FAQ

The LIFE metric, like many other biodiversity indicators, focuses on terrestrial vertebrates—namely mammals, birds, reptiles, and amphibians. These groups are often used as proxies for biodiversity more broadly because they are relatively well-studied, with comprehensive data available on their geographic distributions, habitat associations, and sensitivity to habitat changes. However, it is important to recognize that relying solely on vertebrates may not fully capture the complexity of biodiversity. Other groups, such as plants and invertebrates, can exhibit different patterns of diversity and may respond differently to changes in land cover.

We are working to expand the metric to systematically include additional taxa. Rather than adding individual species on an ad hoc basis—which could introduce global biases—we aim to incorporate entire taxonomic groups for which comprehensive data are available.

This initial focus on human-driven habitat loss is appropriate, as habitat loss remains the largest current and future threat to terrestrial biodiversity. Other threats—including overexploitation, invasive alien species, and climate change—are also important, as they influence whether species can occupy otherwise suitable habitats. However, these pressures are not yet comprehensively mapped at a global scale, making their integration into global metrics challenging. Efforts are underway to map these additional pressures and incorporate them into LIFE.

We use extinction as the common unit, treating all species equally in our calculations. This approach ensures global comparability and avoids introducing subjective weightings. Two species that have lost the same proportion of their habitat to date would therefore face the same extinction risk today.

However, this does not mean all species are treated equally in practice. Species that have already lost significant portions of their habitat start with a higher baseline extinction risk. As a result, any further habitat loss has a greater impact on them, effectively giving higher weight to these species.

LIFE quantifies the probability that a species will go extinct as a result of human activity. To isolate the impact of human actions, LIFE estimates the extinction probability relative to a world without humans. No species persists forever, so extinction probability is tied to a specific time frame. LIFE focuses on a 100-year horizon—a timescale that captures the effects of present-day human actions, aligns with IUCN Red List criteria, and reflects the expected lag between habitat loss and species decline.

Since LIFE uses a probabilistic framework, it cannot be used to determine which species will go extinct due to a given action; our estimates instead capture the cumulative change in likelihood of extinction over a large number of species. 

LIFE assumes that a species’ risk of extinction scales non-linearly with its current risk relative to the original population, with a given change in population (estimated via Area of Habitat, or AOH) having a small impact when a population is close to its original size and a much larger effect when a population has already been greatly reduced. However, the exact shape of this curve is not known. 

In Eyres et al. (2025) we explored the impacts of assuming different relationships by calculating LIFE for exponentials of 0.1, 0.5, and 1.0 as well as for a modified Gompertz curve (which allows for the disproportionate impact of stochasticity when a very small proportion of population remains). The resulting LIFE maps showed broadly similar patterns and are available to users.

As a starting point we recommend using LIFE layers computed using a curve with an exponential of z = 0.25 (based on Thomas et al. (2004)). However it is important to understand the consequences of choosing a particular curve. If a linear curve (z = 1) is used and the assumption of a linear relationship is inappropriate, it risks underestimating the impact of losing the remaining areas of habitat or restoring potential habitat. Conversely, LIFE layers calculated with z = 0.1 or the modified Gompertz curve make remaining habitat much more important for species which have already lost most of their area of habitat.

The reason LIFE values are typically small is two-fold: first, because values represent marginal changes in extinction risk (a single cell at a time) and second, because extinction risk values are divided by the number of species considered (roughly 30,000). It is usually easiest to inspect the order of magnitude and direction (positive or negative) of LIFE scores rather than trying to interpret the particular values.

The real power of the LIFE metric lies in making comparisons—such as assessing which protected site has greater potential to prevent species extinctions, or whether eating meat has a larger impact on biodiversity loss than eating cheese.

LIFE values quantify the change in extinction risk associated with the conversion or restoration of 1 km² of habitat across terrestrial vertebrate species. To make the metric comparable and stable, we standardise these values by dividing by the total number of species assessed globally (roughly 30,000). This ensures that LIFE is independent of the number of species included and remains robust as new species are added. Ultimately, LIFE represents the per-species change in the probability of extinction resulting from 1 km² of land cover change—a measure we refer to as the change in extinction risk.

LIFE-convert and LIFE-restore maps from Eyres et al. (2025) are available in the WGS84 coordinate system at a resolution of 1 arc-minute (approximately 1.8 km × 1.8 km at the equator). An additional LIFE-restore map is provided at 5 arc-minute resolution (approximately 9 km × 9 km at the equator), matching the spatial scale of crop distribution data from GAEZ.

Custom LIFE layers can also be generated using our code, allowing users to specify different resolutions or projections. However, we advise against using resolutions finer than 1 km² as the underlying range maps are relatively coarse estimates of species distributions, and overly fine resolutions may create a misleading impression of precision.

LIFE is calculated by summing across species and so can be aggregated into different taxonomic groups or disaggregated, in principle even to the level of individual species. Published LIFE layers are aggregated across all terrestrial vertebrates as well as for the four taxa separately (amphibians, birds, mammals, and reptiles).

The original LIFE layers published in Eyres et al. (2025) focused on two land-cover change scenarios: conversion to arable land (LIFE-convert) and restoration of non-natural habitats (LIFE-restore). We have since added further LIFE-convert layers for other key changes, including conversion to urban areas and pasture. LIFE-convert can be used to assess the impacts of land-cover change or the benefits of preventing conversion. It is important to choose the layer that matches the likely land-cover change in a region, as species respond differently to various types of habitat conversion.

The LIFE-restore layer estimates the benefits of habitat restoration or the opportunity costs of maintaining current land use, reflecting the foregone reduction in extinction risk. Ball et al. (2025) adapted this layer using data from the GAEZ project to assess opportunity costs related to food production. GAEZ provides highly accurate global crop distribution data. If your analysis needs to align with crop-specific data—for example, in agricultural land-use studies—use the Ball et al. (2025) version. Otherwise, for broader restoration analyses, the original LIFE-restore layer from Eyres et al. (2025) may be preferable due to its finer resolution and compatibility with LIFE-convert layers.

LIFE layers can be used to assess the impact of actions at spatial scales larger or smaller than the native map resolution, without requiring extensive new analyses. This is possible despite the non-linear relationship between a species’ remaining area of habitat and its probability of extinction. We have tested the scalability of our maps and found that they can be used to estimate the approximate per km² impact on extinctions for actions affecting areas from 0.5 km² up to 1,000 km² in size (with less than 10% error) by simply summing cell values. However, if users generate their own layers at different spatial resolutions, this scalability would need to be reassessed.

LIFE data are provided in GeoTIFF format via Zenodo under the terms and conditions of the underlying species' elevation and habitat preference data and distribution polygons as laid out by the IUCN Red List. These data and any derivatives may not be used for commercial or any revenue generating activities without prior written permission from IUCN. All forms of reposting, sub-licensing, reselling or other forms of redistribution of these data in their original format are also prohibited without prior written permission from IUCN. Please refer to the IUCN Red List Terms and Conditions of Use.

Although a variety of LIFE layers exist which offer flexibility it is also likely that users might require bespoke layers to answer specific questions which might require calculation of LIFE layers at a different resolution, projections, for a different curve shape, for different species, or for new land-use change scenarios. All of our code is open accessible on GitHub so users can produce new LIFE layers as required.