The EFI2024 Conference provided the second opportunity for EFI to give out the EFI Futures Outstanding Student Presentation Award. This award is given to promote, recognize, and reward an outstanding student presenter and provides valuable feedback to student presenters on their research and presentation skills.  Awards were given to students who gave both Posters and Oral Presentations. Poster or oral presentations were anonymously reviewed by three volunteer reviewers with no conflicts of interest with the presenters. In addition to being recognized for their outstanding work, award winners get to select any item from the EFI shop. We thank all the students who presented and the volunteers who reviewed the presentations!

Congratulations to this year’s Outstanding Presentation Award recipients!

• Gabrielle Koerich (University of Canterbury) won for her oral presentation, “Modelling bryophyte distributions in Antarctica: unveiling the influence of water availability, sampling bias, and spatially complex dynamics” and
• Nima Farchadi (San Diego State University) won for his poster “Integrating diverse data for robust species distribution models in a dynamic ocean.”

See Gabrielle and Nima’s presentation details below.

Modelling bryophyte distributions in Antarctica: unveiling the influence of water availability, sampling bias, and spatially complex dynamics

Gabrielle Koerich¹ , Hao Ran Lai¹ , Grant Duffy² , Eva B. Nielsen¹ , Jonathan D. Tonkin^1
1University of Canterbury, Christchurch, New Zealand. ²University of Otago, Dunedin, New Zealand

Abstract: In Antarctica, observations of biodiversity are often incomplete and commonly restricted to presence-only data, as sampling in the continent is difficult. Thus, understanding and forecasting the range dynamics of key taxa, such as bryophytes, requires a modelling framework capable of dealing with such challenges. Here, we developed log-Gaussian Cox process models of bryophytes’ across the entire Antarctic continent to (1) assess whether broad-scale bryophytes distributions are driven by water availability, as widely hypothesized; (2) map and forecast their distributions and identify under sampled regions; and (3) determine if there’s a spatial dependency between “patches” of bryophytes, which may be related to their limited dispersal in Antarctica. Results show that the main driver of bryophytes distributions in Antarctica is indeed related to areas where water tends to accumulate (97.5% CI:-0.886;-0.797). Maximum temperature was the second most important predictor (97.5% CI: 0.511; 0.681), signalling the importance of elevated temperatures for bryophytes’ restricted metabolism in this extreme environment. The covariate related to human activity showed a high level of sampling bias, and by accounting for this covariate in predictions, we detected habitat suitability for bryophytes in two under-sampled mountain ranges. Finally, the inclusion of a Gaussian random field to account for spatial autocorrelation increased model performance (pseudo R2 and mapped mean estimated density), indicating a spatial dependency between the presence of mosses. Our study demonstrates that a spatially structured modelling framework can provide robust results and allow for valuable forecasts of biodiversity change in data-poor regions.

Integrating diverse data for robust species distribution models in a dynamic ocean

Nima Farchadi¹ , Camrin Braun² , Martin Arostegui² , Rebecca Lewison^1
1San Diego State University, San Diego, USA. ²Woods Hole Oceanographic Institution, Woods Hole, USA

Abstract: Species distribution models (SDMs) are an important tool for marine conservation and management. Despite the burgeoning use of SDMs, limited guidance is available on how to leverage distinct data types to build robust models. Here we assess whether an integrative model framework improves performance over traditional data pooling or ensemble approaches when synthesizing multiple data types. We trained traditional, correlative SDMs and integrative SDMs (iSDMs) with three distinct data types that represent the distribution of a heavily exploited pelagic fish, the blue shark (Prionace glauca), in the North Atlantic. Weevaluated data pooling and ensembling approaches in a correlative SDM framework and compared performance to a model-based data integration approach designed to explicitly account for data-specific biases while retaining the strengths of each dataset. We found that while each integration approach can result in robust models, there was variation in predictive accuracy among data types, with all models predicting fishery-dependent data more accurately than fishery-independent data. Differences in performance were primarily attributed to each model’s ability to explain the spatiotemporal dynamics of the training data, with iSDMs including spatiotemporal terms to have the most accurate and ecological realistic estimates. Our findings reveal trade-offs in the current techniques for integrating data in SDMs between accurately estimating species distributions, generating ecologically realistic predictions, and practical feasibility. With increasing access to growing and diverse data sources, comparing integration approaches can provide valuable guidance for practitioners navigating diverse data types in SDM development and will help users better understand model biases and estimate error.