A National Application of the Species Habitat Index for Ethiopia Reveals Uneven Habitat Change Across Plant Groups and Ecosystems

Moabe Fernandes, Joseph D.M. White, Ermias Lulekal, Sebsebe Demissew, Jack Plummer, Amy Barker, Shambel Chengere, Liz Brogan, Andrew Budden, Iain Darbyshire, Becca Davis, Kelda Elliott, Tesfanesh Feseha, Bezawit Genanaw, Gabriella Hoban, Hanny Lidetu, Lidet Mehari, Efrata Mekebib, Henry Miller, Paloma Moore, Lynda Murray, Sileshi Nemomissa, Tim Pearce, Seth Ratcliffe, Alexandra Roberts, Harry Smith, Vida Svahnström, Emily Terry, Jenna Willis, James Borrell, Steven Bachman, Carolina Tovar

Accelerating delivery of the Kunming-Montreal Global Biodiversity Framework (GBF) requires indicators that can detect species-level habitat change at policy-relevant scales. The Species Habitat Index (SHI) offers a method to optimise resource allocation and enhance national biodiversity reporting, but its broader utilisation has been constrained by significant data gaps, particularly for taxa that are less well documented than birds and mammals. In this study, we (i) examine how methodological choices impact the assessment of changes in species’ suitable habitat through the SHI, and (ii) develop a national SHI for plants of Ethiopia. We compiled georeferenced occurrence records and IUCN-based habitat preferences for 1,247 plant species (141 families; 340 endemics) and combined these with a 29-year land-cover time series (1992–2020) to quantify SHI. We first compared different geometry-based range-mapping methods with species distribution models (SDMs) to assess their influence on SHI estimates and then applied a combined approach to estimate national SHI trends for Ethiopian plants. SHI trends were broadly robust across methods, with Rapoport’s Mean Propinquity polygons showing closest match with SDMs. Nationally, median SHI increased slightly (+0.51%), but declines were common and disproportionately affected endemic and non-woody species. Declines SHI trends were concentrated in species inhabiting shrubland and grassland land-cover classes and were geographically clustered in highland regions corresponding to the Afro-alpine and Dry Evergreen Afromontane Forest and Grassland Complex vegetation types. These findings demonstrate that SHI can be applied to data-limited floras using simplified range-mapping approaches, providing a scalable tool for biodiversity monitoring. At the same time, the results highlight uneven responses to land-cover change and the need to prioritise vulnerable ecosystems in Ethiopia’s conservation planning.