The Impact of Climate and Land Use Change on Greek Centipede Biodiversity and Conservation

Abstract

Centipedes (Chilopoda, Myriapoda) are crucial soil predators, yet their vulnerability to climate and land use change remains unexplored. We assess the impact of these drivers on Greek centipedes, identify current and future biodiversity hotspots, and evaluate the effectiveness of the Natura 2000 Network of protected areas for their conservation. We used an updated species occurrence database of Greek centipedes, derived from literature reviews and museum collections, and evaluated database completeness and geographic sampling biases. Species Distribution Models were employed to predict future distribution shifts under climate and land use change scenarios. Biodiversity hotspots were identified based on species richness (SR) and corrected-weighted endemism (CWE) metrics. We overlapped SR and CWE metrics against the Natura 2000 Network to assess its effectiveness. We found that sampling effort is highly heterogeneous across Greece. All species are projected to experience range contractions, particularly in the 2080s, with variation across scenarios and taxa. Current biodiversity hotspots are concentrated in the south Aegean islands and mainland mountain ranges, where areas of persistent high biodiversity are also projected to occur. The Natura 2000 Network currently covers 52% of SR and 44% of CWE hotspots, with projected decreases in SR coverage but increases in CWE coverage. Our work highlights the vulnerability of Greek centipedes to climate and land use change and reveals conservation shortfalls within protected areas. We identify priority areas for future field surveys, based on sampling bias and survey completeness assessments, and highlight the need for further research into mechanisms driving centipede responses to global change.

1. Introduction

There is an ongoing biodiversity extinction crisis [1,2], with arthropods being severely affected [3,4]. Human activities are almost entirely responsible for the decline and extirpation of arthropod populations, primarily due to habitat destruction, degradation and fragmentation, climate change, the introduction of alien species, and pollution [3,5]. Consequently, it is imperative to identify and preserve areas where biodiversity can be protected against these threats (i.e., refugia; [6,7]) and more importantly, the ‘Anthropocene refugia’, that is, regions where biodiversity is expected to persist despite long-term anthropogenic pressures [8].

Centipedes have an evolutionary history spanning 420 million years [9] and comprise over 3150 described species worldwide [10,11], classified into five orders: (1) Scutigeromorpha; (2) Lithobiomorpha; (3) Craterostigmomorpha; (4) Geophilomorpha; and (5) Scolopendromorpha [12,13]. Centipedes play a crucial role as predators in soil ecosystems. They are soil dwellers, showing a distinct preference for moist microhabitats [14]. Due to their limited dispersal ability [14], centipede distributions are shaped by habitat specificity, geo-climatic processes [15,16,17,18], and are strongly influenced by geographic barriers [19] and geological events, leading to considerable variation in latitudinal ranges and biogeographic affiliations [10,20]. This makes them particularly vulnerable to sudden environmental and climatic changes [21,22]. In general, centipedes are considered reliable indicators of environmental health [23,24,25], as their sensitivity to habitat changes makes them useful for assessing environmental disturbances [26]. Research examining how climate and land use changes affect centipedes remains scarce. This knowledge gap raises particular concerns in the Mediterranean region, which stands out as a major center of invertebrate biodiversity [27] whilst facing acute threats from climate change and species extinctions (e.g., [28]). Within this context, Greece (Figure 1) harbors remarkable centipede diversity, with 110 (20 endemics) of Europe’s approximately 585 species present within its borders [11,29,30,31,32] (see also Table S1). Greece serves as a vital refuge for centipedes [31] and various other taxonomic groups [33,34].

The distinctive biodiversity patterns observed in Greece stem from its complex topographical features, which include extensive mountain ranges across the mainland and a highly fragmented landscape comprising over 7500 islands and islets in the Aegean archipelago [35,36,37]. The current species count likely understates the true centipede diversity in Greece, given these animals’ small size, secretive behavior, and limited dispersal capabilities. This underestimation is further compounded by uneven sampling efforts between mainland and insular regions [32,38,39].

The conservation of Greek centipedes faces significant challenges due to gaps in distribution data, despite the severe extinction risks threatening invertebrates in the region [40]. Research on the impacts of climate and land use changes on Greek centipedes is lacking, and their conservation needs have received little attention at both regional and global levels, with only a few exceptions (e.g., [41]).

Examining centipede responses to environmental changes and mapping their biodiversity centers through time would strengthen the inclusion of arthropods in Greek conservation frameworks, addressing a current shortcoming in protection measures [40,42]. Moreover, studying areas beyond the Natura 2000 Network boundaries may reveal previously unknown zones of high soil biodiversity value. This research represents the first assessment of how effectively the Natura 2000 Network protects Greek soil arthropods and their diversity centers, building upon similar analyses for other taxonomic groups [34,43,44].

This study pursues four main objectives: (1) To map centers of centipede species richness across Greece, (2) to explore spatial coverage and potential sampling biases in the available distribution data of Greek centipedes, (3) to determine how climate and landscape features shape centipede diversity and distribution patterns under current and future conditions, and, (4) to measure the effectiveness of the Natura 2000 Network in protecting Greek centipedes, both now and under projected future scenarios. Through these analyses, we aim to create a scientific foundation for incorporating these understudied soil arthropods into national Greek conservation strategies.

2. Materials and Methods

2.1. Species Occurrence Data

We compiled a list of centipede accounts belonging to the four orders of Scutigeromorpha, Lithobiomorpha, Scolopendromorpha, and Geophilomorpha that are present in Greece. Information on centipede presence records was drawn from the published literature and from species samples stored at the collections of the Natural History Museum of Crete. We followed the systematic classification of species and nomenclature proposed by ChiloBase v.2 (2016) [11]. Species occurrences were mapped onto a 1 × 1 km spatial grid. The ‘elimCellDups’ function from the ‘enmSdm’ 0.5.3.3 R package [45] was used to remove duplicate entries in each grid cell. For modeling purposes (see below), further spatial thinning was applied using the ‘thin’ function from the ‘spThin’ 0.1.0 R package [46] at 1 km resolution to match the spatial resolution of the predictor variables. We cleaned and spatially thinned the data according to standardized methods [46,47], while adhering to species distribution modeling (SDM) protocols for data quality and spatial resolution [48].

To maintain analytical robustness, from a total of 110 taxa, only taxa with at least five occurrences were included in the species distribution models and analyses, yielding 2147 records across 59 taxa (55 widespread species and four Greek endemics; Table S2). The chosen threshold was supported by [49], who demonstrated that the Ensemble of Small Models (ESMs) framework can produce reliable distribution models with as few as two occurrences per taxon.

Additionally, two alien henicopid centipede species, i.e., Rhodobius lagoi Silvestri, 1933 and Lamyctes emarginatus (Newport, 1844), were excluded from the species list. R. lagoi, native to the Southern Hemisphere, was likely introduced to Rhodes by human activity [29,39], while L. emarginatus, an Australian species introduced to Europe, has been recorded from a single location in northwestern mainland Greece [29]. The precise distribution, population status, and ecological impact of these alien species on local biodiversity remain critical topics for future investigation.

2.2. Sampling Bias and Data Completeness

Geographic sampling biases were examined using the ‘calculate_bias’ function from the ‘sampbias’ 2.0.0 R package, which models sampling rates via a Poisson process within a Bayesian framework [50]. The model assessed how specimen records vary with proximity to rivers, airports, and populated areas, while accounting for inter-factor correlations [50]. This method quantifies bias magnitude and identifies unexplored areas by assuming species presence throughout the study region. The bias effect measures missed records per cell relative to distance from geographical features, with pronounced biases indicated by rapid declines in sampling intensity [50]. We obtained Greek city and town data from the Greek geographical data portal (www.geodata.gov.gr, accessed on 5 January 2025) and road network information from the Humanitarian Data Exchange portal (https://data.humdata.org/, accessed on 6 January 2025).

Using the ‘KnowBR’ 2.2 R package, we evaluated survey completeness and centipede diversity knowledge across Greece. The package generates species accumulation curves for each geographic unit and determines survey coverage from existing data [51]. By fitting these curves to asymptotic functions, the analysis predicts species numbers per unit under infinite sampling effort [51]. The ratio of observed to expected records determines database completeness, indicating both survey effort and knowledge depth [51].

We modified the ‘KnowB’ and ‘KnowBpolygon’ functions to analyze geographic information quality and species diversity knowledge at 1 × 1 km resolution. Species accumulation curves were fitted to the ‘Clench’ function for individual geographic units to assess sampling completeness. Sampling quality was determined using two metrics: the terminal gradient of accumulation curves and completeness values. To ensure analytical robustness, grid cells containing fewer than five specimens were excluded. The ‘SurveyQ’ function from the ‘KnowBR’ 2.2 R package was used to assess sampling effort quality by integrating completeness, the final accumulation curve slope, and the R/S ratio. Well-sampled areas were characterized by a slope < 0.02, completeness > 90%, and an R/S ratio > 15, whereas poorly sampled areas exhibited a slope > 0.3, completeness < 50%, and an R/S ratio < 3 [51].

We then adapted the ‘ignorance_map’ function from the ‘ignobior’ 1.0.0 R package to assess faunistic spatiotemporal uncertainty for each 1 × 1 km grid cell by assigning record-specific spatiotemporal weights. This yielded a relative faunistic ignorance index, with temporal coefficient tau set to 20 and spatial uncertainty to 1000 m, matching the resolution of previous analyses.

The mapping framework proposed by [52] was modified to assess sampling gaps and biases. The analysis incorporated four layers: cells with fewer than six records (lowest decile), cells with poor survey coverage, cells with sampling rates below the 10th percentile, and cells with ignorance values below the 10th percentile. A numerical classification system was applied, assigning values of 1 for sparse records, 3 for poor survey coverage, 5 for low sampling rates, and 11 for high ignorance, generating unique sums for all condition combinations. This approach identified areas where sampling efforts were constrained by limited accessibility, highlighting priority locations for future centipede surveys.

2.3. Environmental Data

We developed a high-resolution (1 km) monthly climate dataset covering 1981–2009, which incorporated 19 WorldClim bioclimatic variables [53] and 16 additional environmental parameters [54]. This timeframe corresponded with the collection period of most occurrence records. The dataset integrated altitude data from the CGIAR Consortium for Spatial Information [55]. Data processing was conducted using ClimateEU v4.63 and the R packages ‘dismo’ 1.1.4 [56] ‘envirem’ 2.2 [54], following established protocols [57,58,59].

The analysis incorporated soil metrics from SoilGrids [60] and dynamic land use/land cover (LULC) data from [61], maintaining consistent resolution across all environmental parameters. We calculated five topographical metrics—aspect, heat load index, slope, topographic position index and terrain ruggedness index—using the aforementioned altitude data and functions from the R packages ‘terra’ 1.8–50 [62] and ‘spatialEco’ 2.0–2 [63].

Climate projections spanned three periods [59]: the 2020s (2011–2040), 2050s (2041–2070), and 2080s (2071–2100). These projections integrated three Global Circulation Models (GCMs)—CCSM4, HadGEM2, and an ensemble of 15 GCMs—with two Intergovernmental Panels on Climate Change Representative Concentration Pathways (RCP45 and RCP85). We included future LULC projections [61] under three Shared Socioeconomic Pathways: SSP1-RCP26, SSP3-RCP70, and SSP5-RCP85 scenarios [64].

From 60 environmental variables, we utilized a taxon-specific curated set of uncorrelated environmental variables, treating topographical and soil parameters as static and bioclimatic and LULC variables as dynamic. This curated taxon-specific set underwent rigorous collinearity testing using the ‘collinear’ 1.1.1 R package [65], applying Spearman rank correlation (threshold < 0.7) and variance inflation factor (threshold < 5) criteria [66].

2.4. Species Distribution Models

Our analyses included species with occurrence-to-predictor ratios below 10:1. We modeled their realized climatic niches using the Random Forest algorithm and the ‘ecospat’ 4.1.2 R package [67], following protocols established by [68,69,70] specifically designed for rare species modeling and [71,72]. Background areas for each taxon were calculated using the alpha hull method via the ‘ConR’ 1.1.1 package [73], given the imprecise knowledge of their Greek distributions.

We categorized taxa into two groups based on occurrence numbers: ≥10 occurrences and 5–9 occurrences [74]. For the first group, pseudo-absences were generated using the ‘sample_pseudoabs’ function from ‘flexsdm’ 1.3.0 [75], incorporating geographical buffering, environmental constraints and k-means clustering. The second group required random pseudo-absences, following rare species protocols [76,77].

For taxa with ≥20 occurrences, we performed optimized spatial cross-validation before model fitting using the ‘part_sblock’ function from ‘flexsdm’ 1.3.0 [75]. Model performance was assessed against null models [78] using multiple evaluation metrics [79,80,81], as recommended by [82,83]. For taxa with 5–19 occurrences, we partitioned occurrences and pseudo-absences using the ‘bm_CrossValidation’ function from ‘biomod2’ 4.2-6-2 [84].

Suitable habitats were identified using models with TSS scores ≥ 0.4. Binary maps were created using metrics that optimized sensitivity and specificity. We produced the final species maps (habitat suitability and binary maps) by excluding the cells that had high extrapolation values. To address potential extrapolation errors, we quantified uncertainty in our predictions using the ‘extra_eval’ function from the ‘flexsdm’ 1.3.3 package [75], and we investigated several distance thresholds for all species regarding their extrapolation values for model prediction truncation [85]. This strategy is robust against potential prediction errors [86]. We then truncated habitat suitability and binary maps by excluding areas with high extrapolation uncertainty.

Future range shifts were analyzed using ‘biomod2’ 4.2-6-2 [84], assuming minimal dispersal ability for Greek endemic taxa. We quantified habitat fragmentation through patch numbers and effective mesh size calculations using the ‘landscapemetrics’ 2.0.0 R package [87].

2.5. Biodiversity Hotspots Detection

For our spatial analyses involving species richness (SR) and corrected-weighted endemism (CWE; [88,89]), we utilized the methodology described in [34]. In line with the procedures detailed in [34,90], we identified biodiversity hotspots based on various taxonomic biodiversity metrics. These hotspots represent areas with the highest 1% values (termed L1 hotspots) for each metric, identified using functions available in the ‘phyloregion’ 1.0.4 R package [91].

Additionally, Priority Hotspots were determined as per [34]. In this context, biodiversity hotspots are defined and referred to as local biodiversity hotspots, which are situated within broader regional biodiversity hotspots [92]. These analyses were replicated across all GCMs, RCPs, SSPs, and periods. We identified Anthropocene refugia across the study area by applying a strict consensus approach to determine which cells maintain their status as Priority Hotspots under all examined Global Circulation Models, Representative Concentration Pathways, Shared Socioeconomic Pathways, and time periods. For each identified refugium, we calculated its area and altitudinal range across all studied taxa.

2.6. Assessment of Protected Area Effectiveness and Conservation Gaps in Greece

Our overlap analysis was confined to terrestrial Greece and the Special Areas of Conservation (which also includes Special Areas of Conservation that are Special Protection Areas) within the Natura 2000 Network of protected areas in Greece. To evaluate the efficacy of the existing protected areas network in Greece, we initially gathered data from the World Database on protected areas using the ‘wdpar’ 1.0.0 R package [93]. Subsequently, we superimposed current and future L1 hotspots for the weighted biodiversity metrics onto the Greek protected areas network using the ‘sf’ 0.8.0 R package [94]. We thus concentrated on the Priority Hotspots as classified by [34], to pinpoint conservation gaps as per [95]. Cells identified as Priority Hotspots in the 99% quantile (L1) in our analyses, either not covered by Special Areas of Conservation or had less than 10% coverage [96], were designated as Priority conservation gaps in accordance with [95].

3. Results

3.1. Species Occurrence Data

A total of 110 centipede species were documented in Greece, representing four orders: Scutigeromorpha (1 species), Lithobiomorpha (51 species, including 9 endemics), Scolopendromorpha (16 species, including 6 endemics), and Geophilomorpha (42 species, including 5 endemics) (see also Table S1). Endemic species comprised 18% (20 species) of the total centipede diversity. Among 3323 unique occurrence records, Lithobiomorpha accounted for 39.8% (1322 records), followed by Geophilomorpha with 36.8% (1223 records), Scolopendromorpha with 19% (631 records), and Scutigeromorpha with 4.4% (147 records) (

Table 1. Summary results on Greek centipede occurrences resulting from a literature search and the study of the material stored at the collections of the Natural History Museum of Crete.

Order	Unique Occurrences	Species Richness	Endemic Richness
Scutigeromorpha	147	1	0
Lithobiomorpha	1322	51	9
Scolopendromorpha	631	16	6
Geophilomorpha	1223	42	5
Total	3323	110	20

).

3.2. Sampling Bias and Data Completeness

The highest number of records was documented in the southeastern Aegean, specifically on the island of Nisyros in the Dodecanese (Figure S1). The spatial coverage of Greece regarding centipede surveys is low and only 1.5% of the cells covering Greece are considered well-surveyed (Table S3). The well-surveyed cells are scattered in the following: (a) the Peloponnese (e.g., Mount Parnonas and Mount Taygetos), (b) central and north Greece (e.g., Pindos mountain massif, Mount Parnassos), (c) the island of Crete, (d) the southeastern Aegean region (e.g., the islands of Rhodos, Karpathos, Kos, Nisyros, and Symi in the Dodecanese), and (e) the central Aegean region (e.g., the islands of Andros, Kea, Sifnos, Tinos, and Chios) (Figure 2 and Figure S2). The sampling effort across Greece is heterogeneous indicated by the varying values of completeness (0% to ca. 80%) of survey efforts (Figure S3). The survey effort quality is better in southern Greece, and stronger collection efforts are required in order to achieve a better representation of species in poorly sampled areas across central and northern Greece (Figures S2 and S4). Poorly studied or rarely surveyed areas are scattered throughout Greece, including regions such as the Pindos mountain massif and Mount Olympos in northern Greece, the border areas with North Macedonia and Bulgaria, the Evros region, Mount Lefka Ori in Crete, the islands of Naxos and Paros in the Cyclades in the central Aegean, and Limnos Island in the northern Aegean (Figure 2, Figures S2 and S3).

3.3. Species Distribution Models

The models achieved very good performance across evaluation metrics (Figure S5, Table S4). The distribution patterns of centipede species showed differential responses to environmental predictors. Temporally static variables (topographic and soil) drove the distributions of 56% of the species, whilst dynamic variables (bioclimatic and LULC) governed the remaining 44%. Topographic parameters were the dominant predictors for 24 species, with aspect (e.g., Lithobius nigripalpis L. Koch, 1867), heat load index [e.g., Pachymerium ferrugineum (C. L. Koch, 1835)], topographic position index (e.g., Scolopendra cingulata Latreille, 1829), and slope orientation (e.g., Lithobius creticus Dobroruka, 1977) showing the strongest influence. For 17 species, bioclimatic variables were the primary drivers, particularly Thornthwaite’s Aridity Index (e.g., Clinopodes flavidus C. L. Koch, 1847), mean monthly Potential Evapotranspiration of the driest quarter (e.g., Lithobius schuleri Verhoeff, 1825), and isothermality [e.g., Eupolybothrus litoralis (L. Koch, 1867)]. Soil characteristics determined the distribution of 10 species, with sand [e.g., Stenotaenia naxia (Verhoeff, 1901)] and clay particle proportions (e.g., Cryptops anomalans Newport, 1844) and soil pH [e.g., Bothriogaster signata (Kessler, 1874)] being the most influential parameters. Land use patterns shaped the distribution of nine species, with forests (e.g., Lithobius erythrocephalus C.L. Koch, 1847), shrubs [e.g., Cryptops hortensis (Donovan, 1810)], urban areas (e.g., Scolopendra cretica Lucas, 1853), and cropland (e.g., Scolopendra canidens Newport, 1844) emerging as the key categories (Figure 3, Table S5).

The predictive models revealed limited habitat fragmentation across most species (Figure S6B). Only a small number of species exhibited high fragmentation levels, as measured by effective mesh size values, in both current and future scenarios. The patch cohesion analysis demonstrated strong spatial continuity, with species typically occurring in adjacent cells, barring a few exceptions (Figure S6C).

The models forecast range contractions for all species, with the most pronounced reductions occurring in the 2080s (Figure S7, Table S6). Eupolybothrus werneri (Attems, 1902) will undergo the most prominent reduction with 85.5% of its current range projected to be lost. Scolopendra canidens will maintain its current range, with a minimum loss of 1.47%. Range contractions remain consistent across all modeled scenarios (Figure S7, Tables S8–S11). The projections indicate a 29.94% reduction by 2020s, escalating to 37.17% and 38.92% by the 2050s and 2080s, respectively (Table S8). The 2050 and 2080 scenarios consistently predict the most severe contractions across all parameter combinations (Table S8). One-third of the analyzed species face range contractions irrespective of the scenario, with mountain ranges in mainland Greece experiencing the highest rates of local extinctions (Figure 4, Figure 5 and Figure S8). Geophilomorpha exhibited the highest range loss of all four orders, but the difference was significant only compared to Scolopendromorpha (Figure S9, Kruskal–Wallis test with Bonferroni corrections P = 0.000074).

Contrary to expectations, 32 species are projected to contract their vertical niche towards lower altitudes (Figure S10). Cryptops hortensis shows the most substantial altitudinal contraction (approximately 500 m), whilst Eupolybothrus transsylvanicus (Latzel, 1882) exhibits the greatest altitudinal expansion (Table S7). The endemic species Scolopendra cretica, Lithobius brignolli (Matic, 1970) and L. nudus (Matic, 1976) are projected to restrict their range towards lower altitudes and the only endemic species to slightly expand its altitudinal range upward is Lithobius creticus.

3.4. Biodiversity Hotspots and Assessment of Protected Area Effectiveness in Greece

Current biodiversity hotspots are concentrated in three main regions: Crete, the southern Aegean islands, and along Greece’s primary mountain ranges. These ranges start from the northern mountains of continental Greece (e.g., Mount Smolikas, Mount Tymfi) to the central mountains (e.g., Mount Tzoumerka, Mount Timfristos, and Mount Parnassos) ending at the southern Peloponnese mountains (Mount Taygetos, Mount Parnonas, and Mount Mainalo) (Figure 4 and Figure S11). The CWE metric identified key hotspots the central part of mainland Greece (e.g., Mount Timfristos), in the Ionian region (e.g., the island of Kerkyra), in the central Aegean archipelago (e.g., Amorgos), in the southeastern Aegean archipelago (e.g., the islands of Karpathos and Rhodos), as well as in Crete (e.g., coastal areas south of Lefka Ori and the three principal mountain ranges namely Mount Lefka Ori, Mount Psiloreitis and Mount Dikti) (Figure S11). It should be noted that we do not provide richness maps for endemic species, as their numbers are too few to warrant separate representation. Priority Hotspots currently align with both species’ richness and CWE hotspots (Figure S12). Whilst some of these Priority Hotspots are expected to shift in lowland and coastal regions, many existing Priority Hotspots will maintain their conservation value. At present, the Greek protected areas network encompasses 52% of species richness (Table S12) and 44% of CWE L1 hotspots (Table S13), with these percentages projected to decrease for species richness across most parameter combinations and increase for CWE over time (Tables S12 and S13).

4. Discussion

Greece, located at the end tip of the Balkan peninsula, is a local biodiversity hotspot of the Mediterranean region, harboring a plethora of taxa, including several hundreds of endemics (e.g., [31,34,36,42,97,98]). Centipedes have a key role in ecosystems as top soil predators and are good ecological indicators of ecosystem alterations (e.g., [99,100]). Despite their importance, so far, centipedes have been largely overlooked in biodiversity studies globally and at the country-level. In Greece, only a handful of works have reported on the centipedes’ biogeography, diversity patterns, and taxonomy (e.g., [101,102] and references therein), and, at least for Greece, almost none exist on their ecological preferences (but see [29,39]). Here, for the first time, we explored the effect of a set of environmental parameters on the distribution of Greek centipedes in the present and the future, we revealed current centipede diversity hotspots, assessed future species range shifts and their effect on the existing hotspots, and evaluated the effectiveness of the Natura 2000 Network for the protection of current and future centipede biodiversity. Our work aims to lay the groundwork for the incorporation of a smaller and often overlooked animal group in planning of conservation schemes in Greece.

4.1. Greek Centipede Diversity

Identifying spatial distribution gaps and sampling biases in biodiversity datasets is essential for the accurate interpretation of ecological research findings. Our study highlights several regions across Greece that remain underexplored or entirely unstudied to date (see [29]). In comparison to other taxa such as arthropods (e.g., [103,104]), land snails [98], amphibians, and reptiles [105], the distributional knowledge of Greek centipedes remains relatively limited (but see [38,39,102]). Nevertheless, for the first time, we can indicate priority areas for centipede field research based on sound estimations of sampling bias and survey completeness efforts.

The orders Lithobiomorpha and Geophilomorpha exhibited the highest species richness and number of recorded occurrences (Table 1), consistent with findings from previous studies (e.g., [31,38,39]) and reflecting the global diversity patterns observed within centipedes (cf. [10]). The frequent occurrence of Lithobiomorpha can be attributed to their ecological adaptability, as they are capable of surviving and maintaining stable populations across a wide range of habitats and environmental gradients, in contrast to Scolopendromorpha and Scutigeromorpha (e.g., [14,106] and references therein). Similarly, Geophilomorpha are known to inhabit diverse biotopes and, depending on the species, can tolerate extreme conditions such as saline or hypersaline environments and temporary flooding ([14] and references therein). In contrast, Scolopendromorpha and Scutigeromorpha were less frequently recorded (Table 1), with the notable exception of Scolopendra cingulata on the Greek mainland and S. cretica in Crete, both of which are considered common and are frequently observed [29,39,107].

The geographic coverage of centipede records in Greece remains limited at the selected spatial resolution (1 × 1 km grid), with significant variation in the number of records per species, reflecting uneven sampling efforts across the country. Both the distributional knowledge of Greek centipedes and the intensity of field surveys are spatially biased, leaving many areas under-sampled or entirely unexplored. Targeted fieldwork in such poorly investigated regions, such as Lefka Ori in Crete, and parts of northern mainland Greece, is essential to improve the completeness of species inventories and reduce sampling heterogeneity. Such efforts may not only fill existing geographic gaps in species distributions but also lead to the discovery of previously unrecorded taxa, as demonstrated by the first documented occurrence of the west-central Mediterranean species Eurygeophilus multistiliger (Verhoeff, 1899) in mainland Greece, discovered during a modern zoological expedition on Mount Parnonas in the eastern Peloponnese [108]. Possibly, the use of a different spatial resolution (e.g., 5 × 5 km grid) would yield a better geographic coverage. Although by doing so, the degree of data scarcity, gap, and coverage would be hidden, leading to erroneous conclusions regarding the areas in need of increased field sampling. Enhancing sampling coverage will contribute significantly to understanding the biogeography and evolutionary history of centipedes in this Mediterranean biodiversity hotspot (e.g., [31,101]).

4.2. Centipede Diversity Hotspots

Current centipede richness hotspots occur across three main regions: the southern Aegean islands, Crete, and the principal mountain ranges of Greece. These mountain ranges extend from the northern Pindos massif, through Mount Timfristos and Mount Parnassos, to the southern Peloponnese, where they include Mount Taygetos, Mount Parnonas, and Mount Mainalo.

The Aegean region is widely recognized as a local biodiversity hotspot, supporting high species richness across various animal and plant taxa, including reptiles and amphibians [105], butterflies [22,109], land snails [98], and angiosperms [22,36]. This exceptional biodiversity is primarily attributed to the region’s long and complex geological and paleogeographic history [21,110,111]. The Aegean archipelago forms part of a transitional biogeographic zone between the Euro-Mediterranean and the Anatolian peninsula, with many islands acting as biogeographic stepping stones ([112] and references therein). Consistent with our findings, the accumulation of species in the southern Aegean islands, combined with relatively low levels of endemism, likely reflects strong biogeographic affinities with adjacent continental regions, including mainland Europe and Asia Minor [101,112,113].

The application of the CWE metric identified several additional areas of high conservation value across the Aegean Sea. Notable among these are Skyros in the northern Aegean, Amorgos in the central region, and Kalymnos, Rhodos, and Karpathos in the southeastern Aegean. The CWE approach proved particularly effective in revealing previously unrecognized centers of species richness, consistent with patterns observed in other taxonomic groups within Greece [34,112].

Crete, located in the southern Aegean, serves as a significant refuge for a broad range of taxa, including vascular plants, mollusks, arthropods, and centipedes [34,97,114]. The island’s long-term geographic isolation from mainland Greece, in combination with its diverse topography and climatic variability, has fostered distinctive patterns of species richness and distribution. Our analysis identified the mountain ranges of Psiloritis, Lefka Ori, and Dikti as major centers of centipede diversity (Figure 4 and Figure 5), where Mediterranean and European elements coexist with endemics restricted to Crete and the southern Aegean [39].

Three Scolopendra lineages—S. canidens, S. clavipes C. L. Koch, 1847, and S. cretica—inhabit the southeastern Aegean islands. Our findings confirm that the Cyclades and the southeastern Dodecanese archipelagos function as regional diversity centers for centipedes [31], mirroring diversity patterns observed in other biological groups [34,97,112].

Among the endemic centipede species, Scolopendra cretica is the most widespread on Crete, occurring across a broad elevational gradient from sea level to approximately 1600 m [38]. In contrast, other endemics such as Lithobius cretaicus Matic, 1980 and Cryptops kosswigi Chamberlin, 1952 exhibit highly restricted distributions, with records confined to only two cave systems [29]. Overall, the relatively low rate of centipede endemism on Crete [38] indicates that, in comparison with other groups such as land snails [98], non-centipede arthropods [42], and vascular plants [36], centipedes contribute less to the island’s endemic richness.

On the mainland, the northwestern Pindos mountain range emerges as a prominent center of centipede diversity, despite remaining poorly surveyed. This finding reinforces its designation as a broader biodiversity hotspot [34]. In the Peloponnese, the mountain ranges of Taygetos, Parnonas, and Erymanthos also support diverse centipede assemblages, in agreement with previous studies on both centipedes [108] and other faunal groups [97]. Similarly, the Rhodope Mountains in northeastern Greece are confirmed to be a significant diversity center [115].

These mainland diversity hotspots share several ecological features conducive to centipede diversity, including elevated summer humidity and the presence of dense deciduous and evergreen forests with abundant leaf litter and ground cover [14]. Such environmental conditions likely facilitate the persistence of rich centipede communities across Greece’s mountainous regions. Nevertheless, substantial knowledge gaps persist, particularly in many areas of the mainland. Targeted field surveys in these understudied regions are essential to validate modeled species distributions and to more fully document the centipede diversity of Greece.

4.3. Climate and Land Use Change Impact on Species Distribution

Centipede distribution patterns reflect influences at multiple scales: paleogeographic events, environmental heterogeneity, and topography shape broad-scale patterns [31], while climate and habitat preferences determine local distributions [18]. Our analyses identified key predictors (Figure 3) that align with known ecological preferences of centipede species [14,18,29].

The results indicate that global change drivers, particularly climate, strongly influence centipede distributions across Greece (Figure 3 and Figure S10; see also [18]). The consistent range contractions observed across all scenarios suggest that centipedes either show resilience to climate change or respond slowly to changing conditions [18]. However, two methodological considerations warrant attention. First, including land use and land-cover variables in our species distribution models might have led to underestimated species extirpation rates and reduced apparent climate change impacts [116]. Second, our chosen analytical resolution (1 × 1 km) might not fully capture the spatial scale at which centipedes respond to climatic variations.

Climate acts as a primary limiting factor for centipede distributions, particularly at range boundaries [18]. Our analyses demonstrate that climatic parameters strongly influence the predicted distributions of 12 out of 16 species for which Greece represents a range limit. For instance, Greece marks the western boundary for Diphyonyx conjungens (Verhoeff, 1898) [117], whilst serving as the southern limit for both Eupolybothrus werneri and Lithobius aeruginosus L. Koch, 1862 [29,118]. These patterns suggest that climatic conditions play a substantial role in determining the current geographical boundaries of centipede species across Greece (Table S5).

Topographic complexity generates diverse habitats that support rich species assemblages and provide refugia. Centipedes require specific microhabitats, particularly stones for shelter and soil with abundant decaying vegetation that provide cover, moisture, and food resources. Our models show that soil characteristics and land use patterns are the primary predictors for nine and ten species, respectively (Figure 3).

The relationship between land use and centipede distribution aligns with previous findings. The endemic Scolopendra cretica, widespread across Crete, shows a strong association with anthropogenic landscapes [100], becoming scarcer at higher elevations where human influence diminishes [38]. Similarly, Eupolybothrus caesar (Verhoeff, 1899), a woodland specialist [29], shows close ties to forested areas (Table S5). Additional research at finer spatial scales would help clarify the environmental factors, including microclimatic conditions, that shape local centipede distributions.

4.4. Mapping Species Range Shifts and Richness Hotspots over Time

Future scenarios project progressive shifts in biodiversity hotspots (Figure 5, Figures S8, S9, S11 and S12), reflecting species range contractions that become more pronounced towards the 2080s. These changes in climate and land use patterns alter biodiversity hotspots through reduced area and species richness, consistent with findings across other taxonomic groups [40,119].

Aridification will likely drive species losses from southern Greece and lower elevations (see also [40]). However, these projections warrant careful interpretation, as the Greek mainland remains understudied, and mountain massifs such as Olympus, Pindos and Rhodope might harbor unexplored refugia. In Crete, even widespread taxa face substantial threats. The common Scolopendra cretica is projected to experience range and altitudinal distribution contractions, a pattern also observed in Cryptops hortensis, despite its broad distribution. The combined effect of range contractions and limited altitudinal shifts suggests local extinctions, matching predictions for other Cretan flora and fauna [42,114].

Habitat specialists show particular vulnerability to climate change. Woodland-dependent species such as Eupolybothrus werneri, E. transsylvanicus, and Lithobius schuleri are projected to shift northward, possibly tracking the recession of forest habitats [120,121].

4.5. Conservation Implications—Greek Natura 2000 Network Coverage

Centipedes and other invertebrates provide vital ecosystem services, particularly in soil systems [122]. Their loss could severely disrupt food chains and ecological cycles, ultimately threatening ecosystem stability [123]. Understanding their current ecological requirements and future risks is therefore essential for developing effective conservation strategies for soil arthropods at national scales [124].

The Natura 2000 Network currently encompasses 52% of species richness hotspots and 44% of CWE L1 hotspots. While this coverage exceeds that of other taxonomic groups, its projected decline over coming decades reveals substantial gaps in Greece’s conservation framework. This situation is particularly concerning given the incomplete survey coverage of centipede diversity, which might mask biodiversity losses.

Our models predict a reduction in species richness hotspots but increased coverage of CWE L1 hotspots by the Natura 2000 Network. This apparent improvement stems from the shrinking size of CWE hotspots, which concentrates endemic species within protected areas. The persistence of Priority Hotspots (Anthropocene refugia) likely reflects both the resilience of widespread centipedes to global change and the dominant influence of stable abiotic factors on their distributions.

These findings support calls for reassessing the Natura 2000 Network [125], incorporating data from this and other arthropod studies. Anthropocene refugia, given their resilience and projected species richness, warrant particular attention in future conservation planning [8].

4.6. Research Limitations and Future Recommendations

The primary limitation of this study is the scarcity of centipede distribution data from mainland Greece, despite some systematic research efforts (e.g., [108]). Although our models generate robust predictions, additional occurrence records would strengthen our findings. Based on robust estimations of sampling bias and data completeness, we propose geographic areas (Figure 2) that should be prioritized in field research. Furthermore, future field studies should target areas identified by both survey completeness assessments and predictive models as species-rich, particularly the northeastern and central Greek mountain ranges, alongside established hotspots.

The shortage of taxonomic experts working on Greek centipedes has led to reduced research activity. This knowledge deficit, especially within the Natura 2000 Network, necessitates extensive field studies and the training of new taxonomists in centipede systematics. Protected areas should, at minimum, encompass persistent biodiversity hotspots and Anthropocene refugia (see also [34]).

The recent Greek Red List of Endangered Species, which evaluates 100 centipede species according to International Union for Conservation of Nature criteria, marks an important advance in understanding Greek centipede biodiversity (https://redlist.necca.gov.gr/, accessed on 5 March 2025). Future research priorities include addressing knowledge gaps, monitoring population trends, and studying evolutionary histories. Such work may reveal species previously unknown in Greece or new to science, whilst also identifying undiscovered richness hotspots.