We understood individual jaguars based on unique spot designs (Silver et al. 2004). Cubs integrated obviously younger and you may young some one submitted which have adult girls. We classified women as reproductive when they have been filed having cubs any kind of time area from inside the data year, so when nonreproductive, if they was indeed never ever submitted having cubs. I managed presence away from cubs since a target standard having evidence out of reproduction. Class off breeding otherwise non was held constant for the whole research several months. Though simplified, we feel it class justified because of the enough time reproductive period out of women jaguars (we.elizabeth., ninety days pregnancy and you can 17 days proper care of cubs) and you may much time (3–4 age) time for you first breeding (Crawshaw and you can Quigley 1991; De Paula mais aussi al. 2013). I result in the assumption you to definitely reproductive females take care of their territories to have extended periods (i.age., years) and any small-name enjoy (i.e., dropping cubs) wouldn’t change its territory dimensions. Also, we essentially filed old cubs (>90 days dated), which may has actually live this new believed very early top within the juvenile mortality documented in other large carnivores (Jedrzejewska ainsi que al. 1996; Palo). The brand new identification process are performed by a couple authors individually (MFP and MA) and you will affirmed by a third (WJ). Unidentifiable grabs was in fact omitted regarding after that analyses. Having grab-recapture activities, i defined every day testing era in a way that we noticed only one take every single day for every single pitfall, i.elizabeth., binomial recognition records (Royle mais aussi al. 2009; Goldberg et al. 2015).
Society thickness estimate for adult jaguars
We applied restrict probability SCR models during the secr dos.10.step 3 R package (Efford ainsi que al. 2004, 2009; Borchers and you may Efford 2008; Efford 2016) so you’re able to guess jaguar densities. Such hierarchical models establish (1) a beneficial spatial brand of the newest shipments from creature hobby facilities and you can (2) a good spatial observation design linked the chances of finding a single on a certain trap for the distance regarding animal’s passion cardio (Efford 2004). With the observance model, we put a danger 50 % of-normal recognition mode:
Gender out of adult some body is actually determined by this new visibility/absence of testicles or nipples or any other reproductive signs
where ? 0 represents the baseline detection probability at an individual’s activity center, ? defines the shape of the decline in detection away from the activity center and can be interpreted in terms of the animal movement distribution, and d specifies the distance between a detector (camera trap) and the activity center (Efford et al. 2009; Efford 2016). This detection model implies a Binomial distribution of detections of an individual at a particular detector (Efford and Fewster 2013; Royle et al. 2014). We used a 15-km buffer around the study area to include the activity centers of any individuals that pling. We checked the adequacy of the buffer size by examining likelihoods and estimates from models with larger buffers. We applied full likelihood models with three sex/reproductive status groups (adult males, adult reproductive females, and adult nonreproductive females) and six shorter sessions as covariates (Borchers and Efford 2008). By doing this, we also fulfilled the assumptions of the closed population model in analyzing our long dataset. We fit models with all possible additive combinations of sex/reproductive status groups and sessions as covariates on density (D), ? 0 , and ?. For density, we always used sex/female reproductive state as a covariate to provide an estimate of population structure and did not consider intercept-only models. We assessed how D, ? 0 , and ? differed across sessions and sex/reproductive status groups and how this variation influenced the overall density estimate. We evaluated models with AICc (corrected Akaike information criterion) and AICc weights (Hurvich and Tsai 1989; Wagenmakers and Farrell 2004). To test the effect of study duration on estimates of all parameters, we compared models that included session covariates in the parameters D, ? 0 , and ? (corresponding to the situation when model parameters were estimated based on separate sessions, as in short-term studies) with the best model that did not include any session covariates.