Hostname: page-component-5db58dd55d-jnbmb Total loading time: 0 Render date: 2026-06-01T09:50:39.517Z Has data issue: false hasContentIssue false
  • English
  • Français

The risk of underestimating generation length for extinction risk assessments

Published online by Cambridge University Press:  27 April 2026

Giordano Mancini Open the ORCID record for Giordano Mancini [Opens in a new window] ,
Valerio Mezzanotte ,
Luca Santini  and
Moreno Di Marco
Giordano Mancini*
Affiliation:
Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy
Valerio Mezzanotte
Affiliation:
Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy
Luca Santini
Affiliation:
Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy
Moreno Di Marco
Affiliation:
Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy
*
Corresponding author: Giordano Mancini; Email: giordano.mancini@uniroma1.it
Rights & Permissions [Opens in a new window]

Abstract

Generation length is a key parameter for extinction risk assessment in the IUCN Red List, defining the biological horizon over which population decline is measured. Generation length data are largely unavailable, and a naive time frame of 10 years is often used to represent the three-generation horizon required for assessing decline in short-lived species. This assumption risks introducing biases in the assessment. Here, we quantified how using a 10-year time frame, instead of one based on predicted generation lengths, affects the extinction risk assessments of amphibians: one of the most imperiled groups worldwide. We measured past deforestation trends for 1,278 forest-dependent amphibian species and predicted future climate-driven habitat change for 1,559 species. Using predicted generation lengths systematically increased the percentage of declining species from deforestation (3–29%) and climate change (1–20%). This was also reflected in a higher proportion of species potentially assessed as threatened. Relying on a 10-year period to assess extinction risk can lead to substantial underestimations across both past and future time frames, compared to using predicted generation length. Predictions of generation length can allow applying IUCN Red List criteria in the absence of empirical data and improve consistency of extinction risk assessments.

Topics structure

Topic(s)

Human factorsMass extinctionSpecies extinction

Subtopic(s)

Global driversScenarios and forecasts

Keywords

amphibiansbiodiversity lossclimate changedeforestationIUCN Red List

Information

Type
Research Article
Information
Cambridge Prisms: Extinction , Volume 4 , 2026 , e5
Creative Commons
Creative Common License - CCCreative Common License - BYCreative Common License - NC
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-NonCommercial licence (http://creativecommons.org/licenses/by-nc/4.0), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original article is properly cited. The written permission of Cambridge University Press or the rights holder(s) must be obtained prior to any commercial use.
Copyright
© The Author(s), 2026. Published by Cambridge University Press

Impact statement

Accurately estimating species extinction risk is fundamental to conserving biodiversity, and this requires considering appropriate time frames of evaluation (e.g., three generations). Many species in the IUCN Red List are evaluated using simplified time frames, such as 10 years, if data on generation length are unavailable. This may distort extinction risk assessment. In this article, we highlight the importance of using biologically meaningful time frames to estimate species status, using amphibians as a case study. We show that even modest extensions beyond the 10-year period can substantially change extinction risk outcomes, especially in regions experiencing high threat levels (e.g., rapid deforestation) and for long-lived species. Our results emphasize the risks of naive assumptions in the absence of generation length data, highlighting the usefulness of trait-based model predictions to fill knowledge gaps and address bias in extinction risk assessments.

Introduction

Monitoring global trends in extinction risk is crucial to plan effective conservation strategies (Pereira et al., Reference Pereira, Martins, Rosa, Kim, Leadley, Popp, van Vuuren, Hurtt, Quoss, Arneth, Baisero, Bakkenes, Chaplin-Kramer, Chini, Di Marco, Ferrier, Fujimori, Guerra, Harfoot, Harwood, Hasegawa, Haverd, Havlík, Hellweg, Hilbers, Hill, Hirata, Hoskins, Humpenöder, Janse, Jetz, Johnson, Krause, Leclère, Matsui, Meijer, Merow, Obersteiner, Ohashi, De Palma, Poulter, Purvis, Quesada, Rondinini, Schipper, Settele, Sharp, Stehfest, Strassburg, Takahashi, Talluto, Thuiller, Titeux, Visconti, Ware, Wolf and Alkemade2024), and the Red List of the International Union for Conservation of Nature (hereafter, IUCN Red List) represents the most adopted standard for this monitoring (Rodrigues et al., Reference Rodrigues, Pilgrim, Lamoreux, Hoffmann and Brooks2006). The IUCN Red List classifies species into extinction risk categories, according to five quantitative criteria based on species’ distribution, abundance and respective trends (IUCN, 2012). Three of the IUCN Red List criteria (criteria A, C and E) require establishing biologically meaningful time frames over which to measure species decline, based on the species generation length, defined as the average age of parents of the current cohort (i.e., parents of newborn individuals in the population; IUCN Standards and Petitions Committee, 2024). Scaling the time frame by generation length allows standardizing the estimate of population across species with very different life histories (Mace et al., Reference Mace, Collar, Gaston, Hilton-Taylor, Akçakaya, Leader-Williams, Milner-Gulland and Stuart2008), ultimately ensuring consistency and comparability of the extinction risk categories across taxa. IUCN Red List Criterion A in particular measures species decline over three generations or 10 years, whichever is longer (IUCN, 2012). Where estimates of generation length are unavailable, but the species is thought to have a short life history (i.e., short-living species), some assessors might naively use a 10-year time frame for assessing population reductions under criterion A (Edgar, Reference Edgar2025). While such an approach could be convenient in the absence of generation length, it risks introducing bias in extinction risk assessments.

Generation length is poorly recorded across the tree of life, with a limited number of species per taxa for which this parameter is known, hampering the consistent application of IUCN Red List criteria. Tetrapods represent an exception, as generation length is available for almost all birds (Bird et al., Reference Bird, Martin, Re¸sit, Akçakaya, Gilroy, Burfield, Garnett, Symes, Taylor, Gan, Sekercio˘ and Butchart2020), mammals (Pacifici et al., Reference Pacifici, Santini, Di Marco, Baisero, Francucci, Marasini, Visconti and Rondinini2013) and recently for amphibians and reptiles (Mancini et al., Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b). This is reflected in the application of IUCN Red List Criterion A (population decline), which is much more common for birds and mammals than for herptiles: the majority of species in the latter group are classified under Criterion B, which does not require generation length. Amphibians have the shortest generation length among tetrapods (mean = 3.9 years; 58% of species have a generation length >3.5 years; Mancini et al., Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b), which might lead to the assumption that the use of a naive 10-year time frame is appropriate for the application of Criterion A. For example, out of 162 amphibians for which population reductions were recently assessed over a generic 10-year time frame (IUCN, 2025), 32% were estimated to have generation lengths >3.5 years (Mancini et al., Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b), therefore requiring an evaluation period of >10 years.

While the lack of generation length data could be a bottleneck for extinction risk assessments, the attempts to overcome such an issue could introduce further bias and inconsistencies on the IUCN Red List. Different studies have tried to assess species extinction risk using fixed time frames (typically from 50 to 100 years) instead of a generation length time frame as required by the IUCN Red List criteria (IUCN, 2012), especially to measure future climate change impacts (see Akçakaya et al., Reference Akçakaya, Butchart, Mace, Stuart and Hilton-Taylor2006 for a review). These approaches likely overestimate extinction risk for short-lived species, as any population reductions would have been incorrectly assessed over a time frame of more than three generations. Similarly, the application of a 10-year time frame for species with a generation length longer than 3 years can underestimate their extinction risk.

In this study, we show the implications of applying a default naive 10-year time frame for extinction risk assessments. We focused on amphibians, a highly threatened group for which generation length data were largely unavailable and have only recently been predicted statistically. We consistently calculated rates of habitat loss from: (i) past deforestation, to quantify suspected past population reduction under IUCN Red List Criterion A2 and (ii) future climate change, to quantify projected future population reduction under IUCN Red List Criterion A3. We performed our estimates over two alternative time frames: a 10-year naive time frame (which simulates a generic time horizon assumption) and a time frame based on predicted species generation length. We compared habitat loss rates and potential extinction risk categories obtained by the two time frames. We expect that relying on a naive 10-year time frame, instead of a generation-based period, systematically underestimates species extinction risk. Finally, we highlight discrepancies between our predicted Red List categories (based on generation length values) and those recently assessed by the IUCN Red List authorities (IUCN, 2025) for the same species, highlighting species that could be considered for reassessment.

Methods

We estimated the potential extinction risk of amphibians under IUCN Red List Criterion A based on population reductions estimated from past deforestation (Criterion A2) or future climate change (Criterion A3). We applied IUCN Red List criteria A2 and A3 using two time frames over which to measure species decline: a 10-year time frame and one based on three generations. We then compared how the potential extinction risk changed based on different time frames considered.

Deforestation analysis

Species selection and distribution data

For the deforestation analysis, we focused on all strictly forest-dependent amphibian species assessed in the IUCN Red List (Tracewski et al., Reference Tracewski, Butchart, Di Marco, Ficetola, Rondinini, Symes, Wheatley, Beresford and Buchanan2016). Following the IUCN Red List habitat classification scheme, we focused on all species occurring in forest only, excluding those with a secondary habitat (e.g., grassland). We also included species occurring in forests and wetlands, as wetlands are a suitable habitat type for most amphibians, including forest-dependent ones. We excluded species for which generation length data were available and already used in the IUCN Red List assessment and all species classified as Data Deficient. We then focused on all species with available range polygons in the IUCN spatial data repository (i.e., Extent of Occurrence; IUCN, 2012) and with a generation length predicted to be >3.5 years (Mancini et al., Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b), as these are the species for which adopting a 10-year time frame would be incorrect and inconsistent with the IUCN Red List Guidelines. We excluded species with a generation length > 8 years because of the annual forest cover layers’ availability (see below). This resulted in a final sample of 1,278 species.

Quantifying past deforestation trends

We used forest cover data from Hansen et al. (Reference Hansen, Potapov, Moore, Hancher, Turubanova, Tyukavina, Thau, Stehman, Goetz, Loveland, Kommareddy, Egorov, Chini, Justice and Townshend2013), which represent the percentage of forest cover at the global scale. Maps were at an original 30 m resolution from 2000 to 2024 in a WGS84 geographic coordinate system. We aggregated all annual forest cover maps to a 300 m resolution in a Mollweide Equal Area projection; such a resolution is sufficient to assess range-restricted species and coarse enough to control for computational power needed to run the analysis. We retained pixels covered by at least 50% of forest to represent suitable habitat for forest-dependent species (Tracewski et al., Reference Tracewski, Butchart, Di Marco, Ficetola, Rondinini, Symes, Wheatley, Beresford and Buchanan2016). We measured potential habitat changes by calculating the percentage of area within the species range that was lost by 2024. This was based on two historical starting points: 2014 (for the 10-year time frame), and the year representing the three past generations (e.g., 2003 for the past 21 years, if generation length was 7 years).

Climate change analysis

Species selection and distribution data

For the climate change analysis, we considered all amphibian species on the IUCN Red List with available polygon ranges (i.e., Extent of Occurrence), excluding species for which generation length data were available and already used in the IUCN Red List assessment and all species classified as Data Deficient. We then retained all species with a predicted generation length of >3.5 years (Mancini et al., Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b). We also excluded all species with a range size <250 km2 (37 species), because the resolution of climate data was too coarse to assess those species (see below). This resulted in a sample of 3,075 amphibian species retained for modeling.

Species distribution modeling

We used the methodology proposed by Mancini et al. (Reference Mancini, Santini, Cazalis, Akçakaya, Lucas, Brooks, Foden and Di Marco2024) to estimate species decline from future climate change according to IUCN Red List Criterion A3. We fit Species Distribution Models in Maxent, using species’ range polygons to sample occurrences (Visconti et al., Reference Visconti, Bakkenes, Baisero, Brooks, Butchart, Joppa, Alkemade, Di Marco, Santini, Hoffmann, Maiorano, Pressey, Arponen, Boitani, Reside, van Vuuren and Rondinini2016), while generating background points randomly from within the ecoregions (Olson et al., Reference Olson, Dinerstein, Wikramanayake, Burgess, Powell, Underwood, D’Amico, Itoua, Strand, Morrison, Loucks, Allnutt, Ricketts, Kura, Lamoreux, Wettengel, Hedao and Kassem2001) where species occurred, both within and outside the species range; we set a 1:10 proportion for presence: background points. Models were calibrated using the yearly 19 bioclimatic variables (CHELSA; Karger et al., Reference Karger, Schmatz, Dettling and Zimmermann2020), available from 2013 to 2100, that we averaged across four Global Circulation Models (GCMs: ACCESS1-3, CESM1-BGC, CMCC-CM, and MIROC5) under the Representative Concentration Pathway (RCP) 4.5 in a WGS84 geographic coordinate system with a spatial resolution of 0.049° (∼5 km at the equator). To represent climate conditions experienced by the species, we averaged each climate variable as each year reflects the average of the 10 previous years (i.e., the year 2025 corresponds to the average between 2016 and 2025). The Maxent algorithm was trained on present data (i.e., 2025), iteratively removing variables whose variance inflation factor >3 (Zuur et al., Reference Zuur, Ieno and Elphick2010). We did not perform ecological variable selection (Mancini et al., Reference Mancini, Di Marco, Carboni, Cerretti, Maiorano and Santini2025a) as, in this case, we were only interested in comparing the same models under the two time horizons considered. We calibrated the models considering linear and quadratic features and scaling the regularization function iteratively by 0.5, 1, 2 and 10 (Radosavljevic and Anderson, Reference Radosavljevic and Anderson2014). We employed a spatial block validation, dividing the sampled data into four blocks of equal size (as per the BIOMOD_Modelling function) and iteratively used three of the four blocks as training set and the remaining block as a testing set. We assessed model performance using the True Skill Statistic (TSS) and retained all models with TSS ≥0.5. Habitat suitability maps were produced for the present (i.e., 2025), for the 10-year future time frame (i.e., 2035) and for the generation length time frame (e.g., 2046 for a species with a 7-year generation time). Continuous probability predictions were binarized using the threshold that maximized the TSS in the validation.

Calculation of habitat change

We calculated potential habitat changes by considering only the areas classified as suitable in the current species’ range (2025) and measuring the percentage of such areas predicted to be lost in the future (either in 10 years and three generations; Mancini et al., Reference Mancini, Santini, Cazalis, Akçakaya, Lucas, Brooks, Foden and Di Marco2024). For example, when a pixel that was classified as suitable in the present becomes not suitable in the future, it is considered lost. Amphibians lack estimates of dispersal; however, they are generally assumed to have very limited dispersal ability (Alex Smith and Green, Reference Alex Smith and Green2005), which affects their adaptive capacity to climate change. Among the major global databases of amphibian traits, only 20 species had this data reported, and just 7 had dispersal >1,500 m (Trochet et al., Reference Trochet, Moulherat, Calvez, Stevens, Clobert and Schmeller2014). Thus, for the sake of our comparison, we only performed a “no-dispersal” scenario where the currently suitable areas could be lost to climate change, but areas outside the species’ range could not be colonized.

Application of IUCN Red List Criteria A2 and A3

We used the potential habitat change to deforestation to apply IUCN Red List Criterion A2, which refers to “population (size) reduction observed, estimated, inferred, or suspected in the past where the causes of reduction may not have ceased or may not be understood or may not be reversible” (IUCN Standards and Petitions Committee, 2024), and we used the potential habitat change to future climate change to apply IUCN Red List Criterion A3, which refers to “population (size) reduction projected, inferred or suspected to be met in the future (up to a maximum of 100 years),” both based on subcriterion c that refers to “a decline in area of occupancy (AOO), extent of occurrence (EOO) and/or habitat quality.” Both criteria have to be measured over 10 years or three generations (whichever is the longest; IUCN Standards and Petitions Committee, 2024). We considered a linear relationship between habitat change and population size as this is the simplest valid assumption in the absence of species-specific information on population size–habitat relationship (IUCN Standards and Petitions Committee, 2024). While such an assumption might not be true for several species, and therefore for actual Red List assessments, it is still useful for our comparative analysis of generation length. IUCN Red List criteria do not provide a threshold for the Near Threaten category; for this study, we considered a loss of 20% of habitat to classify a species as Near Threaten (Mancini et al., Reference Mancini, Santini, Cazalis, Akçakaya, Lucas, Brooks, Foden and Di Marco2024). We finally compared potential extinction risk categories based on the 10-year time frame with those based on the generation length time frame.

We retrieved all the generation length data from Mancini et al. (Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b), which provide statistical predictions of generation length based on a set of life-history traits, such as body mass, reproductive modes, development type (e.g., paedomorphism vs. direct development) and climate variables known to be related to generation length.

The deforestation analysis was performed using Python 3.8.10 (Python Software Foundation, 2025) with the following libraries: rasterio (Gillies, Reference Gillies2013), numpy (Harris et al., Reference Harris, Millman, van der Walt, Gommers, Virtanen, Cournapeau, Wieser, Taylor, Berg, Smith, Kern, Picus, Hoyer, van Kerkwijk, Brett, Haldane, del Río, Wiebe, Peterson, Gérard-Marchant, Sheppard, Reddy, Weckesser, Abbasi, Gohlke and Oliphant2020), pandas (McKinney, Reference McKinney2010) and geopandas (Jordahl et al., Reference Jordahl, den Bossche, Fleischmann, Wasserman, McBride, Gerard, Tratner, Perry, Badaracco, Farmer, Hjelle, Snow, Cochran, Gillies, Culbertson, Bartos, Eubank, Bilogur, Rey, Ren, Arribas-Bel, Wasser, Wolf, Journois, Wilson, Greenhall, Holdgraf, Filipe and Leblanc2020); the climate change analysis and plots were done using R 4.4.3 (R Core Team, 2025) in RStudio 2023.6.0.421 (Posit Team, 2023), with the following packages: “biomod2” (Thuiller et al., Reference Thuiller, Georges, Gueguen, Engler, Breiner, Lafourcade, Patin and Blancheteau2024), “data.table” (Barrett et al., Reference Barrett, Dowle, Srinivasan, Gorecki, Chirico and Hocking2024), “patchwork” (Pedersen TL, Reference Pedersen2023), “sf” (Pebesma, Reference Pebesma2018), “terra” (Hijmans, Reference Hijmans2024) and “tidyverse” (Wickham et al., Reference Wickham, Averick, Bryan, Chang, McGowan, François, Grolemund, Hayes, Henry, Hester, Kuhn, Pedersen, Miller, Bache, Müller, Ooms, Robinson, Seidel, Spinu, Takahashi, Vaughan, Wilke, Woo and Yutani2019).

Results

Extinction risk from deforestation

Our results for deforestation highlight that estimates of population decline were consistently higher when using the time frame based on predicted generation length instead of a naive 10-year time frame. The discrepancy was higher for species with long generation length. We found that 78 (7%) of all species with a three-generation time of 11–15 years were predicted to lose ≥30% of their habitat (i.e., potentially assessed as threatened), compared to 66 (6%) species under a naive 10-year time frame (Figure 1a). Eight species (8%) with a three-generation time of 16–20 years had a habitat loss of ≥30%, while only three of those had ≥30% loss when considering a 10-year time frame (Figure 1a). Finally, we found two species (29%) with a > 20 year generation time frame for which we predicted ≥30% of habitat loss under the generation time frame, while none qualified as ≥30% loss under the 10-year time frame.

Figure 1.

Potential extinction risk under IUCN Red List Criterion A2 due to past deforestation. (a) Habitat loss relative to different time frames is considered. Percentages, the percentage of species that trigger a category of near threatened (dashed line at -20%) or a threatened category (vulnerable or higher; dashed line at -30%). (b) Percentage of species in each potential extinction risk category considering different time frames and relative to current extinction risk categories (specified in the gray bars). N: number of species; numbers in brackets: number of species moved to a higher risk category when using the species’ generation length time frame, instead of 10 years (e.g., 26 more species classified as Near Threaten or higher when using generations instead of 10 years, among the least concern species). LC, least concern; NT, near threatened; VU, vulnerable; EN, endangered; CR, critically endangered.

Our results also reflect on the number of species that might require a category revision, based on the discrepancy between current versus predicted category (Figure 1b). The number of species potentially at higher risk than currently assessed was always higher when the impact of deforestation was measured over three generations compared to the 10-year time frame. Among species currently listed as Least Concern, we found that 26 more species could be uplisted to Near Threaten or higher categories (average habitat loss 23%), when using the generation length time frame rather than 10 years (average habitat loss 17%) (Figure 1b). Notably, the Southern torrent salamander (Rhyacotriton variegatus) is currently listed as Least Concern, but we estimated a potential 30% habitat loss in the last three generations, which would qualify it as Vulnerable, against a 19% habitat loss in the last 10 years, which would qualify it as Least Concern. Among Near Threaten and threatened species, we found two more species that could be potentially moved to a higher category when using the generation length time frame rather than 10 years.

Extinction risk from climate change

For the climate change analysis, 1,559 species out of 3,075 analyzed (ca. 50%) had models with good validation performance (TSS ≥0.5; Supplementary Appendix S1); therefore, application of IUCN Red List Criterion A3 was limited to this subset of species. Similar to the deforestation analysis, assuming a 10-year projection time frame underestimated population decline and influenced the classification into extinction risk categories, especially for species with long generation lengths. For species with short generation time frames (three generations = 11–15 years), we found little difference in habitat change compared to 10-year projections, with only 1% (n = 8) more species predicted to lose ≥30% of habitat (Figure 2a). Effects were higher for species with longer generations, for example, 23% (n = 33) species with a generation time frame of 16–20 years had a projected habitat loss of ≥30% compared to 15% (n = 22) species under a 10-year time frame.

Figure 2.

Potential extinction risk under IUCN Red List Criterion A3 due to future climate change. (a) Habitat loss relative to different time frames is considered. Percentages, the percentage of species that trigger a category of near threatened (dashed line at -20%) or a threatened category (vulnerable or higher; dashed line at -30%). (b) Percentage of species in each potential extinction risk category considering different time frames and relative to current extinction risk categories (specified in the gray bars). N: number of species; numbers in brackets: number of species moved to a higher risk category when using the species’ generation length time frame, instead of 10 years. LC, least concern; NT, near threatened; VU, vulnerable; EN, endangered; CR, critically endangered.

Once again, we found a large number of species having a higher potential category of risk than currently assessed (Figure 2b), even if in this case the difference between the generation time projection and the 10 year projection was less stark compared to the deforestation analysis. Among species currently classified as Least Concern, 177 were predicted to be near threatened or threatened using the generation length projection (mean habitat loss 39.5%) versus 164 species under the 10-year projection (mean habitat loss 40%). Among species already classified as near threatened or threatened, 76 were predicted into higher categories of risk when using the generation time projection versus 70 when using a 10-year time frame; the only difference in this case (six species) was for species currently assessed as Vulnerable.

Finally, we found 743 species that were common to both the deforestation and climate change analyses. This sample of species did not show major differences in habitat loss between the 10-year projection and the generation-time projection, with only 1% more species predicted to be at risk when the generation-length time frame was applied in both analyses (Supplementary Figure S1).

Discussion

Overall, we found that a naive use of a 10-year time frame for calculating habitat loss could lead to the underestimation of extinction risk compared to using predicted generation length time frames. This underestimation was less severe for species with a generation time between 3.5 and 5 years (i.e., decline measured over 11–15 years), which is often the case for amphibians. However, longer generation species and those occurring in highly impacted regions could face much higher underestimation of risk. Importantly, we show that generation length predictions could allow consistent application of IUCN Red List criteria and that many species could be at higher risk than previously thought.

We showed that measuring past deforestation over time frames slightly longer than 10 years (e.g., 12 years) could increase the number of species potentially at risk of extinction. Longer time frames could be crucial for assessing the impact of past drivers of extinction, such as historical deforestation, which has increased due to human activities (Hansen et al., Reference Hansen, Potapov, Moore, Hancher, Turubanova, Tyukavina, Thau, Stehman, Goetz, Loveland, Kommareddy, Egorov, Chini, Justice and Townshend2013; Betts et al., Reference Betts, Wolf, Ripple, Phalan, Millers, Duarte, Butchart and Levi2017; Wiebe and Wilcove, Reference Wiebe and Wilcove2025). For example, we found that the White-lipped Frog (Chalcorana labialis), a species occurring in Malaysia and Singapore, currently classified as Least Concern, potentially lost 25% of its habitat in the past 10 years and 32% of its habitat in the past 13 years (using the generation length time frame). In just 3 years, this species experienced an additional 7% of habitat loss, which triggered the threshold for Vulnerable category, instead of a potential near threaten if measured across 10 years. This is expected for species occurring in highly deforested regions such as tropical forests (Betts et al., Reference Betts, Wolf, Ripple, Phalan, Millers, Duarte, Butchart and Levi2017), particularly in South East Asia (Wilcove et al., Reference Wilcove, Giam, Edwards, Fisher and Koh2013), where the annual trend of deforestation could be dramatically high (Curtis et al., Reference Curtis, Slay, Harris, Tyukavina and Hansen2018). In contexts of exceptionally intense drivers of habitat loss, measuring the impact of threats even across a few more years could be important to distinguish threatened from non-threatened species.

In our climate change analysis, we found similar risk predictions when habitat loss was projected over similar short time frames, that is, a naive 10 year interval versus 11–15 year intervals. This confirms that across short periods of time the difference in climate conditions would be minimal, except for areas of exceptionally high climate exposure. According to future scenarios, the most dramatic changes in climate conditions are predicted to occur beyond the decade of 2040 (Tebaldi et al., Reference Tebaldi, Debeire, Eyring, Fischer, Fyfe, Friedlingstein, Knutti, Lowe, O’Neill, Sanderson, van Vuuren, Riahi, Meinshausen, Nicholls, Tokarska, Hurtt, Kriegler, Lamarque, Meehl, Moss, Bauer, Boucher, Brovkin, Byun, Dix, Gualdi, Guo, John, Kharin, Kim, Koshiro, Ma, Olivié, Panickal, Qiao, Rong, Rosenbloom, Schupfner, Séférian, Sellar, Semmler, Shi, Song, Steger, Stouffer, Swart, Tachiiri, Tang, Tatebe, Voldoire, Volodin, Wyser, Xin, Yang, Yu and Ziehn2021). In fact, when habitat change was predicted over longer periods (>15 years), which also overlaps with the decades in which major climate changes are predicted to occur, we found many more species potentially at risk. This is in line with previous studies assuming that species with slow life histories, indicated by long generation lengths, are generally more vulnerable to climate change because of their increased climate exposure associated with slow recovery rates (Cayuela et al., Reference Cayuela, Monod-Broca, Lemaître, Besnard, Gippet, Schmidt, Romano, Hertach, Angelini, Canessa, Rosa, Vignoli, Venchi, Carafa, Giachi, Tiberi, Hantzschmann, Sinsch, Tournier, Bonnaire, Gollmann, Gollmann, Spitzen-van der Sluijs, Buschmann, Kinet, Laudelout, Fonters, Bunz, Corail, Biancardi, Di Cerbo, Langlois, Thirion, Bernard, Boussiquault, Doré, Leclerc, Enderlin, Laurenceau, Morin, Skrzyniarz, Barrioz, Morizet, Cruickshank, Pichenot, Maletzky, Delsinne, Henseler, Aumaître, Gailledrat, Moquet, Veen, Krijnen, Rivière, Trenti, Endrizzi, Pedrini, Biaggini, Vanni, Dudgeon, Gaillard and Léna2022) and lower adaptability (Foden et al., Reference Foden, Young, Akçakaya, Garcia, Hoffmann, Stein, Thomas, Wheatley, Bickford, Carr, Hole, Martin, Pacifici, Pearce-Higgins, Platts, Visconti, Watson and Huntley2019). For such species, a biologically meaningful time frame to estimate future climate change impact can be very important compared to a naive 10-year time frame. In particular, such information is crucial for taxa assumed to have very limited dispersal ability, such as amphibians (Alex Smith & Green Reference Alex Smith and Green2005), as dispersal is used to account for their adaptive capacity to a changing climate (Bateman et al., Reference Bateman, Murphy, Reside, Mokany and Vanderwal2013; Santini et al., Reference Santini, Cornulier, Bullock, Palmer, White, Hodgson, Bocedi and Travis2016). For example, we predicted a 20% habitat loss from climate change in 10 years for the Chuanan Short-legged Toad (Megophrys chuannanensis), currently assessed as Near Threatened; however, when adjusting the projection time frame to 21 years (based on the predicted 7-year generation length), the habitat loss increased to 58%, potentially qualifying it as Endangered. Overall, our results emphasize that using short time frames to predict climate impact likely underestimates extinction risk for long-living species.

Presenting an extensive application of IUCN Red List criteria A2 and A3 on amphibian species for the first time, we highlight that many of them could be at higher risk than currently classified. More than 20% of the non-threatened species analyzed in this study could be at risk of extinction due to deforestation or climate change. Importantly, such threats might act synergistically (Brook et al., Reference Brook, Sodhi and Bradshaw2008). Synergies between threats mean, for example, that climate change might exacerbate the impact of habitat loss on animal populations (Mantyka-Pringle et al., Reference Mantyka-Pringle, Visconti, Di Marco, Martin, Rondinini and Rhodes2015). We identified 743 species in common between our criterion A2 and A3 analyses. These were species with short estimated generation time (median 3.9 years); therefore, they showed smaller changes in predicted habitat loss when the two time frames were used, compared to the full species sample. However, 20 species showed more than 20% habitat loss in both past deforestation and climate change. Although the two criteria are independent and cannot be cumulated, half of those species are currently listed as Least Concern, and such species could be considered for a reassessment as they might be at higher risk. For example, the Tschenk’s Madagascar frog (Gephyromantis tschenki), classified as Least Concern in 2016 (IUCN SSC Amphibian Specialist Group, 2016), could be considered for reassessment into a threatened category as it experienced a 21% habitat loss within its range in the past three generations, and we predicted a 31% of potential habitat loss due to future climate change.

Our analyses are not exempt from limitations. Some of our methodological choices deviate from IUCN Red List guidelines, especially for the climate change analysis; thus, our results should not be considered actual extinction risk assessments. Our analysis should be complemented by the species-specific expert knowledge of IUCN Red List assessors in terms of ecologically relevant variables to consider in distribution modeling (Mancini et al., Reference Mancini, Di Marco, Carboni, Cerretti, Maiorano and Santini2025a). Neglecting important variables associated with species habitat suitability or physiology, such as the frequency of extreme events (Pottier et al., Reference Pottier, Kearney, Wu, Gunderson, Rej, Rivera-Villanueva, Pollo, Burke, Drobniak and Nakagawa2025), could affect the estimation of the species climatic niche. Additionally, we used occurrence points sampled from species polygons, which can provide robust projections of species responses to climate (Lawler et al., Reference Lawler, Shafer, White, Kareiva, Maurer, Blaustein and Bartlein2009; Visconti et al., Reference Visconti, Bakkenes, Baisero, Brooks, Butchart, Joppa, Alkemade, Di Marco, Santini, Hoffmann, Maiorano, Pressey, Arponen, Boitani, Reside, van Vuuren and Rondinini2016) but could suffer from inherent commission errors of polygons (Ficetola et al., Reference Ficetola, Rondinini, Bonardi, Katariya, Padoa-Schioppa and Angulo2014). Due to the lack of accurate dispersal data for amphibians, we only consider a no-dispersal scenario, which might provide higher extinction risk estimates compared to analyses that consider dispersal (Mancini et al., Reference Mancini, Santini, Cazalis, Akçakaya, Lucas, Brooks, Foden and Di Marco2024). This is a reasonable assumption for amphibians at the scale of our analysis, as amphibians are poor dispersers in general.

In this study, we show how assuming a 10-year time frame in species with unknown generation length can bias extinction risk assessments. While such an assumption allows assessors to overcome data limitations, it can lead to a consistent underestimation of extinction risk depending on the threat and the life history of the species under consideration. The development and application of trait-based models (e.g., Mancini et al., Reference Mancini, Santini, Cazalis, Ficetola, Meiri, Roll, Silvestri, Pincheira-Donoso and Di Marco2025b) can handle data limitations, enhancing comparability across taxa. Applying for the first time the IUCN Red List Criterion A on many amphibian species, we showed that many of them are potentially at risk due to deforestation or climate change (or both) and should therefore be considered for reassessment. New data and tools are increasingly available to help IUCN Red List assessors perform first assessments (Cazalis et al., Reference Cazalis, Di Marco, Zizka, Butchart, González-Suárez, Böhm, Bachman, Hoffmann, Rosati, De Leo, Jung, Benítez-López, Clausnitzer, Cardoso, Brooks, Mancini, Lucas, Young, Akçakaya, Schipper, Hilton-Taylor, Pacifici, Meyer and Santini2024) and identify reassessment priorities (Cazalis et al., Reference Cazalis, Santini, Lucas, González-Suárez, Hoffmann, Benítez-López, Pacifici, Schipper, Böhm, Zizka, Clausnitzer, Meyer, Jung, Butchart, Cardoso, Mancini, Akçakaya, Young, Patoine and Di Marco2023; Henry et al., Reference Henry, Santini, Butchart, González-Suárez, Lucas, Benítez-López, Mancini, Jung, Cardoso, Zizka, Meyer, Akçakaya, Berryman, Cazalis and Di Marco2024; Lucas et al., Reference Lucas, Di Marco, Cazalis, Luedtke, Neam, Brown, Langhammer, Mancini and Santini2024) contributing to a consistent application of the IUCN Red List criteria. The IUCN Red List remains a cornerstone for conserving biodiversity, and consistent and reliable extinction risk assessments are fundamental to tackling the current biodiversity crisis.

Open peer review

To view the open peer review materials for this article, please visit http://doi.org/10.1017/ext.2026.10012.

Supplementary material

The supplementary material for this article can be found at http://doi.org/10.1017/ext.2026.10012.

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article (and/or its Supplementary Materials).

Author contribution

Conceptualization: G.M., L.S., M.D.M.; Data Curation: G.M., Formal Analysis: G.M., V.M.; Investigation: G.M., Methodology: G.M., V.M., L.S., M.D.M.; Supervision: M.D.M.; Visualization: G.M., V.M.; Writing – Original Draft: G.M., V.M.; Writing – Review and Editing: G.M., V.M., L.S., M.D.M.

Financial support

G.M. and M.D.M. acknowledge support from the project GaP, funded by Biodiversa+, the European Biodiversity Partnership under the 2021–2022 BiodivProtect joint call for research proposals, co-funded by the European Commission (GA N°101,052,342) and the Italian Ministry of University and Research (CUP B83C23001960001).

Competing interests

The authors declare none.

References

Akçakaya, HR, Butchart, SHM, Mace, GM, Stuart, SN and Hilton-Taylor, C (2006) Use and misuse of the IUCN red list criteria in projecting climate change impacts on biodiversity. Global Change Biology 12(11), 2037–2043. https://doi.org/10.1111/j.1365-2486.2006.01253.x.Google Scholar
Alex Smith, M and Green, DM (2005) Dispersal and the metapopulation paradigm in amphibsian ecology and conservation: Are all amphibian populations metapopulations? Ecography 28(1), 110–128. https://doi.org/10.1111/j.0906-7590.2005.04042.x.Google Scholar
Barrett, T, Dowle, M, Srinivasan, A, Gorecki, J, Chirico, M and Hocking, T (2024) Data.Table: Extension of ‘Data.Frame’. Available at https://CRAN.R-project.org/package=data.table.Google Scholar
Bateman, BL, Murphy, HT, Reside, AE, Mokany, K and Vanderwal, J (2013) Appropriateness of full-, partial-and no-dispersal scenarios in climate change impact modelling. Diversity and Distributions 19(10), 1224–1234. https://doi.org/10.1111/ddi.12107.Google Scholar
Betts, MG, Wolf, C, Ripple, WJ, Phalan, B, Millers, KA, Duarte, A, Butchart, SHM and Levi, T (2017) Global forest loss disproportionately erodes biodiversity in intact landscapes. Nature 547(7664), 441–444. https://doi.org/10.1038/nature23285.Google Scholar
Bird, JP, Martin, R, Re¸sit, H, Akçakaya, R, Gilroy, J, Burfield, IJ, Garnett, ST, Symes, A, Taylor, J, Gan, Ç, Sekercio˘, H and Butchart, SHM (2020) Generation lengths of the world’s birds and their implications for extinction risk. Conservation Biology 34(5), 1252–1261. https://doi.org/10.1111/cobi.13486.Google Scholar
Brook, BW, Sodhi, NS and Bradshaw, CJA (2008) Synergies among extinction drivers under global change. Trends in Ecology and Evolution 23(8), 453–460. https://doi.org/10.1016/j.tree.2008.03.011.Google Scholar
Cayuela, H, Monod-Broca, B, Lemaître, J-F, Besnard, A, Gippet, JMW, Schmidt, BR, Romano, A, Hertach, T, Angelini, C, Canessa, S, Rosa, G, Vignoli, L, Venchi, A, Carafa, M, Giachi, F, Tiberi, A, Hantzschmann, AM, Sinsch, U, Tournier, E, Bonnaire, E, Gollmann, G, Gollmann, B, Spitzen-van der Sluijs, A, Buschmann, H, Kinet, T, Laudelout, A, Fonters, R, Bunz, Y, Corail, M, Biancardi, C, Di Cerbo, AR, Langlois, D, Thirion, J-M, Bernard, L, Boussiquault, E, Doré, F, Leclerc, T, Enderlin, N, Laurenceau, F, Morin, L, Skrzyniarz, M, Barrioz, M, Morizet, Y, Cruickshank, SS, Pichenot, J, Maletzky, A, Delsinne, T, Henseler, D, Aumaître, D, Gailledrat, M, Moquet, J, Veen, R, Krijnen, P, Rivière, L, Trenti, M, Endrizzi, S, Pedrini, P, Biaggini, M, Vanni, S, Dudgeon, D, Gaillard, J-M and Léna, J-P (2022) Compensatory recruitment allows amphibian population persistence in anthropogenic habitats. Proceedings of the National Academy of Sciences 119(38), e2206805119. https://doi.org/10.1073/pnas.2206805119.Google Scholar
Cazalis, V, Di Marco, M, Zizka, A, Butchart, SHM, González-Suárez, M, Böhm, M, Bachman, SP, Hoffmann, M, Rosati, I, De Leo, F, Jung, M, Benítez-López, A, Clausnitzer, V, Cardoso, P, Brooks, TM, Mancini, G, Lucas, PM, Young, BE, Akçakaya, HR, Schipper, AM, Hilton-Taylor, C, Pacifici, M, Meyer, C and Santini, L (2024) Accelerating and standardising IUCN red list assessments with sRedList. Biological Conservation 298, 110761. https://doi.org/10.1016/j.biocon.2024.110761.Google Scholar
Cazalis, V, Santini, L, Lucas, PM, González-Suárez, M, Hoffmann, M, Benítez-López, A, Pacifici, M, Schipper, AM, Böhm, M, Zizka, A, Clausnitzer, V, Meyer, C, Jung, M, Butchart, SHM, Cardoso, P, Mancini, G, Akçakaya, HR, Young, BE, Patoine, G and Di Marco, M (2023) Prioritizing the reassessment of data-deficient species on the IUCN red list. Conservation Biology 37, e14139. https://doi.org/10.1111/cobi.14139.Google Scholar
Curtis, PG, Slay, CM, Harris, NL, Tyukavina, A and Hansen, MC (2018) Classifying drivers of global forest loss. Science 361(6407), 1108–1111. https://doi.org/10.1126/science.aau3445.Google Scholar
Edgar, GJ (2025) IUCN red list criteria fail to recognise most threatened and extinct species. Biological Conservation 301, 110880. https://doi.org/10.1016/j.biocon.2024.110880.Google Scholar
Ficetola, GF, Rondinini, C, Bonardi, A, Katariya, V, Padoa-Schioppa, E and Angulo, A (2014) An evaluation of the robustness of global amphibian range maps. Journal of Biogeography 41(2), 211–221. https://doi.org/10.1111/jbi.12206.Google Scholar
Foden, WB, Young, BE, Akçakaya, HR, Garcia, RA, Hoffmann, AA, Stein, BA, Thomas, CD, Wheatley, CJ, Bickford, D, Carr, JA, Hole, DG, Martin, TG, Pacifici, M, Pearce-Higgins, JW, Platts, PJ, Visconti, P, Watson, JEM and Huntley, B (2019) Climate change vulnerability assessment of species. WIREs Climate Change 10(1), 1–36. https://doi.org/10.1002/wcc.551.Google Scholar
Gillies, S (2013) Rasterio: geospatial raster I/O for Python programmers. Mapbox. Available at https://github.com/mapbox/rasterio.Google Scholar
Hansen, MC, Potapov, PV, Moore, R, Hancher, M, Turubanova, SA, Tyukavina, A, Thau, D, Stehman, SV, Goetz, SJ, Loveland, TR, Kommareddy, A, Egorov, A, Chini, L, Justice, CO and Townshend, JRG (2013) High-resolution global maps of 21st-century forest cover change. Science 342(6160), 850–853. https://doi.org/10.1126/science.1244693.Google Scholar
Harris, CR, Millman, KJ, van der Walt, SJ, Gommers, R, Virtanen, P, Cournapeau, D, Wieser, E, Taylor, J, Berg, S, Smith, NJ, Kern, R, Picus, M, Hoyer, S, van Kerkwijk, MH, Brett, M, Haldane, A, del Río, JF, Wiebe, M, Peterson, P, Gérard-Marchant, P, Sheppard, K, Reddy, T, Weckesser, W, Abbasi, H, Gohlke, C and Oliphant, TE (2020) Array programming with NumPy. Nature 585(7825), 357–362. https://doi.org/10.1038/s41586-020-2649-2.Google Scholar
Henry, EG, Santini, L, Butchart, SHM, González-Suárez, M, Lucas, PM, Benítez-López, A, Mancini, G, Jung, M, Cardoso, P, Zizka, A, Meyer, C, Akçakaya, HR, Berryman, AJ, Cazalis, V and Di Marco, M (2024) Modelling the probability of meeting IUCN red list criteria to support reassessments. Global Change Biology 30(1), e17119. https://doi.org/10.1111/gcb.17119.Google Scholar
Hijmans, RJ (2024) terra: Spatial Data Analysis. Available at https://CRAN.R-project.org/package=terra.Google Scholar
IUCN (2012) IUCN Red List Categories and Criteria: Version 3.1. Available at https://portals.iucn.org/library/node/10315.Google Scholar
IUCN (2025) The IUCN Red List of Threatened Species. Version 2025-2. Available at https://www.iucnredlist.org (accessed 19 November 2025).Google Scholar
IUCN SSC Amphibian Specialist Group (2016) Gephyromantis tschenki. The IUCN Red List of Threatened Species 2016, e.T57529A67382718. https://doi.org/10.2305/IUCN.UK.2016-1.RLTS.T57529A67382718.en.Google Scholar
IUCN Standards and Petitions Committee (2024) Guidelines for Using the IUCN Red List Categories and Criteria. Version 16. Available at https://nc.iucnredlist.org/redlist/content/attachment_files/RedListGuidelines.pdf.Google Scholar
Jordahl, K, den Bossche, JV, Fleischmann, M, Wasserman, J, McBride, J, Gerard, J, Tratner, J, Perry, M, Badaracco, AG, Farmer, C, Hjelle, GA, Snow, AD, Cochran, M, Gillies, S, Culbertson, L, Bartos, M, Eubank, N, maxalbert, Bilogur, A, Rey, S, Ren, C, Arribas-Bel, D, Wasser, L, Wolf, LJ, Journois, M, Wilson, J, Greenhall, A, Holdgraf, C, Filipe, and Leblanc, F (2020) geopandas/geopandas: v0.8.1. (Version v0.8.1). Zenodo. https://doi.org/10.5281/zenodo.3946761.Google Scholar
Karger, DN, Schmatz, DR, Dettling, G and Zimmermann, NE (2020) High-resolution monthly precipitation and temperature time series from 2006 to 2100. Scientific Data 7(1), 1–10. https://doi.org/10.1038/s41597-020-00587-y.Google Scholar
Lawler, JJ, Shafer, SL, White, D, Kareiva, P, Maurer, EP, Blaustein, AR and Bartlein, PJ (2009) Projected climate-induced faunal change in the Western hemisphere. Ecology 90(3), 588–597. https://doi.org/10.1890/08-0823.1.Google Scholar
Lucas, PM, Di Marco, M, Cazalis, V, Luedtke, J, Neam, K, Brown, MH, Langhammer, PF, Mancini, G and Santini, L (2024) Using comparative extinction risk analysis to prioritize the IUCN red list reassessments of amphibians. Conservation Biology, e14316. https://doi.org/10.1111/cobi.14316.Google Scholar
Mace, GM, Collar, NJ, Gaston, KJ, Hilton-Taylor, C, Akçakaya, HR, Leader-Williams, N, Milner-Gulland, E j and Stuart, SN (2008) Quantification of extinction risk: IUCN’S system for classifying threatened species. Conservation Biology 22(6), 1424–1442. https://doi.org/10.1111/j.1523-1739.2008.01044.x.Google Scholar
Mancini, G, Di Marco, M, Carboni, M, Cerretti, P, Maiorano, L and Santini, L (2025a) On the importance of expert-informed variable selection in species distribution modelling. Journal of Biogeography, e70037. https://doi.org/10.1111/jbi.70037.Google Scholar
Mancini, G, Santini, L, Cazalis, V, Akçakaya, HR, Lucas, PM, Brooks, TM, Foden, W and Di Marco, M (2024) A standard approach for including climate change responses in IUCN red list assessments. Conservation Biology, e14227. https://doi.org/10.1111/cobi.14227.Google Scholar
Mancini, G, Santini, L, Cazalis, V, Ficetola, GF, Meiri, S, Roll, U, Silvestri, S, Pincheira-Donoso, D and Di Marco, M (2025b) Generation length of the world’s amphibians and reptiles. Ecography, e07527. https://doi.org/10.1111/ecog.07527.Google Scholar
Mantyka-Pringle, CS, Visconti, P, Di Marco, M, Martin, TG, Rondinini, C and Rhodes, JR (2015) Climate change modifies risk of global biodiversity loss due to land-cover change. Biological Conservation 187, 103–111. https://doi.org/10.1016/j.biocon.2015.04.016.Google Scholar
McKinney, W (2010) Data structures for statistical computing in python. In Proceedings of the 9th Python in Science Conference, 445. Austin, TX, 51–56.Google Scholar
Olson, DM, Dinerstein, E, Wikramanayake, ED, Burgess, ND, Powell, GVN, Underwood, EC, D’Amico, JA, Itoua, I, Strand, HE, Morrison, JC, Loucks, CJ, Allnutt, TF, Ricketts, TH, Kura, Y, Lamoreux, JF, Wettengel, WW, Hedao, P and Kassem, KR (2001) Terrestrial ecoregions of the world: A new map of life on earth. Bioscience 51(11), 933–938. https://doi.org/10.1641/0006-3568(2001)051%255B0933:TEOTWA%255D2.0.CO;2.Google Scholar
Pacifici, M, Santini, L, Di Marco, M, Baisero, D, Francucci, L, Marasini, GG, Visconti, P and Rondinini, C (2013) Generation length for mammals. Nature Conservation 5, 87–94. https://doi.org/10.3897/natureconservation.5.5734.Google Scholar
Pebesma, E (2018) Simple features for R: Standardized support for spatial vector data. The R Journal 10(1), 439–446.Google Scholar
Pedersen, TL (2023) patchwork: The Composer of Plots. R package version 1.1.3. Available at https://CRAN.R-project.org/package=patchwork.Google Scholar
Pereira, HM, Martins, IS, Rosa, IMD, Kim, H, Leadley, P, Popp, A, van Vuuren, DP, Hurtt, G, Quoss, L, Arneth, A, Baisero, D, Bakkenes, M, Chaplin-Kramer, R, Chini, L, Di Marco, M, Ferrier, S, Fujimori, S, Guerra, CA, Harfoot, M, Harwood, TD, Hasegawa, T, Haverd, V, Havlík, P, Hellweg, S, Hilbers, JP, Hill, SLL, Hirata, A, Hoskins, AJ, Humpenöder, F, Janse, JH, Jetz, W, Johnson, JA, Krause, A, Leclère, D, Matsui, T, Meijer, JR, Merow, C, Obersteiner, M, Ohashi, H, De Palma, A, Poulter, B, Purvis, A, Quesada, B, Rondinini, C, Schipper, AM, Settele, J, Sharp, R, Stehfest, E, Strassburg, BBN, Takahashi, K, Talluto, L, Thuiller, W, Titeux, N, Visconti, P, Ware, C, Wolf, F and Alkemade, R (2024) Global trends and scenarios for terrestrial biodiversity and ecosystem services from 1900 to 2050. Science 384(6694), 458–465. https://doi.org/10.1126/science.adn3441.Google Scholar
Posit Team (2023) RStudio: Integrated Development Environment for R.Google Scholar
Pottier, P, Kearney, MR, Wu, NC, Gunderson, AR, Rej, JE, Rivera-Villanueva, AN, Pollo, P, Burke, S, Drobniak, SM and Nakagawa, S (2025) Vulnerability of amphibians to global warming. Nature 639(8056), 954–961. https://doi.org/10.1038/s41586-025-08665-0.Google Scholar
Python Software Foundation (2025) Python Language Reference, version 3.8.10. Available at http://www.python.org.Google Scholar
R Core Team (2025) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. URL https://www.R-project.org/.Google Scholar
Radosavljevic, A and Anderson, RP (2014) Making better maxent models of species distributions: Complexity, overfitting and evaluation. Journal of Biogeography 41(4), 629–643. https://doi.org/10.1111/jbi.12227.Google Scholar
Rodrigues, ASL, Pilgrim, JD, Lamoreux, JF, Hoffmann, M and Brooks, TM (2006) The value of the IUCN red list for conservation. Trends in Ecology & Evolution 21(2), 71–76. https://doi.org/10.1016/J.TREE.2005.10.010.Google Scholar
Santini, L, Cornulier, T, Bullock, JM, Palmer, SCF, White, SM, Hodgson, JA, Bocedi, G and Travis, JMJ (2016) A trait-based approach for predicting species responses to environmental change from sparse data: How well might terrestrial mammals track climate change?. Global Change Biology 22(7), 2415–2424. https://doi.org/10.1111/gcb.13271.Google Scholar
Tebaldi, C, Debeire, K, Eyring, V, Fischer, E, Fyfe, J, Friedlingstein, P, Knutti, R, Lowe, J, O’Neill, B, Sanderson, B, van Vuuren, D, Riahi, K, Meinshausen, M, Nicholls, Z, Tokarska, KB, Hurtt, G, Kriegler, E, Lamarque, J-F, Meehl, G, Moss, R, Bauer, SE, Boucher, O, Brovkin, V, Byun, Y-H, Dix, M, Gualdi, S, Guo, H, John, JG, Kharin, S, Kim, Y, Koshiro, T, Ma, L, Olivié, D, Panickal, S, Qiao, F, Rong, X, Rosenbloom, N, Schupfner, M, Séférian, R, Sellar, A, Semmler, T, Shi, X, Song, Z, Steger, C, Stouffer, R, Swart, N, Tachiiri, K, Tang, Q, Tatebe, H, Voldoire, A, Volodin, E, Wyser, K, Xin, X, Yang, S, Yu, Y and Ziehn, T (2021) Climate model projections from the scenario model intercomparison project (ScenarioMIP) of CMIP6. Earth System Dynamics 12(1), 253–293. https://doi.org/10.5194/esd-12-253-2021.Google Scholar
Thuiller, W, Georges, D, Gueguen, M, Engler, R, Breiner, F, Lafourcade, B, Patin, R and Blancheteau, H (2024) biomod2: Ensemble Platform for Species Distribution Modeling. Available at https://biomodhub.github.io/biomod2/.Google Scholar
Tracewski, Ł, Butchart, SHM, Di Marco, M, Ficetola, GF, Rondinini, C, Symes, A, Wheatley, H, Beresford, AE and Buchanan, GM (2016) Toward quantification of the impact of 21st-century deforestation on the extinction risk of terrestrial vertebrates. Conservation Biology 30(5), 1070–1079. https://doi.org/10.1111/cobi.12715.Google Scholar
Trochet, A, Moulherat, S, Calvez, O, Stevens, VM, Clobert, J and Schmeller, DS (2014) A database of life-history traits of European amphibians. Biodiversity Data Journal 2, e4123. https://doi.org/10.3897/BDJ.2.e4123.Google Scholar
Visconti, P, Bakkenes, M, Baisero, D, Brooks, T, Butchart, SHM, Joppa, L, Alkemade, R, Di Marco, M, Santini, L, Hoffmann, M, Maiorano, L, Pressey, RL, Arponen, A, Boitani, L, Reside, AE, van Vuuren, DP and Rondinini, C (2016) Projecting global biodiversity indicators under future development scenarios. Conservation Letters 9(1), 5–13. https://doi.org/10.1111/conl.12159.Google Scholar
Wickham, H, Averick, M, Bryan, J, Chang, W, McGowan, L, François, R, Grolemund, G, Hayes, A, Henry, L, Hester, J, Kuhn, M, Pedersen, T, Miller, E, Bache, S, Müller, K, Ooms, J, Robinson, D, Seidel, D, Spinu, V, Takahashi, K, Vaughan, D, Wilke, C, Woo, K and Yutani, H (2019) Welcome to the Tidyverse. Journal of Open Source Software 4(43), 1686. https://doi.org/10.21105/joss.01686.Google Scholar
Wiebe, RA and Wilcove, DS (2025) Global biodiversity loss from outsourced deforestation. Nature 639(8054), 389–394. https://doi.org/10.1038/s41586-024-08569-5.Google Scholar
Wilcove, DS, Giam, X, Edwards, DP, Fisher, B and Koh, LP (2013) Navjot’s nightmare revisited: Logging, agriculture, and biodiversity in Southeast Asia. Trends in Ecology & Evolution 28(9), 531–540. https://doi.org/10.1016/j.tree.2013.04.005.Google Scholar
Zuur, AF, Ieno, EN and Elphick, CS (2010) A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1(1), 3–14. https://doi.org/10.1111/j.2041-210X.2009.00001.x.Google Scholar
Figure 0

Figure 1. Potential extinction risk under IUCN Red List Criterion A2 due to past deforestation. (a) Habitat loss relative to different time frames is considered. Percentages, the percentage of species that trigger a category of near threatened (dashed line at -20%) or a threatened category (vulnerable or higher; dashed line at -30%). (b) Percentage of species in each potential extinction risk category considering different time frames and relative to current extinction risk categories (specified in the gray bars). N: number of species; numbers in brackets: number of species moved to a higher risk category when using the species’ generation length time frame, instead of 10 years (e.g., 26 more species classified as Near Threaten or higher when using generations instead of 10 years, among the least concern species). LC, least concern; NT, near threatened; VU, vulnerable; EN, endangered; CR, critically endangered.

Figure 1

Figure 2. Potential extinction risk under IUCN Red List Criterion A3 due to future climate change. (a) Habitat loss relative to different time frames is considered. Percentages, the percentage of species that trigger a category of near threatened (dashed line at -20%) or a threatened category (vulnerable or higher; dashed line at -30%). (b) Percentage of species in each potential extinction risk category considering different time frames and relative to current extinction risk categories (specified in the gray bars). N: number of species; numbers in brackets: number of species moved to a higher risk category when using the species’ generation length time frame, instead of 10 years. LC, least concern; NT, near threatened; VU, vulnerable; EN, endangered; CR, critically endangered.

Supplementary material: File

Mancini et al. supplementary material

Mancini et al. supplementary material
Download Mancini et al. supplementary material(File)
File 1.8 MB

Author comment: The risk of underestimating generation length for extinction risk assessments — R0/PR1

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr1 [Opens in a new window]
Giordano Mancini [Opens in a new window] Department of Biology and Biotechnologies "Charles Darwin", University of Rome La Sapienza, Italy
Revision round: 0
Role: author

Comments

Dr. John Alroy, Prof. Barry Brook

Cambridge Prisms: Extinction - Editors-in-Chief

The peril of using naive generation lengths to assess species extinction risk

Dear Dr. John Alroy, Prof. Barry Brook,

We are pleased to follow up upon your invitation to submit a manuscript for consideration in Cambridge Prisms: Extinction, and we attach here our work for consideration as a potential Research Article:

“The peril of using naive generation lengths to assess species extinction risk” by Giordano Mancini, Valerio Mezzanotte, Luca Santini and Moreno Di Marco.

Generation length is a key parameter for extinction risk assessment in the IUCN Red List, used to scale the period over which to measure trends in species populations. Generation length is used as a biologically meaningful time frame to assess and compare the extinction risk of species with very different positioning in the slow-fast continuum. Yet, this important information is not always available to IUCN Red List assessors. In the absence of this information, a naive standard 10-year time frame is often used to represent three generations of a species, especially for species assumed to have short life histories like amphibians. This practice can introduce inconsistencies on the IUCN Red List and underestimate species extinction risk. Predictions of generation length for amphibians have recently been published [Mancini et al. (2025) Ecography, https://doi.org/10.1111/ecog.07527], which allows us to estimate the potential implications of assumed vs predicted generation length data. In this study, we compared the extinction risk of amphibians affected by deforestation and climate change, applying IUCN Red List Criteria A2 and A3 (respectively) using a naive 10-year time frame vs a time frame based on predicted generation length.

We found that assessing the threat status of species over a naive 10-year time frame, instead of a 3-generation time frame, consistently underestimates extinction risk of amphibians. Particularly, we found that time frames even slightly longer than 10 years could strongly increase species extinction risk due to deforestation in highly deforested areas. Instead, the extinction risk due to future climate change was similar when measured over short periods of time, but increased when measured over longer periods, which correspond to predicted intensification of this threat.

Our results warn against the use of a naive 10-year time frame for IUCN Red List extinction risk assessments, emphasizing that this practice likely underestimates species extinction risk when looking at both past and future threats. We believe this study aligns well with the goals of Cambridge Prisms: Extinction, being of theoretical interest in the field of extinction risk assessments and with important practical implications.

We confirm that all authors approved the manuscript of this submission and that the manuscript is not under consideration by any other journal.

Yours sincerely,

Giordano Mancini, on behalf of all co-authors.

Review: The risk of underestimating generation length for extinction risk assessments — R0/PR2

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr2 [Opens in a new window]
Date of review:  24 December 2025
Revision round: 0
Role: reviewer
Recommendation/decision: major-revision

Conflict of interest statement

Reviewer declares none.

Comments

This ms examines the risks of Red List errors when generation lengths of amphibians are underestimated for assessing population reductions under criterion A. Apparently, some assessors erroneously assume a minimum time frame of 10 years when estimates of generation length are unavailable or uncertain, contrary to Red List application guidelines. In such cases, the authors suggest that predicted generation lengths could produce more accurate Red List assessment outcomes. The authors explored the consequences for estimated rates of population reduction when the assessment time frame is underestimated by erroneously using 10 years instead of three generation lengths. Unsurprisingly, they found that population reductions were underestimated when the 10 year time frame is used. On one hand, this is hardly a major discovery and the solution (assessors should read and apply the Red List criteria and guidelines correctly) is not novel. On the other hand, there could be benefit in raising awareness of this misapplication and its consequences, both for assessors to correct their practices in the future, and for users of Red List assessments to alert them to possible biases (i.e. underestimates of status) in Red List data. The ms makes this point clearly, but needs some revisions to ensure that the analyses underpinning application of IUCN Red List criterion A has a robust basis in line with other requirements in the IUCN Red List guidelines.

Main comments

1. The title could more clearly indicate the content of the paper. Without context, what is a naïve generation length? Later, we learn that the study actually explores the consequences of underestimating the time frames for assessing population declines in Red Listing species. A title along those lines would be more informative.

2. A comparison of 10 year vs 3-generation time frames is central to the study, yet we are not told how generation lengths were estimated other than a citation of another reference by the same authors. Please add a sentence or phrase at first mention (L59-61) to explain how these predictions were made (e.g. from species with similar life history traits?). The first mention is in L59-6. Some extra detail (a couple of sentences) with the accompanying reference would also be warranted in the Methods section.

3. Although the focus of the study is generation length, the method for estimating population reduction from habitat loss exposes other fallibilities in the Red List assessments that were undertaken to illustrate the consequences of alternative tie frames for assessing population reductions (L124-132).. The methods are imprecise and rely on a number of assumptions that are either unstated or insufficiently justified, and seem likely to compound errors. The key issue are:

a. The analysis assumes a linear relationship between habitat loss and population reduction but no justification is given. The Red List Guidelines are cited, out of context in L197-98, which posit that “acceptable assumption in the absence of species specific information on population size-habitat relationship (IUCN Standards and Petitions Committee 2024).” What the Guidelines (section 5.8) actually say is that this is the simplest assumption, and that it is not often true. They further state “In all cases, an understanding of the taxon and its relationship to its habitat, and the threats facing the habitat is central to making the most appropriate assumptions about habitat loss and subsequent population reduction.” The authors may argue that this is not the main focus of the analysis, however it could affect the results (especially for spp with longer generation lengths, as noted n s5.8 of the Guidelines). Further, a study that aims to correct one misapplication of Red List methods, would not seem very persuasive if it introduces or perpetuates another type of misapplication in doing so. The authors could address this shortcoming provide justifications for a representative subsample of the taxa they assess (in Supplementary Material), and identify the scenarios where their default assumption is likely to be robust and others where it is not. An appropriate caveat would then need to be included in the main text in lieu of text on L197-198.

b. The analysis assumes that proportional decline in tree cover equals proportional decline in habitat suitability and proportional decline in numbers of mature individuals for all amphibian taxa assessed in the study. This assumption needs to be evaluated – how many of the amphibian species in the final data set may not depend on trees for persistence, how many may instead depend on surface water with or without trees (noting species slection was based on forest or wetland habitat (L113-114)?

4. For climate change there are also limitations in linking change in modelled habitat suitability to projected reductions in the number of mature individuals. The differences in modelled outputs are interpreted to represent declines in habitat area (based on unspecified suitability thresholds), and declines in habitat area are assumed to represent declines in mature individuals directly (e.g 322-328). These layered assumptions needs to be explicitly stated and justified, including the assumption that populations of all modelled amphibian species equilibrate rapidly with the changing climate with no ecological lags. Several aspects of the modelling methods described here may not meet standards prescribed in IUCN Guidelines (section 12.1.12) for training data or variable selection, exploration of uncertainties via individual GCMs and RCPs, model evaluation, etc. The key issues are:

a. ‘Species range’ is undefined. Does it mean Extent of Occurrence (sensu IUCN 2012) or something else? What are the implications, assumptions and limitations of using point samples within the species range, as opposed to in situ records of occupancy? Further, the spatial data on change in tree cover degraded to a coarser spatial resolution prior to analysis. Some justification is needed.

b. Sampling of training and validation points from range polygons assumes occurrence and habitat suitability is evenly distributed within range polygons. This seems very unlikely, especially for amphibians, given the strong association of many species with surface water.

c. “We did not perform ecological variable selection (Mancini et al. 2025a) as in this case we were only interested in comparing the same models under the two time horizons considered.” If the models are suboptimal due to naïve or generic variable selection, might different outcomes of comparison between 10-year and 3-generation time frames be expected? It could, depending on their projected rates of change. For many species, extreme events (e.g. heat waves, extended moisture deficit, etc.) may be more proximal predictors of climate change impacts on habitat suitability than average conditions indicated by most BioClim variables. This limitation needs to be documented.

5. The assessment of projected future change is insufficient – it largely identifies software packages without any details of parameterisation. How was the change over 10 yrs actually calculated from what time series of model projections. There needs to be sufficient detail so that others can repeat the analysis.

6. The results are reported separately for subcriteria criteria A2 and A3, which is appropriate because the sets of species included in each analysis are different. However, it would be useful to conclude the Results with a synopsis of the degree to which the same species identified as sensitive to assessment time frame under A2, were also sensitive under A3. This might require focus on the subset of species that were common to both analyses.

Minor points

1. L46-47. “defined as the average age of parents of the current cohort (i.e., newborn individuals in the population…” Would be clearer id expressed as “… i.e. parents of newborn individuals in the population…)”.

2. L52-55. “If generation length is unavailable but the species is thought to have a short life history (i.e., short-living species), a 10-year time frame might be used to apply IUCN Red List criteria (Edgar 2025). This assumption is often unwarranted.” After citing the IUCN Red List guidelines in the preceding sentence, the authors here are describing a perceived practice, which they say is often unwarranted. However, the expression and context of the sentence on L52-54 need clarification so that this practice is not misunderstood by readers to be prescribed or recommended in the IUCN guidelines. For example, “Where estimates of generation length is unavailable but the species is thought to have a short life history (i.e., short-living species), some assessors have (naively) assumed a 10-year time frame for assessing population reductions under criterion A (Edgar 2025).”

3. L67-70, This sentence could also do with greater clarity. “For example, of 162 amphibians for which population reductions were recently assessed over a generic 10-year time frame (according to documentation of their Red List assessments; IUCN 2025), 32% (53) of them were estimated to have generation lengths >3.5 years (Mancini et al. 2025b), and therefore should be assessed over time frames longer than 10 years (IUCN Standards and Petitions Committee 2024).”

4. L78-79, “…as the impact would be measured over multiple generations” suggest reword this phrase to “…as any population reductions would have been incorrectly assessed over a time frame of more than three generations.”

5. L86-88, This sentence needs to be reworded to ensure the correct use of terms defined in IUCN (2012) and the 2024 Guidelines. Suggest “We consistently calculated rates of habitat loss: i) due to past deforestation, to quantify suspected past population reduction for applying IUCN Red List Criterion A2; and ii) due to future climate change quantify projected future population reduction for applying Criterion A3.”

6. L94-95, “…between our predicted Red List categories (…) and those and recently assessed by the Red List authorities (IUCN 2025) for the same species, highlighting…”

7. L101-102, “We estimated the potential extinction risk of amphibians under IUCN Red List Criterion A based on population reductions estimated from past deforestation (Criterion A2) or future climate change (Criterion A3).”

8. L118, incorrect and inconsistent with IUCN Guidelines.

9. L167. Probablistic habitat maps or Habitat suitability maps might be better terminology, preferably the latter as Maxent does not strictly produce probability outputs.

10. L176-177. The ‘no dispersal’ assumption could do with a more explicit justification, as there are some exceptions to the generalisation attributed to Smith & Green 2005). Of the subset of species assessed in this study, how many have dispersal traits?

11. Fig 1 & 2 are well designed but the caption (e.g. on L245 for Fig 1) could clarify that current extinction risk categories are given on the grey bars (correct?). Also can you clarify if ‘current status’ is based on all criteria for which data were available or whether current status refers only to outcomes of criterion A in the IUCN published assessments? I ask because some species currently as LC seem to qualify for higher status, even on a 10 year time frame, if I read it correctly. L333-334 implies that some taxa were not previously assessed against criterion A, but this should be clearly stated in Methods and documented in an appendix.

12. L281. “…using a naive 10-year time frame” perhaps rephrase as “…erroneous use of a 10-year time frame for calculating population reductions”

13. L314 “…species with slow life histories, indicated by long generation lengths,…”

14. L321 Please revise English expression.

Review: The risk of underestimating generation length for extinction risk assessments — R0/PR3

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr3 [Opens in a new window]
Jacinta Guirguis Queen’s University Belfast
Date of review:  05 January 2026
Revision round: 0
Role: reviewer
Recommendation/decision: major-revision

Conflict of interest statement

Reviewer declares none.

Comments

Review of ‘The peril of using naive generation lengths to assess species extinction risk’ for Cambridge Prisms: Extinctions (2026)

The authors represent a novel and important study addressing a limitation in IUCN Red List assessments of nature’s most threatened animals (amphibians). The study uses appropriate methods and has a logical flow to deliver the simple but important message that using metrics with little biological relevance can drastically misidentify extinction risk in species. The manuscript contains relevant citations. However, I observed that improvements to the framing and presentation of study are necessary in order for the text to meet the standard for Cambridge Prisms: Extinctions. This can be addressed by the authors acknowledging more explicitly throughout the text the scale of their findings (e.g. that they implicate a minority of amphibians), the broader context of IUCN extinction risk, and by providing more details regarding species implicated from their findings (as the identification of specific species which could have underestimated extinction risk is a major strength to this study). Furthermore, I suggest altering the title to replace the word ‘peril’ to a more neutral word, perhaps ‘impact’, or something which reflects the main finding (e.g. the use of naive generation lengths can underestimate species extinction risk). Therefore, I recommend major revisions given my observations below.

Introduction

The Introduction flows well and is easy to follow. However, the overall context of the problem presented in the manuscript needs to be described in more detail.

Lines 54-55: While I appreciate the authors point of view, this line appears divisive, because it could be argued that the lack of information is what justified the 10-year timeframe. If estimations of generation length are only recently available, and the assessments were conducted prior, then the 10-year timeframe was justified. I strongly suggest the authors change the tone such that it reflects the fact that the 10-year timeframe has limitations, especially now given estimated generation length is available, which is more objective then asserting that the prior use was unwarranted.

Lines 57-59: For what proportion of species is generation length known? This should be added to the text as it may be of interest to the reader. Also, is it know what percentage of the IUCN Red List assessments apply this 10-year timeframe?

Lines 59-61: This line reads as if to suggest the problem of consistent application of IUCN criteria in mammals and birds and recently amphibians and reptiles has been addressed, rather than the absence of generation length data.

Lines 61-62: Does the 10-year timeframe overestimate extinction risk in birds and mammals? Has this study been done before in birds and mammals? This information should be given, if not here somewhere else in the text, as the authors mention that Criterion A implicate mammals and birds more than amphibians.

Lines 84-86: How is generation length distributed among amphibians? It’s mentioned that amphibians have the shortest generation length among tetrapods (lines 64-66), but what proportion have generation length above or below the 10-year timeframe? This information adds context that helps to gauge the overall impact of the study. If not addressed here, it could be addressed by visualising the data in the Methods.

Methods

How is the predicted generation length in amphibians derived? While this information is available in Mancini et al. 2025B, a summary of how Mancini et al. 2025B estimated generation lengths needs be included in the Methods as it is a key feature of this manuscript.

Moreover, it is unclear whether the species analysed are threatened by Criterion A only, or if multiple Criteria apply. The authors should indicate which it is. If multiple Criteria apply, please discuss how this would impact the Results, either in the Methods to justify decisions or below in the Discussion.

Lines 141-142: How many the Results be impacted by removing these species? Are these species threatened by anyway by other Criteria?

Lines 176-180: I suggest to include how the ‘no-dispersal’ scenario impacts findings.

Lines 197-199: Why is this an acceptable assumption?

Results

The findings which the authors present are directly applicable to conservation efforts. Therefore, all the species that were found to have extinction risk underestimated should be detailed in the Supplementary Material, and more examples of specific species should be given throughout the Results text.

Line 215: Subheading should be changed to something that better reflects the findings, e.g. ‘Impact on extinction risk due to deforestation’

Lines 235-238: While I appreciate the endeavour to provide specific examples, especially as the study is applied, this looks like cherry picking. Does the chosen example Pseudhymenochirus merlini reflect the other 83 species? The authors must provide the context, perhaps selecting examples that represent the range, or the median of the findings.

Line 249: Subheading should be more descriptive of findings (see above)

Lines 251-261: Similar to the above section, the authors should provide specific examples that represent the overall range or median of other characteristics of the findings.

Discussion

The findings could be discussed with more reference given to the overall context of IUCN extinction risk. For instance, the Discussion does not indicate whether the findings point to a broader picture of flaws in the IUCN Red List assessments in the naive 10-year timeframe or whether this issue specific to amphibians. Using the 10-year timeframe, is extinction risk typically over-estimated or under-estimated in other animals? Is the 10-year timeframe most biologically irrelevant for amphibians, or for all species?

Recommendation: The risk of underestimating generation length for extinction risk assessments — R0/PR4

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr4 [Opens in a new window]
Helen Regan [Opens in a new window] University of California Los Angeles, United States
Date of review:  20 January 2026
Revision round: 0
Role: Handling Editor
Recommendation/decision: major-revision

Comments

Both reviewers see value in this paper and I agree with their assessments. They have many suggestions for revision, particularly Reviewer 1, and I recommend that all of these be addressed.

Decision: The risk of underestimating generation length for extinction risk assessments — R0/PR5

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr5 [Opens in a new window]
John Alroy [Opens in a new window] School of Natural Sciences, Macquarie University, Australia
Revision round: 0
Role: Editor in Chief
Recommendation/decision: major-revision

Comments

No accompanying comment.

Author comment: The risk of underestimating generation length for extinction risk assessments — R1/PR6

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr6 [Opens in a new window]
Giordano Mancini [Opens in a new window] Department of Biology and Biotechnologies "Charles Darwin", University of Rome La Sapienza, Italy
Revision round: 1
Role: author

Comments

Dr. John Alroy, Prof. Barry Brook

Cambridge Prisms: Extinction - Editors-in-Chief

Re: Submission of revised manuscript EXT-2025-0024

Dear Dr. John Alroy, Prof. Barry Brook,

Please find attached the following revised manuscript for potential publication in Cambridge Prisms: Extinction: “The risk of underestimating generation length for extinction risk assessments”. We are grateful to the Editors and Reviewers for their insightful comments, and for the opportunity to address these comments in a revised version of our manuscript.

The Reviewers recognized the importance and potential of our paper in highlighting the issue of naive assumptions on the generation length of amphibians, and the implications on extinction risk estimates. The reviewers also raised important points that needed to be resolved. We have now incorporated all the suggestions received on our manuscript through a thorough text revision. The result is a much improved manuscript in terms of clarity, relevance, and potential impact.

The main requests from the Reviewers were to:

- clarify amphibian dependence on forests for the deforestation analysis

- further explain our methodological choices

- better clarify the limitations of our methods

- update the title to better reflect the content of the paper

In response to these requests we have now thoroughly revised the Methods and Results sections. We refined our deforestation analysis to include only strictly forest-dependent species. While this adjustment reduced our sample size from 2,370 to 1,278, the core findings remained quantitatively consistent. Specifically, habitat loss increased over the generation length time frame, leading to a higher number of species classified as potentially at risk of extinction.

Additionally, we clarified that some of our choices derive from our application of IUCN Red List Guidelines (IUCN Standards and Petition Committee, 2024). In particular, looking at trends in habitat quality, as per Red List subcriterion c, is a valid proxy for application of Criterion A. Similarly, our assumption about the linear relationship between habitat and population size, although simplistic, is a common assumption for applying Red List Criterion A in absence of species-specific information. We have now further clarified this point.

We have better explained our methodological choices and the associated limitations in Methods and Discussion of the revised manuscript. We highlight that some methodological choices were based on previous papers on application of Red List Criterion A (Mancini et al., 2024; Tracewski et al., 2016). We also stated where and why we deviated from the Red List Guidelines in terms of methodological choices for the Species Distribution Models, also adding a new paragraph in the Discussion dedicated to such limitations.

Finally, we have changed the title of the paper to “The risk of underestimating generation length for extinction risk assessments”. We believe this new title better reflects the content of our manuscript.

Together with our submission, we attach a detailed rebuttal letter where we address all the comments mentioned above, and all other minor comments raised by both Reviewers. We believe that our revised work has much improved compared to the original version, leaving little doubt about the overall robustness of our results and the potential impact on the conservation of amphibians. We hope that this revised manuscript will meet your expectations.

Yours sincerely,

Giordano Mancini, on behalf of all co-authors

Review: The risk of underestimating generation length for extinction risk assessments — R1/PR7

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr7 [Opens in a new window]
Jacinta Guirguis Queen’s University Belfast
Date of review:  08 April 2026
Revision round: 1
Role: reviewer
Recommendation/decision: accept

Conflict of interest statement

Reviewer declares none.

Comments

I am pleased with how the authors have addressed concerns. In addition to the revised analysis using strictly forest-dependent species, the text itself has been improved for publication. The updated title more accurately reflects the manuscript’s subject. The problem and current knowledge provided in the Introduction is outlined more clearly. The Methods now provide details of the estimated generation length data and IUCN guidance. The Results are presented transparently, as a balanced spread of specific amphibian species assessments impacted are provided, and the number of species in common to both analyses has been indicated. Furthermore, the Discussion has been updated to reflect shortcomings and unrealistic assumptions that contrast with the IUCN guidance.

I am recommending acceptance because I believe the quality of the current ms is sufficient to raise awareness of the issue regarding the application of naive time frames in IUCN assessments.

Recommendation: The risk of underestimating generation length for extinction risk assessments — R1/PR8

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr8 [Opens in a new window]
Helen Regan [Opens in a new window] University of California Los Angeles, United States
Date of review:  13 April 2026
Revision round: 1
Role: Handling Editor
Recommendation/decision: accept

Comments

The reviewer is satisfied that their concerns have been addressed. Therefore, I am happy to accept this manuscript for publication.

Decision: The risk of underestimating generation length for extinction risk assessments — R1/PR9

Published online by Cambridge University Press:  27 April 2026
DOI: https://doi.org/10.1017/ext.2026.10012.pr9 [Opens in a new window]
John Alroy [Opens in a new window] School of Natural Sciences, Macquarie University, Australia
Revision round: 1
Role: Editor in Chief
Recommendation/decision: accept

Comments

No accompanying comment.