Critically Endangered frogs in Madagascar 899
P. pollicaris, illustrating the differences in their calling patterns (Supplementary Fig. 1). Weestimated distance to target species call as either near
(#10 m) or far (.10 m) and we recorded the direction to the target using a compass (the observer always stood facing north). We did this to minimize the risk of double count- ing of individuals from separate points; if two calls could potentially belong to the same individual (e.g. both being re- corded as far and in the same relative direction), one record was discounted from the analysis. At each monitoring site we recorded time, maximum and minimum temperatures (°C), maximum and minimum relative humidity (%) and rainfall (categorical: 0 no rain, 1 light rain, 2 heavy rain) at the beginning and end of every 5-minute acoustic point count survey, and elevation, distance from water and vegetation structure (bamboo and Pandanus numbers with- in a 5-m radius of the observer, and canopy cover within a 1-m radius of the observer).
Data analysis
We estimated occupancy and detection probability using a single-season occupancy model (MacKenzie et al., 2002), and we used the Royle–Nichols model to estimate species abundance and population size (Royle & Nichols, 2003). We modelled species separately (i.e. using a single-species modelling framework). Occupancy models account for detection rates (i.e. chances of detecting a species at a site if present) and estimate the proportion of area occupied by a species, whereas the Royle–Nichols model estimates the occupancy rate when heterogeneity in the detection pro- bability exists as a result of variation in animal abundance. The Royle–Nichols model estimates abundance (i.e. number of individuals at each site) using presence/absence data from repeated occasions, and detection probability represents the likelihood of recording all individuals at a site at a given time (Royle&Nichols, 2003). Statistical analyseswere performed in R 3.4.2 (R Foundation for Statistical Computing, Vienna, Austria) using package unmarked (Fiske & Chandler, 2011). We built models using a stepwise approach, using the
Akaike information criterion (AIC) to rank candidatemod- els and to select covariates for final model fitting. In the first stepwe fitted detection covariates using the variables time of survey, rainfall and maximum temperature. In the second step we fitted covariates for occupancy and abundance using the variables elevation, vegetation structure (bamboo, Pandanus and canopy) and distance from water. In the final step we combined the covariates detection probability and abundance/occupancy to build the final models. We ranked the models based on their AIC (the model with the lowest AIC having the best fit) and weighted them by the pro- bability of being the best model in the set. We considered models with ΔAIC,2 to have strong support (Burnham &Anderson, 2002). Wealso used model weights to estimate
the relative importance of predictors, given by the cumula- tive weight of the best-fittedmodels in which the predictors appear, and interpreted this as the probability of being a component of the best model (Symonds & Moussalli, 2011). We averaged all candidate models to account for model selection uncertainty. Becausewe used an acoustic survey, detectability refers to
the probability of hearing the species in a site if the species is both present and active (i.e. calling). Similarly, population size refers to the number of males present at the surveyed sites in the largest fragment in Ambohitantely Special Reserve. We followed the method of Kéry & Royle (2016) to calculate population size and used a parametric bootstrap method that generates a sampling distribution of the popu- lation based on the best-fitted model.We estimated density within our study fragment (i.e. number of adult males/ha) by dividing the estimated population size by the total sur- veyed area (each sampled site of 25 m radius and 1,963 m2; total surveyed area of 16.5 ha). This was then extrapolated across the entire fragment (1,284 ha) to estimate the popu- lation beyond the surveyed sites.
Results
We obtained 126 acoustic records (i.e. presence) for A. vallani and 69 for A. helenae.We detected individuals of A. vallani in 71 sites (naïve occupancy 0.84) and individuals of A. helenae in 51 sites (naïve occupancy 0.61). Details of model fitting and of the covariates influencing species occupancy, abundance and detectability are presented in Supplementary Material 1. Overall, occupancy of both species was best explained by fea- tures in the vegetation structure: canopy cover for A. vallani and bamboo numbers for A. helenae. Detection probability wasinfluenced bytimeofsurvey(for A. vallani)and rainfall (for A. helenae)(Fig. 2, Supplementary Table 1). Based on model-averaged estimates occupancy of the two species was similar, whereas detection probabilities differed (Supplemen- tary Fig. 2). Occupancy estimates for both species were high (0.93 ± 0.07 for A. vallani; 0.80 ± 0.09 for A. helenae). Species detection rates differed, with A. helenae having a 34 ± SE 0.05% chance of being detected if present, which was lower than that of A. vallani, with a probability of detection of 55 ± SE 0.05%. For both species abundance was best explained by vege-
tation structure, whereas detection was best explained by time of survey for A. vallani and rainfall for A. helenae (Supplementary Table 1). The abundance of A. vallani was best explained by bamboo number and canopy cover, whereas the abundance of A. helenae at each site was best explained by the numbers of both bamboo and Pandanus (Fig. 2). There was an estimated abundance of 10 adult male A. vallani at each monitoring site, with a large standard error (range = 4–24,SE = 7). Overall, for our sampled sites
Oryx, 2022, 56(6), 897–903 © The Author(s), 2022. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605321001034
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132 |
Page 133 |
Page 134 |
Page 135 |
Page 136 |
Page 137 |
Page 138 |
Page 139 |
Page 140 |
Page 141 |
Page 142 |
Page 143 |
Page 144 |
Page 145 |
Page 146 |
Page 147 |
Page 148 |
Page 149 |
Page 150 |
Page 151 |
Page 152 |
Page 153 |
Page 154 |
Page 155 |
Page 156 |
Page 157 |
Page 158 |
Page 159 |
Page 160 |
Page 161 |
Page 162 |
Page 163 |
Page 164
