Natural History Sciences. Atti Soc. it. Sci. nat. Museo civ. Stor. nat. Milano, 10 (1): 7-12, 2023 DOI: 10.4081/nhs.2023.605 A preliminary camera trapping study of mammals of Monti Lepini (Central Italy) Claudio Angelini1,2*, Luigi Corsetti1,3, Luca Faustinella1,4, Reuccio Martellato1,5, Giovanni Mastrobuoni1,6, Daniela Mastrodomenico1,7, Luigi Marozza1, Stefano Petra1,8, Venusta Pietrocini1, Massimiliano Proietti1 1 Rete faunistica dei Monti Lepini. 2 Via G. Marconi 30, 04018 Sezze, Italia. 3 Via Monte Rosa 34, 04100 Latina, Italia. 4 Via Colli II tratto, 04018 Sezze, Italia. 5 Via Ezio 80, 04100 Latina, Italia. 6 Via Migliara 45 dx 34, 04014 Pontinia, Italia. 7 Strada delle Pignette 2860, 04014 Pontinia, Italia. 8 Via degli Aurunci 61, 04100 Latina, Italia. * Corresponding author: oppela@gmail.com © 2023 Claudio Angelini, Luigi Corsetti, Luca Faustinella, Reuccio Martellato, Giovanni Mastrobuoni, Daniela Mastrodomenico, Luigi Marozza, Stefano Petra, Venusta Pietrocini, Massimiliano Proietti Received for publication: 27 December 2021 Accepted for publication: 14 September 2022 Online publication: 4 May 2023 Abstract - We used 52 camera traps to survey mammal species in the Lepini Mountains (Lazio, Central Italy) during three trapping sessions. Thirteen mammal species have been recorded. Compared to data based on previous opportunistic observations (2000-2018), camera trapping has quickly increased knowledge about the distribu- tion of species, especially for elusive ones. We modelled species and average community occupancy also considering habitat covariates in a hierarchical modelling framework. Occupancy at community level was positively impacted by woods and negatively impacted by open habi- tats, indicating that the ongoing reforestation of the area may favour mammal diversity. Open surfaces have a negative effect on the presence of Sciurus vulgaris, a species for which our study confirms extensive recolonization of the area. We have also shown that woods increase the probability of occupation by Canis lupus and Felis silvestris, con- firming that these species could be negatively impacted by increased deforestation. Keywords: central Italy, habitat covariates, Lepini Mountains, mammals, occupancy. Riassunto - Studio preliminare mediante fototrappolaggio dei mammiferi dei Monti Lepini (Lazio, Italia centrale). Per questo studio sui mammiferi dei Monti Lepini abbiamo utiliz- zato 52 postazioni di fototrappolaggio per tre sessioni di (video)cat- tura. È stata registrata la presenza di 13 specie. In confronto con i dati ottenuti da osservazioni opportunistiche raccolte tra il 2000 e il 2018, il fototrappolaggio ha velocemente incrementato le conoscenze sulla distribuzione delle specie, specialmente quelle elusive. La presenza (occupancy) delle singole specie e quella media a livello di comunità è stata modellizzata in un contesto gerarchico (hierarchical modelling) anche considerando le covariate ambientali. L’occupancy a livello di comunità è risultata influenzata positivamente dalla presenza di boschi e negativamente dagli ambienti aperti, facendo ipotizzare che la cor- rente riforestazione dell’area di indagine possa favorire la diversità teriologica. Gli ambienti aperti paiono influenzare negativamente anche la presenza di Sciurus vulgaris, specie per la quale il nostro studio con- ferma l’estesa ricolonizzazione dell’area. Mostriamo anche che i boschi aumentano le probabilità di presenza di Canis lupus e Felis silvestris, confermando che la deforestazione potrebbe evere un effetto negativo su queste specie. Parole chiave: covariate ambientali, Italia centrale, mammiferi, Monti Lepini, occupancy. INTRODUCTION Camera trapping is a powerful tool for wildlife re- search, especially for medium and large mammals (Little- wood et al., 2021). It offers a favourable trade-off between research efforts and outcomes, and it allows to gather from basic species presence data to more particular and richer in biological information (Rovero et al., 2013): the kind of outcomes largely depends on the choice of study de- signs and on the analyses applied to the data. In this note, we report the results from a preliminary camera trapping study carried out in the Lepini Mountains (Lazio, Central Italy) focused on mammals. Knowledge on the mammals of this representative area, although not poor, is mostly based on opportunistic and temporally sparse observa- tions (cfr. Amori et al., 2002; Mastrobuoni & Capizzi, 2020). Our study is based on a spatial and temporal sam- pling design and has the twofold purpose of increasing the knowledge on mammal species distribution available for a local atlas (Corsetti & Marozza, 2020) and assessing their occupancy pattern considering habitat preference. MATERIALS AND METHODS The Lepini Mountains is a mountain range about 50 km South-East of Rome, occupying an area of about 880 km2 and ranging from a few meters to an elevation of 1536 m. It is a carbonate ridge, with both underground and surface karst features; the deep Carpineto-Montelanico tectonic valley identifies an eastern and a western chain (Cosen- tino et al., 1993). In the western chain, climate is miti- gated by the proximity of the Tyrrhenian Sea, whereas the 8 ANGELINI, CORSETTI, FAUSTINELLA, MARTELLATO, MASTROBUONI, MASTRODOMENICO, MAROZZA, PETRA, PIETROCINI, PROIETTI eastern one has a more continental climate (Tomaselli et al., 1973). The Carpineto-Montelanico valley, along with many secondary valleys, creates local climates that deter- mine different plant associations: garrigue, maquis shrub- land, grassland, holm oak wood, heathland, broad-leaved forest with Acer spp., Carpinus spp., Fraxinus spp., and beech Fagus sylvatica L. subsp, sylvatica (Copiz et al., 2018). Due to the limestone nature of the area, there are small scattered (though sometimes numerous) water bod- ies and a few seasonal streams. One Special Protection Area and nine Special Areas of Conservation (mostly in- cluded in the SPA) protect about 470 km2 of the whole Lepini Mountains area. Camera-trapping data were collected between Febru- ary 2019 and March 2020. The area was divided into 52 squares of 12.25 km2 (side 3.5 km), and the cameras were placed as close as possible to the centre of each square, unless the centre was in an urban area. The final aver- age distance between each camera and the adjacent ones was 4.2±0.1 SE km (1.4 - 7.4 km). The cameras placed on the perimeter, delineated a polygonal area of 454 km2. Camera traps were utilized in two sequential arrays, each containing 26 traps, across three trapping sessions (Tab. 1). The traps were arranged in a chequerboard pattern to avoid any spatial biases (Fig. 1). Each set of 26 camera traps operated for a period of 21 consecutive days, followed immediately by another ar- ray. In total, the camera traps were active for a duration of 3,276 days, with intervals of 77 and 182 days between the last and first days of the two consecutive sessions, re- spectively. The short time interval between the placement of the two consecutive groups of traps during the same session al- lowed us to consider them as having worked simultaneously. We used camera traps equipped with PIR and invisible IR flash (Apeman H68 [n = 16]), Apeman H45 [n = 2], Acorn 5310A, Boskon Guard BG529, Crenova rd1000, HC-800A Trail Camera, Trail Camera 3G 3.0CG HD, Suntek HC 700G, Toguard H45, Victure HC400; nominal trigger times range from 0.3 to 0.8 sec; cameras were nev- er used for the same location across different sessions), set for recording 20’’ videos during the whole day. The few videos that did not allow discrimination between Canis lupus / C. l. familiaris, Felis silvestris / F. catus and Mar- tes foina / M. martes were discarded. Information obtained from camera traps was used in two ways in the analyses. First, we compared the occur- rence data from the videos with data based on opportun- istic observations collected for the faunal atlas edited by Corsetti & Marozza (2020) from 2000 to 2018. In the atlas, data were plotted on a grid consisting of 44 squares of 5×5 km, of which 27 were occupied by 1 to 4 camera traps (1.9 ±0.2 SE). Below, we refer only to these 27 squares and evaluate where the presence of each species has been con- firmed or recorded for the first time by camera trapping. Second, we analysed the data in a hierarchical model- ling framework to take into account species detection prob- ability (p) for estimating species occupancy (ψ) (MacKen- zie et al., 2002). In doing so, we summarized the data from two consecutive three-weeks arrays as a unique sampling occasion, thus obtaining an amount of three sampling oc- casions. We used a single-species/single season occupancy model for each species detected by camera traps (except for those with too sparse data). Further, in order to evaluate whether the average occupancy at community level was affected by the environment (i.e. the buffer area around the camera traps: see below), we used a latent factor multi- species occupancy model (Doser et al., 2022a). We also evaluated the effect of habitat on occupancy, by using as covariates the CORINE land cover 2018 (100 m resolution, CLC in the following) occurring in a cir- cular buffer area of about 1 km2 (565 m radius) centred on the camera trap placement point. We used four covari- ates: artificial surfaces (CLC class 1), agricultural areas (CLC class 2), forest and maquis (CLC classes 3.1 and Tab. 1 - Operational periods of the camera traps during the two consecutive arrays of each session. / Periodi di funzionamento delle fototrappole durante le due disposizioni consecutive (array) di ogni sessione. Session Array Start Finish 1 A 24 February 2019 16 March 2019 B 17 March 2019 7 April 2019 2 A 23 June 2019 13 July 2019 B 14 July 2019 4 August 2019 3 A 2 February 2020 22 February 2020 B 23 February 2020 15 March 2020 Fig. 1 - Geographical location of the Lepini Mountains area and positions of the camera traps (the different colours group the two arrays of cameras). / Collocazione geografica dei Monti Lepini e schema di posizionamento delle fototrappole (i colori differenti sono per le due disposizioni consecutive di ogni sessione). 9CAMERA TRAPPING IN THE LEPINI MOUNTAINS 3.2.3.1; wood in the following), open space and shrub or herbaceous vegetation (CLC classes 3.3 and 3.2 – except 3.2.3.1; open space in the following) (no other CLC class- es occurred in the buffer areas). These classes occurred in the 52 buffer areas and in the Lepini Mountains as a whole (in brackets) with the following total percentages: artificial surface 2.2% (2.5%), agricultural areas 19.5% (41.2%), woods 56% (37.5%), open areas 22.3% (18.8%). Each kind of CLC class was used separately as a covariate on occupancy. This procedure has been carried out using QGIS ver. 3.16.7 (< www.qgis.org >). We modelled species occupancy as constant or impact- ed by covariates, and detection probabilities as constant or time dependent, thus obtaining ten candidate models for each species. Program PRESENCE 13.10 (Hines, 2006) has been used for running all models, and the Akaike In- formation Criterion (AIC) has been used for model selec- tion. The goodness of fit of the most global model was evaluated by the variance inflation factor (ĉ) assessed us- ing a bootstrap method (10000 replicates); when ĉ > 1 data were considered over-dispersed, and its value was used for obtaining the quasi-likelihood AIC (QAIC), that was then used for model selection (MacKenzie & Bailey, 2004). Models with ΔAIC (or ΔQAIC) > 2 were considered as having little or no support; however, we always used mod- el averaging for the estimation of detection probability and occupancy. Occupancy at community level was analysed by mean of multi-species occupancy model (MSOM) with species correlations, using the function lfMsPGocc (with two latent factors and default values) implemented in the R package spOccupancy (Doser et al., 2022a, 2022b; R Core Team, 2022). Comparison among the ten candidate MSOM models was based on the Widely Available Infor- mation Criterion (WAIC). Hierarchical modelling for esti- mation of detection probability and thus of occupancy is nowadays widely implemented in mammal studies based on camera trapping (e.g. O’Connell et al., 2011). RESULTS During the study period, camera traps recorded 13 species (Tab. 2). Also, Apodemus spp. were recorded, but it was not possible to distinguish between A. flavi- collis and A. sylvaticus, the two species occurring in the area (Amori et al., 2002; Capizzi et al., 2012; Corsetti & Marozza 2020). Videos of unidentifiable micromammals were also recorded. Six domesticated mammal species were recorded too. Compared with the data already col- lected before our camera trap study, the camera trap study resulted in new spatial records for 11 wild species, with an average of 5.1±1 squares per species, and it confirms the occurrence in 7.9±2.3 squares per species. Basing on model fitting, most species substantially showed constant detection probabilities among capture sessions, but M. meles, S. scrofa and S. vulgaris had vari- able detection probabilities as a function of time (Tabs. 3, 4). Capture probabilities ranged from 0.12 to 0.77. Occu- pancy probability ranged from 0.1 (L. europaeus) to 0.99 (V. vulpes). The species with the main differences between naïve and estimated occupancy are C. lupus and F. silves- tris. Occupancy of V. vulpes, M. meles, S. scrofa and L. eu- ropaeus were only marginally affected by environmental covariates; woods positively affected the occupancy of C. lupus, F. silvestris and S. vulgaris; open spaces negatively affected the occupancy of C. lupus and S. vulgaris; agricul- tural surfaces positively impacted the occupancy of M. foina and R. rattus; artificial surfaces positively affected the occu- pancy of H. cristata, and negatively that of C. lupus (Tab. 4). Tab. 2 - Comparison among species’ distribution collected from 2000 to 2018 in 27 5x5-km-squares basing on opportunistic searches (Corsetti & Marozza, 2020) and distribution data from camera trapping (this study) in the Lepini Mountains. (increases due to camera trapping with reference to 2000-2018 are reported as percentages in the last column). / Confronto fra i dati sulle distribuzioni delle specie raccolti in 27 maglie di 5x5 km in modo opportunistico tra il 2000 e il 2018 (Corsetti & Marozza, 2020) e i dati raccolti tramite questo studio basato sul fototrappolaggio nei Monti Lepini. Nell’ultima colonna sono riportati gli incrementi percentuali in rapporto ai dati del 2000-2018). Species 2000-2018 squares Squares confirmed by camera trapping Squares added by camera trapping Erinaceus europaeus Linnaeus, 1758 25 6 (24%) 0 Canis lupus Linnaeus, 1758 8 0 5 (62.5%) Vulpes vulpes (Linnaeus, 1758) 20 19 (95%) 7 (35%) Felis silvestris Schreber, 1777 5 0 7 (140%) Martes foina (Erxleben, 1777) 15 13 (86.7%) 12 (80%) Martes martes (Linnaeus, 1758) 1 0 2 (200%) Meles meles (Linnaeus, 1758) 18 15 (83.3%) 9 (50%) Sus scrofa Linnaeus, 1758 26 23 (88.5%) 0 Myodes glareolus (Schreber, 1780) 7 0 3 (42.9)% Hystrix cristata Linnaeus, 1758 18 15 (83.3%) 9 (50%) Rattus rattus (Linnaeus, 1758) 11 3 (27.3%) 6 (54.6%) Sciurus vulgaris Linnaeus, 1758 21 9 (42.3%) 3 (14.3%) Lepus europaeus Pallas, 1778 19 0 4 (21%) 10 Tab. 3 - Naïve occupancy and estimation of detection probability (p) and occupancy (ψ) (with Standard Error, SE) obtained by model averaging from ten candidate models for each species of mammals detected by camera trapping in the Lepini Mountains. / Naïve occupancy e stima della probabilità di rilevamento (p) e dell’occupancy (ψ) (con i relativi errori standard, SE) ottenute tramite model averaging da 10 modelli differenti per ciascuna specie rilevata tramite fototrappolaggio nei Mont Lepini. Species Naïve occupancy p (SE) ψ (SE) Erinaceus europaeus 0.11 too sparse data Canis lupus 0.11 0.12 (0.12) - 0.15 (0.14) 0.33 (0.08) Vulpes vulpes 0.96 0.76 (0.04) - 0.77 (0.04) 0.99 (0.00) Felis silvestris 0.21 0.21 (0.11) - 0.29 (0.14) 0.38 (0.22) Martes foina 0.79 0.56 (0.07) - 0.61 (0.07) 0.88 (0.04) Martes martes 0.02 too sparse data Meles meles 0.75 0.53 (0.1) - 0.69 (0.09) 0.8 (0.02) Sus scrofa 0.85 0.58 (0.1) - 0.72 (0.08) 0.88 (0.01) Myodes glareolus 0.08 too sparse data Hystrix cristata 0.75 0.64 (0.07) - 0.69 (0.07) 0.79 (0.04) Rattus rattus 0.23 0.41 (0.12) - 0.43 (0.12) 0.29 (0.19) Sciurus vulgaris 0.35 0.35 (0.11) - 0.60 (0.15) 0.41 (0.05) Lepus europaeus 0.08 0.43 (0.21) - 0.45 (0.22) 0.10 (0.01) Model selection for MSOM (Tab. 5) showed that the agricultural surfaces positively affected the occupancy at community level, and also woods had a secondary posi- tive effect on average occupancy, while open habitats only marginally affected it negatively. DISCUSSION Camera trapping allowed for the detection of 13 spe- cies, representing 43.3% of wild mammal species occur- ring in the Lepini Mountains (30 species, with the excep- tion of bats) (Mastrobuoni & Capizzi, 2020). Out of the ten micromammal species detected in the previous study, only Myodes glareolus was positively identified. However, other micromammals were observed in the video record- ings, but unfortunately, we were unable to identify them. The previous study has reported the presence of fossorial Talpa caeca and T. romana, as well as the arboreal Glis glis, which could not be captured in the video traps due to the positioning of the devices just above ground level. Among the species known to occur in the Lepini Mountains, only Mustela nivalis, M. putorius, Rattus norvegicus and L. cor- sicanus were actually not recorded. M. putorius and L. cor- sicanus are relatively rare and localised species in the study area, as reported by Mastrobuoni (2020a) and Mastrodo- menico (2020). On the other hand, R. norvegicus is main- ly found in anthropised habitats (Mastrobuoni, 2020b), which we excluded from our camera trapping efforts. M. nivalis, as other small and fast-moving mammals, is not likely to be detected by camera traps (Mos & Hof- meester, 2020) and we think this is a possible explanation for the absence of this species from the footage, despite its apparent abundance in the area (Amori et al., 2002; Mas- trobuoni, 2020c). We advance a similar explanation also for the small, but slow E. europaeus, which was filmed only a few times, compared to its apparent abundance in the area (Pietrocini, 2020) (Tab. 1). With respect to the data collected in the period 2000- 2018, our lower duration camera trap study confirmed the presence of species in 42.6% of the area on average and recorded species for the first time in 57.7% of the area on average (Tab. 2). Furthermore, camera trapping was particularly effective for elusive species, with a substan- tial increase of the previously assessed local ranges of C. lupus (in more than 62.5% of the squares), F. silvestris (140%) and M. martes (200%). These results are impres- sive, considering that finding species in new parts of the area was not our primary target, since the camera trapping protocol we used was tailored for an occupancy study and was limited by logistic and economic resources. Such out- come supports the effectiveness of camera trapping and suggests that in relatively small areas (such as the Le- pini Mountains), if adequate resources are available, in relatively short time, camera trapping could improve the knowledge of the local ranges of several species. In con- trast, it should be considered that very common species could escape detection, as mentioned above. In general, we found that agricultural areas and woods positively impacted the occupancy at community level, while open habitats negatively impacted it. This result is interesting in perspective, since the area is undergoing reforestation (Buccomino, 2005). Concerning the single species, the presence of S. scrofa and V. vulpes was not af- fected by any environmental covariate, consistently with their wide ecological adaptability (Apollonio, 2003; Boi- tani & Ciucci, 2003), and confirmed also by the high oc- cupancy probability. Similarly, the presence of M. meles was seemingly not affected by the habitat, contrary to the expected preference for woods with low or any human disturbance (De Marinis & Genovesi, 2003). A possible explanation may be the limestone substrate of the area, of- ten with a thin or absent soil layer, which limits the exten- sion of setts, something that forces badgers to move fre- quently among small setts (Brøseth et al., 1997). Future research could address if the high occupancy rate depends on the density of this species. Interestingly, the occupancy of two opportunistic species such as M. foina and espe- cially R. rattus is favoured by agricultural surfaces. ANGELINI, CORSETTI, FAUSTINELLA, MARTELLATO, MASTROBUONI, MASTRODOMENICO, MAROZZA, PETRA, PIETROCINI, PROIETTI 11 Open habitats negatively impact on the presence of S. vulgaris more than wooded areas impact on this spe- cies positively, consistently with its sensitivity to forest fragmentation (Koprowoski, 2005). The species is re- colonizing its former range and expanding its range in Central Italy (Battisti et al., 2013). Apparently, the spe- cies went extinct in the Lepini Mountains since the early 1950s (Amori et al., 2002), but isolated records have been reported in the late 1970s (various personal communica- tions to CA), in the 1980s (Amori et al., 2002) and in the 1990s (Esposito, 2013). Then, since 2008 (Esposito, 2013) an increasing number of records have been report- ed. In the last decades grazing activities decreased in the study area (pers. obs.), with consequent natural reforesta- tion and reduction of forest fragmentation (Buccomino, 2005). This could explain the recent re-colonization of the Lepini Mountains area by this fragmentation-sensitive species, not unlikely starting from local residual very small population. Conversely, logging has dramatically increased in the Lepini Mountains in the last decade. This likely rep- resents a direct threat for the elusive C. lupus, positively associated with woods and negatively impacted by ar- tificial areas (Ciucci & Boitani, 2003), and F. silves- tris, which is strictly associated with woods (Genovesi, 2003). This is suggesting an environmental manage- ment outcome from our study. A second management Tab. 4 - Single-species single season occupancy model selection summary table of mammals detected by camera trapping in the Lepini Mountains. Reported values are the ΔAIC or ΔQAIC and relative model weight (second line, in italics) of each model for each species. Codes for environmental covariates for occupancy: “ag”: agricultural land, “ar”: artificial surface, “o”: open habitats, “w”: woods; “+” or “-” indicate if the covariate impacts positively or negatively the occupancy. Models having Δ(Q)AIC > 3 are not reported. / Sommario dei modelli di occupancy per le singole specie di mammiferri rilevate tramite fototrappolaggio nei Monti Lepini. Sono riportati i valori di ΔAIC o ΔQAIC e il relativo peso del modello (seconda linea, in corsivo) per ogni modello per ciascuna specie. Abbreviazioni per le covariate ambientali applicate all’occupancy: “ag”: aree agricole, “ar”: superfici antropiche, “o”: ambienti aperti, “w”: boschi, “+” o “-” indicano se la covariata ha un effetto positivo o negativo sull’occupancy. I valori dei modelli con Δ(Q)AIC > 3 non sono riportati. Model Species ψ(.) p(.) ψ(ag) p(.) ψ(ar) p(.) ψ(o) p(.) ψ(w) p(.) ψ(.) p(t) ψ(ag) p(t) ψ(ar) p(t) ψ(o) p(t) ψ(w) p(t) C. lupus ΔAIC 0.6 0.18 2.38 - 0.07 1.02 - 0.14 1.11 - 0.14 0 + 0.24 3 0.05 2.43 + 0.07 V. vulpes ΔQAIC 1.56 0 0.31 1.24 + 0.17 2 + 0.11 2 - 0.11 2 - 0.11 F. silvestris ΔAIC 0 + 0.53 1.32 + 0.27 M. foina ΔQAIC 1.03 0.38 0.17 0 + 0.20 0.75 + 0.14 0.84 - 0.13 2.26 + 0.07 2.13 0.07 1.79 + 0.08 2.43 + 0.06 2.56 - 0.06 M. meles ΔQAIC 1.24 0.96 0.14 2.87 + 0.05 2.74 + 0.06 2.15 - 0.08 2.75 + 0.06 0 0.22 1.89 + 0.09 1.77 + 0.09 1.19 - 0.12 1.83 + 0.09 S. scrofa ΔQAIC 1.43 0.69 0.16 2.62 - 0.06 2.45 + 0.07 2.5 - 0.06 2.49 + 0.06 0 0.22 1.95 - 0.08 1.73 + 0.09 1.77 - 0.09 1.8 + 0.09 H. cristata ΔAIC 0.99 0.15 1.1 + 0.14 0 + 0.24 2.09 - 0.09 2.85 - 0.06 2.47 0.07 2.69 + 0.06 1.42 + 0.12 R. rattus ΔQAIC 1.88 0 + 0.64 S. vulgaris ΔAIC 0.16 0.25 2.01 + 0.10 2.07 - 0.10 0 - 0.27 0.96 + 0.17 L. europaeus ΔQAIC 2.05 0 0.33 2 + 0.12 1.5 - 0.15 1.74 - 0.14 1.75 + 0.14 Tab. 5 - Model selection for the multi-species occupancy model from camera trapping of mammals in the Lepini Mountains at community level (average occupancy), based on the Widely Applicable Information Criterion (WAIC). Codes for environmental covariates for abundance: “ag”: ag- ricultural areas, “ar”: artificial surface, “o”: open areas, “w”: woods; “+” or “-” indicate if the covariate impacts positively or negatively the occupancy. / Selezione dei modelli multi- species occupancy dai dati del fototrappolaggio nei Monti Lepini al livello di comunità (occupancy media), basata sul Widely Applicable Information Criterion (WAIC). Abbrevi- azioni per le covariate ambientali applicate alla numerosità delle specie: “ag”: aree agricole, “ar”: superfici antropiche, “o”: ambienti aperti, “w”: boschi, “+” o “-” indicano se la covariata ha un effetto positivo o negativo sull’occupancy. MSOM model ΔWAIC Covariate ψ(ag) p(t) 0 + ψ(w) p(t) 2.04 + ψ(ag) p(.) 4.44 + ψ(.) p(t) 4.61 ψ(o) p(t) 4.94 - ψ(w) p(.) 6.03 + ψ(o) p(.) 6.4 - ψ(an) p(t) 6.88 + ψ(an) p(.) 9.41 + ψ(.) p(.) 9.57 CAMERA TRAPPING IN THE LEPINI MOUNTAINS 12 outcome results from the general avoidance of open areas by the species detected by camera traps, as con- firmed also at community level. This raises a question over the appropriateness of the actions planned for the Special Protection Areas (“Monti Lepini” SPA covers 469.25 km2, more than half of the territory), including the artificial maintenance of open areas such as pas- tures even in the absence of current grazing (Agenzia Regionale Parchi, 2012). Such actions are specifically tailored for bird conservation, but while conservation of small pastures could increase overall habitat diversity, it should be considered that artificial maintenance of large pastures could not foster non-avian species, such as C. lupus, F. silvestris and S. vulgaris. A further and longer camera trap study, which will consider habitat manage- ment (including logging), as well as other sources of environmental change, could likely help in understand- ing if and how these factors impact on the presence of mammals. Acknowledgments This study has been partly funded by Regione Lazio and Compagnia dei Lepini in the framework of the “Pro- getto speciale di sviluppo locale dei Monti Lepini”. Giulia Malizia and Giorgio Chiozzi kindly reviewed the English language and style of the manuscript. Two anonymous reviewers stimulated a reflection on the manuscript, and thus its improvement from the first to the third draft. REFERENCES Agenzia Regionale Parchi, 2012 − Indirizzi di gestione e misure di con- servazione della ZPS “Monti Lepini” (IT6030043). (Accessed 29 August 2021). Amori G., Corsetti L. & Esposito C., 2002 − Mammiferi dei Monti Lepini. Ministero dell’Ambiente, Istitoto Nazionale per la Fauna Selvatica. Quaderni di Conservazione della Natura, 11. Apollonio M., 2003 − Sus scrofa (Linnaeus, 1758). Notizie ecologiche ed etologiche. In: Fauna d’Italia, vol. 38. Mammalia III: Carnivora - Artiodactyla. Boitani L., Lovari S. & Vigna Taglianti A. (eds.). Calderini, Bologna: 267-270. Battisti C., Amori G., Angelici F. M., Capizzi D., De Felici S., Mortel- liti A., Schiavano A., Verucci P. & Zapparoli M., 2013 – Evidence of a local range expansion in a fragmentation-sensitive species: the case of red squirrel (Sciurus vulgaris) in Central Italy. Vie et Milieu, 63: 59-65. Boitani L. & Ciucci P., 2003 − Vulpes vulpes (Linnaeus, 1758). Notizie ecologiche ed etologiche. In: Fauna d’Italia, vol. 38. Mammalia III: Carnivora - Artiodactyla. Boitani L., Lovari S. & Vigna Taglianti A. (eds.). Calderini, Bologna: 63-66. Brøseth H., Knutsen B. & Bevanger K., 1997 − Spatial organization and habitat utilization of badgers Meles meles: effects of food patch dispersion in the boreal forest of central Norway. Zeitschrift für Säugetierkunde, 62: 12-22. Buccomino G., 2005 – I Monti Lepini. Rapporto conoscitivo del pa- trimonio ambientale. Tesi di master in “Pianificazione e gestione delle aree protette”. Università di Camerino, a.a. 2004/2005. Capizzi D., Mortelliti A., Amori G., Colangelo P. & Rondinini C. (eds.), 2012 − I mammiferi del Lazio. Distribuzione, ecologia e conserva- zione. Edizioni ARP, Roma. Ciucci P. & Boitani L., 2003 – Canis lupus Linnaeus, 1758. Notizie ecologiche ed etologiche. In: Fauna d’Italia, vol. 38. Mammalia III: Carnivora - Artiodactyla. Boitani L., Lovari S. & Vigna Taglianti A. (eds.). Calderini, Bologna: 33-42. Copiz R., Iberite M., Lucchese F., Nicolella G. & Petriglia B., 2018 − Atlante della Biodiversità Floristica dei Monti Lepini. Edizione Belvedere, Latina. Corsetti L. & Marozza L., 2020 − Atlante della biodiversità faunistica dei Monti Lepini. Edizioni Belvedere, Latina. Cosentino D., Parotto M. & Praturlon A., 1993 – Guida geologica del Lazio. BeMa editrice, Milano. De Marinis A.M. & Genovesi P., 2003 − Meles meles (Linnaeus, 1758). Notizie ecologiche ed etologiche. In: Fauna d’Italia, vol. 38. Mam- malia III: Carnivora - Artiodactyla. Boitani L., Lovari S. & Vigna Taglianti A. (eds.). Calderini, Bologna: 162-166. Doser J. W., Finley A. O. & Banerjee S., 2022a − Joint species distribu- tion models with imperfect detection for high-dimensional spatial data. arXiv preprint. Doser J. W., Finley A. O., Kéry M. & Zipkin E. F., 2022b − spOccupan- cy: An R package for single-species, multi-species, and integrated spatial occupancy models. Methods in Ecology and Evolution, 13 (8): 1670-1678. Esposito C., 2013 – Lo scoiattolo comune (Sciurus vulgaris) sui Monti Lepini (Italia, Lazio): evidenze di una ricolonizzazione. Biologia Ambientale, 30: 39-44. Genovesi P., 2003 − Felis silvestris (Schreber, 1777). Notizie ecologi- che ed etologiche. In: Fauna d’Italia, vol. 38. Mammalia III: Car- nivora - Artiodactyla. Boitani L., Lovari S. & Vigna Taglianti A. (eds.). Calderini, Bologna: 216-219. Hines J. E., 2006 – PRESENCE - Software to estimate patch occupancy and related parameters. Koprowski J. L., 2005 − The response of tree squirrels to fragmen- tation: a review and synthesis. Animal Conservation, 8: 369-376. Littlewood N. A., Hancock M. H., Newey S., Shackelford G. & Toney R., 2021 − Use of a novel camera trapping approach to measure small mammal responses to peatland restoration. European Journal of Wildlife Research, 67: 12. MacKenzie D. I. & Bailey L. L., 2004 − Assessing the fit of site-oc- cupancy models. Journal of Agricultural, Biological and Environ- mental Statistics, 9: 300-318. MacKenzie D. I., Nichols J., Lachman G., Droege S., Royle J. & Lang- timm C. A., 2002 − Estimating site occupancy rates when detection probabilities are less than one. Ecology, 83: 2248-2255. Mastrobuoni G., 2020a − Puzzola europea Mustela putorius. In: Atlante della biodiversità faunistica dei Monti Lepini. Corsetti L. & Maroz- za L. (eds.). Edizioni Belvedere, Latina: 215. Mastrobuoni G., 2020b – Ratto delle chiaviche Rattus norvegicus. In: Atlante della biodiversità faunistica dei Monti Lepini. Corsetti L. & Marozza L. (eds.). Edizioni Belvedere, Latina: 226. Mastrobuoni G., 2020c – Donnola Mustela nivalis. In: Atlante della biodiversità faunistica dei Monti Lepini. Corsetti L. & Marozza L. (eds.). Edizioni Belvedere, Latina: 214. Mastrobuoni G. & Capizzi D., 2020 − I Mammiferi. In: Atlante della biodiversità faunistica dei Monti Lepini. Corsetti L. & Marozza L. (eds.). Edizioni Belvedere, Latina: 329-347. Mastrodomenico D., 2020 − Lepre italica Lepus corsicanus. In: Atlante della biodiversità faunistica dei Monti Lepini. Corsetti L. & Maroz- za L. (eds.). Edizioni Belvedere, Latina: 229. Mos J. & Hofmeester T. R., 2020 – The Mostela: an adjusted camera trapping device as a promising non-invasive tool to study and mon- itor small mustelids. Mammal Research, 65: 843-853. O’Connell A., Nichols J. D. & Karanth K. U. (eds.), 2011 − Camera traps in animal ecology. Methods and analyses. Springer, Tokyo. Pietrocini V., 2020 – Riccio europeo Erinaceus europaeus. In: Atlante della biodiversità faunistica dei Monti Lepini. Corsetti L. & Maroz- za L. (eds.). Edizioni Belvedere, Latina: 183. R Core Team, 2022 − R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. Rovero F., Zimmermann F., Becci D. & Meek P., 2013 – “Which cam- era trap type and how many do I need?” A review of camera fea- tures and study designs for a range of wildlife research applica- tions. Hystrix. The Italian Journal of Mammalogy, 24: 148-156. Tomaselli R., Balduzzi A & Filipello S., 1973 − Carta bioclimatica d’Italia. Collana Verde, 33. Ministero dell’agricoltura e delle foreste, Roma. ANGELINI, CORSETTI, FAUSTINELLA, MARTELLATO, MASTROBUONI, MASTRODOMENICO, MAROZZA, PETRA, PIETROCINI, PROIETTI