Next Article in Journal
Impact of Lower Airway Inflammation on Fitness Parameters in Standardbred Racehorses
Next Article in Special Issue
Climate Change Habitat Model Forecasts for Eight Owl Species in the Southwestern US
Previous Article in Journal
Implications of Salinity and Acidic Environments on Fitness and Oxidative Stress Parameters in Early Developing Seahorses Hippocampus reidi
Previous Article in Special Issue
Body Condition in the Tawny Owl Strix aluco near the Northern Limit of Its Range: Effects of Individual Characteristics and Environmental Conditions
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Animals
Volume 12
Issue 22
10.3390/ani12223226
animals-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Environmental Niche Modelling Predicts a Contraction in the Potential Distribution of Two Boreal Owl Species under Different Climate Scenarios

by
Kristina Cerman
1,*,
Draženko Rajković
2,
Biljana Topić
3,
Goran Topić
3,
Peter Shurulinkov
4,
Tomaž Mihelič
5 and
Juan D. Delgado
6
1
Faculty of Environmental Sciences, Czech University of Life Sciences Prague, Kamýcká 129, Suchdol, 165 00 Prague, Czech Republic
2
Center for Biodiversity Research, Maksima Gorkog 40/3, 21000 Novi Sad, Serbia
3
Ornithological Society ‘‘Naše Ptice’’, Semira Frašte 6/14, 71000 Sarajevo, Bosnia and Herzegovina
4
National Museum of Natural History—Sofia, Bulgarian Academy of Sciences, 1 Tsar Osvoboditel Blvd., 1000 Sofia, Bulgaria
5
DOPPS-BirdLife Slovenia, Tržaška Cesta 2, 1000 Ljubljana, Slovenia
6
Department of Physical, Chemical and Natural Systems, Ecology Area, Faculty of Experimental Sciences, Universidad Pablo de Olavide, E-41013 Sevilla, Spain
*
Author to whom correspondence should be addressed.
Animals 2022, 12(22), 3226; https://doi.org/10.3390/ani12223226
Submission received: 4 October 2022 / Revised: 13 November 2022 / Accepted: 16 November 2022 / Published: 21 November 2022
(This article belongs to the Special Issue Owls' Responses to Environmental Challenges)
Browse Figures
Versions Notes

Abstract

:

Simple Summary

Studying species distribution modelling in the face of climate change provides more insight into how endangered species are affected by these changes. Therefore, we studied two locally endangered owl species, the Boreal and Eurasian Pygmy Owl, in the Balkan Peninsula to better understand their current and future distribution. We aimed to perform species distribution modelling for these two targeted owl species in current climate and future predicted climate scenarios. We quantified highly suitable areas for both species currently and in future climate scenarios. Additionally, we looked at the size of the areas of future species’ refugia where environmental factors might be suitable for the species. Results showed that the future highly suitable area for Boreal Owl shrunk compared to the current area in all climate scenarios; however, for Eurasian Pygmy Owl, the results did not follow such a clear trend. Our study is important from the species’ conservation perspective and fills a knowledge gap about species distribution in the Balkan Peninsula.

Abstract

Studying current and future geographic distribution is essential for conserving endangered species such as the Boreal Owl and Eurasian Pygmy Owl. The main aim of this study was to determine the potential distribution of both species in the Balkan Peninsula by using spatial distribution models (SDMs) in MaxEnt. We used data from field surveys, the scientific and grey literature, and an online database. We considered the current time and two future periods, 2041–2060 and 2061–2080. For future periods, we included different climate scenarios (SSP 126, 245, 370, and 585) in studying the potential geographic distribution of both species. We identified two types of potential future refugia for species: in situ and ex situ. Our study shows the highly suitable area for the Boreal Owl increased during the 2041–2060 period compared with the current area in all scenarios, except in SSP 585. However, during the 2061–2080 period, the highly suitable areas contracted. For the Eurasian Pygmy Owl, highly suitable areas decreased during 2041–2060, but during the 2061–2080 period, it was larger than the current area. Our study is of importance for conservation and preserving areas of potential distribution and refugia for Boreal and Eurasian Pygmy Owls in the face of climate change.

1. Introduction

Biodiversity is a fundamental component of planet life-support systems, and human well-being depends on nature services, such as essential material goods, underpinning functions, and nonmaterial benefits [1,2,3]. However, biodiversity on our planet has been declining at an alarming rate in recent decades. This rate is predicted to be 100 to 1000 times bigger than natural background extinction rates [4,5,6] and is expected to continue at an increasing pace in the forthcoming decades [2,3,4].
Overall, five underlying key drivers cause biodiversity loss and species extinction via many pathways across different physical and temporal scales. These drivers include habitat loss, invasive alien species, overexploitation of natural resources, environmental pollution, and global climate change [7]. Among them, climate change is perceived as the major environmental issue of the 21st century and is anticipated to have vast negative consequences on the planet’s biosphere [8]. The Intergovernmental Panel on Climate Change (IPCC) report predicts that global warming temperatures will likely reach 1.5 °C above preindustrial levels by 2040. Additionally, it is projected to grow by nearly 0.2 °C per decade [8]. Climate change strongly impacts biodiversity at various levels. It shifts species distribution [9,10,11] and migration phenology [12], affects population dynamics [13,14], changes community structure and composition [15], and influences the functioning of entire ecosystems [16,17].
Climate and species geographical distribution are causally related. Predicted global warming is expected to significantly impact the spatial distribution of biota worldwide. For instance, in environments closer to the Equator (i.e., tropics) or that are mountainous, species can be forced to shrink their distributions toward poles or move upslope to higher altitudes to escape warming temperatures and other unsuitable climatic conditions [18,19,20]. These changes in a species’ distribution may jeopardise its persistence by reducing its range or fragmenting the population, leading to population size declines or risk of extinction [21,22,23]. Further, forecasts indicate that the population of habitat-specialised species is decreasing at a notably greater rate than habitat generalists [24,25]. Thus, relatively small biodiversity hotspots could be heavily threatened by climate change [26].
Therefore, over the last 30 years, scientists started studying species distribution modelling, also known as environmental (or ecological) niche modelling (ENM) [27,28,29,30]. This approach is based on mathematical algorithms that use data from presence/absence records and the environmental conditions at occurrence localities [29]. Specifically, modelling is applied but not restricted to predicting the potential geographical distribution [31,32] to recognise habitat suitability and priority areas for conservation [33,34,35], and, more recently, is used to study changes in geographic distribution concerning climate change [24,25,36]. One of the most used species distribution modelling approaches is the machine-learning algorithm MaxEnt (Maximum Entropy), American Museum of Natural History, New York, USA (for details, see [37,38,39,40]). MaxEnt is a favoured and widely applied tool because it demands only presence data, can utilise categorical as well as continuous variables, includes interactions between predictor variables, shows a satisfactory predictive performance, and generally outperforms other SDMs [41,42,43].
The Boreal (Aegolius funereus) and Eurasian Pygmy Owls (Glaucidium passerinum) are small, forest-dwelling avian predators belonging to the Siberian–Canadian faunal type [44,45]. Consequently, both species are confined to boreal climatic zones and high-mountain regions in the Palearctic (Eurasian Pygmy Owl) and Holarctic (Boreal Owl) realms. In Europe, they are almost sympatric inhabitants of the taiga belt in the northern parts of the continent. At the same time, several small, disjunct populations occur in high-mountain forests in the central and southern parts of the continent [44,45]. Across the European distribution range, both species are highly dependent on old growth (>80 years old), and primarily, coniferous forest stands, choosing dry and dead trees with cavities for breeding and food storage [45]. In Southern Europe, particularly on the Balkan Peninsula, both species prefer higher elevations, north-faced slopes, and medium-to-dense forests with a cold and humid climate [46,47,48]. Therefore, it can be assumed that Boreal and Eurasian Pygmy Owls are stenovalent habitat specialists with a narrow tolerance range and few possibilities of adaptation, which can only survive in the specific, above-mentioned environmental conditions. According to BirdLife International [49,50], there are less than ten thousand mature individuals of Boreal Owls (around 3% of the European population) and no more than six thousand mature individuals of Eurasian Pygmy Owls (about 2% of the European population) in the whole Balkan Peninsula. Knowledge about the spatial distribution range is limited, especially for the Eurasian Pygmy Owl, as well as information on population trends, except for a few countries where it is known that the numbers are decreasing (e.g., Serbia). Additionally, in almost all Balkan countries, both species are assessed as vulnerable or endangered with significant threats, such as forest exploitation and fragmentation, the development of ski resorts, and other human disturbances [51,52]. Further, due to global climate warming, the area comprising Norway spruce (Picea abies) in the central Balkan Peninsula, a primary habitat of both species, is expected to decrease, and the range will shift to higher altitudes [53]. Thus, the projected climate change may have a negative impact on the habitat suitability of both species, which may lose remarkable portions of their primary niche. Accordingly, determining the optimal forest habitat patches of both owl species is necessary to understand the role of topographic and climate factors in their potential habitat suitability under present and future climate scenarios.
The aims of this study were: (1) to define the potential current distribution through the development of an SDM and a set of environmental predictor variables; (2) to evaluate which environmental factor(s) influence spatial distribution; (3) to consider the potential impact of climate scenarios on the future distribution; and (4) to recognise potential refugial areas of Boreal and Eurasian Pygmy Owls in the Balkan Peninsula using MaxEnt modelling.

2. Materials and Methods

Our study area, the Balkan Peninsula, extends from Central Europe in the north to the Eastern Mediterranean region in the south, covering approximately 667,000 km2, and is surrounded by the Adriatic, Ionian, Aegean, and Black Seas [54] (Figure 1a). Although belted by four seas, the Mediterranean climate is only present on the coast, with mountain ranges preventing warm air from penetrating into other parts of the peninsula [55]. Therefore, the rest of the peninsula is characterised by an alpine climate with strong altitudinal changes in precipitation and temperature, and by the continental climate in the river valleys and lowlands [56]. Due to the variety of climatic conditions, the Balkan Peninsula is one of Europe’s endemism and biodiversity hotspots, as well as a glacial refuge for flora and fauna [57].
Regarding the vegetation cover, at an altitude of 0–700 m, forests comprise the mixed Fagus and Carpinus communities, with montane forest communities including mostly Fagus species [57]. At an altitude of 700–1700 m, the forest community comprises conifers such as Abies, Picea, and Pinus. Above this altitude is alpine vegetation with Pinus, Juniperus, and Alnus [57].
To compile the Boreal and Eurasian Pygmy Owls occurrence data (the geographic coordinates) from across their natural range in the Balkan Peninsula, we used three different sources: (1) an online database [58]; (2) the scientific and “grey” literature; and (3) records from targeted field surveys using GPS devices, which provided most data (>80%) used in this study. The IUCN Red List criteria for the size of the last three generations of a species were followed to provide biologically meaningful data. Therefore, the collected data related to the period from 2002 to 2020 (3 generations = 18 years) for the Boreal Owl and 2008–2020 (3 generations = 12 years) for the Eurasian Pygmy Owl were used. We derived 883 and 584 occurrence points of Boreal Eurasian Pygmy Owls, respectively. It is important to mention that differences in data collecting approaches were not expected to substantially impact the final model results because Maximum Entropy modelling is particularly well suited to handle all kinds of presence-only data [31].
After this initial step, we carefully cross-checked the data, deleted all duplicate records, and discarded data with obvious georeferencing errors. To avoid spatial autocorrelation in occurrence localities, we performed a filtering process of the rest of the occurrence data using the ArcGIS 10.7.1 software [59]. The spatial filter of occurrence localities was limited to 30 arc s between each other (ca. 1 × 1 km resolution at ground level), which is consistent with published data related to the territory density [47,48,49,50,51,52,53,54,55,56,57,58,59,60] and home range size [61,62] of both owl species. Thus, we left only one occurrence record within each grid cell of 1 × 1 km. Additionally, we used the Global Moran’s coefficient for an additional recheck if there was a potential problem with spatial autocorrelation in the occurrence dataset [63]. This index represents the widely used multidimensional and multidirectional statistical tool for measuring spatial autocorrelation in ecological studies [64,65]. We employed the “Spatial Autocorrelation (Global Moran’s I)” tool from ArcGIS (Esri, Redlands, CA, USA) software to calculate the Global Moran index using the nearest neighbour approach. We did not detect autocorrelated data in either the Boreal Owl (Moran’s I = 0.047 p = 0.573) or the Eurasian Pygmy Owl (Moran’s I = 0.189 p = 0.748). Finally, 439 and 235 precise occurrences of Boreal and Eurasian Pygmy Owls, respectively, were left (Table S1); these data were utilised to create the SDMs and the detailed distribution map (Figure 1b,c).
For modelling the current distribution of the species, we used recent bioclimatic variables, such as elevation, aspect and slope of the mountain, soil classification, snow cover, human footprint index, and land-use type. The sources of the environmental variables are available in Table S2.
Based on the published literature and the authors’ assessment, climate and other predictor variables were selected according to their relevance and importance to owls’ life cycles. For instance, it is generally known that most species, including Boreal and Eurasian Pygmy Owls, inhabit a specific bioclimatic niche which is predominantly regulated by main climatic factors such as temperature and precipitation [66,67]. In this case, both owl species across distributional ranges are associated with cold and humid boreal and high-mountain climate conditions [44,45]. Therefore, we decided to use all 19 different bioclimate variables from Global Climate Data–WorldClim version 2.1 [68] in the initial baseline model (present time, 1950–2000). These variables represent a crucial, ecologically meaningful, and the most applied set of high-resolution global climate layers in SDMs and related ecological modelling techniques [30,69]. In Southeast Europe, particularly in the Balkan Peninsula, Boreal and Eurasian Pygmy Owls, as postglacial relicts, inhabit high mountain areas, preferably above 1000 m a.s.l. Furthermore, they are cold-adapted forest-dwelling species that prefer north-facing, steep, often rocky slopes at higher altitudes covered with old-growth mixed and coniferous forests, usually grown in shallow soil [46,47,48,51,60,70]. In addition to bioclimatic variables, we included the digital elevation model (DEM), slope gradient, aspect, soil type, and hill shade in the initial modelling in this study (Table S2). We did not include other potentially useful layers, such as land use or land cover, due to their high potential variability in time and space, making them unrealistic and irrelevant for modelling distribution patterns in future scenarios. To determine the future distribution of both owl species under contrasting climate scenarios, we used datasets of future climate predictions from Global Climate Data–WorldClim version 2.1 [68]. Four representatives of the Shared Socioeconomic Pathways (SSP126, SSP245, SSP370, and SSP585) ratified by the Intergovernmental Panel on Climate Change (IPCC) [8] were considered in modelling processes related to future climate scenarios and the habitat suitability distribution of both owl species. These SSPs are a part of the Coupled Model Intercomparison Project, Phase 6 (CMIP6) [8]. The four SSPs are defined by the predicted range of radiative forcing values [8]. Predicting suitable species distributions under climate change scenarios involved climate data for the next two periods: 2041–2060 and 2061–2080. All used layers were converted into a spatial resolution of 30 arc seconds (ca. 1 × 1 km resolution at ground level) and trimmed to the Balkan Peninsula shape using ArcGIS software.
We predicted the potential distribution of Boreal and Eurasian Pygmy Owls under different climate change scenarios by applying MaxEnt version 3.4.4. [38]. The MaxEnt program settings undoubtedly significantly influence model performance and prediction power [40]. Although the MaxEnt software can be successfully utilised for SDM purposes with the default settings [31], later studies have convincingly demonstrated that employing automatic features will not generally result in the best prediction model [71,72,73]. Therefore, respecting the calls for prudence and following general recommendations [73,74], we tried to achieve potentially the best combination of feature classes and a regularisation multiplier (β coefficient) to express the best fitting model adequately.
To model habitat suitability for each species (Tables S3 and S4), we developed a comprehensive set of initial models with all 19 BioClim variables plus 5 topographic variables and a β coefficient changing from 0 to 5 in increments of 0.2, resulting in 26 models per owl species. Except for the β coefficient, other MaxEnt parameter settings were kept as the default. Tuning the β coefficient (regularisation multiplier) between 0 and 5 was a standard procedure that aimed to sufficiently reduce overfitting to reasonable levels [39,73]. For each species and each initial model, we used the sample-size-adjusted Akaike information criterion (AICc) [75,76] to determine the most appropriate variable combination and to tune model complexity [32,36]. We retained only the model with the lowest AICc from the initial set for each species, creating a baseline model. Moreover, we calculated the MaxEnt contribution scores for each environmental variable from each baseline model. Predictor variables indicating no remarkable effect on species occurrence with percent contribution scores ≤1% in the baseline model were eliminated. Then, the variable with the highest score was retained and added to the final variable set [77]. All other variables strongly correlated with the retained predictor variable at a pairwise Pearson correlation coefficient of | r | > 0.70 [36,78] were deleted. This process was replicated until all variables were switched to the baseline model set or discarded. Next, we checked the newly established set of the baseline model variables for multicollinearity with the help of a widely used diagnostic quotient: the variance inflation factor (VIF). All variables with a VIF score ≥6 [77,79] were eliminated from further processing, starting with the one with the highest VIF score. This process was repeated until all the remaining variables scored lower than 6. Altogether, 9 predictor variables for the Boreal Owl and 12 for the Eurasian Pygmy Owl were retained as inputs for MaxEnt modelling of Boreal and Eurasian Pygmy Owls in the Balkan Peninsula.
To reduce overfitting and simplify the interpretation [32,34], we only employed linear (L) and quadratic (Q) features and their combination (L + Q) in the finishing stage of the SDMs. This procedure resulted in generating three models per species. As in the previous steps, we retained the model with the lowest AICc to simulate the current and future distributions of the Boreal and Eurasian Pygmy Owls in the Balkan Peninsula. We set the maximum number of iterations to 1000 to allocate the models sufficient time to converge [35]. We applied “maximum training sensitivity plus specificity”, which represents a pretty satisfactory method for threshold selection in the case when only presence data are available [80]. The random test data were 25% of the sample data, and the training data were the remaining 75% of the sample data selected randomly. The habitat suitability curves of each predictor variable were calculated, as were the contributions of each predictor variable using the jack-knife test. All other MaxEnt parameter settings were kept as the default. We used the AUC (area under the ROC curve) to determine which models performed better than others. AUC values range from 0 to 1, with 0 being the lowest performance of the model and 1 being the highest performance of the model.
As metrics for quantifying the similarity among SDMs are important for testing patterns of niche evolution, we calculated the similarity statistic I [28]. It ranges from 0 (no overlap) to 1 (identical niche models). The mathematical formula is available in a study by Warren et al. [28].
All statistical tests were performed in RStudio [81].
After choosing the final models, we imported them into ArcGIS and divided habitat suitability into four levels according to the AUC values: unsuitable habitat (0–0.05), poorly suitable habitat (0.05–0.33), moderately suitable habitat (0.33–0.67), and highly suitable habitat (0.67–1). Various studies have different approaches in determining “highly suitable habitat” classification, where some are too strict (0.8–1) [82] and others are more accepting (0.6–1) [83,84,85]. Therefore, we decided to use a classification that would meet the requirements in the middle. According to these levels, we calculated the area of each species distribution under each climatic scenario and for each period, as well as an area of species distribution within each country.
We calculated areas of potential climate refugia for both species by looking at the highly suitable habitats in the current and future species distribution models. We followed the methodology of Brambilla et al. [32], where two types of refugia were identified: type 1 refugia are habitats suitable in both current and future conditions (in situ sites), and type 2 refugia are habitats that are not suitable in current conditions but provide suitable conditions in all future predictions (ex situ sites).

3. Results

The current species distribution prediction accuracy for Boreal and Eurasian Pygmy Owls was considered “excellent”, where AUCmean = 0.91 for both species (Table 1 and Table 2). Regarding the environmental variables for both species, bio5 (maximum temperatures of the warmest month) contributed the most to the MaxEnt models (74%). Interestingly, the Boreal Owl was absent in cells with maximum temperatures of the warmest month higher than 31 °C, whereas the Eurasian Pygmy Owl was absent in cells with maximum temperatures of the warmest month higher than 34 °C. The rest of the environmental variables all had less than a 10% contribution to the MaxEnt models. Regarding the current predicted distribution for the Boreal Owl, highly suitable areas cover 261 km2, moderately suitable areas cover 447 km2, and low suitable areas cover 1992 km2 of the entire Balkan Peninsula (Table 3) (see Table S1 per country). For the Eurasian Pygmy Owl, highly suitable areas cover 233 km2, moderately suitable areas cover 385 km2, and low suitable areas cover 1271 km2 of the entire Balkan Peninsula (Table 4) (see Table S2 per country). Both species had the largest areas of highly suitable habitats in Serbia and Bosnia and Herzegovina. Note that the alpine parts of Slovenia are excluded from the analysis since this area does not belong to the Balkan Peninsula.
When looking at the future species distribution models for both species, all four scenarios (SSP 126, 245, 370, and 585) and both periods (2041–2060 and 2061–2080) were considered either “very good” or “excellent” (Table 1). The environmental variable for the Boreal Owl that contributed the most to the model was bio5 (maximum temperatures of the warmest month), with one exception for SSP 370 in 2041–2060, when bio9 contributed the most (69%). However, for the Eurasian Pygmy Owl, apart from bio5, elevation majorly contributed to the MaxEnt models. Regarding the area changes (Figure 2 and Figure 3), specifically for the Boreal Owl, the highly suitable habitat in comparison to the current distribution was only positive, i.e., the area was larger than the current distribution, during the 2041–2060 period for SSP 126, 245, and 370. However, this was not true for SSP 585, where the changes were negative, i.e., the area was smaller than the current distribution. Furthermore, for the entire period of 2061–2080, we found changes to be negative, i.e., smaller than the current distribution. When looking at the area changes for the Eurasian Pygmy Owl, the highly suitable habitat in comparison to the current distribution was only negative during the 2041–2060 period for all scenarios. However, the 2061–2080 models predicted a positive change for all scenarios except for SSP 370.
Despite some changes in spatial distribution between the current and future predictions for both species, an ANOVA did not show statistically significant changes in the DEM (Boreal Owl p-value = 0.77, Eurasian Pygmy Owl p-value = 0.55).
Results obtained from the similarity statistic I showed that the Boreal Owl’s current niche highly overlapped with SSP 126 and 245 in the 2041–2060 period (0.926 and 0.991, respectively). However, when looking at the 2061–2080 period, the current species distribution overlapped highly with all except SSP 585 (0.719) (Table 5). Regarding the Eurasian Pygmy Owl, its current species niche moderately overlapped with all SSPs from the 2041–2060 period, but for the 2061–2080 period, its species niche highly overlapped with all SSPs (Table 5).
Our results show that type 1 refugia (in situ) of the Boreal Owl in the periods of 2041–2060 and 2061–2080 reduced its area among the different SSPs (Table 6) (Figure 4). Type 2 refugia (ex situ) followed the same pattern (Table 6). However, for the Eurasian Pygmy Owl, the area of the type 1 refugia in the 2041–2060 period was larger in SSP 126 and 585, and in 2061–2080 it was the largest in SSP 370 and 585 (Table 7). Furthermore, the Eurasian Pygmy Owl had a larger area of type 2 refugia in the 2041–2060 period in SSP 245 and 370, and in the 2061–2080 period, the largest in SSP 126 and 585 (Table 7).

4. Discussion

Our results provide the first look at current and future species potential distributions of Boreal and Eurasian Pygmy Owls covering the entire Balkan Peninsula by using MaxEnt modelling. Additionally, our study provides more insight into the environmental and climate variables affecting current and future species distributions. Furthermore, we calculated species area changes and potential refugia at varying temporal scales for these two locally endangered boreal owl species in the face of climate change in the Balkan Peninsula. The outcomes of this study can be utilised to build future conservation strategies, and habitat restoration and management plans for these key, flagship predators of high-mountain habitats in the Balkan Peninsula.
The maximum temperature of the warmest month (bio5) represents the environmental variable that contributes the most to and notably shapes the Boreal Owl’s and Eurasian Pygmy Owl’s habitat suitability and spatial distribution. This is not very surprising since it is well known that high temperatures have a significant influence on boreal species, their distribution, and physiology. When looking at specific temperatures for each species in the current distributions, the Boreal Owl is more sensitive to higher temperatures than the Eurasian Pygmy Owl due to its absence in areas with higher temperatures than 31 °C. Similar results have been reported in a study from the Czech Republic, proving that Boreal owls prefer colder temperatures and higher altitudes [86], and which provides further evidence that species in southern populations, such as in the Balkan Peninsula, are a postglacial relict. The next environmental variable that contributes the most to spatial distribution for the Eurasian Pygmy Owl was elevation. The species prefer higher altitudes, which contradicts the results from the study in the Czech Republic [86]. This is most probably because the Balkan Peninsula has a tree line at higher altitudes than the Czech Republic. Therefore, there is more forest area to inhabit. Altogether, these results suggest a high sensitivity of Boreal and Eurasian Pygmy Owl populations to maximum temperatures of the warmest month. Thus, any significant change in temperatures in the Balkan Peninsula and, probably, through a wider area might affect species potential distributions, as shown in other research for other avian species and geographical areas [87,88].
With global climate change, it is expected that some species will move close to the poles or high elevations [89,90] whereas other species might adapt to these changes [91]. However, in our study, when looking at the changes between current and future species potential distributions, we did not register statistically significant results. We can speculate that this is due to the tree line preventing species from moving to higher altitudes in the future and must consider the limiting factor for both species: higher temperatures at lower altitudes.
All projected distribution models, without exception, show narrow ecological adaptability in both owl species. When looking at the change in future highly suitable areas of the Boreal Owl, a positive change, i.e., the area increases in comparison to the current distribution, is overall present in the 2041–2060 period, except in SPP 585. This was expected, since SSP 585 is considered the worst-case climatic scenario in which CO2 emissions rapidly increase until 2080, and then reach the peak at which the trend stabilises [8]. Furthermore, in the period of 2061–2080, only a negative change occurs, meaning the highly suitable area of the species distribution is reduced in comparison to the current species distribution. Regarding future highly suitable area changes of the Eurasian Pygmy Owl, the models showed that for the period of 2041–2060, the area would shrink in its size for each SSP. We can speculate that due to the increased temperatures caused by higher CO2 emissions, both species’ highly suitable areas will shrink, since, as it was previously discussed, the species are prone to avoid temperatures above 31 °C and 34 °C. Furthermore, a relatively new study carried out by researchers in the Bulgarian mountains showed that Boreal and Eurasian Pygmy Owls are avoiding inhabiting managed forests and young forests [70]. Even though Serbia and Bosnia and Herzegovina are facing urbanisation of mountainous areas with the development of ski slopes and touristic accommodations that require forest clear cuts [52], these countries still have the largest areas of highly suitable habitats for both species. Unfortunately, the combination of factors such as deforestation and increased temperatures might just be the reason for the loss of highly suitable habitats for these endangered species.
We calculated the type 1 refugia (in situ) of the Boreal Owl, the areas where the species is present currently and where it might be present in the future, under different climate scenarios. These areas are the most important for species conservation since they can enhance populations’ resilience [92]. Our results showed that type 1 refugia would be increasingly contract with the different SSPs toward the worst-case scenario: SSP 585. This result was expected due to the increase in CO2 emissions and higher temperatures. Furthermore, the ex situ refugia, type 2, where a species is not present currently but might be in the future, are important for the species’ future redistribution [92]. Our models showed that the Boreal Owl’s potential type 2 refugia would also contract with the different SSPs. However, both types of refugia of the Eurasian Pygmy Owl did not show such a clear trend along the SSPs. Even though our models showed that the future areas of both types of refugia are reducing, these areas are the key habitats for species protection and should be considered targets for conservation. Consequently, declaring these areas as protected areas and managing them accordingly could help support species’ resilience to climate change.
With our study, we filled a knowledge gap regarding both researched species’ current distribution in the Balkan Peninsula. Currently, there are several studies on Boreal and Eurasian Pygmy Owls’ distributions in Serbia [52,93] and Bulgaria [48,51], with unpublished data from Bosnia and Herzegovina, Montenegro, and Croatia. However, little is known about the Boreal and Eurasian Pygmy Owls’ population sizes and distributions in Albania and North Macedonia. Hence, our models of the current distribution of the species are beneficial for species mapping in these areas.

5. Conclusions

To safeguard Boreal and Eurasian Pygmy Owls, regular monitoring, habitat preservation, and sustainable management in the Balkan Peninsula are highly required. Special care must be paid to the core areas of both species, i.e., type 1 refugia which can be critical habitat patches for the future survival of both species. In addition, further detailed research is needed to determine how anthropogenic activities affect these two species’ capacity to adapt to changing climatic circumstances.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani12223226/s1, Table S1. Boreal (Aegolius funereus) and Eurasian Pygmy Owl (Glaucidium passerinum) occurrence coordinates obtained from (1) an online database [5], (2) scientific and grey literature and (3) records from targeted field surveys using GPS devices which includes most data (>80%) used in this study.; Table S2. List of the six predictor variables, used for predicting future distribution of the Boreal and the Eurasian Pygmy Owl, with their sources form where they were obtained.; Table S3. Boreal Owl’s current and future area (km2) of highly suitable habitats in each Balkan country.; Table S4. Eurasian Pygmy Owl’s current and future area (km2) of highly suitable habitats in each Balkan country.

Author Contributions

Conceptualisation, K.C. and D.R.; methodology, K.C. and D.R.; software, K.C.; validation, K.C., D.R. and J.D.D.; formal analysis, D.R.; investigation, D.R.; data curation, B.T., G.T., P.S., T.M. and D.R.; writing—original draft preparation, K.C. and D.R.; writing—review and editing, K.C., D.R., J.D.D., P.S. and T.M.; visualisation, K.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are available in the Supplementary Materials, Table S1.

Acknowledgments

The authors would like to thank Dimitrije Radišić, Draško Grujić, Radislav Mirić, Miloš Radaković, Milan Ružić, Đurđe Đuković, Nikola Stanojević, Marko Janković, Marko Ostojić, Srđan Čuturilov, Milivoj Vučanović, Ivan Đorđević, Ivan Medenica, Dragomir Damnjanović, Vjekoslav Joksimović, Ilhan Dervović (deceased), Radosav Novčić, Andrey Ralev, Girgina, Daskalova, and Petko Tzvetkov for assistance during field surveys and their valuable comments on initial ideas and methodology. Special thanks to Josip Turkalj (BIOM, Croatia), Marija Šoškić, Bojan Zeković, Nikola Novović, Igor Stojović, Borut Rubinić (CZIP, Montenegro), Milena Bataković (ANEP, Montenegro), Marko Raković (DBIWP, University of Belgrade, Serbia), and DOPPS BirdLife Slovenia volunteers for providing the occurrence records for both owl species, and Goran Vučićević for help with data curation.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. MEA—Millennium Ecosystem Assessment. Ecosystems and Human Well-Being: Synthesis; Island Press: Washington, DC, USA; p. 7815. Available online: www.islandpress.org (accessed on 28 September 2022).
  2. Michael, R.W.R.; William, M.A.; Bennun, L.; Stuart, H.M.B.; Clements, A.; Coomes, D.; Entwistle, A.; Hodge, I.; Kapos, V.; Jörn, P.W.S.; et al. Biodiversity Conservation: Challenges Beyond 2010. Science 2010, 329, 1298–1303. [Google Scholar]
  3. Mace, G.M.; Norris, K.; Fitter, A.H. Biodiversity and ecosystem services: A multilayered relationship. Trends Ecol. Evol. 2011, 27, 19–26. [Google Scholar] [CrossRef] [PubMed]
  4. Ceballos, G.; Ehrlich, P.R.; Barnosky, A.D.; García, A.; Pringle, R.M.; Palmer, T.M. Accelerated modern human–induced species losses: Entering the sixth mass extinction. Environ. Sci. 2015, 1, 1–5. [Google Scholar] [CrossRef] [Green Version]
  5. De Vos, J.M.; Joppa, L.N.; Gittleman, J.L.; Stephens, P.R.; Pimm, S.L. Estimating the normal background rate of species extinction. Conserv. Biol. 2015, 29, 452–462. [Google Scholar] [CrossRef] [Green Version]
  6. Pimm, S.L.; Jenkins, C.N.; Abell, R.; Brooks, T.M.; Gittleman, J.L.; Joppa, L.N.; Raven, P.H.; Roberts, C.M.; Sexton, J.O. The biodiversity of species and their rates of extinction, distribution, and protection. Science 2014, 34, 1–10. [Google Scholar] [CrossRef]
  7. IPBES. The Global Assessment Report on Biodiversity and Ecoystem Services, Summary for Policymakers. Published online 2019. Available online: www.ipbes.net (accessed on 28 September 2022).
  8. Skea, J.; Sjukla, P.; Reisinger, A.; Slade, R.; Pathak, M.; Some, P.; Vyas, P.; Fradera, R.; Belkacemi, M.; Hasija, A.; et al. IPCC Climate Change 2022—Mitigation of Climate Change—Working Group III; Cambridge Univ. Press: Cambridge, UK, 2022; p. 1454. [Google Scholar]
  9. Loarie, S.R.; Carter, B.E.; Hayhoe, K.; McMahon, S.; Moe, R.; Knight, C.A.; Ackerly, D.D. Climate change and the future of California’s endemic flora. PLoS ONE 2008, 3, 1–10. [Google Scholar] [CrossRef]
  10. Lehikoinen, A.; Johnston, A.; Massimino, D. Climate and land use changes: Similarity in range and abundance changes of birds in Finland and Great Britain. Ornis Fernica 2021, 98, 1–15. [Google Scholar]
  11. Pélissié, M.; Johansson, F.; Hyseni, C. Pushed Northward by Climate Change: Range Shifts With a Chance of Co-occurrence Reshuffling in the Forecast for Northern European Odonates. Environ. Entomol. 2022, 51, 910–921. [Google Scholar] [CrossRef] [PubMed]
  12. Workie, T.G.; Debella, H.J. Climate change and its effects on vegetation phenology across ecoregions of Ethiopia. Glob. Ecol. Conserv. 2018, 13, e00366. [Google Scholar] [CrossRef]
  13. Diele-Viegas, L.M.; Werneck, F.P.; Rocha, C.F.D. Climate change effects on population dynamics of three species of Amazonian lizards. Comp. Biochem. Physiol. -Part A Mol. Integr. Physiol. 2019, 236, 110530. [Google Scholar] [CrossRef]
  14. Van De Pol, M.; Vindenes, Y.; Sæther, B.E.; Engen, S.; Ens, B.J.; Oosterbeek, K.; Tinbergen, J.M. Effects of climate change and variability on population dynamics in a long-lived shorebird. Ecology 2010, 91, 1192–1204. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  15. Yang, H.; Wu, M.; Liu, W.; Zhang, Z.; Zhang, N.; Wan, S. Community structure and composition in response to climate change in a temperate steppe. Glob. Chang. Biol. 2011, 17, 452–465. [Google Scholar] [CrossRef]
  16. Heidari, H.; Arabi, M.; Warziniack, T. Effects of climate change on natural-caused fire activity in western U.S. national forests. Atmosphere 2021, 12, 981. [Google Scholar] [CrossRef]
  17. Grimm, N.B.; Chapin, F.S.; Bierwagen, B.; Gonzalez, P.; Groffman, P.M.; Luo, Y.; Melton, F.; Nadelhoffer, K.; Pairis, A.; Raymond, P.A.; et al. The impacts of climate change on ecosystem structure and function. Front. Ecol. Environ. 2013, 11, 474–482. [Google Scholar] [CrossRef] [Green Version]
  18. Peh, K.S.H. Potential effects of climate change on elevational distributions of tropical birds in Southeast Asia. Condor 2007, 109, 437–441. [Google Scholar] [CrossRef]
  19. Freeman, B.G.; Class Freeman, A.M. Rapid upslope shifts in New Guinean birds illustrate strong distributional responses of tropical montane species to global warming. Proc. Natl. Acad. Sci. USA 2014, 111, 4490–4494. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  20. Couet, J.; Marjakangas, E.L.; Santangeli, A.; Kålås, J.A.; Lindström, Å.; Lehikoinen, A. Short-lived species move uphill faster under climate change. Oecologia 2022, 198, 877–888. [Google Scholar] [CrossRef]
  21. Levinsky, I.; Skov, F.; Svenning, J.C.; Rahbek, C. Potential impacts of climate change on the distributions and diversity patterns of European mammals. Biodivers. Conserv. 2007, 16, 3803–3816. [Google Scholar] [CrossRef]
  22. Dullinger, S.; Gattringer, A.; Thuiller, W.; Moser, D.; Zimmermann, N.E.; Guisan, A.; Willner, W.; Plutzar, C.; Leitner, M.; Mang, T.; et al. Extinction debt of high-mountain plants under twenty-first-century climate change. Nat. Clim. Chang. 2012, 2, 619–622. [Google Scholar] [CrossRef]
  23. Jiguet, F.; Gregory, R.D.; Devictor, V.; Green, R.E.; Vořšek, P.; Van Strien, A.; Couvet, D. Population trends of European common birds are predicted by characteristics of their climatic niche. Glob. Chang. Biol. 2010, 16, 497–505. [Google Scholar] [CrossRef]
  24. Koh, L.P.; Sodhi, N.S.; Brook, B.W. Ecological correlates of extinction proneness in tropical butterflies. Conserv. Biol. 2004, 18, 1571–1578. [Google Scholar] [CrossRef]
  25. Manes, S.; Costello, M.J.; Beckett, H.; Debnath, A.; Devenish-Nelson, E.; Grey, K.A.; Jenkins, R.; Khan, T.M.; Kiessling, W.; Krause, C.; et al. Endemism increases species’ climate change risk in areas of global biodiversity importance. Biol. Conserv. 2021, 257, 109070. [Google Scholar] [CrossRef]
  26. Trew, B.T.; Maclean, I.M.D. Vulnerability of global biodiversity hotspots to climate change. Glob. Ecol. Biogeogr. 2021, 30, 768–783. [Google Scholar] [CrossRef]
  27. Guisan, A.; Zimmermann, N.E. Predictive habitat distribution models in ecology. Ecol. Modell. 2000, 135, 147–186. [Google Scholar] [CrossRef]
  28. Warren, D.L.; Glor, R.E.; Turelli, M. Environmental Niche Versus Conservatism: Quantitative Approaches to Niche Evolution. Evolution 2008, 62, 2868–2883. [Google Scholar] [CrossRef]
  29. Elith, J.; Leathwick, J.R. Species distribution models: Ecological explanation and prediction across space and time. Annu. Rev. Ecol. Evol. Syst. 2009, 40, 677–697. [Google Scholar] [CrossRef]
  30. Booth, T.H.; Nix, H.A.; Busby, J.R.; Hutchinson, M.F. Bioclim: The first species distribution modelling package, its early applications and relevance to most current MaxEnt studies. Divers. Distrib. 2014, 20, 1–9. [Google Scholar] [CrossRef]
  31. Phillips, S.J.; Anderson, R.P.; Schapire, R.E. Maximum entropy modeling of species geographic distributions. Ecol. Modell. 2006, 190, 231–259. [Google Scholar] [CrossRef] [Green Version]
  32. Brambilla, M.; Rubolini, D.; Appukuttan, O.; Calvi, G.; Karger, D.N.; Kmecl, P.; Mihelič, T.; Sattler, T.; Seaman, B.; Teufelbauer, N.; et al. Identifying climate refugia for high-elevation Alpine birds under current climate warming predictions. Glob. Chang. Biol. 2022, 28, 4276–4291. [Google Scholar] [CrossRef]
  33. Araújo, M.B.; Williams, P.H. Selecting areas for species persistence using occurrence data. Biol. Conserv. 2000, 96, 331–345. [Google Scholar] [CrossRef] [Green Version]
  34. Sun, X.; Long, Z.; Jia, J. A multi-scale Maxent approach to model habitat suitability for the giant pandas in the Qionglai mountain, China. Glob. Ecol. Conserv. 2021, 30, e01766. [Google Scholar] [CrossRef]
  35. Spiers, J.A.; Oatham, M.P.; Rostant, L.V.; Farrell, A.D. Applying species distribution modelling to improving conservation based decisions: A gap analysis of trinidad and tobago’s endemic vascular plants. Biodivers. Conserv. 2018, 27, 2931–2949. [Google Scholar] [CrossRef]
  36. Warren, D.L.; Wright, A.N.; Seifert, S.N.; Shaffer, H.B. Incorporating model complexity and spatial sampling bias into ecological niche models of climate change risks faced by 90 California vertebrate species of concern. Divers. Distrib. 2014, 20, 334–343. [Google Scholar] [CrossRef]
  37. Phillips, S.B.; Aneja, V.P.; Kang, D.; Arya, S.P. Modelling and analysis of the atmospheric nitrogen deposition in North Carolina. Int. J. Glob. Environ. Issues 2006, 6, 231–252. [Google Scholar] [CrossRef] [Green Version]
  38. Phillips, S.J.; Dudik, M.; Schapire, R.E. Maxent Software for Modeling Species Niches and Distributions (Version 3.4.1). Available online: http://biodiversityinformatics.amnh.org/open_source/maxent/ (accessed on 18 September 2022).
  39. Elith, J.; Phillips, S.J.; Hastie, T.; Dudík, M.; Chee, Y.E.; Yates, C.J. A statistical explanation of MaxEnt for ecologists. Divers. Distrib. 2011, 17, 43–57. [Google Scholar] [CrossRef]
  40. Merow, C.; Smith, M.J.; Silander, J.A. A practical guide to MaxEnt for modeling species’ distributions: What it does, and why inputs and settings matter. Ecography 2013, 36, 1058–1069. [Google Scholar] [CrossRef]
  41. Elith, J.; H. Graham, C.P.; Anderson, R.; Dudík, M.; Ferrier, S.; Guisan, A.; J. Hijmans, R.; Huettmann, F.; R. Leathwick, J.; Lehmann, A.; et al. Novel methods improve prediction of species’ distributions from occurrence data. Ecography 2006, 29, 129–151. [Google Scholar] [CrossRef] [Green Version]
  42. Wisz, M.S.; Hijmans, R.J.; Li, J.; Peterson, A.T.; Graham, C.H.; Guisan, A.; Elith, J.; Dudík, M.; Ferrier, S.; Huettmann, F.; et al. Effects of sample size on the performance of species distribution models. Divers. Distrib. 2008, 14, 763–773. [Google Scholar] [CrossRef]
  43. Aguirre-Gutiérrez, J.; Carvalheiro, L.G.; Polce, C.; van Loon, E.E.; Raes, N.; Reemer, M.; Biesmeijer, J.C. Fit-for-Purpose: Species Distribution Model Performance Depends on Evaluation Criteria—Dutch Hoverflies as a Case Study. PLoS ONE 2013, 8, 1–11. [Google Scholar] [CrossRef] [Green Version]
  44. Mikkola, H. Owls of Europe; T & AD Poyser Ltd. (A & C Black): London, UK, 1983. [Google Scholar]
  45. Korpimäki, E.; Hakkarainen, H. The Boreal Owl: Ecology, Behaviour and Conservation of a Forest-Dwelling Predator; Cambridge University Press: Cambridge, UK, 2012. [Google Scholar]
  46. Vrezec, A.; Davorin, T.; Vrezec, A.; Tome, D. Habitat selection and patterns of distribution in a hierarchic forest owl guild. Ornis Fenn. 2004, 81, 109–118. [Google Scholar]
  47. Rajković, D.; Grujić, D.; Novčić, R.; Mirić, R. Population of Tengmalm’s Owl Aegolius funereus in Kopaonik National Park (Central Serbia). Acrocephalus 2013, 34, 27–32. [Google Scholar] [CrossRef] [Green Version]
  48. Shurulinkov, P.; Stoyanov, G.; Komitov, E.; Daskalova, G.; Ralev, A. Contribution to the Knowledge on Distribution, Number and Habitat Preferences of Rare and Endangered Birds in Western Rhodopes Mts, Southern Bulgaria. Strigiformes and Piciformes. Acta Zool. Bulg. 2012, 64, 43–56. Available online: https://www.researchgate.net/publication/232712650 (accessed on 28 September 2022).
  49. BirdLife International Glaucidium passerinum. IUCN Red List Threat. Species 2016, 82, 1–8. Available online: https://www.iucnredlist.org/species/22689194/86868363 (accessed on 6 September 2022).
  50. BirdLife International Boreal Owl (Aegolius funereus). IUCN Red List Threat. Species 2021, 8235, 1–8. Available online: https://www.iucnredlist.org/species/22689362/166227347 (accessed on 6 September 2022).
  51. Rajković, D.; Puzović, S.; Raković, M.; Grubač, B.; Simović, A.; Đorđević, V.M. Records of Boreal Owl Aegolius funereus in Serbia. Ciconia 2010, 19, 131–140. [Google Scholar]
  52. Shurulinkov, P.; Ralev, A.; Daskalova, G.; Chakarov, N. Distribution, numbers and habitat of Pigmy Owl Glaucidium passerinum in Rhodopes Mts (S Bulgaria). Acrocephalus 2007, 135, 161–165. [Google Scholar]
  53. Obratov-Petković, D.; Beloica, J.; Čavlović, D.; Djurdjević, V.; Simić, S.B.; Bjedov, I. Modelling Response of Norway Spruce Forest Vegetation to Projected Climate and Environmental Changes in Central Balkans Using Different Sets of Species. Forests 2022, 13, 566. [Google Scholar] [CrossRef]
  54. Cvijić, J. Balkansko Poluostrvo i Južnoslovenske Zemlje; SANU: Književne novine: Zavod za udžbenike i nastavna sredstva: Belgrade, Serbia, 1987; pp. 12–25. [Google Scholar]
  55. Reed, J.M.; Kryštufek, B.; Eastwood, W.J. The Physical Geography of The Balkans and Nomenclature of Place Names. In Balkan Biodiversity; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2004; pp. 9–22. [Google Scholar] [CrossRef]
  56. Furlan, D. The Climate of Southeast Europe. In Climates of Central and Southern Europe. World Survey of Climatology; Elsevier: Amsterdam, The Netherlands, 1977; Volume 6, pp. 185–235. [Google Scholar]
  57. Griffiths, H.I.; Kryštufek, B.; Reed, J.M. Balkan Biodiversity; Springer Science + Business Media: Berlin/Heidelberg, Germany, 2004; ISBN 978-90-481-6732-6. [Google Scholar]
  58. Global Biodiversity Information Facility—GBIF. Available online: https://www.gbif.org/ (accessed on 7 September 2022).
  59. ESRI ArcGIS desktop: Release 10. 2011. Available online: https://www.esri.com/en-us/home (accessed on 6 September 2022).
  60. Pačenovský, S.; Shurulinkov, P. Latest data on distribution of the Pygmy Owl (Glaucidium passerinum) in Bulgaria and Slovakia including population density comparison. Slovak Raptor J. 2008, 2, 91–106. [Google Scholar] [CrossRef]
  61. Sorbi, S. Size and use of Tengmalm’s Owl Aegolius funereus home range in the high Belgian Ardennes: Results from radio-tracking. Alauda 2003, 71, 215–220. [Google Scholar]
  62. Santangeli, A.; Hakkarainen, H.; Laaksonen, T.; Korpimäki, E. Home range size is determined by habitat composition but feeding rate by food availability in male Tengmalm’s owls. Anim. Behav. 2012, 83, 1115–1123. [Google Scholar] [CrossRef]
  63. Moran, P.A.P. The Interpretation of Statistical Maps. J. R. Stat. Soc. Ser. B 1948, 2, 243–251. [Google Scholar] [CrossRef]
  64. Kissling, W.D.; Carl, G. Spatial autocorrelation and the selection of simultaneous autoregressive models. Glob. Ecol. Biogeogr. 2008, 17, 59–71. [Google Scholar] [CrossRef]
  65. Valcu, M.; Kempenaers, B. Spatial autocorrelation: An overlooked concept in behavioral ecology. Behav. Ecol. 2010, 21, 902–905. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  66. Beaumont, L.; Hughes, L.; Poulsen, M. Predicting species distributions: Use of climatic parameters in BIOCLIM and its impact on predictions of species’ current and future distributions. Ecol. Modell. 2005, 2, 251–270. [Google Scholar] [CrossRef]
  67. Zhao, Y.; Cao, H.; Xu, W.; Chen, G.; Lian, J.; Du, Y.; Ma, K. Contributions of precipitation and temperature to the large scale geographic distribution of fleshy-fruited plant species: Growth form matters. Sci. Rep. 2018, 8, 1–9. [Google Scholar] [CrossRef] [Green Version]
  68. WorldClim. Published 2022. Available online: https://www.worldclim.org/ (accessed on 1 October 2021).
  69. Wei, B.; Wang, R.; Hou, K.; Wang, X.; Wu, W. Predicting the current and future cultivation regions of Carthamus tinctorius L. using MaxEnt model under climate change in China. Glob. Ecol. Conserv. 2018, 16, e00477. [Google Scholar] [CrossRef]
  70. Nikolov, B.P.; Zlatanov, T.; Groen, T.; Stoyanov, S.; Hristova-Nikolova, I.; Lexer, M.J. Habitat requirements of Boreal Owl (Aegolius funereus) and Pygmy Owl (Glaucidium passerinum) in rear edge montane populations on the Balkan Peninsula. Avian Res. 2022, 13, 100020. [Google Scholar] [CrossRef]
  71. Elith, J.; Kearney, M.; Phillips, S. The art of modelling range-shifting species. Methods Ecol. Evol. 2010, 1, 330–342. [Google Scholar] [CrossRef]
  72. Syfert, M.M.; Smith, M.J.; Coomes, D.A. The Effects of Sampling Bias and Model Complexity on the Predictive Performance of MaxEnt Species Distribution Models. PLoS ONE 2013, 8, e55158. [Google Scholar] [CrossRef]
  73. Radosavljevic, A.; Anderson, R.P. Making better Maxent models of species distributions: Complexity, overfitting and evaluation. J. Biogeogr. 2013, 41, 629–643. [Google Scholar] [CrossRef]
  74. Morales, N.S.; Fernández, I.C.; Baca-González, V. MaxEnt’s parameter configuration and small samples: Are we paying attention to recommendations? A systematic review. PeerJ 2017, 2017, e3093. [Google Scholar] [CrossRef] [PubMed]
  75. Akaike, H. A new look at the statistical model identification. IEEE Trans Automat Contr. 1974, 19, 716–723. [Google Scholar] [CrossRef]
  76. Burnham, K.P.; Anderson, D.R. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach, 2nd ed.; Springer-Verlag: Berlin/Heidelberg, Germany, 2002. [Google Scholar]
  77. Kline, R.B. Principles and Practice of Structural Equation Modeling; The Guilford Press: New York, NY, USA, 1998. [Google Scholar]
  78. Grimmett, L.; Whitsed, R.; Horta, A. Presence-only species distribution models are sensitive to sample prevalence: Evaluating models using spatial prediction stability and accuracy metrics. Ecol. Modell. 2020, 431, 109194. [Google Scholar] [CrossRef]
  79. Quinn, G.; Keough, M. Experimental Design and Data Analysis for Biologists; Cambridge University Press: Cambridge, UK, 2002; ISBN 9780521009768. [Google Scholar] [CrossRef] [Green Version]
  80. Liu, C.; White, M.; Newell, G. Selecting thresholds for the prediction of species occurrence with presence-only data. J. Biogeogr. 2013, 40, 778–789.8. [Google Scholar] [CrossRef]
  81. rStudio Team. rStudio: Integrated Development Environment for R. rStudio. 2021. Available online: https://posit.co/download/rstudio-desktop/ (accessed on 3 September 2022).
  82. Zhu, B.; Wang, B.; Zou, B.; Xu, Y.; Yang, B.; Yang, N.; Ran, J. Assessment of habitat suitability of a high-mountain Galliform species, buff-throated partridge (Tetraophasis szechenyii). Glob. Ecol. Conserv. 2020, 24, e01230. [Google Scholar] [CrossRef]
  83. Zhang, J.; Jiang, F.; Li, G.; Qin, W.; Li, S.; Gao, H.; Cai, Z.; Lin, G.; Zhang, T. Maxent modeling for predicting the spatial distribution of three raptors in the Sanjiangyuan National Park, China. Ecol. Evol. 2019, 9, 6643–6654. [Google Scholar] [CrossRef] [Green Version]
  84. Obunga, G.; Siljander, M.; Maghenda, M.; Pellikka, P.K.E. Habitat suitability modelling to improve conservation status of two critically endangered endemic Afromontane forest bird species in Taita Hills, Kenya. J. Nat. Conserv. 2022, 65, 126111. [Google Scholar] [CrossRef]
  85. Meza Mori, G.; Rojas-Briceño, N.B.; Cotrina Sánchez, A.; Oliva-Cruz, M.; Olivera Tarifeño, C.M.; Hoyos Cerna, M.Y.; Ramos Sandoval, J.D.; Torres Guzmán, C. Potential Current and Future Distribution of the Long-Whiskered Owlet (Xenoglaux loweryi) in Amazonas and San Martin, NW Peru. Animals 2022, 12, 1794. [Google Scholar] [CrossRef] [PubMed]
  86. Ševčík, R.; Riegert, J.; Šťastný, K.; Zárybnický, J.; Zárybnická, M. The effect of environmental variables on owl distribution in Central Europe: A case study from the Czech Republic. Ecol. Inform. 2021, 64, 101375. [Google Scholar] [CrossRef]
  87. Flousek, J.; Telenský, T.; Hanzelka, J.; Reif, J. Population trends of central European montane birds provide evidence for adverse impacts of climate change on high-altitude species. PLoS ONE 2015, 10, 1–14. [Google Scholar] [CrossRef] [PubMed]
  88. Alba, R.; Kasoar, T.; Chamberlain, D.; Buchanan, G.; Thompson, D.; Pearce-Higgins, J.W. Drivers of change in mountain and upland bird populations in Europe. Ibis 2022, 164, 635–648. [Google Scholar] [CrossRef]
  89. Reif, J.; Flousek, J. The role of species’ ecological traits in climatically driven altitudinal range shifts of central European birds. Oikos 2012, 121, 1053–1060. [Google Scholar] [CrossRef]
  90. Tellería, J.L. Long-term altitudinal change in bird richness in a Mediterranean mountain range: Habitat shifts explain the trends. Reg. Environ. Chang. 2020, 20, 9. [Google Scholar] [CrossRef]
  91. Chamberlain, D.; Arlettaz, R.; Caprio, E.; Maggini, R.; Pedrini, P.; Rolando, A.; Zbinden, N. The altitudinal frontier in avian climate impact research. Ibis 2012, 154, 205–209. [Google Scholar] [CrossRef] [Green Version]
  92. Brambilla, M.; Caprio, E.; Assandri, G.; Scridel, D.; Bassi, E.; Bionda, R.; Celada, C.; Falco, R.; Bogliani, G.; Pedrini, P.; et al. A spatially explicit definition of conservation priorities according to population resistance and resilience, species importance and level of threat in a changing climate. Divers. Distrib. 2017, 23, 727–738. [Google Scholar] [CrossRef]
  93. Puzović, S.; Radišić, D.; Ružić, M.; Rajković, D.; Radaković, M.; Pantović, U.; Janković, M.; Stojnić, N.; Šćiban, M.; Tucakov, M.; et al. Birds of Serbia: Breeding Population Estimates and Trend for the period 2008–2013; Department of Biology and Ecology, Faculty of Sciences, University of Novi Sad: Novi Sad, Serbia, 2015; ISBN 9788691130305. [Google Scholar]
Animals 12 03226 g001 550
Figure 1. Study area (a) at a larger scale. Map of the study area with elevation of the Balkan Peninsula at a smaller scale with points where Boreal Owl (Aegolius funereus) (b) and Eurasian Pygmy Owl (Glaucidium passerinum) were present (c).
Figure 1. Study area (a) at a larger scale. Map of the study area with elevation of the Balkan Peninsula at a smaller scale with points where Boreal Owl (Aegolius funereus) (b) and Eurasian Pygmy Owl (Glaucidium passerinum) were present (c).
Animals 12 03226 g001
Animals 12 03226 g002 550
Figure 2. Predicted highly suitable habitat of Boreal Owl (Aegolius funereus) under projected future climate scenarios (SSP 126, 245, 370, 585) in two different periods: 2041–2060 and 2061–2080. Colour coding: beige = unsuitable habitat; light yellow = poorly suitable habitat; dark yellow = moderately suitable habitat; red = highly suitable habitat).
Figure 2. Predicted highly suitable habitat of Boreal Owl (Aegolius funereus) under projected future climate scenarios (SSP 126, 245, 370, 585) in two different periods: 2041–2060 and 2061–2080. Colour coding: beige = unsuitable habitat; light yellow = poorly suitable habitat; dark yellow = moderately suitable habitat; red = highly suitable habitat).
Animals 12 03226 g002
Animals 12 03226 g003 550
Figure 3. Predicted highly suitable habitat of Eurasian Pygmy Owl (Glaucidium passerinum) under projected future climate scenarios (SSP 126, 245, 370, 585) in two different periods: 2041–2060 and 2061–2080. Colour coding: beige = unsuitable habitat; light yellow = poorly suitable habitat; dark yellow = moderately suitable habitat; red = highly suitable habitat).
Figure 3. Predicted highly suitable habitat of Eurasian Pygmy Owl (Glaucidium passerinum) under projected future climate scenarios (SSP 126, 245, 370, 585) in two different periods: 2041–2060 and 2061–2080. Colour coding: beige = unsuitable habitat; light yellow = poorly suitable habitat; dark yellow = moderately suitable habitat; red = highly suitable habitat).
Animals 12 03226 g003
Animals 12 03226 g004 550
Figure 4. Type 1—in situ refugia for Boreal Owl (Aegolius funereus) and for Eurasian Pygmy Owl in the Balkan Peninsula in 2041–2060 and 2061–2080. Colour coding: grey = no refugia; blue = refugia for Boreal Owl; purple = refugia for Eurasian Pygmy Owl.
Figure 4. Type 1—in situ refugia for Boreal Owl (Aegolius funereus) and for Eurasian Pygmy Owl in the Balkan Peninsula in 2041–2060 and 2061–2080. Colour coding: grey = no refugia; blue = refugia for Boreal Owl; purple = refugia for Eurasian Pygmy Owl.
Animals 12 03226 g004
Table
Table 1. Mean AUC values (AUCmean) and standard deviation of the mean AUC values (AUCmeanSD) for the current and future MaxEnt models for Boreal Owl’s distribution under different SSP scenarios.
Table 1. Mean AUC values (AUCmean) and standard deviation of the mean AUC values (AUCmeanSD) for the current and future MaxEnt models for Boreal Owl’s distribution under different SSP scenarios.
PeriodsSSPAUCmeanAUCmeanSD
Current 0.910.015
2041–20601260.930.015
2450.890.021
3700.910.023
5850.890.024
2061–20801260.880.023
2450.0890.023
3700.860.026
5850.870.025
Table
Table 2. Mean AUC values (AUCmean) and standard deviation of the mean AUC values (AUCmeanSD) for the current and future MaxEnt models for Eurasian Pygmy Owl’s distribution under different SSP scenarios.
Table 2. Mean AUC values (AUCmean) and standard deviation of the mean AUC values (AUCmeanSD) for the current and future MaxEnt models for Eurasian Pygmy Owl’s distribution under different SSP scenarios.
PeriodsSSPAUCmeanAUCmeanSD
Current 0.910.025
2041–20601260.870.037
2450.880.037
3700.920.016
5850.90.017
2061–20801260.920.017
2450.940.009
3700.90.025
5850.920.019
Table
Table 3. Extent of predicted three different categories of suitable habitats (km2) for Boreal Owl (Aegolius funereus) in current time, and in two periods: 2041–2060 and 2061–2080, in different climate scenarios (SSP 126, 245, 370, 585). Additionally, we calculated changes (%) from current time to future periods.
Table 3. Extent of predicted three different categories of suitable habitats (km2) for Boreal Owl (Aegolius funereus) in current time, and in two periods: 2041–2060 and 2061–2080, in different climate scenarios (SSP 126, 245, 370, 585). Additionally, we calculated changes (%) from current time to future periods.
YearsScenariosPredicted Area (km2)Changes in Area (%)
Total poorly suitable habitatTotal moderately suitable habitatTotal highly suitable habitatTotal poorly suitable habitatTotal moderately suitable habitatTotal highly suitable habitat
Current-1192447261
2041–2060ssp 126135544027613.67−1.575.75
ssp 24512704052666.54−9.401.92
ssp 37012664522806.211.127.28
ssp 5851107386222−7.13−13.65−14.94
2061–2080ssp 12612553972495.29−11.19−4.60
ssp 24512483992484.70−10.74−4.98
ssp 3701128391247−5.37−12.53−5.36
ssp 5851147386233−3.78−13.65−10.73
Table
Table 4. Extent of predicted three different categories of suitable habitats (km2) for Eurasian Pygmy Owl (Glaucidium passerinum) in current time, and in two periods: 2041–2060 and 2061–2080, in different climate scenarios (SSP 126, 245, 370, 585). Additionally, we calculated changes (%) from current time to future periods.
Table 4. Extent of predicted three different categories of suitable habitats (km2) for Eurasian Pygmy Owl (Glaucidium passerinum) in current time, and in two periods: 2041–2060 and 2061–2080, in different climate scenarios (SSP 126, 245, 370, 585). Additionally, we calculated changes (%) from current time to future periods.
YearsScenariosPredicted Area (km2) Changes in Area (%)
Total poorly suitable habitatTotal moderately suitable habitatTotal highly suitable habitatTotal poorly suitable habitatTotal moderately suitable habitatTotal highly suitable habitat
Current-1271385233---
2041–2060ssp 12612793382140.63−12.21−8.15
ssp 24513573592126.77−6.75−9.01
ssp 370164140420629.114.94−11.59
ssp 58513093452152.99−10.39−7.73
2061–2080ssp 126158839923824.943.642.15
ssp 245164145423829.1117.922.15
ssp 370145139123214.161.56−0.43
ssp 585145740125014.634.167.30
Table
Table 5. Results of the similarity between species distribution models (SDMs) performed by calculating I statistics for Eurasian Pygmy Owl (Glaucidium passerinum) and Boreal Owl (Aegolius funereus). Comparison of current SDM with each SSP (126, 245, 370, 585) from each period (2041–2060, 2061–2080). I statistic ranges from 0–1, 0 being no similarity, and 1 being complete similarity of niche models.
Table 5. Results of the similarity between species distribution models (SDMs) performed by calculating I statistics for Eurasian Pygmy Owl (Glaucidium passerinum) and Boreal Owl (Aegolius funereus). Comparison of current SDM with each SSP (126, 245, 370, 585) from each period (2041–2060, 2061–2080). I statistic ranges from 0–1, 0 being no similarity, and 1 being complete similarity of niche models.
PeriodClimatic ScenariosEurasian Pygmy OwlBoreal Owl
I statisticI statistic
2041–2060Current vs 1260.8570.926
Current vs 2450.8250.991
Current vs 3700.7060.882
Current vs 5850.8720.442
2061–2080Current vs 1260.9910.95
Current vs 2450.9910.941
Current vs 3700.9990.932
Current vs 5850.9860.719
Table
Table 6. Extent of two types of refugia (km2) for Boreal Owl (Aegolius funereus). Area of type 1 refugium (in-situ refugium) and type 2 refugium (ex-situ refugium) for each period (2041–2060, 2061–2080) and each SSP (126, 245, 370, 585).
Table 6. Extent of two types of refugia (km2) for Boreal Owl (Aegolius funereus). Area of type 1 refugium (in-situ refugium) and type 2 refugium (ex-situ refugium) for each period (2041–2060, 2061–2080) and each SSP (126, 245, 370, 585).
RefugiumPeriodSSPArea (km2)
type 12041–2060126232
245218
370221
585196
type 12061–2080126214
245206
370205
585207
type 22041–206012645
24519
37014
5857
type 22061–208012635
24523
37015
5859
Table
Table 7. Extent of two types of refugia (km2) for Eurasian Pygmy Owl (Glaucidium passerinum). Area of type 1 refugium (in-situ refugium) and type 2 refugium (ex-situ refugium) for each period (2041–2060, 2061–2080) and each SSP (126, 245, 370, 585).
Table 7. Extent of two types of refugia (km2) for Eurasian Pygmy Owl (Glaucidium passerinum). Area of type 1 refugium (in-situ refugium) and type 2 refugium (ex-situ refugium) for each period (2041–2060, 2061–2080) and each SSP (126, 245, 370, 585).
RefugiumPeriodSSPArea (km2)
type 12040–2061126186
245168
370168
585180
type 12061–2080126184
245190
370203
585198
type 22040–206112628
24544
37038
58535
type 22061–208012654
24548
37029
58552
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Cerman, K.; Rajković, D.; Topić, B.; Topić, G.; Shurulinkov, P.; Mihelič, T.; Delgado, J.D. Environmental Niche Modelling Predicts a Contraction in the Potential Distribution of Two Boreal Owl Species under Different Climate Scenarios. Animals 2022, 12, 3226. https://doi.org/10.3390/ani12223226

AMA Style

Cerman K, Rajković D, Topić B, Topić G, Shurulinkov P, Mihelič T, Delgado JD. Environmental Niche Modelling Predicts a Contraction in the Potential Distribution of Two Boreal Owl Species under Different Climate Scenarios. Animals. 2022; 12(22):3226. https://doi.org/10.3390/ani12223226

Chicago/Turabian Style

Cerman, Kristina, Draženko Rajković, Biljana Topić, Goran Topić, Peter Shurulinkov, Tomaž Mihelič, and Juan D. Delgado. 2022. "Environmental Niche Modelling Predicts a Contraction in the Potential Distribution of Two Boreal Owl Species under Different Climate Scenarios" Animals 12, no. 22: 3226. https://doi.org/10.3390/ani12223226

APA Style

Cerman, K., Rajković, D., Topić, B., Topić, G., Shurulinkov, P., Mihelič, T., & Delgado, J. D. (2022). Environmental Niche Modelling Predicts a Contraction in the Potential Distribution of Two Boreal Owl Species under Different Climate Scenarios. Animals, 12(22), 3226. https://doi.org/10.3390/ani12223226

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop