Introduction

Shorebirds, by virtue of their life history, behaviour, migratory and foraging habits are important indicators of the integrity of coastal environments, which are currently pinpointed as one of the most threatened ecosystems, as well as being prone to biodiversity loss1,2,3. Coastal sandy ecosystems have undergone severe environmental alterations imposed primarily by human activities4. Many shorebird populations are facing a global decline, driven mainly by human disturbance, though other impacts such as climate change, antagonistic species interactions, and predation have been implicated2,5.

The Kentish Plover, Anarhynchus alexandrinus(Linnaeus, 1758), is a small ground–nesting shorebird in the family Charadriidae that breeds in Europe, Asia and Northern Africa, mainly in coastal habitats, especially sandy beaches6. This species has an extremely large geographic distribution range and is currently evaluated as “Least Concern” (LC) by the International Union for the Conservation of Nature (IUCN), both globally and at European level7. Although Kentish Plovers do not currently approach the threshold for the population decline criterion of the IUCN Red List7, their populations are known to be declining in much of their range2. For instance, the last Red List of Italian breeding birds evaluated the species as “Endangered (EN)” in Italy8, due to the sharp reduction in numbers of the national population (i.e., -50% for the years 2000–2010)9. Because of its conservation status, in Europe the Kentish Plover is included in Annex I of the Birds Directive 2009/147/EEC, Appendix II of the Bern Convention, Appendix II of the Bonn Convention, and it is classified as a Species of European Concern (SPEC) at level ‘3’, with negative population trends10,11,12. In Italy, the species is protected by Art. 2 of L. 157/928,13.

The ongoing decline in the Kentish Plover population in Italy, like other Mediterranean areas, is driven by coastal alterations due to urban expansion, seashore erosion, tourism, and recreational activities, as well as by predation of eggs and chicks (e.g., by crows, rats, yellow–legged gulls, dogs, and cats)9,14. This highlights the urgent need for effective conservation strategies accounting for the most vulnerable phases of the Kentish Plover life cycle. Since the Kentish Plover tends to choose the same nidification sites across the years, monitoring their nests is crucial for its conservation and management15,16. Furthermore, analyzing nest spatial distribution can enhance the understanding of the species’ reproductive ecology and the relationship between nest location and environmental and threat factors. Although standardized nest monitoring protocols are widely implemented17, new research efforts are needed to further develop spatially explicit models that describe nesting habitat preferences and the related threats over time. Previous studies in Europe, Africa, and North America have proposed models depicting the relationship between nesting habitat and environmental and anthropogenic variables derived from remotely sensed data2,4,5,15,16,18. While these studies promoted the use of remotely sensed products as potentially valuable environmental predictors for modelling suitable habitats of various bird species, including the Kentish Plover, they were mainly focused on current habitat conditions. Starting from this evidence, we propose a study aimed at exploring the potential role of environmental variables derived from remote sensing to model current and future Kentish Plover nesting habitat suitability in Southern Italy (i.e., Molise region), explicitly accounting for future scenarios of coastal dynamics. In particular, we focused on addressing the following questions: (a) is there a combination of environmental variables that is particularly conducive to nesting? (b) how would shoreline future dynamics influence the distribution and extent of suitable nesting locations?

In order to answer these questions, we set an analytical framework (Fig. 1) around the following three main objectives:

  1. 1.

    Collect georeferenced observations of Kentish Plover nests and map landscape characteristics derived from remote sensing data along Central Adriatic beaches.

  2. 2.

    Quantify and map species nesting habitat suitability under an ecological niche model (ENM) framework, quantifying the importance of environmental variables in defining nesting preferences.

  3. 3.

    Predict 2040 nesting habitat suitability considering future scenarios of coastline modifications due to climate change and anthropogenic pressure, also calculating range net change (RNC) between current and future modelled nesting habitats.

Fig. 1
figure 1

Flow chart showing the methodological framework used in this research. Maps Data: Google, ©2024 Airbus.

Full size image

Methods

Study species

The Kentish Plover is a ground nesting shorebird that belongs to the Anarynchus genus19.

It belongs to the family Charadriidae, which includes plovers, dotterels, and lapwings, usually with compact bodies and short, thick necks20. The Kentish Plover is a small bird within its family, with sizes ranging between 15 and 17 cm for a weight of about 39–48 g, making it one of the smallest breeding plovers in Italy16. The species follows the typical colouring pattern that characterizes the family Charadriidae, sandy upperparts and candy white lower parts16,21. Adult males and females have dimorphic plumage; males having incomplete black breast–bands, black eye–stripes, and a black frontal head bar, whereas these areas are pale brown in females (Fig. 2)22. The cryptic colouration of the Kentish Plover makes the species difficult to distinguish from the substrate on which it nidifies and feeds16.

The main source of food of the Kentish Plover consists of small aquatic and terrestrial invertebrates such as insects, larvae, arachnids, small crustaceans, bivalve and univalve molluscs, and polychaete worms6,16. The bird usually finds these invertebrates on the coastline, lakes or lagoons, where it picks them up directly from the ground or digs small wells in the sand16,23. Although most of the Kentish Plover populations are fully migratory, this species is partially sedentary in Molise7,24. It is a bird that exhibits a high degree of loyalty to its nesting site, resulting in the pairs consistently constructing their nests in the same locations as in prior years15. The nesting period begins around March and ends in August. The modal clutch size is three small eggs, very mimetic, sand–greyish colour spotted of black (Fig. 2), that hatch after 25–26 days of incubation. Both sexes incubate the eggs; females incubate mostly during the day, and males incubate mostly during the night16,22. The chicks are nidifugous and after about 2 h from hatching are ready to leave the nest to follow the movements of the parents that attend, brood, and defend them for about 4–5 weeks, but they do not provide food for them15,22.

Fig. 2
figure 2

Adult male (a) and adult female (b) of Kentish Plover; typical Kentish Plover nest with three eggs (c). Pictures by Rosario Balestrieri.

Full size image

Study area

The Molise coast (Southern Italy) extends for over 30 km along the Adriatic Sea and falls entirely in the physiographic unit of Punta Penna (Abruzzo region) – Punta Pietre Nere (Apulia region)25,26. This area is mainly composed of sandy beaches, a few river mouths and channels, and one rocky promontory. Dunes occupy a narrow strip (approximately 50–70 m large) parallel to the seashore, being low, simple in structure, and relatively recent (formed in the Holocene period25,27). Along a sea–to–inland gradient, the typical vegetation zonation ranges from embryonic dunes in the seashore, followed by mobile dunes with perennial herbaceous vegetation, fixed dunes covered by evergreen shrub and small sclerophyllous trees and, in the inner sectors, by wooded dunes covered by coniferous forests1,27,28. These dune ecosystems are currently threatened by coastal erosion and anthropogenic activities, such as beach mechanical cleaning and tourism25,29. Moreover, especially in the first half of the 20th century, the area has been prone to urban expansion, and spread of agriculture and afforestation activities25,29, making the dunes particularly squeezed between progressively eroded coastline and anthropic infrastructures placed in the inland (e.g., highway and railway lines). Nonetheless, the coastal dunes of Molise still host many EU Directive 92/43 Habitat types and, for this reason, this area is largely included in the European Natura 2000 system28,29 (Fig. 3). Between the 1950s and the 1990s, these coastal areas underwent considerable transformation30. Yet, following the establishment of protected areas spanning 20 km of coastal dunes two decades ago, the pace of urbanization and agricultural growth has either progressively decreased or come to a complete halt31.

Fig. 3
figure 3

Zoomed section of the study area over the extent of the projection area. The main map depicts an example of the sampling transects surveyed to detect Kentish Plover’s nests, superimposed to the land cover categories used in the modelling phase (labels reported in the legend with an asterisk refer to land cover categories occurring in the study area but not within the extent of the zoomed section). Due to scale limitations, we represented parallel transects as single orange belts (a zoomed detail of the parallel transects is provided in the circular inset). The inset in the upper right corner shows the spatial reference of the study area (i.e., the red line), along with the location of the detail reported in the main map (i.e., the yellow dot). The numbered polygons in green refer to the three Natura 2000 sites insisting over the study area (i.e., 1: Foce Trigno – Marina di Petacciato [IT7228221]; 2: Foce Biferno – Litorale di Campomarino [IT7222216]; 3: Foce Saccione – Bonifica Ramitelli [IT7222217]). Maps Data: Google, ©2024 Airbus.

Full size image

Fieldwork and data collection

In Molise region, a monitoring programme has been conducted over the last twenty years on the development of the Kentish Plover breeding population23,32,33. However, these studies are limited to quantifying the number of nests and surviving nestlings, and they do not focus on reproductive ecology or nesting habitat preferences. To calibrate ENMs, we specifically considered data from the monitoring of the Kentish Plover nests carried out from year 2020 to 2023 along the entire coast of Molise region. This monitoring was organized every year by the “National Committee for the Conservation of the Kentish Plover” (represented by its regional reference “Ambiente Basso Molise”) and was based on the monitoring protocol established by the Italian Institute for Environmental Protection and Research (Istituto Superiore per la Protezione e la Ricerca Ambientale, ISPRA13, which involves volunteers as well as specialists. To develop this monitoring, 18 beach stretches of 1–2 km were identified, in which two parallel transects were defined, one along the coastline and the other along the strip separating the annual vegetation and the dune (Figs. 1 and 3). Both transects were walked by two operators in parallel by 11 a.m., with the help of binoculars to be able to recognize the individuals of the species from a long distance. During the walk, the surrounding area was observed to identify adult individuals and examine the sand to avoid trampling and because sandy dunes often concentrate footprints of the species, debris, and driftwood around the nests22. In the case of observation of adult individuals, their movements were followed for 15–30 minutes to verify the presence of a nest, as the behaviour of incubating parents is different from the non–incubating individuals22. Incubating parents show nervous behaviour or alarm calls, while the behaviour of non–incubating individuals is more relaxed13,22. If a nest was found along the transect, the operator observed the surrounding area to verify the absence of adult individuals and geo–referenced the nest. In case of identification of a nest from a distance, one of the operators fixed the point with the binoculars, while the other operator left the transect and approached the nest to take the coordinates. This approach to the nest was only carried out if the adult was far away. If an individual was detected running away or demonstrating nervous behaviour, the operator moved away, and waited for the individual to return. The monitoring was carried out in the reproductive period (March–July) with periodic visits (following the scheme produced by ISPRA), during which each nest was georeferenced and rendered in shapefile format on QGIS version 3.34.034. We collected a total of 77 Kentish Plover nests, which were subsequently thinned as to factor out any effect of sampling bias, according to Aiello–Lammens et al35.. After thinning data considering a maximum distance between nests of 100 m36, we obtained a final calibration dataset of 35 records.

Variable preparation

Using coeval Google images (i.e., year 2022), we digitized the current coastline and visually interpreted a fine scale land cover map (1:800) within a 100 m buffer area generated around the current coastline. This radius was chosen considering that the sandy beach and dune habitat in the study area adds up to an average width of ca. 50 m (10–150 m, as measured from the Google images). Moreover, such radius represents a reasonable measure to exclude feeding ground from breeding sites, since the former usually occurs at a median distance of ca. 160 m from nesting areas37. The land cover categories were those already established for the coastal dune mosaics in the Central Adriatic coast by Malavasi et al30., i.e., open sand, water bodies, herbaceous vegetation, shrub vegetation, forested areas, artificial surfaces (i.e., any impervious surface including piers and breakwaters), urban green areas, agricultural areas, beach accesses, and coastline (Fig. 3). These mapped categories follow a standardized approach developed for the detailed mapping of coastal dunes25. Most of the mapped natural and seminatural categories can be referred to Habitats of European concern as defined by the Habitat Directive (Table 1).

Table 1 Correspondence between natural and seminatural land cover categories and habitat type according to the Habitat Directive.
Full size table

The legend conforms to the European standard CORINE land cover expanded to a fourth level of detail for natural and semi–natural areas, which is the most accurate level that can be produced through aerial photos of coastal dune ecosystems (see25for details). As for possible tidal variation effects on coastline position, they can be considered negligible, since tidal amplitude along the Molise coast adds up to ca. 6–8 cm38. We then computed Euclidean distances from each of these categories using ‘rgrass’ R package39. Additionally, to map habitat diversity, we computed Shannon’s diversity index on the different land cover categories using a moving window of size 25 m using ‘rasterdiv’ R package40. This window size was chosen according to the average 50 m width of the sandy beach and dune habitat in the study area, as a smaller size would overly detail the landscape around nests, while a larger radius would cause significant margin effect issues. We rendered all resulting raster images at a spatial resolution of 2 m and assured the absence of multicollinearity (VIF ≤ 541;). To account for future coastline change, we relied on the 2040 projection presented in Di Paola et al26.. This forecasting is based on the Digital Shoreline Analysis System (DSAS) with the Kalman filter methodology. In particular, historical shoreline data from 1869 to 2019 were utilized to create a linear regression model, which projected shoreline positions by combining observed and model–derived data, estimating changes every tenth of a year. Additionally, the study incorporated sea–level rise projections from the IPCC sixth report42, specifically the SSP5 8.5 scenario, which anticipates a 0.22 m rise by 2050 due to climate change. This comprehensive approach allows for identifying areas undergoing accretion or at risk of erosion and inundation, taking into account climate change projections. Variables for ENM calibration were prepared for the entire Molise region coast, while predictors for current and future model projections were generated within a projection area corresponding to the southernmost branch of Molise coast (representing ca. 43% of the entire coastline), where 2040 coastline forecasted by Di Paola et al26.. is available. While for current projection variables we simply cut the calibration ones within the boundaries of the projection area, for 2040 variables we proceeded as follows: i) set the 2040 coastline by Di Paola et al26.. as the new coastal border of the projection area; ii) attributed all the coastal accretion areas (i.e., zones currently underwater but predicted as emerged in 2040) to the open sand land cover category; and iii) re–calculated all the variables considered for current time (i.e., Euclidean distance from each land cover category and coastline, and Shannon diversity index) within this new spatial domain26.

Ecological niche models

To calibrate ENMs, we used the so–called “ensemble of small models” approach to avoid model overfitting43, which is a problem when there are few occurrences compared to the number of environmental variables44. Accordingly, we calibrated a set of bivariate models, i.e., considering all possible combinations of predictor pairs and then averaging the results of each model. Bivariate models were trained by using two algorithms: Generalized Boosting Model (GBM) and Maximum Entropy (MAXENT), which were listed among the most performing algorithms within an “ensemble of small models” setup45. To sample background environments, we generated 10,000 random points in the study area46, following the modelling setup proposed as optimal for the “ensemble of small models” framework45. To assess the predictive performance of the models, we applied a bootstrap cross–validation approach by randomly dividing the data into 80%, used to calibrate the models, and the remaining part for evaluation purposes47. The predictive performance of each model was assessed by measuring the area under the receiver operating characteristic curve (AUC48) and the continuous Boyce index (CBI49). We dropped poorly calibrated models (i.e., achieving AUC < 0.7 46) from the subsequent analyses. Ensemble models were obtained by averaging the projections of individual models, weighted by their respective AUC scores50. Ensemble model spatially explicit predictions were generated within the projection area (ca. 176 ha) considering current time and 2040 coastal environment characteristics (as described above)26. Moreover, current and future model projections were converted into presence/absence values selecting three alternative binarization threshold (i.e., ‘equalize sensitivity and specificity’, ‘maximize TSS’ and ‘10th percentile’51). The impact of future coastal dynamics on Kentish Plover nesting habitat suitability was assessed by calculating the range net change metric on binary maps generated for each model52. In particular, the range net change metric was calculated according to the formula (b – c) / (a + c), where a represents the stable habitat area (i.e., predicted suitable in both present and future), b is the gained habitat area (i.e., predicted suitable in future but not present), and crefer to the loss habitat area (i.e., predicted suitable in present but not future)52. Lastly, we calculated the relative variable importance from the ensemble models through permutations, as implemented in the “biomod2” package53. We also generated partial dependence plots to explore the shape of the relationship between Kentish Plover nesting habitat suitability and environmental variables. All these analyses were carried out using the ‘ecospat’ R package54.

Results

ENMs achieved excellent predictive performances (sensu Swets55), with a mean AUC = 0.923. (s.d. = 0.04) and a mean CBI = 0.799 (s.d. = 0.09). Models showed that the key environmental predictors (i.e., variable importance > 10%) in defining Kentish Plover nesting preferences were Euclidean distances from open sand, artificial surfaces, coastline, herbaceous vegetation, and forested areas (Fig. 4).

Fig. 4
figure 4

Barplot in the upper left corner depicts the percentage of relative importance of the environmental variables. All variables are Euclidean distances except Shannon Index. Curves in partial dependence plots report the relation between Kentish Plover nesting habitat and environmental variables. The horizontal axes of all the panels but Shannon index indicate Euclidean distances and are expressed in meters.

Full size image

.

Partial dependence plots evidenced that the suitability of the nesting habitat was highest at 0 m from the open sand, drastically decreasing toward higher distances. Suitability also increased a few meters from artificial surfaces and gradually decreased further on (Fig. 4). As for coastline distance, suitability showed a peak between 25 and 40 m. Regarding vegetation, our results indicated that suitability increased at a short distance from herbaceous vegetation and at a greater distance from forested areas (Fig. 4).

Model projections indicated that current suitable nesting habitat for Kentish Plover occupies 4.4–14.2% of the projection area (ca. 6 ha to 20 ha; depending on the binarization threshold). As for 2040, the suitable nesting habitat will cover between 3.6% and 10.1% of the projection area (ca. 5 ha to 14 ha). Range net change calculation reported values between − 16% to -28% (depending on the binarization threshold), indicating the Kentish Plover will lose approximately 22% on average of its current nesting habitat by 2040 (Fig. 5; see also Figures S1 and S2). This figure represents 0.4 – 2.4% of the projection area (ca. 0.6 ha to 3.5 ha). With specific reference to the processes determining coastal dynamics, ca. 18.5% (ca. 32 ha) of the projection area will undergo coastal accretion in 2040, yet only < 1% (< 2 ha) of these new coastal areas will be gained as suitable nesting habitat for the species (Fig. 6). By contrast, approximately 15.5% (ca. 27 ha) of the projection area will be subject to erosion. Following this erosion process, ca. 1.5% (< 3 ha) of the projection area will no longer include suitable nesting habitat for Kentish plover. Moreover, ENM predicted an increase in suitable nesting habitat of ca. 1% (ca. 2 ha) of the projection area, and a decrease of approximately 2% (ca. 3.5 ha), in those parts of the projection area that will not be involved in either accretion or erosion processes (Fig. 6).

Fig. 5
figure 5

Model projection (EPSG:32633) showing the predicted range modifications under 2040 based on the coastline shift along a stretch of Molise coast. The presented binary map was created considering the ‘maximize TSS” binarization threshold. Maps Data: Google, ©2024 Airbus.

Full size image
Fig. 6
figure 6

Mean percentage of the projection area covered by each type of landscape and Kentish Plover nesting habitat modification generated by overlaying current and future ENM projections according to the coastline changes predicted for 2040. Bar height refers to the mean percentage cover values calculated from the three binary maps (i.e., ‘equalize sensitivity and specificity’, ‘maximize TSS’ and ‘10th percentile’), while error whiskers indicate standard deviation values.

Full size image

Discussion

Data from field monitoring combined with remotely sensed environmental variables proved able to accurately predict Kentish Plover nesting habitat suitability along the Central Adriatic coast, pinpointing to a number of important environmental predictors. Importantly, our results indicated coastal alterations induced by climate change and anthropogenic activities will lead to an average loss of more than 22% in potential nesting habitat for the Kentish Plover in the projection area. The distance from open sand resulted as the most important environmental variable in defining nesting preferences. The area that comprises an extensive mass of open sand spans approximately from the breaker area to the embryonic dunes, coinciding with the upper beach habitat (EU Habitat 1210 Annual vegetation of drift lines). This habitat is generally located in the lower sectors of the dune system, on slightly upwardly concave to flat surfaces and covered by annual vegetation56, all properties that characterize these sectors as open areas. Accordingly, our results are coherent with the evidence that the species prefers open areas with low elevation and scarce vegetation2,57. The preference for open sites probably indicates the benefits of places that allow good long-distance visibility for the detection of approaching predators and guarantee the mobility required for ground feeding adults and precocial chicks2,16,58. In addition, Treccani16 mentioned that cryptic coloration of the Kentish Plover makes it difficult to distinguish it from the substrate where it nests and feeds, thus reducing the possibility of being preyed. This could also explain the high importance of open sand in our study, since the species presents a strong mimesis with this substrate, particularly eggs, as well as juveniles and females.

Our model also highlighted that distance from artificial surfaces have a great importance in defining nesting preferences, with close distances positively affecting Kentish Plover breeding suitability. This result is not in keeping with previous research showing that Kentish Plovers distribution was negatively influenced by built–up areas2,15,16. Noteworthily, the actual adverse effects on biodiversity occurring close to human establishments are highly debated and not univocal4,5. For instance, in the study of Li et al59., it was explained how the survival rate of nestlings was slightly higher in unprotected areas close to anthropic zones than in protected areas, probably due to the reduced presence of predators. This phenomenon may explain why the Kentish Plover nesting suitability was higher around man–made areas, since several cases of predation have been reported elsewhere in the region23,33,60,61. Although our model did not provide evidence of the detrimental effects of artificial surfaces, it must be acknowledged that our analysis did not take into account data regarding the hatching success of nests or chick fledging rates. Therefore, our results may not adequately account for human behaviours that have potentially harmful impacts on the species, such as mechanical beach cleaning, which has been frequently cited as the primary cause of nest and chick loss23,24,33,60,61. The fact that the species reported a high nesting suitability near the artificial surfaces supports what Treccani16 mentioned about the possible coexistence between Kentish Plovers and human activities when appropriate measures are taken. This pattern could be further investigated by e.g., accounting for the intensity of artificial light at night as a metric of human pressure.

The coastline is an area of the beach that offers a high availability of food resources for Kentish Plover, as the species takes advantage of tidal flow and wave motion that expose potential prey above the shoreline16. This can explain the high importance of the distance from the coastline in our study, as shorebird distributions are known to be strongly influenced by the availability and abundance of specific food resources2. However, our results showed that the nesting habitat suitability increases with a distance from the coastline between 25 and 40 m, which might be due to the flooding risk for tides or storms, as other studies have evidenced that is one of the main causes of nest failure4,18,62,63. Coastal erosion has been identified as one of the main causes of the loss of seashore breeding birds64,65. This could be another reason that contributes to the nesting preferences of the Kentish Plover in the Molise region. As a matter of fact, much of the Molise coastline presents a high susceptibility to erosion due to the topographically low–lying dune systems66. A range between 25 m and 45 m would allow the species to be close enough to the coastline for feeding, avoiding at the same time the risk of flooding by storms or coastal erosion.

Prior research has indicated that the Kentish Plover exhibits a preference for nesting in places with sparse vegetation, particularly consisting of herbaceous or low–growing plants2,5,15,57. In keeping with this, we found that the distance from nests to herbaceous vegetation and forested areas is of great relevance in defining the nesting preferences of the species. Specifically, we found that nesting habitat suitability increases within a few metres from herbaceous vegetation and decreases markedly from forested areas. This may be explained by the importance of annual vegetation for nesting, as it is used in some cases for nest construction, offers a higher probability of finding trophic resources, and it is essential to provide protection for nests15. The selection of sparsely vegetated areas by the Kentish Plover, as an anti–predator behaviour, has been reported in several studies2,57,67, as these types of sites allow them to detect approaching predators and provide a flight path to escape from them58.

The projection of nesting habitat along the 2040 coastline evidenced an average loss of > 22% of the current breeding habitat as predicted by our model, likely attributable to climate change and anthropogenic factors65,68. This loss is based on the coastline modification generated by coastal erosion and flooding26, which are the main hazards affecting coastal areas, especially low–lying ones66. Besides harshening the potential impact of these threats, climate change also plays a significant role in increasing coastal vulnerability, which can be influenced by multiple factors, especially topography and sediment dynamics26,66. Accordingly, our finding can be explained by the fact that the Molise coastline presents a high susceptibility to erosion mainly due to the low–lying dune systems, which are fragmented and damaged by wave action and anthropic activities33. Moreover, not only eroded areas affected Kentish Plover nesting habitat suitability. In fact, even though the accretion margins of the projection area resulted larger than the erosion ones, we evidenced an overall detrimental effect on the species nesting habitat, which is thus attributable to the whole coastal and landscape dynamics, instead of deriving from single processes. Human expansion and activities along the coast also contribute to coastal vulnerability, altering natural coastal processes, disrupting ecosystems, and increasing exposure to hazards26. Consistent with this, Malavasi et al30. observed that the composition and structure of the typical Molise coastal dune landscape have been drastically modified by human impacts. Particularly, the most relevant drivers are introduction of alien species, trampling, and mechanical beach cleaning, where the latter is known to usually produce biomass or sand removal69. Moreover, Molise coast is also threatened by the strengthening of the so–called coastal “squeezing” process, which occurs between the erosion front of the shoreline and the fixed human infrastructures (e.g., highway and railway lines) on the land30,56. Therefore, the infrastructure systems from the land would certainly not allow dune retreat even in scenarios of shoreline change, which makes our results even more alarming. From this perspective, many conservation measures should be implemented to protect coastal landscapes and eventually the potential nesting habitat of the Kentish Plover, such as e.g., substituting the mechanical beach cleaning procedure with manual collection, or avoiding trampling. In particular, the sustainable management of annual vegetation of drift lines (EU Habitat 1210) may play a key role in the conservation of the Kentish Plover. Indeed, these habitats could act as “umbrella habitat”, as their protection triggers a sort of “chain protection” from the individual habitat to the entire landscape31. These habitats should be preserved not only for their intrinsic floristic value, but also as providers of optimal habitat for the Kentish Plover. Conservation efforts for these habitats could benefit from using the Kentish Plover as a flagship species, which could drive public interest and awareness towards these often neglected vegetation communities.

Near–future developed scenarios regarding coastal vulnerability highlight that, without specific interventions, large sectors of the Molise coast will most probably undergo significant erosion and local permanent inundation26. This situation is not only affecting our target species but is one of the main factors jeopardizing the distribution and reproduction of many shorebirds70,71. Ivajnšič et al65. suggested that inundation due to sea level rise will result in the loss of habitats that are essential breeding sites for bird species, which will subsequently decrease their breeding success. The awareness that erosion and inundation phenomena could significantly increase in the mid–long term due to sea level rise, should push public administrations and stakeholders towards a strategic approach of integrated coastal zone management1,66. Otherwise, not only a great part of the Kentish Plover habitat will be lost, but also the biodiversity and all the ecological services associated with coastal ecosystems.

Variables for 2040 ENM projection were prepared relying on the assumption that all the land cover categories but coastline and open sand will remain stable within the 16 years period between current time and 2040. This assumption might seem to somewhat undermine the overall plausibility of the outcomes presented in this study. Indeed, the specific context where we set the study, given its peculiar long–term landscape dynamics, allows us to consider this assumption as realistic. In fact, Molise coastline experienced most of its strongest transformations between 1950s and 1990s30. The progressive institution of urban planning constraints (due to the proximity of a highway and a railway) that characterized the second half of this period, along with the establishment of protected areas encompassing more than 20 km of coastal dunes ca. 20 years ago, determined urbanization and agricultural expansion to significantly slow down or cease entirely31. In keeping with that, we recently reported how the percentage cover of new vegetated areas that colonized open sand in the 1986–2022 period adds up to less than 1.5% of the study area (i.e., an increase of ca. 0.04% per year; pers. comm.). Translating this trend in the future (i.e., from 2024 to 2040), we would obtain an overall average increase in new vegetated areas on open sand of ca. 0.64%. Although we acknowledge our approach may have overlooked some minor transformations that could insist on current land cover categories up to 2040, by and large it seems rather reasonable to consider all the land cover categories but coastline and open sand as predominantly stable within the 2024–2040 time period. Arguably, the 100 m large coastal belt we used to delimit our study area may have introduced a flaw by relying on Euclidean distances to describe the spatial pattern of land cover. Specifically, our approach could have overlooked the contribution of certain land cover categories adjacent to the study area, while favouring more distant landscape elements located within it, thus overemphasising the along-beach dimension rather than the cross-beach one. That said, the Kentish Plover is well known to nest almost exclusively within coastal sandy belts, typically moving further inland only for foraging activities, and mainly when lagoons or saltpans are nearby37. Moreover, several stretches of the Molise coastline are bordered inland by highways and railways, which act as barriers and no–habitat areas. In light of that, any potential effect of land cover categories beyond the 100 m threshold can be considered almost negligible. That said, the width of the buffer considered to identify the extent of the modelling area should be set according to the specific characteristics of the dunes under investigation. In areas where dunes are large, it could be possible to consider wider buffers, while for narrow, squeezed dunes (as those analysed here), small buffers should be preferrable.

The modelling approach used in our study showed that the Kentish Plover on the beaches of the Molise region prefers to nest on open sand, 25 to 45 m from the coastline, and at a short distance from herbaceous vegetation. Furthermore, we found that man–made areas do not seem to exert a negative impact on nesting habitat suitability due to the reduced distance where they build their nests. Finally, and most importantly, our results predicted a > 22% average loss of Kentish Plover potential nesting habitat as a consequence of anthropogenic and climate change–induced coastal alterations. Conservation management plans should consider the nesting preferences of the species under study to define more precisely new protected areas or buffer areas. Also, new guidelines for coexistence with humans should be redacted, as the intense tourism activities are still increasing. In addition, the negative trend predicted by our model reinforces the need for action to mitigate the threats of climate change, as not only part of the Kentish Plover’s nesting habitat will be lost, but also the ecosystem services associated with coastal landscapes, which are essential for safety and human well–being.