Roe deer population size has traditionally represented a difficult task for population ecologists, due to the preference of wood and elusive behaviour of the species. The Alpine habitat, characterised by patchiness and a poorly accessible terrain, may even enhance the difficulties. In this paper, we assessed the feasibility and efficacy of pellet group count with distance sampling as opposed to censusing methods commonly applied in game management (e.g. drives, spot counts, block censuses). The study was conducted in a 9500ha area (Monte Bondone, Trento Province, Italy) where a sample of the roe deer population was tracked by means of GPS-telemetry. Systematic grids of line transects were randomly over imposed to the study area and sampled in 2006, both in spring (winter density) and autumn (summer density), for a total survey length of 43200m and 216 transects. From October 2005 to October 2006, we carried out and experimental assessment of pellet group decay rate, placing fresh samples in different habitats and analysing monthly disappearance rates by logistic regression. In autumn, we carried out thermal imaging distance sampling and mark-resight observations in a portion of the study area. Mean monthly time to decay was mainly influenced by habitat, slope and weather condition at deposition, particularly snow cover in winter months. In spring, pellet group encounter rate was higher than in autumn (0.020 pellet groups/m and 0.009 pellet groups/m, respectively), which led to a lower %CV of pellet group density estimate (13% and 26% respectively). This corresponded to a longer time to decay in spring than in autumn (242.9 days and 57.2 days respectively) and to halved roe deer density values (2.5 deer/100ha and 5.5deer/100ha respectively). However, the disproportion between the decrease of encounter rate and time to decay between the two seasons (2 fold vs 4 fold), is likely due to a high variability of the spring mean times to decay, which may have led to biased spring density values. In autumn, pellet group count density was comparable to thermal imaging distance sampling and mark-resight estimates and consistent with hunting data. Therefore, density estimates proved more robust when pellet groups decayed in a short time, comparable to the length of the sampling period. Pellet group count with distance sampling seemed to encompass randomness, replicability and possibility to control a priori the threshold precision/effort. However, if routinely used, a continuous monitoring of decay rate is strongly suggested.
|Citation:||Cagnacci, F.; Pedrotti, L.; Tagliabò, A.; Luchesa, L.; Franzetti, B.; Focardi, S. (2007). Monitoring roe deer populations in the Alps: the pellet group count with distance sampling approach. In: 8th Roe deer meeting, Velenje, Slovenia, 25-29 June 2007. handle: http://hdl.handle.net/10449/20854|
|Organization unit:||Centre for Alpine Ecology # CEA_1993-2007|
|Authors:||Cagnacci, F.; Pedrotti, L.; Tagliabò, A.; Luchesa, L.; Franzetti, B.; Focardi, S.|
|Title:||Monitoring roe deer populations in the Alps: the pellet group count with distance sampling approach|
|Appears in Collections:||03 - Conference object|