Fatores que afetam a detectabilidade da toninha (Pontoporia blainvillei) em estudos de estimativas populacionais

Abstract

The franciscana dolphin (Pontoporia blainvillei) is considered the most threatened cetacean species in the southwestern Atlantic Ocean. Therefore, robust abundance estimates are necessary to improve conservation and management effort for this species. The goal of the present work was improve abundance estimates of franciscana dolphins by estimating correction factors to compensate for visibility bias in aerial surveys. Field experiments were made from a Robinson R44 helicopter in Babitonga Bay, Santa Catarina State, to estimate the time during which a franciscana group remained available to be detected from an aircraft. The surfacing interval was defined as the period of time during which one or more franciscanas in a group were available to the observer in the helicopter. A surface-dive cycle was defined as the period of time from the beginning of one surfacing to the next. Generalized mixed effect models (GLMMs) were used to investigate the influence of environmental and biological predictors on the proportion of time franciscana groups spend ate the surface. Availability bias was then estimated following the methods proposed by LAAKE et al. (1997). A total of 15 hours were flow during the helicopter experiment. After filtering the whole dataset 248 complete surface-dive cycles from a total of 101 samples remained. The most supported GLMM model show a positive relationship between the proportion of time at surface and the size of franciscana groups. The average time franciscana groups spent at the surface and in a dive was 16.10 (SE = 9.74) and 39.77 (SE = 29.06), respectively. Assuming a time window of 5.77 seconds, the estimated availability bias was 0.38 (SE = 0.01). The fraction of individuals available at the surface that were missed by the observers (perception bias) was estimated using mark-recapture-distance-sampling methods (MRDS) assuming point independence. Sightings were recorded through aerial surveys carried out between 2011 and 2014 with an Aerocommander 500B and the same observers, which were locaed at the same observing positions. Because of differences between front (bubble) and rear (flat) windows field of view, only sighting made between 60º-30º were used in the analysis (n = 191). The MRDS model that best fit the data had distance, area (factor), and side (factor) as covariates in the mark-recapture component, and distance and Beaufort (factor) as covariates in the distance sampling component. The estimated probability of detecting a franciscana group that was visible at the transect line was 0.38 (SE = 0.12). Results of this study showed that surfacing time recorded from helicopter were 13.42 times greater than time recorded from surface platforms. Therefore, availability bias estimates for aerial surveys of franciscana dolphins must be estimated from dive parameters computed from aerial platforms. Estimates of 14 perception bias indicated that the number of franciscana sightings missed by observers can be significant and that correction factors to compensate for this bias are necessary. However, methodological improvements in the sampling procedures are needed to estimate this parameter in a more robust way.