Improving inferences and predictions of species environmental responses with occupancy data

Abstract

1. Occupancy models represent a useful tool to estimate species distribution throughout the landscape. Among them, MacKenzie et al. 's model (2002, MC), is frequently used to infer species environmental responses. However, the assumption that detection probability is homogeneous or fully explained by covariates may limit its performance. Species should be more easily observed at sites with a higher number of individuals. We simulated data following occupancy model (RN) that accounts for abundance-driven heterogeneous detection and two variants with overdispersion in the detection probability and local abundances. Then, we compared the performance of the MC model against that of RN.2. In addition to model misspecifications, insufficient information in data (i.e. infrequent detections) can limit our ability to detect existing effects with affordable sampling designs. To deal with this source of error, we extended RN approach to a community-level joint species model (RN-JSM), where species responses and detectability depended on their traits and phylogeny. Then, we tested RN-JSM performance in simulated and out-of-sample field data.3. High abundance-driven heterogeneity in detection (i.e. common and secretive species) limited the ability of the MC model to quantify covariate effects; especially, when the number of visits was low. Both models (MC and RN), often failed to detect existing effects when data were overdispersed. Moreover, the RN model consistently lacked sufficient power when analysing data from uncommon species (even when simulations and model specifications perfectly matched). This problem was solved by our RN-JSM, which yielded more precise and accurate estimates of species environmental responses. Increased accuracy in rare species held when the RN-JSM was tested with real and out-of-sample datasets. 4. In the light of our results, we propose: (i) for common and secretive species analyse occupancy data with the RN model and prioritize revisiting sites; (ii) for species that may have overdispersed detectability or local abundances (e.g. with correlated behaviours or occurring in clusters), apply RN extensions that account | 1541 Methods in Ecology and Evoluঞon MORÁN-LÓPEZ et aL.

Replicate DNA metabarcoding can discriminate seasonal and spatial abundance shifts in river macroinvertebrate assemblages

“…Taylor's law: Giometto et al, 2015). Indeed, recent simulation-based studies have shown the Royle-Nichols model was more robust to overdispersion when zero-inflation was included, but, as was the case here, to remain identifiable studies should be designed to maximize detectability (Morán-López et al, 2022).…”

“…Instead, introducing a zero‐inflation parameter to control for within‐stream spatial aggregation helped to resolve the estimation of detectability, and captured the within‐habitat heterogeneity that underpins observed patterns such as the rate of species accumulation (see Figure 2a) and past studies of mean–variance relationships (i.e., Taylor's law: Giometto et al, 2015). Indeed, recent simulation‐based studies have shown the Royle–Nichols model was more robust to overdispersion when zero‐inflation was included, but, as was the case here, to remain identifiable studies should be designed to maximize detectability (Morán‐López et al, 2022).…”

“…We considered the possibility that in addition to imperfect relationship between samples and a population, low detectability could stem from aggregated distributions of particular taxa within the stream, which would limit the proportion of taxa that were “available” for detection in every sample (Joseph et al, 2009). We included a zero‐inflation parameter, ϕ i , to describe the degree to which each taxon was evenly distributed within the stream, and subsequently interpret the predicted estimates of abundance Z ij as average densities in this section of the stream (Morán‐López et al, 2022). Sensitivity analysis indicated the choice of prior had a negligible effect on the model, and we therefore chose flat normal distributions as they are the uninformative priors used by most hierarchical community models (Kéry & Royle, 2015).…”

Timing of invasive species removal influences nonnative biotic resistance and trajectories of community reassembly

“…HMSC is a type of joint species model and, as such, estimates responses to the environment for all species present in communities. However, in HMSC, species responses are not treated as a random term but explicitly modelled as a function of their traits and phylogeny (Morán‐López et al, 2022). Such an approach allows inferring the main ecological mechanisms behind species responses to the environment (i.e.…”

Bird species responses to rangeland management in relation to their traits: Rio de la Plata Grasslands as a case study

“…Here we use an HMSC occupancy model, based on the one recently developed by Mor an-L opez et al (2022), to assess individual bird species responses to rangeland management in the Rio de la Plata Grasslands (RPG). The RPG ecoregion is considered critically endangered and needs urgent attention (Hoekstra et al, 2005) due to the conversion of extensive areas of native grasslands into croplands and increased stocking rates that affect vegetation composition and structure (Modernel et al, 2016).…”