Species Distribution Modelling (SDM), also known as Habitat Distribution Modelling, is a widely used framework for estimating a species’ realised ecological niche based on occurrence records and relevant environmental variables. Once the niche has been characterised, SDMs are then used to identify where and when similar environmental conditions occur, allowing projections across geographic regions and time periods. These models are increasingly used to assess the potential impacts of future environmental change, supporting conservation planning, habitat restoration, and biodiversity management.

A major limitation of many SDM approaches is how they treat time. Species occurrence data are often collected over decades (e.g., GBIF) or even centuries (e.g., museum collections), yet most available software does not explicitly incorporate the temporal dimension or account for how both land use and climate have changed in the meantime. As a result, estimates of a species’ realised niche may be biased, potentially reducing the reliability of model predictions.

In this talk, we will present two complementary approaches to address this challenge. The first is tidysdm, an R package designed to perform SDMs using time-series data. By explicitly pairing long-term species observations with historical climatic information, tidysdm enables more robust reconstruction of realised niches while reducing biases arising from anthropogenic habitat alteration and ongoing climate change, both of which can differently affect species observations over time.

The second approach is a workflow we developed to investigate how species’ realised niches change through time in response to significant climatic fluctuations. Using case studies spanning tens of thousands of years, we will demonstrate how this framework can be used to reconstruct niche dynamics in the face of substantial climate change. We will then discuss its relevance for understanding species responses to the unprecedented environmental changes occurring today.