The effectiveness of species distribution models in predicting local abundance depends on model grain size.

The use of species distribution models (SDMs) to predict local abundance has been often proposed and contested. We tested whether SDMs at different spatiotemporal resolutions may predict the local density of 14 bird species of open/semi-open habitats. SDMs were built at 1 ha and 1 km, and with long-term versus a mix of current and long-term climatic variables. The estimated environmental suitability was used to predict local abundance obtained by means of 275 linear transects. We tested SDM ability to predict abundance for all sampled sites versus occurrence sites, using N-mixture models to account for imperfect detection. Then, we related the R2 of N-mixture models to SDM traits. Fine-grain SDMs appeared generally more robust than large-grain ones. Considering the all-transects models, for all species environmental suitability displayed a positive and highly significant effect at all the four combinations of spatial and temporal grains. When focusing only on occurrence transects, at the 1 km grain only one species showed a significant and positive effect. At the 1 ha grain, 62% of species models showed (over both climatic sets) a significant or nearly significant positive effect of environmental suitability on abundance. Grain was the only factor significantly affecting the model's explanatory power: 1 km grain led to lower amounts of variation explained by models. Our work re-opens the debate about predicting abundance using SDM-derived suitability, emphasizing the importance of grains and of spatiotemporal resolution more in general. The incorporation of local variables into SDMs at fine grains is key to predict local abundance. SDMs worked out at really fine grains, approaching the average size of territory or home range of target species, are needed to predict local abundance effectively. This may result from the fact that each single cell may represent a potential territory/home range, and hence a higher suitability over a given area means that more potential territories occur there.

Molecular investigation on infection by haemosporidians in three Western Palearctic species of swift (Apodidae) and their ectoparasitic louse flies.

Swifts (Apodidae) are an unusual group of birds that spend most of their lives in flight, landing only when breeding. Although this aerial lifestyle greatly reduces their likelihood of being bitten by vectors and infected by vector-born parasites, swifts can still be heavily infested during breeding by nest-based vectors such as louse flies (Hippoboscidae). Here, we investigated host, vector, and vector-borne parasite relationships in the three most widespread swift species in the Western Palearctic (WP): common swifts (Apus apus), pallid swifts (A. pallidus), and alpine swifts (Tachymarptis melba), their nest-based louse flies (Crataerina pallida and C. melbae) and avian haemosporidians (genera Haemoproteus, Plasmodium, and Leucocytozoon). Studies of haemosporidian infections in Apodidae remain limited, with clear evidence of infection found to date in just four Neotropical and one Australasian species. The possible role of louse flies in transmitting haemosporidian infections has never been tested in swifts. We assessed the occurrence of haemosporidian infection by PCR screenings of DNA from blood samples from 34 common swifts and 44 pallid swifts from Italy, and 45 alpine swifts from Switzerland. We also screened 20 ectoparasitic louse flies present on 20 birds and identified them by both morphological features and cytochrome oxidase subunit 1 (COI) barcodes. Our results provide no evidence of haemosporidian infection in the 123 swifts tested or in the two louse fly species we identified. Our findings are consistent with available knowledge showing no haemosporidian occurrence in WP swift species and that the most likely infection route for these highly aerial species (via louse fly ectoparasites during nesting) is unlikely.

The hidden cost of following currents: Microplastic ingestion in a planktivorous seabird.

Microplastics are increasingly pervasive pollutants, particularly abundant in the neuston where they drift with currents. We assessed dietary microplastic ingestion in the Mediterranean storm petrel (Hydrobates pelagicus melitensis), a small pelagic seabird that forages on plankton and inhabit the Mediterranean sea, one of the most polluted seas worldwide. We collected spontaneous regurgitates from 30 chick-rearing individuals and used GPS tracking data from 7 additional individuals to locate foraging areas. Birds foraged in pelagic areas characterized by water stirring and mixing, and regurgitates from 14 individuals (i.e. 45 %) contained microplastics. Fibers were the dominant shape (56 %), with polyester, polyethylene and nylon being the most frequent polymers. Our findings highlight the potential sensitivity of this species of conservation interest to plastic pollution and suggest that storm petrel regurgitates can be a valuable matrix to investigate microplastic ingestion in planktonic foragers, providing a characterization of spatio-temporal patterns of microplastic exposure in pelagic environments.

Italian odonates in the Pandora's box: A comprehensive DNA barcoding inventory shows taxonomic warnings at the Holarctic scale.

The Odonata are considered among the most endangered freshwater faunal taxa. Their DNA-based monitoring relies on validated reference data sets that are often lacking or do not cover important biogeographical centres of diversification. This study presents the results of a DNA barcoding campaign on Odonata, based on the standard 658-bp 5' end region of the mitochondrial COI gene, involving the collection of 812 specimens (409 of which barcoded) from peninsular Italy and its main islands (328 localities), belonging to all the 88 species (31 Zygoptera and 57 Anisoptera) known from the country. Additional BOLD and GenBank data from Holarctic samples expanded the data set to 1,294 DNA barcodes. A multi-approach species delimitation analysis involving two distance (OT and ABGD) and four tree-based (PTP, MPTP, GMYC and bGMYC) methods was used to explore these data. Of the 88 investigated morphospecies, 75 (85%) unequivocally corresponded to distinct molecular operational units, whereas the remaining ones were classified as 'warnings' (i.e. showing a mismatch between morphospecies assignment and DNA-based species delimitation). These results are in contrast with other DNA barcoding studies on Odonata showing up to 95% of identification success. The species causing warnings were grouped into three categories depending on if they showed low, high or mixed genetic divergence patterns. The analysis of haplotype networks revealed unexpected intraspecific complexity at the Italian, Palearctic and Holarctic scale, possibly indicating the occurrence of cryptic species. Overall, this study provides new insights into the taxonomy of odonates and a valuable basis for future DNA and eDNA-based monitoring studies.

Habitat, climate, topography and management differently affect occurrence in declining avian species: Implications for conservation in changing environments.

Climate and land-use change are the most severe threats to biodiversity; their effects are often intermingled, also with those of landscape/habitat management. Birds of mountain grassland are declining throughout Europe. Disentangling climate effects from those of land-cover and management on their occurrence is essential to identify distribution drivers, potential impacts of climate/land-use changes, and effective conservation strategies. We investigated the occurrence of water pipit (elevation specialist), skylark and red-backed shrike (elevation generalists) in Central Apennines, Italy (750-2130 m asl), using point counts. Topographic/climatic, land-cover and management fine-scale variables were considered as potential occurrence predictors in Generalized Linear Models. For all species, combining different types of predictors led to the most accurate models, but the relative importance of single-groups varied: land cover was the most important for skylark, climate/topography for water pipit, all three groups had similar support for red-backed shrike. Skylark was positively affected by solar radiation and grassland cover, and negatively by bare ground, hedgerows, rocks, shrubland, ski-pistes and buildings, confirming sensitivity to anthropic alteration of semi-natural grassland. Water pipit was favoured by grazing and negatively impacted by shrubland and average temperature (most important predictor). Red-backed shrike was affected negatively by broadleaved forest and grazing occurrence, quadratically by isolated shrubs and positively by grassland cover. Climate was a fundamental determinant of water pipit occurrence, but not for the other species. Land-cover was important for all species and also management factors were invariably included in models. Climate, habitat and management factors differently contributed to occurrence patterns in these declining species. Conservation strategies need to embrace landscape planning to preserve grassland extents/mosaics, identify climate refugia for water pipit and implement dedicated management (preventing new ski-pistes over areas suitable for birds and carefully planning grazing). It should be feasible to combine local, sustainable economies with biodiversity conservation into landscape planning for Central Apennines.

Species interactions and climate change: How the disruption of species co-occurrence will impact on an avian forest guild.

Interspecific interactions are crucial in determining species occurrence and community assembly. Understanding these interactions is thus essential for correctly predicting species' responses to climate change. We focussed on an avian forest guild of four hole-nesting species with differing sensitivities to climate that show a range of well-understood reciprocal interactions, including facilitation, competition and predation. We modelled the potential distributions of black woodpecker and boreal, tawny and Ural owl, and tested whether the spatial patterns of the more widespread species (excluding Ural owl) were shaped by interspecific interactions. We then modelled the potential future distributions of all four species, evaluating how the predicted changes will alter the overlap between the species' ranges, and hence the spatial outcomes of interactions. Forest cover/type and climate were important determinants of habitat suitability for all species. Field data analysed with N-mixture models revealed effects of interspecific interactions on current species abundance, especially in boreal owl (positive effects of black woodpecker, negative effects of tawny owl). Climate change will impact the assemblage both at species and guild levels, as the potential area of range overlap, relevant for species interactions, will change in both proportion and extent in the future. Boreal owl, the most climate-sensitive species in the guild, will retreat, and the range overlap with its main predator, tawny owl, will increase in the remaining suitable area: climate change will thus impact on boreal owl both directly and indirectly. Climate change will cause the geographical alteration or disruption of species interaction networks, with different consequences for the species belonging to the guild and a likely spatial increase of competition and/or intraguild predation. Our work shows significant interactions and important potential changes in the overlap of areas suitable for the interacting species, which reinforce the importance of including relevant biotic interactions in predictive climate change models for increasing forecast accuracy.

Combining habitat requirements of endemic bird species and other ecosystem services may synergistically enhance conservation efforts.

Biodiversity conservation and the optimisation of other ecosystem service delivery as a contribution to human well-being are often tackled as mutually alternative targets. Modern agriculture is a great challenge for the fulfilment of both. Here, we explore the potential benefits of integrating biodiversity conservation and the preservation of wider ecosystem services, considering the conservation of an endemic species (Moltoni's warbler Sylvia subalpina; Aves: Sylvidae) and soil erosion control (a final ecosystem service) in intensive vineyards in Italy. We modelled factors affecting warbler occurrence and abundance at 71 study plots by means of N-mixture models, and estimated soil erosion at the same plots by means of the Universal Soil Loss Equation. Shrub cover had positive effects on both warbler abundance and soil retention, whereas higher slopes promote warbler abundance as well as soil erosion. Creating shrub patches over sloping sites would be at the same time particularly suited for warblers and for soil retention. We simulated three alternative conservation strategies: exclusive focus on warbler conservation (1), exclusive focus on soil preservation (2), integration of the two targets (3). Strategies assumed the creation of 1.5-ha shrub patches over 5% of the total area covered by plots and targeted either at wildlife or soil conservation. The exclusive strategies would allow an increase of 105 individuals and the preservation of 783 tons ha-1year-1, respectively. Each individual strategy would ensure benefits for the other target corresponding to 61-64% of the above totals. The integrated strategy would allow for the achievement of 91-93% of the benefits (96 warblers and 729 tons ha-1year-1) of the individual strategies. The integration of the two approaches could provide important synergies, allowing to broaden the effects of conservation strategies, such as agri-environmental schemes that could be drawn from our results (and which are particularly urgent for intensive permanent crops).