Using single visits into integrated occupancy models to make the most of existing monitoring programs

Abstract

24 1. A major challenge in applied ecology consists in integrating knowledge from 25 different datasets to produce robust ecological indicators. To estimate species distribution, 26 occupancy models are a flexible framework that can accommodate several datasets obtained 27 from different sampling methods. However, repeating visits at sampling sites is a prerequisite 28 for using standard occupancy models, which may limit their use. Under certain conditions, 29 detection/non-detection data collected during single visit can be analysed with occupancy 30 models. To date however, single-visit occupancy models have never been used to combine 31 several different datasets. 32 2. Here, we developed an approach that combines multi-method and single-visit 33 occupancy models. As a case study, we estimated the distribution of Bottlenose dolphins 34 (Tursiops truncatus) over the North-western Mediterranean Sea by combining 24,624 km of 35 aerial surveys and 21,464 km of at-sea monitoring. We compared the outputs of single-vs. 36 repeated-visit multi-method occupancy models, and that of single-method occupancy models. 37 3. Multi-method models allowed a better sampling coverage in both coasts and high 38 seas and provided a better precision for occupancy estimates than single-method occupancy 39 models using aerial surveys or at-sea surveys in isolation. 40 4. Overall, single-and repeated-visit multi-method occupancy models produced 41 similar inference about the distribution of bottlenose dolphins. This suggests that single-visit 42 occupancy models provide robust occupancy estimates, which open promising perspectives 43 for the use of non-standardized datasets. 44 5. Synthesis and applications: Single-visit multi-method occupancy models can help 45 making the best out of ecological monitoring programs by optimizing cost effectiveness 46 through the formal combination of datasets. 47 48 Keywords 49 Bottlenose dolphins, Ecological monitoring, Integrated species distribution models, Multi-50 method, Occupancy models, Single-visit 51 52 SuppInfo_Lauret_et_al.html 595

Using integrated multispecies occupancy models to map co-occurrence between bottlenose dolphins and fisheries in the Gulf of Lion, French Mediterranean Sea

“…We considered 4 sampling occasions 𝑗 with similar sampling effort for each monitoring program (winter, spring, summer, and autumn). We extended the observation process of the multispecies occupancy model of Rota et al (2016a) to integrate two datasets in the spirit of Lauret et al (2021).…”

“…To overcome data scarcity, several authors have suggested to combine multiple datasets into an integrated modelling framework (see Kéry & Royle (2020), Chapter 10, for a review). In that spirit, we previously developed a single-species integrated occupancy model to map the distribution of bottlenose dolphins over the Northwestern Mediterranean Sea (Lauret et al 2021).…”

spOccupancy: An R package for single-species, multi-species, and integrated spatial occupancy models

“…Hence, when integrating multiple detection-nondetection data sources in an IOM, each data source has its own unique detection model defined by Equations 3 and 4 that are conditional on a shared latent occupancy process defined by Equations 1 and 2. This joint-likelihood approach enables explicit estimation of different covariate effects (and intercepts) on the detection processes of each data source (Miller et al, 2019;Lauret et al, 2021;Zulian et al, 2021). The spOccupancy function intPGOcc fits IOMs.…”