 0.32)

− 2.65

*

− 0.09 (± 0.25)

− 2.18

*

The statistics refer to single-sample t tests that tested the significance of change in suitable habitat cover of both thermal limits and each climate scenario. Significant differences (P < 0.05) are indicated by one asterisk, and those significant after the Bonferroni correction are indicated by two asterisks (total SPA P < 0.000657; ‘small’ SPA P < 0.00129; ‘large’ SPA P < 0.00135).

Significant habitat loss is projected for both climatic scenarios in geographical distribution too. Habitat decrease shows an uneven distribution, with larger decline noted in the southeastern and northwestern parts of the country (Fig. 2 vs Figs. 4, 5). When comparing habitat loss with the current spatial distribution, one can observe an overlap of areas most affected by habitat loss with areas of high concentration of the species according to our model. The scale of the range decrease, however, is not similar in the ‘small’ and ‘large’ SPAs, with the ‘small’ SPAs losing more habitat than the ‘large’ SPAs, this difference being even significant in the case of scenario B2 for the thermal limit of 13.5 °C (Table 3). By extending our model presumptions to the whole of the protected area network in Romania (including all protected areas in the country, not only the SPA network), using the presumption that Lesser Grey Shrikes may relocate to any suitable habitat available—the results remain nearly the same (less than 1% difference between the observed areas for both thermal limits and scenarios, not shown here).

Discussion

Range modifications are predicted for most bird species under future climatic conditions, and these changes will affect especially species with reduced range or lowland distribution (Huntley et al. 2007; Hamer 2010; Araujo et al. 2011; Unander et al. 2016). A number of empirical studies have demonstrated that carefully implemented bioclimatic models can be used to track possible directions of range changes driven by climate (D’Alba et al. 2010; Sauter et al. 2010; Conroy et al. 2011; Morelli and Tryjanowski 2017). Our results suggest that one of the largest Lesser Grey Shrike populations of Europe (BirdLife International 2004b) may be highly vulnerable to climate change, even in areas where the species is currently abundant. Of course, there are a number of uncertainties with regard to the magnitude of modelled range changes, as these may be correlated with several unmeasured factors, such as changes in the distribution and intensity of agricultural use, interdependence between different species or results of climatic events in migratory/wintering areas, all of which may have cascading effects and may cause adverse impacts on the species (Fischer et al. 2011; Conroy et al. 2011; Bertelsmeier et al. 2013; Chamberlain et al. 2013). However, evidence suggests that even rough models can adequately mirror the tendency of range changes with reasonable accuracy (La Sorte and Jetz 2010; Pearce-Higgins et al. 2011).

The results of our modelling are consistent with the general pattern observed in the case of a number of other bird species. Recent studies show that species’ distributions are already responding to the changing climate (Hamer 2010; Koleček et al. 2010; Marcer et al. 2013) and that the rate at which they do so may increase in the future (Huntley et al. 2007; Bertelsmeier et al. 2013). Although our study is based on a rather small set of occurrence data (only 6.1% of points hold breeding Lesser Grey Shrikes), it is still one of the most comprehensive data sets for any bird species breeding in grassland habitats in Romania. The low encounter rate is possibly due to both rarity of the species and the often clustered distribution as the species may breed semicolonially (Lefranc and Worfolk 1997). Still, our modelled distribution overlaps greatly with the potential habitat where the species was ever recorded. As our modelled distribution overlaps with more than 90% of the known distribution in the country, we consider that the proxy (the dry grasslands) used models well the species’ niche. Moreover, using the habitat as a surrogate for the modelling may reduce the bias caused by the clustered distribution.

The forecasted habitat decrease and range contraction is consistent with the climate scenario used and will be most intense in the case of scenario A1. Obviously, we cannot have a real prediction of the density distribution in case of future time series, as there are so many unknowns in the evolution of the factors limiting the Lesser Grey Shrike density distribution (such as habitat alteration, population changes independent of climate effects and changes in agricultural policies). However, assuming that the species habitat selection will remain conservative, range reduction caused by a decrease in available habitat will affect the population size even more severely, as the main habitat loss will be in the areas of maximum density (Fig. 6). This decrease will be general, showing low resilience potential against climate change for the current protected area system in Romania. This is consistent with the conclusions drawn by Araujo et al. (2011), who found that while protected areas in mountains may offer reliable climate refugia, those designated primarily on flatlands will incur proportionally larger range losses. On the basis of our model the Lesser Grey Shrike is primarily an open-country grassland specialist, with a distribution largely on lowlands, and thus the projected range changes may be greater in lowland areas than in hilly areas. This was shown by our results as well, where the western and southeastern (lowland) areas have a higher habitat-decrease rate than the hillier central part (Fig. 5).
Fig. 6

Projected range change for Lesser Grey Shrike habitats inside protected areas designated for the species, both scenarios, both thermal limits (13.5 and 14.5 °C multiannual temperature means)

The level of forecasted habitat decrease ranges between 8% and 41% (higher temperature limit set to 14.5 °C) and between 62 and 75% (lower temperature limit set to 13.5 °C) at the national level. Although the magnitude of this range decrease is smaller in the case of the higher temperature limit, even in this case the decrease may be important inside the protected areas designated for the conservation of the species (SPAs). The projected 60–84% decrease (scenario A1, Fig. 6) may jeopardise most conservation efforts targeting the species. Currently Lesser Grey Shrike conservation is based purely on the existence of specially designated protected areas, namely the Natura 2000 network, with no alternative conservation measure in force in Romania (Iojă et al. 2010). This coupled with the fact that there is a large-scale decline of the species due to recent changes in land use practices (Sándor and Domşa unpublished data) means there is an urgent need to assess the efficiency of these sites in terms of climate resilience.

We are aware that our approach is merely a simplification of complex ecological processes, and despite the ever-advancing methodological fine-tuning one needs to exercise caution when trying to derive conclusions with policy relevance. However, we think that our modelled results may point to important issues. Although the current policy of reduction of greenhouse gas emissions may help mitigate climate change impacts on biodiversity, maintaining healthy populations of certain bird species may require further efforts (Harrison et al. 2006; Araújo et al. 2011; Fischer et al. 2011). This is especially true for Romania, where biodiversity conservation relies only on reserve delimitations, in most cases without any management policy or activity (Iojă et al. 2010; Popescu et al. 2013). Moreover, because small and isolated areas have a higher capacity for loose breeding populations, the current reserve network does not seem to fit the long-term conservation objectives for which they were designated. We argue that for the conservation of Lesser Grey Shrikes there is a need for the implementation of mechanisms for integrated management of agricultural areas (especially grasslands with scattered tall trees), primarily to facilitate the dispersion of the species breeding populations between conservation areas and to increase the potential of climate resilience of the protected area network through the designation of new areas acting as corridors or temporary buffers where the species may survive (Caroll et al. 2010; Strange et al. 2011). Moreover, similar results were found for a number of other species across Europe, suggesting that conservation based only on a single type of measure will not bring the forecasted results (Wiens et al. 2011). The combination of the protected area management with alternative ways of conservation such as agri-environmental schemes (Arponen et al. 2013) or a different approach to agricultural management at landscape level (Troupin and Carmel 2014) may contribute to the long-term sustainability of species specific to grasslands. This will require a major shift in current conservation policies regionally or locally, and such modelling exercises may provide very basic guidance for starting this process.

Notes

Acknowledgements

The Common Bird Monitoring Programme in Romania is implemented by the Romanian Ornithological Society, the Milvus Group and Babes-Bolyai University, Cluj, and was supported by the Royal Society for the Protection of Birds and the European Bird Census Council (through the Pan-European Common Bird Monitoring Scheme), and the Ministry of Agriculture and Rural Development (2009–2010). We thank all the volunteers who participated in the field surveys. András Báldi commented on the manuscript, significantly improving its contents.

References

  1. Albuquerque FS, Assunção-Albuquerque MJT, Cayuela L, Zamora R, Benito BM (2013) European bird distribution is “well” represented by special protected areas: mission accomplished? Biol Conserv 159:45–50CrossRefGoogle Scholar
  2. Araújo MB, Guisan A (2006) Five (or so) challenges for species distribution modelling. J Biogeogr 33:1677–1688CrossRefGoogle Scholar
  3. Araujo MB, Alagador D, Cabeza M, Nogues-Bravo D, Thuiller W (2011) Climate change threatens European conservation areas. Ecol Lett 14:484–492CrossRefPubMedPubMedCentralGoogle Scholar
  4. Arponen A, Heikkinen RK, Paloniemi R, Pöyry J, Similä J, Kuussaari M (2013) Improving conservation planning for semi-natural grasslands: integrating connectivity into agri-environment schemes. Biol Conserv 160:234–241CrossRefGoogle Scholar
  5. Bertelsmeier C, Luque GM, Courchamp F (2013) The impact of climate change changes over time. Biol Conserv 167:107–115CrossRefGoogle Scholar
  6. BirdLife International (2004a) Birds in Europe: population estimates, trends and conservation status. BirdLife International, CambridgeGoogle Scholar
  7. BirdLife International (2004b) Birds in the European union: a status assessment. BirdLife International, WageningenGoogle Scholar
  8. Brotons L, Thuiller W, Araújo MB, Hirzel AH (2004) Presence-absence versus presence-only modelling methods for predicting bird habitat suitability. Ecography 27:437–448CrossRefGoogle Scholar
  9. Brotons L, De Cáceres M, Fall A, Fortin MJ (2012) Modelling bird species distribution change in fire prone Mediterranean landscapes: incorporating species dispersal and landscape dynamics. Ecography 35:458–467CrossRefGoogle Scholar
  10. Carroll C, Dunk JR, Moilanen A (2010) Optimizing resiliency of reserve networks to climate change: multispecies conservation planning in the Pacific Northwest, USA. Global Change Biol 16:891–904CrossRefGoogle Scholar
  11. Chamberlain DE, Negro M, Caprio E, Rolando A (2013) Assessing the sensitivity of alpine birds to potential future changes in habitat and climate to inform management strategies. Biol Conserv 167:127–135CrossRefGoogle Scholar
  12. Cheung WWL, Lam VWY, Sarmiento JL, Kearney K, Watson R, Zeller D, Pauly D (2010) Large-scale redistribution of maximum fisheries catch potential in the global ocean under climate change. Global Change Biol 16:24–35CrossRefGoogle Scholar
  13. Conroy MJ, Runge MC, Nichols JD, Stodola KW, Cooper RJ (2011) Conservation in the face of climate change: the roles of alternative models, monitoring, and adaptation in confronting and reducing uncertainty. Biol Conserv 144:1204–1213CrossRefGoogle Scholar
  14. Cramp S, Perrins CM (eds) (1993) Handbook of the birds of Europe, the Middle East, and North Africa. The birds of the Western Palearctic volume VII: flycatchers to shrikes. Oxford University Press, OxfordGoogle Scholar
  15. D’Alba L, Monaghan P, Nager RG (2010) Advances in laying date and increasing population size suggest positive responses to climate change in common Eiders Somateria mollissima in Iceland. Ibis 152:19–28CrossRefGoogle Scholar
  16. del Barrio G, Harrison PA, Berry PM, Butt N, Sanjuan ME, Pearson RG, Dawson P (2006) Integrating multiple modelling approaches to predict the potential impacts of climate change on species’ distributions in contrasting regions: comparison and implications for policy. Environ Sci Policy 9:129–147CrossRefGoogle Scholar
  17. Elith J, Graham CH (2009) Do they? How do they? Why do they differ? On finding reasons for differing performances of species distribution models. Ecography 32:66–77CrossRefGoogle Scholar
  18. Elith J, Graham H, Anderson CP, Dudík R, Ferrier M, Guisan S, Hijmans AJ, Huettmann R, Leathwick FR, Lehmann J, Li A, Lohmann JG, Loiselle A, Manion B, Moritz G, Nakamura C, Nakazawa M, Overton CNM, Townsend Peterson J, Phillips AJ, Richardson S, Scachetti-Pereira K, Schapire RE, Soberón R, Williams J, Wisz SS, Zimmermann EM (2006) Novel methods improve prediction of species’ distributions from occurrence data. Ecography 29:129–151CrossRefGoogle Scholar
  19. Fischer J, Batáry P, Bawa KS, Brussaard P, Chappell MJ, Clough Z, Daily GC, Dorrough J, Hartel T, Jackson LE, Klein AM, Kremen C, Juemmerle T, Lindenmayer DB, Mooney HA, Perfecto I, Philpott SM, Tscharntke T, Vandermeer J, Wanger TC, Wehrden H (2011) Conservation: limits of land sparing. Science 334:593CrossRefPubMedGoogle Scholar
  20. Gamero A, Brotons L, Brunner A, Foppen R, Fornasari L, Gregory RD, Lehikoinen A (2017) Tracking progress towards EU biodiversity strategy targets: EU policy effects in preserving its common farmland birds. Conserv Lett 10:395–402.  https://doi.org/10.1111/conl.12292 CrossRefGoogle Scholar
  21. Gaston KJ, Jackson SF, Nagy A, Cantú-Salazar L, Johnson M (2008) Protected areas in Europe. Principle and practice. Ann N Y Acad Sci 1134:97–119CrossRefPubMedGoogle Scholar
  22. Giralt D, Brotons LL, Valera F, Kristin A (2008) The role of natural habitats in agricultural systems for bird conservation: the case of the threatened Lesser Grey Shrike. Biodivers Conserv 17:1997–2012CrossRefGoogle Scholar
  23. Guisan A, Zimmermann NE (2000) Predictive habitat distribution models in ecology. Ecol Model 135:147–186CrossRefGoogle Scholar
  24. Hamer KC (2010) The search for winners and losers in a sea of climate change. Ibis 152:3–5CrossRefGoogle Scholar
  25. Harrison PA, Berry PM, Butt N, New M (2006) Modelling climate change impacts on species’ distributions at the European scale: implications for conservation policy. Environ Sci Policy 9:116–128CrossRefGoogle Scholar
  26. Hastie TJ, Pregibon D (1992) Generalized linear models. In: Chambers J, Hastie T (eds) Statistical models in S. Wadsworth, Pacific Grove, pp 195–247Google Scholar
  27. Hernandez PA, Graham CH, Master LL, Albert DL (2006) The effect of sample size and species characteristics on performance of different species distribution modeling methods. Ecography 29:773–785CrossRefGoogle Scholar
  28. Herremans M (1998) Monitoring the world population of the Lesser Grey Shrike (Lanius minor) on the non-breeding grounds in southern Africa. J Ornithol 139:485–493CrossRefGoogle Scholar
  29. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. Int J Climatol 25:1965–1978CrossRefGoogle Scholar
  30. Huntley B, Green RE, Collingham YC, Willis SG (2007) A climatic Atlas of European breeding birds. Lynx, BarcelonaGoogle Scholar
  31. Iojă CI, Pătroescu M, Rozylowicz L, Popescu VD, Vergheleţ M, Zotta MI, Felciuc M (2010) The efficacy of Romania’s protected areas network in conserving biodiversity. Biol Conserv 143:2468–2476CrossRefGoogle Scholar
  32. Jetz W, Wilcove DS, Dobson AP (2007) Projected impacts of climate and land-use change on the global diversity of birds. PLoS Biol 5:e157CrossRefPubMedPubMedCentralGoogle Scholar
  33. Kleijn D, Schekkerman H, Dimmers WJ, van Kats RJM, Melman D, Teunissen WA (2010) Adverse effects of agricultural intensification and climate change on breeding habitat quality of Black-tailed Godwits Limosa l. limosa in the Netherlands. Ibis 152:475–486CrossRefGoogle Scholar
  34. Koleček J, Reif J, Šťastný K, Bejček V (2010) Changes in bird distribution in a central European country between 1985–1989 and 2001–2003. J Ornithol 151:923–932CrossRefGoogle Scholar
  35. Krištín A (2008) Lesser Grey Shrike (Lanius minor). In: Del Hoyo J, Elliot A, Christie DA (eds) Handbook of the birds of the world, vol 13. Lynx, Barcelona, p 785Google Scholar
  36. Krištín A, Hoi H, Valera F, Hoi C (2007) Philopatry, dispersal patterns and nest-site reuse in Lesser Grey Shrikes (Lanius minor). Biodivers Conserv 16:987–995CrossRefGoogle Scholar
  37. La Sorte FA, Jetz W (2010) Avian distributions under climate change: towards improved projections. J Exp Biol 213:862–869CrossRefPubMedGoogle Scholar
  38. Leathwick JR, Elith J, Hastie T (2006) Comparative performance of generalized additive models and multivariate adaptive regression splines for statistical modelling of species distributions. Ecol Model 199:188–196CrossRefGoogle Scholar
  39. Lefranc N (1995) Decline and current status of the Lesser Grey Shrike (Lanius minor) in Western Europe. Proc West Found Vertebr Zool 6:93–97Google Scholar
  40. Lefranc N, Worfolk T (1997) Shrikes: a guide to the shrikes of the world. Pica, RobertsbridgeGoogle Scholar
  41. López-López P, García-Ripollés C, Soutullo Á, Cadahía L, Urios V (2007) Are important bird areas and special protected areas enough for conservation? The case of Bonelli’s eagle in a Mediterranean area. Biodivers Conserv 16:3755–3780CrossRefGoogle Scholar
  42. Marcer A, Sáez L, Molowny-Horas R, Pons X, Pino J (2013) Using species distribution modelling to disentangle realised versus potential distributions for rare species conservation. Biol Conserv 166:221–230CrossRefGoogle Scholar
  43. Morelli F, Tryjanowski P (eds) (2017) Birds as useful indicators of high nature value farmlands. Springer, BerlinGoogle Scholar
  44. Papp T, Sándor AD (eds) (2007) Arii de Importanţă Avifaunistică din România/Important Bird Areas in Romania. Societatea Ornitologică Română and Asociaţia pentru Protecţia Păsărilor şi a Naturii “Grupul Milvus”, Târgu Mureş, p 252Google Scholar
  45. Pearce-Higgins JW, Dennis P, Whittingham MJ, Yalden DW (2010) Impacts of climate on prey abundance account for fluctuations in a population of a northern wader at the southern edge of its range. Glob Chang Biol 16:12–23CrossRefGoogle Scholar
  46. Pearce-Higgins JW, Bradbury RB, Chamberlain DE, Drewitt A, Langston RHW, Willis SG (2011) Targeting research to underpin climate change adaptation for birds. Ibis 153:207–211CrossRefGoogle Scholar
  47. Popescu VD, Rozylowicz L, Cogălniceanu D, Niculae IM, Cucu AL (2013) Moving into protected areas? Setting conservation priorities for Romanian reptiles and amphibians at risk from climate change. PLoS ONE 8:e79330CrossRefPubMedPubMedCentralGoogle Scholar
  48. R Development Core Team (2009) R: a language and environment for statistical computing R Foundation for Statistical Computing. Vienna. http://wwwR-project.org
  49. Ramirez J, Jarvis A (2008) High resolution statistically downscaled future climate surfaces. International Center for Tropical Agriculture, CGIAR Research Program on Climate Change, Agriculture and Food Security, CaliGoogle Scholar
  50. Regos A, D’Amen M, Herrando S, Guisan A, Brotons L (2015) Fire management, climate change and their interacting effects on birds in complex Mediterranean landscapes: dynamic distribution modelling of an early-successional species—the near-threatened Dartford Warbler (Sylvia undata). J Ornithol 156:275–286CrossRefGoogle Scholar
  51. Ruiz-Labourdette D, Nogués-Bravo D, Ollero HS, Schmitz MF, Pineda FD (2012) Forest composition in Mediterranean mountains is projected to shift along the entire elevational gradient under climate change. J Biogeogr 39:162–176CrossRefGoogle Scholar
  52. Sándor AD, Domșa C (2012) Special protected areas (SPA) for the conservation of Romania’ forest birds: status assessment and possible expansion using predictive tools. Acta Zool Bulg 64:367–374Google Scholar
  53. Saunders DA, Wintle BA, Mawson PR, Dawson R (2013) Egg-laying and rainfall synchrony in an endangered bird species: implications for conservation in a changing climate. Biol Conserv 161:1–9CrossRefGoogle Scholar
  54. Sauter A, Korner-Nievergelt F, Jenni L (2010) Evidence of climate change effects on within-winter movements of European Mallards Anas platyrhynchos. Ibis 152:600–609CrossRefGoogle Scholar
  55. Schwartz MW (2012) Using niche models with climate projections to inform conservation management decisions. Biol Conserv 155:149–156CrossRefGoogle Scholar
  56. Seavy NE, Dybala KE, Snyder MA (2008) Climate models and ornithology. Auk 125:1–10CrossRefGoogle Scholar
  57. Smallegange IM, Van Der Meer J, Fiedler W (2011) Population dynamics of three songbird species in a nestbox population in central Europe show effects of density, climate and competitive interactions. Ibis 153:806–817CrossRefGoogle Scholar
  58. Strange N, Jellesmark Thorsen B, Bladt J, Wilson KA, Rahbek C (2011) Conservation policies and planning under climate change. Biol Conserv 144:2968–2977CrossRefGoogle Scholar
  59. Thomas CD, Cameron A, Green RE, Bakkenes M, Beaumont LJ, Collingham YC, Erasmus BFN, Ferreira de Siqueira M, Grainger A, Hannah L, Hughes L, Huntley B, van Jaarsveld AS, Midgley GF, Miles L, Ortega-Huerta MA, Townsend Peterson A, Phillips OL, Williams SE (2004) Extinction risk from climate change. Nature 427:145–148CrossRefPubMedGoogle Scholar
  60. Triviño M, Cabeza M, Thuiller W, Hickler T, Araújo MB (2013) Risk assessment for Iberian birds under global change. Biol Conserv 168:192–200CrossRefGoogle Scholar
  61. Troupin D, Carmel Y (2014) Can agro-ecosystems efficiently complement protected area networks? Biol Conserv 169:158–166CrossRefGoogle Scholar
  62. Tryjanowski P, Sparks TH, Kuczyński L, Kuźniak S (2004) Should avian egg size increase as a result of global warming? A case study using the red-backed shrike (Lanius collurio). J Ornithol 145:264–268CrossRefGoogle Scholar
  63. Tryjanowski P, Sparks TH, Profus P (2005) Uphill shifts in the distribution of the white stork Ciconia ciconia in southern Poland: the importance of nest quality. Divers Distrib 11:219–223CrossRefGoogle Scholar
  64. Unander S, Pedersen ÅØ, Soininen EM, Descamps S, Hörnell-Willebrand M, Fuglei E (2016) Populations on the limits: survival of Svalbard rock ptarmigan. J Ornithol 157:407–418CrossRefGoogle Scholar
  65. Wiens JA, Seavy NE, Jongsomjit D (2011) Protected areas in climate space: what will the future bring? Biol Conserv 144:2119–2125CrossRefGoogle Scholar
  66. Wisz MS, Hijmans RJ, Li J, Peterson AT, Graham CH, Guisan A, NCEAS Predicting Species Distributions Working Group (2008) Effects of sample size on the performance of species distribution models. Divers Distrib 14:763–773CrossRefGoogle Scholar
  67. Zar JH (1996) Biostatistical analysis. Prentice Hall, Upper Saddle RiverGoogle Scholar

Copyright information

© Dt. Ornithologen-Gesellschaft e.V. 2017

Authors and Affiliations

  1. 1.Department of Parasitology and Parasitic DiseasesUniversity of Agricultural Sciences and Veterinary Medicine Cluj-NapocaCluj-NapocaRomania
  2. 2.Romanian Ornithological SocietyBucharestRomania

Personalised recommendations