Modeling distribution and abundance of multiple species: Different pooling strategies produce similar results

Abstract

Abstract. Quantifications of spatial distribution and abundance of animals are essential to identifying key landscape characteristics and targeting locations for conservation action. Since conservation decisions often focus on multiple species aggregated in groups, e.g., guild-level, rather than individual species, predictions of species group abundance are of central importance. However, areas chosen for conservation action may differ if results from various modeling strategies also differ. Therefore, we compared three different strategies for modeling species group distribution and abundance: predict first, assemble later (PA); assemble first, predict later (AP); and the combined assemble, predict, then assemble (APA). All strategies were performed using Boosted Regression Trees (BRTs), which were fit to individual species data and then grouped after modeling, or fit to datasets that were grouped before modeling. Modeling strategies produced very similar results in terms of statistical performance assessed through four evaluation metrics and in spatial patterns in predicted abundance. To further assess potential functional implications of any numeric differences to conservation planning, we examined the relative proportion of the predicted population within existing Canadian protected areas. This metric further confirmed similarity in predictions from the three modeling strategies. Our results suggest that locations targeted for conservation action would be highly consistent among modeling strategies. Slight differences we observed in spatial predictions may be due to data coverage across species ranges, data quality, and the flexibility of the BRTs.

Evaluating the waterfowl breeding population and habitat survey for scaup

“…Use of the WBPHS data to infer spatial changes in breeding populations of scaup is clearly not appropriate (e.g., Afton and Anderson ; Ross et al ; Barker et al , ) because, as documented (this study, Finger et al ), they are still migrating to their breeding sites during the WBPHS. We detected that nearly half (48.8% of the total migrations) of our marked scaup using the traditional survey area were available to be counted more than once as they migrated north with the WBPHS crew.…”

“…Further, estimates produced by the WBPHS are based on the assumption that ducks migrate ahead of survey crews, and subsequently settle at breeding sites where they are then available to be counted once the area is surveyed (Smith ). However, only 14.7% of migrations by our marked scaup across years met that criteria, indicating that use of stratum‐scale breeding population estimates of scaup for determining their breeding distribution, changes in distributions, and locations for focused conservation efforts is clearly problematic (e.g., Hobson et al ; Ross et al ; Barker et al ,). We also found that the number of times a marked scaup was available to be counted varied positively with date of settling during the breeding period.…”

Avian SDMs: current state, challenges, and opportunities

“…By modeling Swiss breeding birds, the authors found that different functional groups affect model predictions and suggested that species-or group-specific environmental predictor sets should be incorporated to improve species richness models (Zurell et al 2016). Within such groups, species richness patterns can be accurately predicted, for example as shown for Canadian waterfowl assemblages (Barker et al 2014b). …”

Drifting baited stereo‐videography: a novel sampling tool for surveying pelagic wildlife in offshore marine reserves

“…au/topics/marine/marine-reserves) which, at 3.1 million km 2 , is currently the world's largest. This new development provided a unique opportunity to test both (1) the relevance of mid-water stereovideography as a field data collection tool for the monitoring of offshore marine reserves, and (2) the validity of camera-derived sightings as input to species distribution models, a suite of analytical tools in growing demand to support conservation decision-making and aid reserve mapping and management (Guisan et al 2013, Barker et al 2014, Marshall et al 2014. Specifically, we designed a new generation of drifting baited pelagic cameras suited for deployment in regions of complex seabed topography and depths greater than 100 m. These were used to document the occurrence of pelagic species within and around the formerly eastern-most national park of the Perth Canyon CMR (hereafter ''the eastern park'', for brevity; see Methods: Study area for details regarding updated zoning arrangements).…”