Highway construction often leads to habitat fragmentation, altering the genetic features of wild populations. While compensatory measures are commonly implemented to mitigate their negative demographic and genetic impacts, their effectiveness remains species-dependent and difficult to assess. We analysed the short-term impact of a newly built highway in eastern France and the associated restoration measures on population genetic structure in the protected southern damselfly (Coenagrion mercuriale). We genotyped 46 populations collected after the highway construction using microsatellite and SNP markers over two successive years to detect potential genetic discontinuities across the highway. We also evaluated the effectiveness of watercourse restoration and identified the likely origin of recolonising individuals. We found no significant barrier effect of the highway construction on gene flow. However, these encouraging results should be interpreted with caution, considering the likely time-lag in populations’ responses to genetic drift effects. We observed a gradual recolonisation of restored watercourses that had undergone compensatory measures, suggesting positive effects of reproductive habitat recovery. Recolonisation involved nearby populations but also long-distance dispersal events occurring overland. Detected migrant individuals may come from geographically distant populations belonging to a different watercourse. This mirrors a migrant-pool model of colonisation that leaves high levels of genetic diversity and no evident spatial patterns in populations located close to the highway. Altogether, our study highlights both the potential and limitations of genetic tools to assess the impacts of linear infrastructure on gene flow and for understanding watercourse recolonisation processes after ecological management restoration.

ABSTRACT The recent acceleration of global climate warming has created an urgent need for reliable projections of species distributions, widely used by natural resource managers. Such projections have been mainly produced by species distribution models with little information on their performances in novel climates. Here, we hindcast the range shifts of forest tree species across Europe over the last 12,000 years to compare the reliability of three different types of models. We show that in the most climatically dissimilar conditions, process‐explicit models (PEMs) tend to outperform correlative species distribution models (CSDMs), and that PEM projections are likely to be more reliable than those made with CSDMs by the end of the 21st century. These results demonstrate for the first time the often promoted albeit so far untested idea that explicit description of mechanisms confers model robustness, and highlight a new avenue to increase model projection reliability in the future.

Genomic markers are essential tools for studying species of conservation concern, yet nonmodel species often lack a reference genome. Here we describe a methodology for identifying and genotyping thousands of SNP loci in the southern damselfly (Coenagrion mercuriale), a bioindicator of freshwater stream quality classified as near-threatened, with locally declining populations. We used a hybrid approach combining reduced representation sequencing and target enrichment. First, we identified putative SNP loci using ddRADseq and de novo assembly. Then, single primer enrichment technology targeted 6000 of these SNPs across 1920 individuals. Challenges encountered included sequence recapture failure, coverage depth discrepancies, and aberrant FIS values. We provide recommendations to address such issues. After multiple filtering steps, 2092 SNPs were retained and used to analyze the genetic structure of 131 individuals belonging to 11 populations in France, comparing central and marginal populations. Genetic differentiation was lower among central populations, with no sign of inbreeding. As compared with microsatellite loci, SNPs exhibited greater resolution in detecting fine-scaled genetic structure, and identifying putative hybrids in adjacent populations. In this study, we emphasize the difficulties of large-scale SNP genotyping in nonmodel species via a hybrid method that ultimately did not offer the expected cost and time-saving compared with classical ddRAD approaches. However, SNPs showed greater power than previously available markers in identifying conservation units or admixture events, and the panel of reusable probes we describe here offers the potential to improve conservation efforts through future diachronic studies or finer estimations of key parameters like effective population size.

Social Impact Statement Cassava is a lifeline for millions of people across the tropics, but its productivity is at risk from a disease called Cassava Bacterial Blight (CBB). In this study, we explore how the everyday practices of farmers exchanging cassava cuttings-a long-standing tradition-can shape how the disease spreads and evolves in the Colombian Caribbean. By combining knowledge from socio-cultural studies, plant pathology, and genetics, we reveal an often-overlooked reality: plant diseases can spread through human activities. Our findings shed new light on CBB epidemiology and emphasize the importance of incorporating farmers' practices and local economies into the development of effective, sustainable disease control strategies.Summary Although our knowledge about the prevalence of CBB and the diversity of Xpm populations has deepened in recent years, the role played by the seed circulation system on pathogen dispersal is still unknown. In this paper, our objective is to determine the influence of the social network shaping seed circulation on the dispersion and diversity of the pathogen at the local scale. Cassava is particularly interesting as planting material is often scarce at the time of the planting season, and diseases may not be fully visible, making it hard to control the phytosanitary condition of the propagative material. We used interdisciplinary methods based on ethnoecology, epidemiology, ecology, social network analysis, and genetics that included semi-structured interviews with farmers, coupled with molecular genotyping of cassava plants and Xpm. Our findings highlight two main effects of local seed circulation on Xpm spread and diversity: (1) a high diversity of the pathogen at the scale of the evaluated village, and (2) greater genetic similarities of the pathogen between farmers that exchanged cuttings, regardless of the geographic distance between the fields. This study reveals the multifactorial nature of Xpm diversity and spread within a single village. By combining genetic, agricultural, and social data, it shows the value of interdisciplinary approaches to understand plant pathogen dynamics. The findings support the design of locally adapted disease control strategies. Unlike studies focused solely on genetic data, this work emphasizes the importance of integrating sociocultural factors, particularly for seedborne pathogens and crops where seed circulation systems are a determining factor. La yuca es el sustento de millones de personas en los tr & oacute;picos, pero su productividad est & aacute; en peligro a causa de una enfermedad llamada bacteriosis vascular de la yuca (CBB, por sus siglas en ingl & eacute;s). En este estudio, exploramos c & oacute;mo las pr & aacute;cticas cotidianas de los agricultores que intercambian esquejes de yuca, como parte de sus tradiciones, pueden influir en la propagaci & oacute;n y evoluci & oacute;n de la enfermedad en el Caribe colombiano. Combinando conocimientos de estudios socioculturales, fitopatolog & iacute;a y gen & eacute;tica, revelamos una realidad que a menudo se pasa por alto: las enfermedades de las plantas pueden propagarse a trav & eacute;s de actividades humanas. Nuestros hallazgos arrojan nueva luz sobre la epidemiolog & iacute;a de CBB y subrayan la importancia de incorporar las pr & aacute;cticas locales de los agricultores y sus medios de vida en el desarrollo de estrategias eficaces y sostenibles de control de la enfermedad.

Genomic markers are essential tools for studying species of conservation concern, yet non-model species often lack a genome reference. Here we describe a methodology for identifying and genotyping thousands of SNP loci in the southern damselfly ( Coenagrion mercuriale ), a bioindicator of freshwater stream quality classified as near-threatened, with locally declining populations. We used a hybrid approach combining reduced representation sequencing and target enrichment. First, we identified putative SNP loci using ddRADseq and de novo assembly. Then, single primer enrichment technology targeted 6,000 of these SNPs across 1,920 individuals. Challenges encountered included sequence recapture failure, coverage depth discrepancies, and aberrant F IS values. We provide recommendations to address such issues. After multiple filtering, we retained 2,092 SNPs. We used them to characterise rear-edge populations of the southern damselfly in Northern France, a region where populations are sparsely distributed. Previous surveys utilising microsatellite markers allowed comparison of genetic diversity and differentiation estimates. Consistent with prior findings, genetic diversity estimates were similar across the studied populations that showed no sign of inbreeding. SNP markers exhibited greater resolution in detecting fine-scaled genetic structure, identifying two putative hybrids in adjacent populations, a feat unattainable with microsatellite loci. Altogether, this study highlighted the ongoing challenge of large-scale SNP genotyping using target sequencing techniques in non-model species to set conservation guidelines. Nonetheless, these new markers showed greater statistical power in identifying conservation units and offered the promise of greater precision in the identification of admixture events or the estimation of key population parameters such as effective population size. ### Competing Interest Statement The authors have declared no competing interest.

The effective population size (N e) is a key parameter in conservation and evolutionary biology, reflecting the strength of genetic drift and inbreeding. Although demographic estimations of N e are logistically and time-consuming, genetic methods have become more widely used due to increasing data availability. Nonetheless, accurately estimating N e remains challenging, with few studies comparing N e estimates across molecular markers types and estimators such as single-sample methods based on linkage disequilibrium or sibship analyses versus methods based on temporal variance in allele frequencies. This study aims at bridging this gap by analysing single-sample and temporally spaced populations in the southern damselfly (Coenagrion mercuriale), a bioindicator Odonata species of conservation concern found in southwestern Europe's freshwater stream networks. A total of 77 local populations were sampled from a semi-urbanised area located in eastern France near Strasbourg city, yielding 2842 individuals that were genotyped with microsatellites and 958 of which were also genotyped for 2092 SNPs. Spatial genetic structure was stable over time, suggesting porosity between alternate-year cohorts. When accounting for spatial genetic structure, single-sample and temporal estimations of N e were consistent for each set of molecular markers. Biologically meaningful results were obtained when the effect of migration was minimising by considering metapopulation N e estimates based on the level of genetic differentiation and population boundaries. In terms of applied conservation and management, most depicted metapopulations displayed large N e, indicating no immediate need for conservation measures to mitigate anthropogenic pressures, provided that a continuous suitable freshwater network is maintained. However, urbanisation negatively impacted N e levels in populations close to Strasbourg city. Because N e is used to inform conservation decisions, caution is crucial in interpreting N e estimates, especially in continuously distributed populations undergoing migration. Altogether, our study highlights the challenge of obtaining robust N e estimates and the necessity of careful interpretation to set relevant conservation guidelines.

Aim: Human-induced environmental changes result in habitat loss and fragmentation, impacting wildlife population genetic structure and evolution. Urbanised and geographically peripheral areas often represent unfavourable environments, reducing connectivity among populations and causing higher population genetic differentiation and lower intra-population genetic diversity. We examined how geographic peripherality and anthropogenic pressures affect genetic diversity and genetic differentiation in the protected southern damselfly (Coenagrion mercuriale, Odonata), which has low dispersal capabilities and specific habitat requirements and whose populations are declining. Location: We studied two areas: one in semi-natural habitats at the periphery of the species geographic range (northern France) and the other more central to the species' range, in an urbanised area surrounding the city of Strasbourg (Alsace, eastern France). Methods: We genotyped 2743 individuals from 128 populations using 11 microsatellite loci. We analysed the spatial distribution of neutral genetic diversity (allelic richness, heterozygosity, levels of inbreeding and genetic relatedness), the extent of genetic differentiation and population affiliations (sPCAs) within the two areas. We also examined fine-scale patterns of gene flow in the urbanised area of Alsace by investigating patterns of isolation by distance and estimating effective migration surfaces (EEMS) method. Results: Northern peripheral populations showed lower levels of genetic diversity and higher levels of genetic differentiation than central Alsacian populations. Although located in anthropised habitats, geographically central Alsacian populations showed high levels of gene flow, with dispersal events mainly occurring overland and not restricted to watercourses. However, the highly urbanised city of Strasbourg negatively impacted nearby populations by reducing levels of genetic diversity and increasing population genetic differentiation. Main Conclusions: These results showed the need for management action by restoring breeding sites and creating migratory corridors for peripheral southern damselfly populations. However, our results also highlighted the resilience of southern damselfly in central range populations facing strong urbanisation pressures.

The recent acceleration of global climate warming has created an urgent need for reliable projections of species distributions, widely used by natural resource managers. Such projections, however, are produced using various modeling approaches with little information on their relative performances under expected novel climatic conditions. Here, we hindcast the range shifts of five forest tree species across Europe over the last 12,000 years to compare the performance of three different types of species distribution models and determine the source of their robustness. We show that the performance of correlative models (CSDMs) decreases twice as fast as that of process-based models (PBMs) when climatic dissimilarity rises, and that PBM projections are likely to be more reliable than those made with CSDMs, at least until 2060 under scenario SSP245. These results demonstrate for the first time the well-established albeit so far untested idea that explicit description of mechanisms confers models robustness, and highlight a new avenue to improve model projections in the future. ### Competing Interest Statement The authors have declared no competing interest.

Anthropogenic changes drive shifts in species' geographic distributions and increase the occurrence of leading or trailing‐edge marginal populations. Theoretical predictions and empirical observations indicate substantial changes in life‐history traits in marginal populations, often involving dispersal and reproductive abilities. Using a common garden experiment, we studied the variation of life‐history traits of populations sampled on spatial gradients extending from range‐core to range‐edge habitats for three expanding (miner's lettuceClaytonia perfoliata, Danish scurvygrassCochlearia danicaand rock samphireCrithmum maritimum) and one receding plant species (dune pansyViola tricolorsubs.curtisii). We monitored life‐history traits related to dispersal, phenology, survival, reproductive output and selfing ability. Significant shifts in life‐history traits between central and marginal populations strongly differed among species. Marginal populations of the three expanding species displayed modified seed weight in natura, suggesting increased dispersal abilities in leading‐edge populations. Discarding unassessed maternal effects, this trait modification can be due to phenotypic plasticity or to genetic differentiation. In miner's lettuce, marginal expanding populations show advanced phenology and higher reproductive output, that may potentially influence their colonization ability. In rock samphire, life‐history traits showed large intra‐ and inter‐population variability that did not follow a core‐to‐edge geographic trend, except for seed size. Finally, the receding populations of the dune pansy displayed a shift towards a plant architecture maximizing survival but reducing individual reproductive success. Altogether, our results indicated a common trend for increased dispersal abilities in marginal populations of expanding species. However, shifts in species' distributions may drive idiosyncratic changes in other life‐history traits, for which we observed no general evolutionary syndrome at range edges. These findings go along a stochastic view of trait evolution during range expansion, and question how to draw predictive projections of species' distribution shifts under current global change.

In a response to gravitropic stress, G-layers (gelatinous layers) were deposited in xylem cell walls of tilted flax plants. G-layers were produced in both tension wood (upper side) as expected but were also observed in opposite wood (lower side). Raman spectral profiles were acquired for xylem G-layers from the tension and opposite side as well as from the G-layer of bast fibers grown under non-tilted conditions. Statistical analysis by principal component analysis (PCA) and partial least square-discriminant analysis (PLS-DA) clearly distinguished bast fiber G-layers from xylem G-layers. Discriminating bands were observed for cellulose (380-1150-1376 cm-1), hemicelluloses (517-1094-1126-1452 cm-1) and aromatics (1270-1599-1658 cm-1). PCA did not allow separation of G-layers from tension/opposite-wood sides. In contrast, the two types of xylem G-layers could be incompletely discriminated through PLS-DA. Overall, the results suggested that while the architecture (polymer spatial distribution) of bast fibers G-layers and xylem G-layers are similar, they should be considered as belonging to a different cell wall layer category based upon ontogenetical and chemical composition parameters.

Cassava ( Manihot esculenta Crantz) is a primary crop for food security of millions of people worldwide. In Colombia, the Caribbean region contributes about half of the national cassava production, despite major socioeconomic constraints such as unequal land property, omnipresence of middlemen, low and unstable prices, armed conflict, climate change and phytosanitary issues. Among the latter is Cassava Bacterial Blight (CBB), a disease caused by the bacterial pathogen Xanthomonas phaseoli pv. manihotis ( Xpm ) that leads to irreversible damage to plants, impeding growth and productivity. In 2016, we analyzed the role of sociocultural and agricultural practices on CBB prevalence in small-scale fields of a village of the Colombian Caribbean region, where farmers live almost exclusively from the sale of their cassava production. Semi-structured interviews (48) were conducted with all farmers who cultivated cassava to document individual sociodemographic characteristics, cassava farming practices, and perceptions about CBB occurrence. Cassava Bacterial Blight was diagnosed in the field and the presence of Xpm was further confirmed upon laboratory analysis of collected diseased leaf samples. Our data show that (i) according to the risks perceived by farmers, CBB is the main disease affecting cassava crops in the village and it could indeed be detected in about half of the fields visited; (ii) CBB occurrence depends strongly on land property issues, likely because of an inadequate phytosanitary control during acquisition of cuttings when farmers are forced to rent the land; and (iii) there is a strong positive correlation between the use of commercial fertilizers and the occurrence of CBB in the village of Villa López.

Flax is an important fiber crop that is subject to lodging. In order to gain more information about the potential role of the bast fiber cell wall in the return to the vertical position, 6-week-old flax plants were subjected to a long-term (6 week) gravitropic stress by stem tilting in an experimental set-up that excluded autotropism. Stress induced significant morphometric changes (lumen surface, lumen diameter, and cell wall thickness and lumen surface/total fiber surface ratio) in pulling- and opposite-side fibers compared to control fibers. Changes in the relative amounts and spatial distribution of cell wall polymers in flax bast fibers were determined by Raman vibrational spectroscopy. Following spectra acquisition, datasets (control, pulling- and opposite sides) were analyzed by principal component analysis, PC score imaging, and Raman chemical cartography of significant chemical bonds. Our results show that gravitropic stress induces discrete but significant changes in the composition and/or spatial organization of cellulose, hemicelluloses and lignin within the cell walls of both pulling side and opposite side fibers.

Biomonitoring ecosystems is necessary in order to evaluate risks and to efficiently manage ecosystems and their associated services. Agrosystems are the target of multiple stressors that can affect many species through effects cascading along food webs. However, classic biomonitoring, focused on species diversity or indicator species, might be a poor predictor of the risk of such whole-ecosystem perturbations. Thanks to high-throughput sequencing methods, however, it might be possible to obtain sufficient information about entire ecological communities to infer the functioning of their associated interaction networks, and thus monitor more closely the risk of the collapse of entire food webs due to external stressors. In the course of the `next-generation biomonitoring' project, we collectively sought to experiment with this idea of inferring ecological networks on the basis of metabarcoding information gathered on different systems. We here give an overview of issues and preliminary results associated with this endeavour and highlight the main difficulties that such next-generation biomonitoring is still facing. Going from sampling protocols up to methods for comparing inferred networks, through biomolecular, bioinformatic, and network inference, we review all steps of the process, with a view towards generality and transferability towards other systems.

Ecosystems of Lake Victoria and riparian communities have been strongly disrupted by the introduction of the invasive Nile perch and its fishing industry. Beyond this invasion and other recent anthropogenic stressors, the Lake Victoria ecoregion also underwent phases of pronounced aridity over the Late Pleistocene, lastly during the Last Glacial Maximum (LGM). The consequences of recent and historic environmental change have been canvassed for the adaptive radiation of haplochromine cichlids occupying the ecoregion, but their effect on freshwater invertebrate diversity remains largely unknown. Here, we use 15 microsatellite loci and approximate Bayesian computation to test whether viviparid gastropods experienced a population bottleneck during the LGM, as did cichlids. Clustering analyses support three viviparid gene pools in the Lake Victoria ecoregion, gathering specimens from 1) Lake Albert and the White Nile, 2) the Victoria Nile and Lake Kyoga and 3) Lake Victoria and tributaries. The last group contains the highest genetic diversity, but all groups have a considerable number of private alleles and are inferred to predate the LGM. Examinations of demographic history reveal a 190‐ to 500‐fold population decline that started ~ 125–150 ka ago, thus substantially before the LGM bottleneck documented in haplochromine cichlids. Population collapses in viviparids are an order of magnitude more severe than declines in cichlids and have not been halted by the re‐establishment of freshwater ecosystems since the LGM. Recent anthropogenic ecosystem deterioration is causing homogenization of previously diversified microhabitats, which may contribute to (local) extinction and enhanced gene flow among species within gene pools.

Aim The aims of this study were to reconstruct the phylogeny of Manihot, a Neotropical genus restricted to seasonally dry areas, to yield insight into its biogeographic history and to identify the closest wild relatives of a widely grown, yet poorly known, crop: cassava (Manihot esculenta). Location Dry and seasonally dry regions of Mesoand South America. Methods We collected 101 samples of Manihot, representing 52 species, mostly from herbaria, and two outgroups (Jatropha gossypiifolia and Cnidoscolus urens). More than half of the currently accepted Manihot species were included in our study; our sampling covered the whole native range of the genus, and most of its phenotypic and ecological variation. We reconstructed phylogenetic relationships among Manihot species using sequences for two nuclear genes and a noncoding chloroplast region. We then reconstructed the history of traits related to growth form, dispersal ecology, and regeneration ability. Results Manihot species from Mesoamerica form a grade basal to South American species. The latter species show a strong biogeographic clustering: species from the cerrado form well-defined clades, species from the caatinga of northeastern Brazil form another, and so do species restricted to forest gaps along the rim of the Amazon basin. Vine and tree growth habits evolved repeatedly in the genus, as did fruit indehiscence and loss of ant-mediated seed dispersal. Main conclusions The genus Manihot probably originated in Mesoamerica, where it diversified prior to colonizing South America. Within South America, several groups then radiated within all kinds of seasonally dry lowland habitats southwards and eastwards. Some species also adapted to more humid environments, such as forest gaps, around the rim of the Amazon basin. Given the limited dispersal abilities of Manihot species, we propose that this radiation is most likely to have occurred only after, or shortly before, the completion of the Isthmus of Panama, around 3.5 Ma. Our results are in agreement with the past existence of a corridor of dry vegetation through Amazonia or along the eastern South American coast. In addition, our phylogeny allows identification of cassava’s closest wild relatives.

Species' geographical ranges are often restricted due to niche limitation resulting in geographical isolation and reduced population size at range margins. Under the "abundant center" paradigm, static marginal populations are thus expected to show higher genetic differentiation and lower genetic diversity than core populations. Low mate availability may also drive shifts toward higher propensity for selfing in geographically marginal populations. However, these predictions remain to be validated for contemporary range shifts occurring under current environmental change. This study is devoted to bridging this gap and assesses the spatial patterns of genetic structure and mating system across the geographical range of two coastal plant species characterized by contrasting contemporary range dynamics: the receding myrmecochorous Dune pansy (Viola tricolor subsp. curtisii) and the widespread expanding hydrochorous Rock samphire (Crithmum maritimum). Both species exhibited high propensity for selfing, with indications of inbreeding depression acting at early life stages. In Dune pansy, a biogeographical break was observed between core and marginal populations, with trailing-edge populations showing higher levels of genetic differentiation, reduced genetic diversity, and higher levels of selfing estimated through progeny arrays. In contrast, genetic structuring was weak in Rock samphire and no clear spatial trends were observed in genetic diversity nor in mating system, likely the result of efficient long-distance seed dispersal by sea-surface currents. Our study highlights that key species differences in life-history traits related to dispersal and/or mate limitation modify the expectations of genetic diversity loss and mating system shift in contemporary range-expanding populations, as compared with historical core populations.

Identifying spatial patterns of genetic differentiation across a species range is critical to set up conservation and restoration decision-making. This is especially timely, since global change triggers shifts in species’ geographic distribution and in the geographical variation of mating system and patterns of genetic differentiation, with varying consequences at the trailing and leading edges of a species’ distribution. Using 454 pyrosequencing, we developed nuclear microsatellite loci for two plant species showing a strictly coastal geographical distribution and contrasting range dynamics: the expanding rock samphire (Crithmum maritimum, 21 loci) and the highly endangered and receding dune pansy (Viola tricolor subsp. curtisii, 12 loci). Population genetic structure was then assessed by genotyping more than 100 individuals from four populations of each of the two target species. Rock samphire displayed high levels of genetic differentiation (FST = 0.38), and a genetic structure typical of a mostly selfing species (FIS ranging from 0.16 to 0.58). Populations of dune pansy showed a less pronounced level of population structuring (FST = 0.25) and a genotypic structure more suggestive of a mixed-mating system when excluding two loci with heterozygote excess. These results demonstrate that the genetic markers developed here are useful to assess the mating system of populations of these two species. They will be tools of choice to investigate phylogeographical patterns and variation in mating system over the geographical distribution ranges for two coastal plant species that are subject to dynamic evolution due to rapid contemporary global change.

As climate change leads to global warming and modified precipitation patterns, the distribution of forest biomes and tree species is expected to shift towards higher latitudes and altitudes. Such shifts are currently projected by species distribution models fitted to different climate change scenarios. Field validation of these models for several life stages of different tree species is a necessity to adapt forest management and understand the future of forest ecosystems. The study presented here aims to assess whether signs of the projected near-future range shift, contraction or expansion of ten European forest tree species are already observable in survival and growth of their seedlings at the core, leading and trailing edges of their distribution. The results show limited validation of the projected near-future changes in spatial distribution: seedling survival and growth paralleled modelled near-future habitat suitability for the three Mediterranean/Southern species, whereas cold-adapted species showed limited validation of model projections with lower growth and survival at their trailing edge; and widespread species showed inconsistent performance with model projections. Individuals of contrasting provenances did not show strong differences in survival; however they did show substantial differences in growth. The role of extreme events and biotic interactions might prove to be more important factors into shaping the future realized niche and the distribution of these species and thus should be investigated to complete the current studies on latitudinal and altitudinal shifts in relation to climate change.

Summary In a rapidly changing world, ecology has the potential to move from empirical and conceptual stages to application and management issues. It is now possible to make large‐scale predictions up to continental or global scales, ranging from the future distribution of biological diversity to changes in ecosystem functioning and services. With these recent developments, ecology has a historical opportunity to become a major actor in the development of a sustainable human society. With this opportunity, however, also comes an important responsibility in developing appropriate predictive models, correctly interpreting their outcomes and communicating their limitations. There is also a danger that predictions grow faster than our understanding of ecological systems, resulting in a gap between the scientists generating the predictions and stakeholders using them (conservation biologists, environmental managers, journalists, policymakers). Here, we use the context provided by the current surge of ecological predictions on the future of biodiversity to clarify what prediction means, and to pinpoint the challenges that should be addressed in order to improve predictive ecological models and the way they are understood and used. Synthesis and applications . Ecologists face several challenges to ensure the healthy development of an operational predictive ecological science: (i) clarity on the distinction between explanatory and anticipatory predictions; (ii) developing new theories at the interface between explanatory and anticipatory predictions; (iii) open data to test and validate predictions; (iv) making predictions operational; and (v) developing a genuine ethics of prediction.