The capercaillie (Tetrao urogallus) and the hazel grouse (Bonasa bonasia) truly play the role of umbrella species of the forest biocenosis of the Alps. The knowledge of the characteristics of their habitats, and their identification from a cartographic point of view, are essential tools to preserve their presence in the Alps. The present study aims at exploiting the opportunistic observations of capercaillie and black grouse made by the hunting and environmental surveillance staff operating in the Trento province, in order to implement the knowledge on the habitat selection of these two species at the provincial scale. The opportunistic observations detected, in the period 2019-2021, were collected and cataloged using specially created smartphone applications and, subsequently, organized into a dataset containing 703 observations of capercaillie and 226 of hazel grouse. To analyze habitat selection, environmental variables were measured and compared within cells of species presence and random cells of availability scattered throughout the provincial territory. In order to do that, the entire provincial territory (6.200 km2) was divided into two different sampling grids, each with cells of area comparable to the home range of the two species. In particular, cells of 900 ha (3 km2) were used for the capercaillie and 100 ha (1 km2) for the hazel grouse. The 27 environmental variables used considered in each cell: the main characteristics of the geomorphology, the length of the forest roads present, the percentage coverage by other anthropic structures (buildings, main road network, etc.), and the percentage coverage by the various forest types. The values of these environmental variables, measured in the cells of ascertained presence of capercaillie and hazel grouse, were compared with those of an equal number of cells of availability, randomly selected in the study area. The Mann-Whitney U-test (Legendre and Legendre, 1998) was used to test for significant differences between presence and availability cells. Subsequently, presence-availability models (Boyce et al. 2002; Manly et al. 2002) were performed using binary logistic regression analysis to assess environmental suitability. Significant differences were found for 19 of the environmental variables, i.e., elevation, forest viability, and 17 different forest types, while logistic regression analysis yielded 11 models, with a total of 16 component variables in the mean model. In the case of francolin, significant differences between presence cells and availability cells were found for 16 environmental variables: altitude, slope, roughness, forest road system, anthropization and 11 different forest types, while binary logistic regression analysis produced 55 models, with a total of 15 component variables of the mean model. Finally, based on the models obtained, a probability of occurrence prediction map was created for each species, which was then converted into a suitable/unsuitable binomial map (Liu et al. 2005). Some aspects related to the selection of the two species for certain forest types were substantially confirmed in the present study but, on the other hand, the influence of forest roads on the distribution of the species introduces new questions given the positive selection shown by both species towards cells with a greater presence of forest roads.
Il gallo cedrone (Tetrao urogallus) ed il francolino di monte (Bonasa bonasia) rivestono a pieno titolo il ruolo di specie ombrello delle biocenosi forestali dell’arco alpino. La conoscenza delle caratteristiche dei loro habitat, e l’identificazione degli stessi da un punto di vista cartografico, sono strumenti essenziali per preservare la loro presenza sulle Alpi. Il presente studio si pone l'obiettivo di sfruttare le osservazioni opportunistiche di gallo cedrone e francolino di monte, rilevate dal personale di vigilanza venatoria ed ambientale operante in provincia di Trento, per implementare le conoscenze sulla selezione dell’habitat a scala provinciale di queste due specie. Le osservazioni opportunistiche rilevate, nel periodo 2019-2021, dal personale di vigilanza operante in Provincia di Trento sono state raccolte e catalogate mediante applicativi per smartphone appositamente creati e, successivamente, organizzate in un dataset contenente 703 osservazioni di gallo cedrone e 226 e di francolino di monte. Per analizzare la selezione dell’habitat sono state misurate e confrontate le variabili ambientali all’interno di celle di presenza della specie e di celle casuali di disponibilità sparse su tutto il territorio provinciale. A tal fine l’intero territorio provinciale (6.200 km2) è stato suddiviso in due differenti maglie di campionamento, ognuna con celle di superficie paragonabile all’home range di una delle due specie, in particolare sono state utilizzate celle di 900 ha (3 km2) per il cedrone e 100 ha (1 km2) per il francolino. Le 27 variabili ambientali utilizzate consideravano in ciascuna cella: le principali caratteristiche della: geomorfologia, della lunghezza delle strade forestali presenti, della copertura percentuale da parte delle altre strutture antropiche (edifici, rete viaria principale, ecc.), e la copertura percentuale da parte dei vari tipi forestali. I valori di queste variabili ambientali, rilevati nelle celle di presenza accertata del gallo cedrone e del francolino di monte, sono stati confrontati con quelli di un egual numero di celle di disponibilità, selezionate casualmente nell’area di studio. Per verificare l’esistenza di differenze significative tra celle di presenza e di disponibilità è stato usato il test U di Mann-Whitney (Legendre e Legendre, 1998). Successivamente, per valutare l’idoneità ambientale sono stati realizzati, attraverso l’analisi di regressione logistica binaria (ARLB), modelli di presenza-disponibilità (Boyce et al. 2002; Manly et al. 2002). Per quanto riguarda il cedrone le analisi univariate hanno riscontrato differenze significative per 19 delle variabili ambientali, vale a dire quota, viabilità forestale e 17 differenti tipi forestali, mentre l’analisi di regressione logistica ha prodotto 11 modelli, con un totale di 16 variabili componenti il modello medio. Nel caso del francolino, sono state riscontrate differenze significative tra celle di presenza e celle di disponibilità per 16 variabili ambientali: altitudine, pendenza, asperità, viabilità forestale, antropizzazione e 11 differenti tipi forestali, nel contempo, l’analisi di regressione logistica binaria ha prodotto 55 modelli, con un totale di 15 variabili componenti il modello medio. Infine, sulla base dei modelli ottenuti, per ciascuna specie è stata creata una mappa di predizione della probabilità di presenza che si è poi convertita in una mappa binomiale idoneo/non idoneo (Liu et al. 2005). Alcuni aspetti relativi alla selezione da parte delle due specie per determinati tipi forestali sono stati sostanzialmente confermati anche nel presente studio ma, di contro, l’influenza delle strade forestali sulla distribuzione della specie introduce dei nuovi interrogativi vista la selezione positiva, dimostrata da ambe le specie, nei confronti delle celle con maggior presenza di strade forestali.
Selezione dell'habitat di gallo cedrone (Tetrao urogallus) e francolino di monte (Bonasa bonasia) in Trentino
ZAMBOTTI, SANDRO
2020/2021
Abstract
The capercaillie (Tetrao urogallus) and the hazel grouse (Bonasa bonasia) truly play the role of umbrella species of the forest biocenosis of the Alps. The knowledge of the characteristics of their habitats, and their identification from a cartographic point of view, are essential tools to preserve their presence in the Alps. The present study aims at exploiting the opportunistic observations of capercaillie and black grouse made by the hunting and environmental surveillance staff operating in the Trento province, in order to implement the knowledge on the habitat selection of these two species at the provincial scale. The opportunistic observations detected, in the period 2019-2021, were collected and cataloged using specially created smartphone applications and, subsequently, organized into a dataset containing 703 observations of capercaillie and 226 of hazel grouse. To analyze habitat selection, environmental variables were measured and compared within cells of species presence and random cells of availability scattered throughout the provincial territory. In order to do that, the entire provincial territory (6.200 km2) was divided into two different sampling grids, each with cells of area comparable to the home range of the two species. In particular, cells of 900 ha (3 km2) were used for the capercaillie and 100 ha (1 km2) for the hazel grouse. The 27 environmental variables used considered in each cell: the main characteristics of the geomorphology, the length of the forest roads present, the percentage coverage by other anthropic structures (buildings, main road network, etc.), and the percentage coverage by the various forest types. The values of these environmental variables, measured in the cells of ascertained presence of capercaillie and hazel grouse, were compared with those of an equal number of cells of availability, randomly selected in the study area. The Mann-Whitney U-test (Legendre and Legendre, 1998) was used to test for significant differences between presence and availability cells. Subsequently, presence-availability models (Boyce et al. 2002; Manly et al. 2002) were performed using binary logistic regression analysis to assess environmental suitability. Significant differences were found for 19 of the environmental variables, i.e., elevation, forest viability, and 17 different forest types, while logistic regression analysis yielded 11 models, with a total of 16 component variables in the mean model. In the case of francolin, significant differences between presence cells and availability cells were found for 16 environmental variables: altitude, slope, roughness, forest road system, anthropization and 11 different forest types, while binary logistic regression analysis produced 55 models, with a total of 15 component variables of the mean model. Finally, based on the models obtained, a probability of occurrence prediction map was created for each species, which was then converted into a suitable/unsuitable binomial map (Liu et al. 2005). Some aspects related to the selection of the two species for certain forest types were substantially confirmed in the present study but, on the other hand, the influence of forest roads on the distribution of the species introduces new questions given the positive selection shown by both species towards cells with a greater presence of forest roads.È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: unitesi@unipv.it.
https://hdl.handle.net/20.500.14239/14151