The ecological forecast horizon, and examples of its uses and determinants

A cloudy forecast for species distribution models: Predictive uncertainties abound for California birds after a century of climate and land-use change

Correlative species distribution models are widely used to quantify past shifts in ranges or communities, and to predict future outcomes under ongoing global change. Practitioners confront a wide range of potentially plausible models for ecological dynamics, but most specific applications only consider a narrow set. Here, we clarify that certain model structures can embed restrictive assumptions about key sources of forecast uncertainty into an analysis. To evaluate forecast uncertainties and our ability to explain community change, we fit and compared 39 candidate multi- or joint species occupancy models to avian incidence data collected at 320 sites across California during the early 20th century and resurveyed a century later. We found massive (>20,000 LOOIC) differences in within-time information criterion across models. Poorer fitting models omitting multivariate random effects predicted less variation in species richness changes and smaller contemporary communities, with considerable variation in predicted spatial patterns in richness changes across models. The top models suggested avian environmental associations changed across time, contemporary avian occupancy was influenced by previous site-specific occupancy states, and that both latent site variables and species associations with these variables also varied over time. Collectively, our results recapitulate that simplified model assumptions not only impact predictive fit but may mask important sources of forecast uncertainty and mischaracterize the current state of system understanding when seeking to describe or project community responses to global change. We recommend that researchers seeking to make long-term forecasts prioritize characterizing forecast uncertainty over seeking to present a single best guess. To do so reliably, we urge practitioners to employ models capable of characterizing the key sources of forecast uncertainty, where predictors, parameters and random effects may vary over time or further interact with previous occurrence states.

Forecasting insect dynamics in a changing world

Predicting how insects will respond to stressors through time is difficult because of the diversity of insects, environments, and approaches used to monitor and model. Forecasting models take correlative/statistical, mechanistic models, and integrated forms; in some cases, temporal processes can be inferred from spatial models. Because of heterogeneity associated with broad community measurements, models are often unable to identify mechanistic explanations. Many present efforts to forecast insect dynamics are restricted to single-species models, which can offer precise predictions but limited generalizability. Trait-based approaches may offer a good compromise that limits the masking of the ranges of responses while still offering insight. Regardless of the modeling approach, the data used to parameterize a forecasting model should be carefully evaluated for temporal autocorrelation, minimum data needs, and sampling biases in the data. Forecasting models can be tested using near-term predictions and revised to improve future forecasts.

Space-for-time substitutions in climate change ecology and evolution

In an epoch of rapid environmental change, understanding and predicting how biodiversity will respond to a changing climate is an urgent challenge. Since we seldom have sufficient long-term biological data to use the past to anticipate the future, spatial climate-biotic relationships are often used as a proxy for predicting biotic responses to climate change over time. These 'space-for-time substitutions' (SFTS) have become near ubiquitous in global change biology, but with different subfields largely developing methods in isolation. We review how climate-focussed SFTS are used in four subfields of ecology and evolution, each focussed on a different type of biotic variable - population phenotypes, population genotypes, species' distributions, and ecological communities. We then examine the similarities and differences between subfields in terms of methods, limitations and opportunities. While SFTS are used for a wide range of applications, two main approaches are applied across the four subfields: spatial in situ gradient methods and transplant experiments. We find that SFTS methods share common limitations relating to (i) the causality of identified spatial climate-biotic relationships and (ii) the transferability of these relationships, i.e. whether climate-biotic relationships observed over space are equivalent to those occurring over time. Moreover, despite widespread application of SFTS in climate change research, key assumptions remain largely untested. We highlight opportunities to enhance the robustness of SFTS by addressing key assumptions and limitations, with a particular emphasis on where approaches could be shared between the four subfields.

Digital twins: dynamic model-data fusion for ecology

Digital twins (DTs) are an emerging phenomenon in the public and private sectors as a new tool to monitor and understand systems and processes. DTs have the potential to change the status quo in ecology as part of its digital transformation. However, it is important to avoid misguided developments by managing expectations about DTs. We stress that DTs are not just big models of everything, containing big data and machine learning. Rather, the strength of DTs is in combining data, models, and domain knowledge, and their continuous alignment with the real world. We suggest that researchers and stakeholders exercise caution in DT development, keeping in mind that many of the strengths and challenges of computational modelling in ecology also apply to DTs.

Structural asymmetry in biotic interactions as a tool to understand and predict ecological persistence

A universal feature of ecological systems is that species do not interact with others with the same sign and strength. Yet, the consequences of this asymmetry in biotic interactions for the short- and long-term persistence of individual species and entire communities remains unclear. Here, we develop a set of metrics to evaluate how asymmetric interactions among species translate to asymmetries in their individual vulnerability to extinction under changing environmental conditions. These metrics, which solve previous limitations of how to independently quantify the size from the shape of the so-called feasibility domain, provide rigorous advances to understand simultaneously why some species and communities present more opportunities to persist than others. We further demonstrate that our shape-related metrics are useful to predict short-term changes in species' relative abundances during 7 years in a Mediterranean grassland. Our approach is designed to be applied to any ecological system regardless of the number of species and type of interactions. With it, we show that is possible to obtain both mechanistic and predictive information on ecological persistence for individual species and entire communities, paving the way for a stronger integration of theoretical and empirical research.

Progress and opportunities in advancing near-term forecasting of freshwater quality

Near-term freshwater forecasts, defined as sub-daily to decadal future predictions of a freshwater variable with quantified uncertainty, are urgently needed to improve water quality management as freshwater ecosystems exhibit greater variability due to global change. Shifting baselines in freshwater ecosystems due to land use and climate change prevent managers from relying on historical averages for predicting future conditions, necessitating near-term forecasts to mitigate freshwater risks to human health and safety (e.g., flash floods, harmful algal blooms) and ecosystem services (e.g., water-related recreation and tourism). To assess the current state of freshwater forecasting and identify opportunities for future progress, we synthesized freshwater forecasting papers published in the past 5 years. We found that freshwater forecasting is currently dominated by near-term forecasts of water quantity and that near-term water quality forecasts are fewer in number and in the early stages of development (i.e., non-operational) despite their potential as important preemptive decision support tools. We contend that more freshwater quality forecasts are critically needed and that near-term water quality forecasting is poised to make substantial advances based on examples of recent progress in forecasting methodology, workflows, and end-user engagement. For example, current water quality forecasting systems can predict water temperature, dissolved oxygen, and algal bloom/toxin events 5 days ahead with reasonable accuracy. Continued progress in freshwater quality forecasting will be greatly accelerated by adapting tools and approaches from freshwater quantity forecasting (e.g., machine learning modeling methods). In addition, future development of effective operational freshwater quality forecasts will require substantive engagement of end users throughout the forecast process, funding, and training opportunities. Looking ahead, near-term forecasting provides a hopeful future for freshwater management in the face of increased variability and risk due to global change, and we encourage the freshwater scientific community to incorporate forecasting approaches in water quality research and management.

Constraining nonlinear time series modeling with the metabolic theory of ecology

Forecasting the response of ecological systems to environmental change is a critical challenge for sustainable management. The metabolic theory of ecology (MTE) posits scaling of biological rates with temperature, but it has had limited application to population dynamic forecasting. Here we use the temperature dependence of the MTE to constrain empirical dynamic modeling (EDM), an equation-free nonlinear machine learning approach for forecasting. By rescaling time with temperature and modeling dynamics on a "metabolic time step," our method (MTE-EDM) improved forecast accuracy in 18 of 19 empirical ectotherm time series (by 19% on average), with the largest gains in more seasonal environments. MTE-EDM assumes that temperature affects only the rate, rather than the form, of population dynamics, and that interacting species have approximately similar temperature dependence. A review of laboratory studies suggests these assumptions are reasonable, at least approximately, though not for all ecological systems. Our approach highlights how to combine modern data-driven forecasting techniques with ecological theory and mechanistic understanding to predict the response of complex ecosystems to temperature variability and trends.

Patterns, sources, and consequences of variation in age-specific vital rates: Insights from a long-term study of Weddell seals

Variations in the reproductive and survival abilities of individuals within a population are ubiquitous in nature, key to individual fitness, and affect population dynamics, which leads to strong interest in understanding causes and consequences of vital-rate variation. For long-lived species, long-term studies of large samples of known-age individuals are ideal for evaluating vital-rate variation. A population of Weddell seals in Erebus Bay, Antarctica, has been studied each Austral spring since the 1960s. Since 1982, all newborns have been tagged each year and multiple capture-mark-recapture (CMR) surveys have been conducted annually. Over the past 20 years, a series of analyses have built on results of earlier research by taking advantage of steady improvements in the project's long-term CMR data and available analytical methods. Here, I summarize progress made on four major topics related to variation in age-specific vital rates for females: early-life survival and age at first reproduction, costs of reproduction, demographic buffering, and demographic senescence. Multistate modelling found that age at first reproduction varies widely (4-14 years of age) and identified contrasting influences of maternal age on survival and recruitment rates of offspring. Subsequent analyses of data for females after recruitment revealed costs of reproduction to both survival and future reproduction and provided strong evidence of demographic buffering. Recent results indicated that important levels of among-individual variation exist in vital rates and revealed contrasting patterns for senescence in reproduction and survival. Sources of variation in vital rates include age, reproductive state, year, and individual. The combination of luck and individual quality results in strong variation in individual fitness outcomes: ~80% of females born in the population produce no offspring, and the remaining 20% vary strongly in lifetime reproductive output (range: 1-23 pups). Further research is needed to identify the specific environmental conditions that lead to annual variation in vital rates and to better understand the origins of individual heterogeneity. Work is also needed to better quantify the relative roles of luck, maternal effects, and environmental conditions on variation in vital rates and to learn the importance of such variation to demographic performance of offspring and on overall population dynamics.

Intermittent instability is widespread in plankton communities

Chaotic dynamics appear to be prevalent in short-lived organisms including plankton and may limit long-term predictability. However, few studies have explored how dynamical stability varies through time, across space and at different taxonomic resolutions. Using plankton time series data from 17 lakes and 4 marine sites, we found seasonal patterns of local instability in many species, that short-term predictability was related to local instability, and that local instability occurred most often in the spring, associated with periods of high growth. Taxonomic aggregates were more stable and more predictable than finer groupings. Across sites, higher latitude locations had higher Lyapunov exponents and greater seasonality in local instability, but only at coarser taxonomic resolution. Overall, these results suggest that prediction accuracy, sensitivity to change and management efficacy may be greater at certain times of year and that prediction will be more feasible for taxonomic aggregates.

Pathways to global-change effects on biodiversity: new opportunities for dynamically forecasting demography and species interactions

In structured populations, persistence under environmental change may be particularly threatened when abiotic factors simultaneously negatively affect survival and reproduction of several life cycle stages, as opposed to a single stage. Such effects can then be exacerbated when species interactions generate reciprocal feedbacks between the demographic rates of the different species. Despite the importance of such demographic feedbacks, forecasts that account for them are limited as individual-based data on interacting species are perceived to be essential for such mechanistic forecasting-but are rarely available. Here, we first review the current shortcomings in assessing demographic feedbacks in population and community dynamics. We then present an overview of advances in statistical tools that provide an opportunity to leverage population-level data on abundances of multiple species to infer stage-specific demography. Lastly, we showcase a state-of-the-art Bayesian method to infer and project stage-specific survival and reproduction for several interacting species in a Mediterranean shrub community. This case study shows that climate change threatens populations most strongly by changing the interaction effects of conspecific and heterospecific neighbours on both juvenile and adult survival. Thus, the repurposing of multi-species abundance data for mechanistic forecasting can substantially improve our understanding of emerging threats on biodiversity.

Forecasting the future of life in Antarctica

Antarctic ecosystems are under increasing anthropogenic pressure, but efforts to predict the responses of Antarctic biodiversity to environmental change are hindered by considerable data challenges. Here, we illustrate how novel data capture technologies provide exciting opportunities to sample Antarctic biodiversity at wider spatiotemporal scales. Data integration frameworks, such as point process and hierarchical models, can mitigate weaknesses in individual data sets, improving confidence in their predictions. Increasing process knowledge in models is imperative to achieving improved forecasts of Antarctic biodiversity, which can be attained for data-limited species using hybrid modelling frameworks. Leveraging these state-of-the-art tools will help to overcome many of the data scarcity challenges presented by the remoteness of Antarctica, enabling more robust forecasts both near- and long-term.

Towards the fully automated monitoring of ecological communities

High-resolution monitoring is fundamental to understand ecosystems dynamics in an era of global change and biodiversity declines. While real-time and automated monitoring of abiotic components has been possible for some time, monitoring biotic components-for example, individual behaviours and traits, and species abundance and distribution-is far more challenging. Recent technological advancements offer potential solutions to achieve this through: (i) increasingly affordable high-throughput recording hardware, which can collect rich multidimensional data, and (ii) increasingly accessible artificial intelligence approaches, which can extract ecological knowledge from large datasets. However, automating the monitoring of facets of ecological communities via such technologies has primarily been achieved at low spatiotemporal resolutions within limited steps of the monitoring workflow. Here, we review existing technologies for data recording and processing that enable automated monitoring of ecological communities. We then present novel frameworks that combine such technologies, forming fully automated pipelines to detect, track, classify and count multiple species, and record behavioural and morphological traits, at resolutions which have previously been impossible to achieve. Based on these rapidly developing technologies, we illustrate a solution to one of the greatest challenges in ecology: the ability to rapidly generate high-resolution, multidimensional and standardised data across complex ecologies.

Near-term phytoplankton forecasts reveal the effects of model time step and forecast horizon on predictability

As climate and land use increase the variability of many ecosystems, forecasts of ecological variables are needed to inform management and use of ecosystem services. In particular, forecasts of phytoplankton would be especially useful for drinking water management, as phytoplankton populations are exhibiting greater fluctuations due to human activities. While phytoplankton forecasts are increasing in number, many questions remain regarding the optimal model time step (the temporal frequency of the forecast model output), time horizon (the length of time into the future a prediction is made) for maximizing forecast performance, as well as what factors contribute to uncertainty in forecasts and their scalability among sites. To answer these questions, we developed near-term, iterative forecasts of phytoplankton 1-14 days into the future using forecast models with three different time steps (daily, weekly, fortnightly), that included a full uncertainty partitioning analysis at two drinking water reservoirs. We found that forecast accuracy varies with model time step and forecast horizon, and that forecast models can outperform null estimates under most conditions. Weekly and fortnightly forecasts consistently outperformed daily forecasts at 7-day and 14-day horizons, a trend that increased up to the 14-day forecast horizon. Importantly, our work suggests that forecast accuracy can be increased by matching the forecast model time step to the forecast horizon for which predictions are needed. We found that model process uncertainty was the primary source of uncertainty in our phytoplankton forecasts over the forecast period, but parameter uncertainty increased during phytoplankton blooms and when scaling the forecast model to a new site. Overall, our scalability analysis shows promising results that simple models can be transferred to produce forecasts at additional sites. Altogether, our study advances our understanding of how forecast model time step and forecast horizon influence the forecastability of phytoplankton dynamics in aquatic systems and adds to the growing body of work regarding the predictability of ecological systems broadly.

Predictability and transferability of local biodiversity environment relationships

Background

Biodiversity varies in space and time, and often in response to environmental heterogeneity. Indicators in the form of local biodiversity measures-such as species richness or abundance-are common tools to capture this variation. The rise of readily available remote sensing data has enabled the characterization of environmental heterogeneity in a globally robust and replicable manner. Based on the assumption that differences in biodiversity measures are generally related to differences in environmental heterogeneity, these data have enabled projections and extrapolations of biodiversity in space and time. However so far little work has been done on quantitatively evaluating if and how accurately local biodiversity measures can be predicted.

Methods

Here I combine estimates of biodiversity measures from terrestrial local biodiversity surveys with remotely-sensed data on environmental heterogeneity globally. I then determine through a cross-validation framework how accurately local biodiversity measures can be predicted within ("predictability") and across similar ("transferability") biodiversity surveys.

Results

I found that prediction errors can be substantial, with error magnitudes varying between different biodiversity measures, taxonomic groups, sampling techniques and types of environmental heterogeneity characterizations. And although errors associated with model predictability were in many cases relatively low, these results question-particular for transferability-our capability to accurately predict and project local biodiversity measures based on environmental heterogeneity. I make the case that future predictions should be evaluated based on their accuracy and inherent uncertainty, and ecological theories be tested against whether we are able to make accurate predictions from local biodiversity data.

Genetic architecture of dispersal and local adaptation drives accelerating range expansions

Contemporary evolution has the potential to significantly alter biotic responses to global change, including range expansion dynamics and biological invasions. Models predicting range dynamics often make highly simplifying assumptions about the genetic architecture underlying relevant traits. However, genetic architecture defines evolvability and higher-order evolutionary processes, which determine whether evolution will be able to keep up with environmental change or not. Therefore, we here study the impact of the genetic architecture of dispersal and local adaptation, two central traits of high relevance for range expansions, on the dynamics and predictability of invasion into an environmental gradient, such as temperature. In our theoretical model we assume that dispersal and local adaptation traits result from the products of two noninteracting gene-regulatory networks (GRNs). We compare our model to simpler quantitative genetics models and show that in the GRN model, range expansions are accelerating and less predictable. We further find that accelerating dynamics in the GRN model are primarily driven by an increase in the rate of local adaptation to novel habitats which results from greater sensitivity to mutation (decreased robustness) and increased gene expression. Our results highlight how processes at microscopic scales, here within genomes, can impact the predictions of large-scale, macroscopic phenomena, such as range expansions, by modulating the rate of evolution.

Forecasting in the face of ecological complexity: Number and strength of species interactions determine forecast skill in ecological communities

The potential for forecasting the dynamics of ecological systems is currently unclear, with contrasting opinions regarding its feasibility due to ecological complexity. To investigate forecast skill within and across systems, we monitored a microbial system exposed to either constant or fluctuating temperatures in a 5-month-long laboratory experiment. We tested how forecasting of species abundances depends on the number and strength of interactions and on model size (number of predictors). We also tested how greater system complexity (i.e. the fluctuating temperatures) impacted these relations. We found that the more interactions a species had, the weaker these interactions were and the better its abundance was predicted. Forecast skill increased with model size. Greater system complexity decreased forecast skill for three out of eight species. These insights into how abundance prediction depends on the connectedness of the species within the system and on overall system complexity could improve species forecasting and monitoring.

A triad of kicknet sampling, eDNA metabarcoding, and predictive modeling to assess richness of mayflies, stoneflies and caddisflies in rivers

Monitoring biodiversity is essential to understand the impacts of human activities and for effective management of ecosystems. Thereby, biodiversity can be assessed through direct collection of targeted organisms, through indirect evidence of their presence (e.g. signs, environmental DNA, camera trap, etc.), or through extrapolations from species distribution and species richness models. Differences in approaches used in biodiversity assessment, however, may come with individual challenges and hinder cross-study comparability. In the context of rapidly developing techniques, we compared three different approaches in order to better understand assessments of aquatic macroinvertebrate diversity. Specifically, we compared the community composition and species richness of three orders of aquatic macroinvertebrates (mayflies, stoneflies, and caddisflies, hereafter EPT) obtained via eDNA metabarcoding and via traditional in situ kicknet sampling to catchment-level based predictions of a species richness model. We used kicknet data from 24 sites in Switzerland and compared taxonomic lists to those obtained using eDNA amplified with two different primer sets. Richness detected by these methods was compared to the independent predictions made by a statistical species richness model, that is, a generalized linear model using landscape-level features to estimate EPT diversity. Despite the ability of eDNA to consistently detect some EPT species found by traditional sampling, we found important discrepancies in community composition between the kicknet and eDNA approaches, particularly at a local scale. We found the EPT-specific primer set fwhF2/EPTDr2n, detected a greater number of targeted EPT species compared to the more general primer set mlCOIintF/HCO2198. Moreover, we found that the species richness measured by eDNA from either primer set was poorly correlated to the richness measured by kicknet sampling (Pearson correlation = 0.27) and that the richness estimated by eDNA and kicknet were poorly correlated with the prediction of the species richness model (Pearson correlation = 0.30 and 0.44, respectively). The weak relationships between the traditional kicknet sampling and eDNA with this model indicates inherent limitations in upscaling species richness estimates, and possibly a limited ability of the model to meet real world expectations. It is also possible that the number of replicates was not sufficient to detect ambiguous correlations. Future challenges include improving the accuracy and sensitivity of each approach individually, yet also acknowledging their respective limitations, in order to best meet stakeholder demands and address the biodiversity crisis we are facing.

Spatiotemporal dynamics drive synergism of land use and climatic extreme events in insect meta-populations

Ecosystems are increasingly threatened by co-occurring stressors associated with anthropogenic global change. Spatial stressor patterns range from local to regional to global, and temporal stressor patterns from discrete to continuous. To date, most multiple stressor studies covered short periods and focused on local effects and interactions. However, it remains largely unknown how stressors with different spatiotemporal patterns interact in their effects over longer periods. In particular, at higher spatial scales, biotic dynamics in ecological networks complicate the understanding of stressor interactions. We used a spatially explicit meta-population model for a generic freshwater insect, parameterized based on traits of the European damselfly Coenagrion mercuriale, to simulate scenarios of discrete climatic extreme events and continuous land use-related stress. Climatic extreme events were modeled as recurring mortality in all patches, whereas land use permanently influenced meta-populations via patch qualities and network connectivity. We found that the risk of discrete climatic extreme events to meta-populations depended strongly on the proportion of land use types, with effects ranging from negligible to extinction. Land use-related stress limited recovery in meta-populations from effects of climatic extreme events, resulting in synergistic stressor interactions. Moreover, the spatial configuration of land use type influenced the combined stressor effects with clustered configurations resulting in lower effects compared to a random configuration. Finally, we found that combined stressor effects can vary with the time point at which they were determined, indicating that inconclusive results in multiple stressor research can partly be due to differences in the time of determination. We conclude that conservation should focus on regional landscape management to mitigate risks on meta-populations from future, intensified extreme climate events. Reducing land use effects, thus improving patch quality and network connectivity, can compensate for effects of additional discrete stressors and, in turn, synergistic interactions.

PERFICT: A Re-imagined foundation for predictive ecology

Making predictions from ecological models—and comparing them to data—offers a coherent approach to evaluate model quality, regardless of model complexity or modelling paradigm. To date, our ability to use predictions for developing, validating, updating, integrating and applying models across scientific disciplines while influencing management decisions, policies, and the public has been hampered by disparate perspectives on prediction and inadequately integrated approaches. We present an updated foundation for Predictive Ecology based on seven principles applied to ecological modelling: make frequent Predictions, Evaluate models, make models Reusable, Freely accessible and Interoperable, built within Continuous workflows that are routinely Tested (PERFICT). We outline some benefits of working with these principles: accelerating science; linking with data science; and improving science‐policy integration.

Simultaneously estimating food web connectance and structure with uncertainty

Food web models explain and predict the trophic interactions in a food web, and they can infer missing interactions among the organisms. The allometric diet breadth model (ADBM) is a food web model based on the foraging theory. In the ADBM, the foraging parameters are allometrically scaled to body sizes of predators and prey. In Petchey et al. (Proceedings of the National Academy of Sciences, 2008; 105: 4191), the parameterization of the ADBM had two limitations: (a) the model parameters were point estimates and (b) food web connectance was not estimated.

The novelty of our current approach is: (a) We consider multiple predictions from the ADBM by parameterizing it with approximate Bayesian computation, to estimate parameter distributions and not point estimates. (b) Connectance emerges from the parameterization, by measuring model fit using the true skill statistic, which takes into account prediction of both the presences and absences of links.

We fit the ADBM using approximate Bayesian computation to 12 observed food webs from a wide variety of ecosystems. Estimated connectance was consistently greater than previously found. In some of the food webs, considerable variation in estimated parameter distributions occurred and resulted in considerable variation (i.e., uncertainty) in predicted food web structure.

These results lend weight to the possibility that the observed food web data is missing some trophic links that do actually occur. It also seems likely that the ADBM likely predicts some links that do not exist. The latter could be addressed by accounting in the ADBM for additional traits other than body size. Further work could also address the significance of uncertainty in parameter estimates for predicted food web responses to environmental change.

Increased adoption of best practices in ecological forecasting enables comparisons of forecastability

Near-term iterative forecasting is a powerful tool for ecological decision support and has the potential to transform our understanding of ecological predictability. However, to this point, there has been no cross-ecosystem analysis of near-term ecological forecasts, making it difficult to synthesize diverse research efforts and prioritize future developments for this emerging field. In this study, we analyzed 178 near-term (≤10-yr forecast horizon) ecological forecasting papers to understand the development and current state of near-term ecological forecasting literature and to compare forecast accuracy across scales and variables. Our results indicated that near-term ecological forecasting is widespread and growing: forecasts have been produced for sites on all seven continents and the rate of forecast publication is increasing over time. As forecast production has accelerated, some best practices have been proposed and application of these best practices is increasing. In particular, data publication, forecast archiving, and workflow automation have all increased significantly over time. However, adoption of proposed best practices remains low overall: for example, despite the fact that uncertainty is often cited as an essential component of an ecological forecast, only 45% of papers included uncertainty in their forecast outputs. As the use of these proposed best practices increases, near-term ecological forecasting has the potential to make significant contributions to our understanding of forecastability across scales and variables. In this study, we found that forecastability (defined here as realized forecast accuracy) decreased in predictable patterns over 1-7 d forecast horizons. Variables that were closely related (i.e., chlorophyll and phytoplankton) displayed very similar trends in forecastability, while more distantly related variables (i.e., pollen and evapotranspiration) exhibited significantly different patterns. Increasing use of proposed best practices in ecological forecasting will allow us to examine the forecastability of additional variables and timescales in the future, providing a robust analysis of the fundamental predictability of ecological variables.

A niche for null models in adaptive resource management

As global systems rapidly change, our collective ability to predict future ecological dynamics will become increasingly important for successful natural resource management. By merging stakeholder objectives with system uncertainty, and by adapting actions to changing systems and knowledge, adaptive resource management (ARM) provides a rigorous platform for making sound decisions in a changing world. Critically, however, applications of ARM could be improved by employing benchmarks (i.e., points of reference) for determining when learning is occurring through the cycle of monitoring, modeling, and decision-making steps in ARM. Many applications of ARM use multiple model-based hypotheses to identify and reduce systematic uncertainty over time, but generally lack benchmarks for gauging discovery of scientific evidence and learning. This creates the danger of thinking that directional changes in model weights or rankings are indicative of evidence for hypotheses, when possibly all competing models are inadequate. There is thus a somewhat obvious, but yet to be filled niche for including benchmarks for learning in ARM. We contend that carefully designed "ecological null models," which are structured to produce an expected ecological pattern in the absence of a hypothesized mechanism, can serve as suitable benchmarks. Using a classic case study of mallard harvest management that is often used to demonstrate the successes of ARM for learning about ecological mechanisms, we show that simple ecological null models, such as population persistence (Nt +1 = Nt ), provide more robust near-term forecasts of population abundance than the currently used mechanistic models. More broadly, ecological null models can be used as benchmarks for learning in ARM that trigger the need for discarding model parameterizations and developing new ones when prevailing models underperform the ecological null model. Identifying mechanistic models that surpass these benchmarks will improve learning through ARM and help decision-makers keep pace with a rapidly changing world.

Understanding mast seeding for conservation and land management

Masting, the intermittent and synchronous production of large seed crops, can have profound consequences for plant populations and the food webs that are built on their seeds. For centuries, people have recorded mast crops because of their importance in managing wildlife populations. In the past 30 years, we have begun to recognize the importance of masting in conserving and managing many other aspects of the environment: promoting the regeneration of forests following fire or other disturbance, conserving rare plants, conscientiously developing the use of edible seeds as non-timber forest products, coping with the consequences of extinctions on seed dispersal, reducing the impacts of plant invasions with biological control, suppressing zoonotic diseases and preventing depredation of endemic fauna. We summarize current instances and future possibilities of a broad set of applications of masting. By exploring in detail several case studies, we develop new perspectives on how solutions to pressing conservation and land management problems may benefit by better understanding the dynamics of seed production. A lesson common to these examples is that masting can be used to time management, and often, to do this effectively, we need models that explicitly forecast masting and the dynamics of seed-eating animals into the near-term future. This article is part of the theme issue 'The ecology and evolution of synchronized seed production in plants'.

Integrated population modelling reveals potential drivers of demography from partially aligned data: a case study of snowy plover declines under human stressors

Knowledge of demography is essential for understanding wildlife population dynamics and developing appropriate conservation plans. However, population survey and demographic data (e.g., capture-recapture) are not always aligned in space and time, hindering our ability to robustly estimate population size and demographic processes. Integrated population models (IPMs) can provide inference for population dynamics with poorly aligned but jointly analysed population and demographic data. In this study, we used an IPM to analyse partially aligned population and demographic data of a migratory shorebird species, the snowy plover (Charadrius nivosus). Snowy plover populations have declined dramatically during the last two decades, yet the demographic mechanisms and environmental drivers of these declines remain poorly understood, hindering development of appropriate conservation strategies. We analysed 21 years (1998-2018) of partially aligned population survey, nest survey, and capture-recapture-resight data in three snowy plover populations (i.e., Texas, New Mexico, Oklahoma) in the Southern Great Plains of the US. By using IPMs we aimed to achieve better precision while evaluating the effects of wetland habitat and climatic factors (minimum temperature, wind speed) on snowy plover demography. Our IPM provided reasonable precision for productivity measures even with missing data, but population and survival estimates had greater uncertainty in years without corresponding data. Our model also uncovered the complex relationships between wetland habitat, climate, and demography with reasonable precision. Wetland habitat had positive effects on snowy plover productivity (i.e., clutch size and clutch fate), indicating the importance of protecting wetland habitat under climate change and other human stressors for the conservation of this species. We also found a positive effect of minimum temperature on snowy plover productivity, indicating potential benefits of warmth during night on their population. Based on our results, we suggest prioritizing population and capture-recapture surveys for understanding population dynamics and underlying demographic processes when data collection is limited by time and/or financial resources. Our modelling approach can be used to allocate limited conservation resources for evidence-based decision-making.

Linking species traits and demography to explain complex temperature responses across levels of organization

Microbial communities regulate ecosystem responses to climate change. However, predicting these responses is challenging because of complex interactions among processes at multiple levels of organization. Organismal traits that determine individual performance and ecological interactions are essential for scaling up environmental responses from individuals to ecosystems. We combine protist microcosm experiments and mathematical models to show that key traits-cell size, shape, and contents-each explain different aspects of species' demographic responses to changes in temperature. These differences in species' temperature responses have complex cascading effects across levels of organization-causing nonlinear shifts in total community respiration rates across temperatures via coordinated changes in community composition, equilibrium densities, and community-mean species mass in experimental protist communities that tightly match theoretical predictions. Our results suggest that traits explain variation in population growth, and together, these two factors scale up to influence community- and ecosystem-level processes across temperatures. Connecting the multilevel microbial processes that ultimately influence climate in this way will help refine predictions about complex ecosystem-climate feedbacks and the pace of climate change itself.

Prediction of habitat suitability for Patrinia sibirica Juss. in the Southern Urals

The paper presents the results of predictions of the habitat persistence for rare relict of the Pleistocene floristic complex Patrinia sibirica (L.) Juss. in the Southern Urals under various forecasted climate change scenarios. Climate variables from CHELSA BIOCLIM, elevation data (GMTED2010) and coarse fragment content in the top level of soil were used as predictors for modeling in the MaxEnt software. The impact of climate change on P. sibirica habitats under the RCP4.5 and RCP8.5 scenarios calculated from an ensemble of four general circulation models has been analyzed. The modeling has shown that the changes in the habitat suitability depend on the altitude. Deterioration of the habitats could be attributed to a temperature increase in mountain forest locations, and to a precipitation of driest quarter increase in mountain forest-steppe locations. In both cases, this leads to the expansion of forest and shrub vegetation. Monitoring of the habitat persistence of P. sibirica and other relict species of the Pleistocene floristic complex can play a major role in predictions, as their massive decline would constitute that climatic changes exceed the ranges of their fluctuations in the Holocene.

Challenges in updating habitat suitability models: An example with the lesser prairie-chicken

Habitat loss from land-use change is one of the top causes of declines in wildlife species of concern. As such, it is critical to assess and reassess habitat suitability as land cover and anthropogenic features change for both monitoring and developing current information to inform management decisions. However, there are obstacles that must be overcome to develop consistent assessments through time. A range-wide lek habitat suitability model for the lesser prairie-chicken (Tympanuchus pallidicinctus), currently under review by the U. S. Fish and Wildlife Service for potential listing under the Endangered Species Act, was published in 2016. This model was based on lek data from 2002 to 2012, land cover data ranging from 2001 to 2013, and anthropogenic features from circa 2011, and has been used to help guide lesser prairie-chicken management and anthropogenic development actions. We created a second iteration model based on new lek surveys (2015 to 2019) and updated predictors (2016 land cover and cleaned/updated anthropogenic data) to evaluate changes in lek suitability and to quantify current range-wide habitat suitability. Only three of 11 predictor variables were directly comparable between the iterations, making it difficult to directly assess what predicted changes resulted from changes in model inputs versus actual landscape change. The second iteration model showed a similar positive relationship with land cover and negative relationship with anthropogenic features to the first iteration, but exhibited more variation among candidate models. Range-wide, more suitable habitat was predicted in the second iteration. The Shinnery Oak Ecoregion, however, exhibited a loss in predicted suitable habitat that could be due to predictor source changes. Iterated models such as this are important to ensure current information is being used in conservation and development decisions.

Advancing estuarine ecological forecasts: seasonal hypoxia in Chesapeake Bay

Ecological forecasts are quantitative tools that can guide ecosystem management. The coemergence of extensive environmental monitoring and quantitative frameworks allows for widespread development and continued improvement of ecological forecasting systems. We use a relatively simple estuarine hypoxia model to demonstrate advances in addressing some of the most critical challenges and opportunities of contemporary ecological forecasting, including predictive accuracy, uncertainty characterization, and management relevance. We explore the impacts of different combinations of forecast metrics, drivers, and driver time windows on predictive performance. We also incorporate multiple sets of state-variable observations from different sources and separately quantify model prediction error and measurement uncertainty through a flexible Bayesian hierarchical framework. Results illustrate the benefits of (1) adopting forecast metrics and drivers that strike an optimal balance between predictability and relevance to management, (2) incorporating multiple data sources in the calibration data set to separate and propagate different sources of uncertainty, and (3) using the model in scenario mode to probabilistically evaluate the effects of alternative management decisions on future ecosystem state. In the Chesapeake Bay, the subject of this case study, we find that average summer or total annual hypoxia metrics are more predictable than monthly metrics and that measurement error represents an important source of uncertainty. Application of the model in scenario mode suggests that absent watershed management actions over the past decades, long-term average hypoxia would have increased by 7% compared to 1985. Conversely, the model projects that if management goals currently in place to restore the Bay are met, long-term average hypoxia would eventually decrease by 32% with respect to the mid-1980s.

A Test of Species Distribution Model Transferability Across Environmental and Geographic Space for 108 Western North American Tree Species

Predictions from species distribution models (SDMs) are commonly used in support of environmental decision-making to explore potential impacts of climate change on biodiversity. However, because future climates are likely to differ from current climates, there has been ongoing interest in understanding the ability of SDMs to predict species responses under novel conditions (i.e., model transferability). Here, we explore the spatial and environmental limits to extrapolation in SDMs using forest inventory data from 11 model algorithms for 108 tree species across the western United States. Algorithms performed well in predicting occurrence for plots that occurred in the same geographic region in which they were fitted. However, a substantial portion of models performed worse than random when predicting for geographic regions in which algorithms were not fitted. Our results suggest that for transfers in geographic space, no specific algorithm was better than another as there were no significant differences in predictive performance across algorithms. There were significant differences in predictive performance for algorithms transferred in environmental space with GAM performing best. However, the predictive performance of GAM declined steeply with increasing extrapolation in environmental space relative to other algorithms. The results of this study suggest that SDMs may be limited in their ability to predict species ranges beyond the environmental data used for model fitting. When predicting climate-driven range shifts, extrapolation may also not reflect important biotic and abiotic drivers of species ranges, and thus further misrepresent the realized shift in range. Future studies investigating transferability of process based SDMs or relationships between geodiversity and biodiversity may hold promise.

Evaluating probabilistic ecological forecasts

Probabilistic near-term forecasting facilitates evaluation of model predictions against observations and is of pressing need in ecology to inform environmental decision-making and effect societal change. Despite this imperative, many ecologists are unfamiliar with the widely used tools for evaluating probabilistic forecasts developed in other fields. We address this gap by reviewing the literature on probabilistic forecast evaluation from diverse fields including climatology, economics, and epidemiology. We present established practices for selecting evaluation data (end-sample hold out), graphical forecast evaluation (times-series plots with uncertainty, probability integral transform plots), quantitative evaluation using scoring rules (log, quadratic, spherical, and ranked probability scores), and comparing scores across models (skill score, Diebold-Mariano test). We cover common approaches, highlight mathematical concepts to follow, and note decision points to allow application of general principles to specific forecasting endeavors. We illustrate these approaches with an application to a long-term rodent population time series currently used for ecological forecasting and discuss how ecology can continue to learn from and drive the cross-disciplinary field of forecasting science.

Non-additive effects of foundation species determine the response of aquatic ecosystems to nutrient perturbation

Eutrophication is a persistent threat to aquatic ecosystems worldwide. Foundation species, namely those that play a central role in the structuring of communities and functioning of ecosystems, are likely important for the resilience of aquatic ecosystems in the face of disturbance. However, little is known about how interactions among such species influence ecosystem responses to nutrient perturbation. Here, using an array (N = 20) of outdoor experimental pond ecosystems (15,000 L), we manipulated the presence of two foundation species, the macrophyte Myriophyllum spicatum and the mussel Dreissena polymorpha, and quantified ecosystem responses to multiple nutrient disturbances, spread over two years. In the first year, we added five nutrient pulses, ramping up from 10 to 50 μg P/L over a 10-week period from mid-July to mid-October, and in the second year, we added a single large pulse of 50 μg P/L in mid-October. We used automated sondes to measure multiple ecosystems properties at high frequency (15-minute intervals), including phytoplankton and dissolved organic matter fluorescence, and to model whole-ecosystem metabolism. Overall, both foundation species strongly affected the ecosystem responses to nutrient perturbation, and, as expected, initially suppressed the increase in phytoplankton abundance following nutrient additions. However, when both species were present, phytoplankton biomass increased substantially relative to other treatment combinations: non-additivity was evident for multiple ecosystem metrics following the nutrient perturbations in both years but was diminished in the intervening months between our perturbations. Overall, these results demonstrate how interactions between foundation species can cause surprisingly strong deviations from the expected responses of aquatic ecosystems to perturbations such as nutrient additions.

A practical guide to selecting models for exploration, inference, and prediction in ecology

Selecting among competing statistical models is a core challenge in science. However, the many possible approaches and techniques for model selection, and the conflicting recommendations for their use, can be confusing. We contend that much confusion surrounding statistical model selection results from failing to first clearly specify the purpose of the analysis. We argue that there are three distinct goals for statistical modeling in ecology: data exploration, inference, and prediction. Once the modeling goal is clearly articulated, an appropriate model selection procedure is easier to identify. We review model selection approaches and highlight their strengths and weaknesses relative to each of the three modeling goals. We then present examples of modeling for exploration, inference, and prediction using a time series of butterfly population counts. These show how a model selection approach flows naturally from the modeling goal, leading to different models selected for different purposes, even with exactly the same data set. This review illustrates best practices for ecologists and should serve as a reminder that statistical recipes cannot substitute for critical thinking or for the use of independent data to test hypotheses and validate predictions.

Iterative model predictions for wildlife populations impacted by rapid climate change

To improve understanding and management of the consequences of current rapid environmental change, ecologists advocate using long-term monitoring data series to generate iterative near-term predictions of ecosystem responses. This approach allows scientific evidence to increase rapidly and management strategies to be tailored simultaneously. Iterative near-term forecasting may therefore be particularly useful for adaptive monitoring of ecosystems subjected to rapid climate change. Here, we show how to implement near-term forecasting in the case of a harvested population of rock ptarmigan in high-arctic Svalbard, a region subjected to the largest and most rapid climate change on Earth. We fitted state-space models to ptarmigan counts from point transect distance sampling during 2005-2019 and developed two types of predictions: (1) explanatory predictions to quantify the effect of potential drivers of ptarmigan population dynamics, and (2) anticipatory predictions to assess the ability of candidate models of increasing complexity to forecast next-year population density. Based on the explanatory predictions, we found that a recent increasing trend in the Svalbard rock ptarmigan population can be attributed to major changes in winter climate. Currently, a strong positive effect of increasing average winter temperature on ptarmigan population growth outweighs the negative impacts of other manifestations of climate change such as rain-on-snow events. Moreover, the ptarmigan population may compensate for current harvest levels. Based on the anticipatory predictions, the near-term forecasting ability of the models improved nonlinearly with the length of the time series, but yielded good forecasts even based on a short time series. The inclusion of ecological predictors improved forecasts of sharp changes in next-year population density, demonstrating the value of ecosystem-based monitoring. Overall, our study illustrates the power of integrating near-term forecasting in monitoring systems to aid understanding and management of wildlife populations exposed to rapid climate change. We provide recommendations for how to improve this approach.

Towards robust statistical inference for complex computer models

Ecologists increasingly rely on complex computer simulations to forecast ecological systems. To make such forecasts precise, uncertainties in model parameters and structure must be reduced and correctly propagated to model outputs. Naively using standard statistical techniques for this task, however, can lead to bias and underestimation of uncertainties in parameters and predictions. Here, we explain why these problems occur and propose a framework for robust inference with complex computer simulations. After having identified that model error is more consequential in complex computer simulations, due to their more pronounced nonlinearity and interconnectedness, we discuss as possible solutions data rebalancing and adding bias corrections on model outputs or processes during or after the calibration procedure. We illustrate the methods in a case study, using a dynamic vegetation model. We conclude that developing better methods for robust inference of complex computer simulations is vital for generating reliable predictions of ecosystem responses.

Predictability of Biotic Stress Structures Plant Defence Evolution

To achieve ecological and reproductive success, plants need to mitigate a multitude of stressors. The stressors encountered by plants are highly dynamic but typically vary predictably due to seasonality or correlations among stressors. As plants face physiological and ecological constraints in responses to stress, it can be beneficial for plants to evolve the ability to incorporate predictable patterns of stress in their life histories. Here, we discuss how plants predict adverse conditions, which plant strategies integrate predictability of biotic stress, and how such strategies can evolve. We propose that plants commonly optimise responses to correlated sequences or combinations of herbivores and pathogens, and that the predictability of these patterns is a key factor governing plant strategies in dynamic environments.

Ecological forecasts reveal limitations of common model selection methods: predicting changes in beaver colony densities

Over the past two decades, there have been numerous calls to make ecology a more predictive science through direct empirical assessments of ecological models and predictions. While the widespread use of model selection using information criteria has pushed ecology toward placing a higher emphasis on prediction, few attempts have been made to validate the ability of information criteria to correctly identify the most parsimonious model with the greatest predictive accuracy. Here, we used an ecological forecasting framework to test the ability of information criteria to accurately predict the relative contribution of density dependence and density-independent factors (forage availability, harvest, weather, wolf [Canis lupus] density) on inter-annual fluctuations in beaver (Castor canadensis) colony densities. We modeled changes in colony densities using a discrete-time Gompertz model, and assessed the performance of four models using information criteria values: density-independent models with (1) and without (2) environmental covariates; and density-dependent models with (3) and without (4) environmental covariates. We then evaluated the forecasting accuracy of each model by withholding the final one-third of observations from each population and compared observed vs. predicted densities. Information criteria and our forecasting accuracy metrics both provided strong evidence of compensatory density dependence in the annual dynamics of beaver colony densities. However, despite strong within-sample performance by the most complex model (density-dependent with covariates) as determined using information criteria, hindcasts of colony densities revealed that the much simpler density-dependent model without covariates performed nearly as well predicting out-of-sample colony densities. The hindcast results indicated that the complex model over-fit our data, suggesting that parameters identified by information criteria as important predictor variables are only marginally valuable for predicting landscape-scale beaver colony dynamics. Our study demonstrates the importance of evaluating ecological models and predictions with long-term data and revealed how a known limitation of information criteria (over-fitting of complex models) can affect our interpretation of ecological dynamics. While incorporating knowledge of the factors that influence animal population dynamics can improve population forecasts, we suggest that comparing forecast performance metrics can likewise improve our knowledge of the factors driving population dynamics.

Seasonality and uncertainty in global COVID-19 growth rates

The virus causing COVID-19 has spread rapidly worldwide and threatens millions of lives. It remains unknown, as of April 2020, whether summer weather will reduce its spread, thereby alleviating strains on hospitals and providing time for vaccine development. Early insights from laboratory studies and research on related viruses predicted that COVID-19 would decline with higher temperatures, humidity, and ultraviolet (UV) light. Using current, fine-scaled weather data and global reports of infections, we develop a model that explains 36% of the variation in maximum COVID-19 growth rates based on weather and demography (17%) and country-specific effects (19%). UV light is most strongly associated with lower COVID-19 growth. Projections suggest that, without intervention, COVID-19 will decrease temporarily during summer, rebound by autumn, and peak next winter. Validation based on data from May and June 2020 confirms the generality of the climate signal detected. However, uncertainty remains high, and the probability of weekly doubling rates remains >20% throughout summer in the absence of social interventions. Consequently, aggressive interventions will likely be needed despite seasonal trends.

Spatial structure in protected forest-grassland mosaics: Exploring futures under climate change

In mosaic ecosystems, multiple land types coexist as alternative stable states exhibiting distinct spatial patterns. Forest-grassland mosaics are ecologically valuable, due to their high species richness. However, anthropogenic disturbances threaten these ecosystems. Designating protected areas is one approach to preserving natural mosaics. Such work must account for climate change, yet there are few spatially explicit models of mosaics under climate change that can predict its effects. We construct a spatially explicit simulation model for a natural forest-grassland mosaic, parameterized for Southern Brazil. Using this model, we investigate how the spatial structure of these systems is altered under climate change and other disturbance regimes. By including local spatial interactions and fire-mediated forest recruitment, our model reproduces important spatial features of protected real-world mosaics, including the number of forest patches and overall forest cover. Multiple concurrent changes in environmental conditions have greater impacts on tree cover and spatial structure in simulated mosaics than single changes. This sensitivity reflects the narrow range of conditions under which simulated mosaics persist and emphasizes their vulnerability. Our model predicts that, in protected mosaics, climate change impacts on the fire-mediated threshold to recruitment will likely result in substantial increases in forest cover under Representative Concentration Pathway (RCP) 8.5, with potential for mosaic loss over a broad range of initial forest cover levels. Forest cover trajectories are similar until 2150, when cover increases under RCP 8.5 outpace those under RCP 2.6. Mosaics that persist under RCP 8.5 may experience structural alterations at the patch and landscape level. Our simple model predicts several realistic aspects of spatial structure as well as plausible responses to likely regional climate shifts. Hence, further model development could provide a useful tool when building strategies for protecting these ecosystems, by informing site selection for conservation areas that will be favourable to forest-grassland mosaics under future climates.

Ecological forecasts to inform near-term management of threats to biodiversity

Ecosystems are being altered by rapid and interacting changes in natural processes and anthropogenic threats to biodiversity. Uncertainty in historical, current and future effectiveness of actions hampers decisions about how to mitigate changes to prevent biodiversity loss and species extinctions. Research in resource management, agriculture and health indicates that forecasts predicting the effects of near-term or seasonal environmental conditions on management greatly improve outcomes. Such forecasts help resolve uncertainties about when and how to operationalize management. We reviewed the scientific literature on environmental management to investigate whether near-term forecasts are developed to inform biodiversity decisions in Australia, a nation with one of the highest recent extinction rates across the globe. We found that forecasts focused on economic objectives (e.g. fisheries management) predict on significantly shorter timelines and answer a broader range of management questions than forecasts focused on biodiversity conservation. We then evaluated scientific literature on the effectiveness of 484 actions to manage seven major terrestrial threats in Australia, to identify opportunities for near-term forecasts to inform operational conservation decisions. Depending on the action, between 30% and 80% threat management operations experienced near-term weather impacts on outcomes before, during or after management. Disease control, species translocation/reintroduction and habitat restoration actions were most frequently impacted, and negative impacts such as increased species mortality and reduced recruitment were more likely than positive impacts. Drought or dry conditions, and rainfall, were the most frequently reported weather impacts, indicating that near-term forecasts predicting the effects of low or excessive rainfall on management outcomes are likely to have the greatest benefits. Across the world, many regions are, like Australia, becoming warmer and drier, or experiencing more extreme rainfall events. Informing conservation decisions with near-term and seasonal ecological forecasting will be critical to harness uncertainties and lower the risk of threat management failure under global change.

Structural forecasting of species persistence under changing environments

The persistence of a species in a given place not only depends on its intrinsic capacity to consume and transform resources into offspring, but also on how changing environmental conditions affect its growth rate. However, the complexity of factors has typically taken us to choose between understanding and predicting the persistence of species. To tackle this limitation, we propose a probabilistic approach rooted on the statistical concepts of ensemble theory applied to statistical mechanics and on the mathematical concepts of structural stability applied to population dynamics models - what we call structural forecasting. We show how this new approach allows us to estimate a probability of persistence for single species in local communities; to understand and interpret this probability conditional on the information we have concerning a system; and to provide out-of-sample predictions of species persistence as good as the best experimental approaches without the need of extensive amounts of data.

When, why and how tumour clonal diversity predicts survival

The utility of intratumour heterogeneity as a prognostic biomarker is the subject of ongoing clinical investigation. However, the relationship between this marker and its clinical impact is mediated by an evolutionary process that is not well understood. Here, we employ a spatial computational model of tumour evolution to assess when, why and how intratumour heterogeneity can be used to forecast tumour growth rate and progression-free survival. We identify three conditions that can lead to a positive correlation between clonal diversity and subsequent growth rate: diversity is measured early in tumour development; selective sweeps are rare; and/or tumours vary in the rate at which they acquire driver mutations. Opposite conditions typically lead to negative correlation. In cohorts of tumours with diverse evolutionary parameters, we find that clonal diversity is a reliable predictor of both growth rate and progression-free survival. We thus offer explanations-grounded in evolutionary theory-for empirical findings in various cancers, including survival analyses reported in the recent TRACERx Renal study of clear-cell renal cell carcinoma. Our work informs the search for new prognostic biomarkers and contributes to the development of predictive oncology.