Publications

• Boyd, Robin J. et al. (2022) ROBITT: A tool for assessing the risk‐of‐bias in studies of temporal trends in ecology. In Methods in Ecology and Evolution, 13(7) p.1497

“Risk-of-bias” tools help researchers to establish whether biases in their data might threaten the validity of their conclusions. Such tools are common in disciplines such as medicine and psychology, but are incredibly rare in biodiversity research. In this paper, we developed ecology’s first risk-of-bias assessment tool. The tool, ROBITT, was designed specifically for analysts who intend to estimate trends in species’ abundances and distributions, but can be applied to all areas of biodiversity trend research.

• Bourhis, Y. et al. (2023) Explainable neural networks for trait-based multispecies distribution modelling – A case study with butterflies and moths. In Methods in Ecology and Evolution, 14(6) p.1531 (as corrected)

In this study we developed a trait-based machine learning approach to species distribution modelling. It is an artificial neural network that use species response traits to produce community-wide predictions. Building on trait mediation to model environmental filtering has several benefits, including (1) improving the predictions of poorly recorded species, (2) model parsimony as environmental drivers’ effects are tied to traits rather than species, as well as (3) explanatory purposes. Indeed, our approach (thanks to SHAP ) provides explanation of the model’s predictions, highlighting the key drivers and traits in the form of species niches and 4^th corner matrices.

• Evans, L. et al. (2024) Population links between an insectivorous bird and moths disentangled through national-scale monitoring data. In Ecology Letters, 27(1).

This recent study has uncovered that insect abundance directly impacts songbird numbers from year to year. By bring together data sets from the Rothamsted Insect Survey and the Breeding Bird Survey, we’ve revealed that increased moth populations correlate with increased numbers of blue tits.

• Wilkes, M. et al. (2024) Predicting nature recovery for river restoration planning and ecological assessment: A case study from England, 1991-2042. In River Research and Applications.

This study modelled spatial and temporal changes in the abundance of 188 riverine invertebrate taxa, using long term records held by the Environment Agency. The models are then used to both inform river restoration planning and set expected values for use in ecological assessment. Using a range of physical and chemical driver datasets, including new LTLS data generated in DRUID, catchment-specific restoration strategies that combine one or more actions involving the removal of channel modifications, reductions in nitrate concentrations and reductions in total dissolved phosphorus concentrations are modelled to identify the most likely combination needed to deliver species abundance targets across three joint climate–socioeconomic scenarios. By hindcasting species abundances under alternative target frameworks, we also demonstrate a new approach to setting expected values in ecological assessment, accounting for changes in water temperature and hydrology.

Parliamentary Committees

• Several members of the DRUID team contributed to a hearing of the Science, Innovation and Technology Committee on Insect Decline and UK Food Security in June 2023. (For more information, see the recording, transcript and final report.)