Data from: Joint spatial modeling of cluster size and density for a heavily hunted primate persisting in a heterogeneous landscape
Data files
Nov 11, 2024 version files 14.75 KB
-
data_censored.csv
6.66 KB
-
README.md
2.40 KB
-
transects.csv
5.69 KB
Abstract
Shared landscapes in which humans and wildlife coexist, are increasingly recognized as integral to conservation. Fine-scale data on the distribution and density of threatened wildlife are therefore critical to promote long-term coexistence. Yet, the spatial complexity of habitat, anthropic threats and animal behaviour in shared landscapes challenges conventional survey techniques. For social wildlife in particular, the size of sub-groups or clusters is likely to both vary in space and influence detectability, biasing density estimation and spatial prediction. Using the R package inlabru, we develop a full-likelihood joint log-Gaussian Cox process to simultaneously perform spatial distance sampling and model a spatially varying cluster size distribution, which we condition upon detection probability to mitigate cluster-size detection bias. We accommodate spatial dependencies by incorporating a non-stationary Gaussian Markov random field, enabling the explicit inclusion of geographical barriers to wildlife dispersal. We demonstrate this model using 136 georeferenced detections of Campbell’s monkey (Cercopithecus campbelli) clusters, collected with 398.56-km of line transects across a shared agroforest landscape mosaic (1067-km2) in Guinea-Bissau. We assess a suite of anthropogenic and environmental spatial covariates, finding that normalized difference vegetation index (NDVI) and proximity to mangroves are both powerful spatial predictors of density. We captured strong spatial variation in cluster size, likely driven by fission-fusion in response to the complex distribution of resources and risk in the landscape. If left unaccounted for under existing approaches, such variation may bias density surface estimation. We estimate a population of 10,301 (95% CI [7606-14,104]) individuals and produce a fine-scale predictive density map, revealing the importance of mangrove-habitat interfaces for the conservation of this heavily hunted primate. This work demonstrates a powerful, widely applicable approach for monitoring socially flexible wildlife and informing evidence-based conservation in complex, heterogeneous landscapes moving forward.
README
Data context
Line transect survey data for Campbell's monkeys (Cercopithecus campbelli) were collected between January 19th and June 18th, 2021 by co-authors A.S. and M.J., as part of a recently established and ongoing biodiversity monitoring program (BMP). The BMP is implemented as part of a Darwin Initiative project (26-018) by the University of Exeter and the Institute of Biodiversity and Protected Areas (IBAP) of Guinea-Bissau. Following standardised protocols, 64 individual line transects approximately 1-km in length were walked up to 6 times each for a total of 354 line transect repeats, with a cumulative walked effort of 398.56-km. Observers moved at a constant pace of 1 to 1.5-km/h along each line transect and for each Campbell's monkey cluster detected, recorded the number of individuals (cluster size), the perpendicular distance between the centre of the cluster and the transect line (detection distance), and took GPS coordinates. Sampling was conducted between the hours of 07:06 and 19:18. To mitigate possible variation in detectability related to activity levels, survey start times were randomised within this window. GPS coordinates are censored (replaced with NA values) in this dataset due to conservation concerns. Please see the related article for more information and to request access to an uncensored geo-referenced version of this survey dataset.
File description
Survey data are provided as a comma-separated value (CSV) text file (censored_data.csv) as follows: "transect" is a unique transect identifier; "longitude" and "latitude" are censored as NA values; "distance" is the detection distance recorded in metres; "cluster" is the cluster size or number of individuals per observation; "date" is the day, month and year of the observation formatted as DD-MMM-YY; and "time" is the time of day formatted as HH:MM:SS.
Transect data are provided as a CSV text file (transect.csv) as follows: "transect" is a unique transect identifier; "survey1" to "survey6" contain the date of a survey (if conducted) during survey occasions 1 to 6, formatted as day, month and year DD-MMM-YY; "repeats" is the total number of surveys conducted; "length" is the length of the transect in kilometres; and "total" is the total length surveyed at that transect in kilometres. NULL values indicate that a transect was not walked during that particular survey occasion.