The power of forecasts to advance ecological theory

Using neural ordinary differential equations to predict complex ecological dynamics from population density data

Simple models have been used to describe ecological processes for over a century. However, the complexity of ecological systems makes simple models subject to modelling bias due to simplifying assumptions or unaccounted factors, limiting their predictive power. Neural ordinary differential equations (NODEs) have surged as a machine-learning algorithm that preserves the dynamic nature of the data (Chen et al. 2018 Adv. Neural Inf. Process. Syst.). Although preserving the dynamics in the data is an advantage, the question of how NODEs perform as a forecasting tool of ecological communities is unanswered. Here, we explore this question using simulated time series of competing species in a time-varying environment. We find that NODEs provide more precise forecasts than autoregressive integrated moving average (ARIMA) models. We also find that untuned NODEs have a similar forecasting accuracy to untuned long-short term memory neural networks and both are outperformed in accuracy and precision by empirical dynamical modelling . However, we also find NODEs generally outperform all other methods when evaluating with the interval score, which evaluates precision and accuracy in terms of prediction intervals rather than pointwise accuracy. We also discuss ways to improve the forecasting performance of NODEs. The power of a forecasting tool such as NODEs is that it can provide insights into population dynamics and should thus broaden the approaches to studying time series of ecological communities.

Ecological forecasts of insect range dynamics: a broad range of taxa includes winners and losers under future climate

Species distribution models are the primary tools to project future species' distributions, but this complex task is influenced by data limitations and evolving best practices. The majority of the 53 studies we examined utilized correlative models and did not follow current best practices for validating retrospective or future environmental data layers. Despite this, a summary of results is largely unsurprising: shifts toward cooler regions, but otherwise mixed dynamics emphasizing winners and losers. Harmful insects were more likely to show positive outcomes compared with beneficial species. Our restricted ability to consider mechanisms complicates interpretation of any single study. To improve this area of modeling, more classic field and lab studies to uncover basic ecology and physiology are crucial.

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.

Planktonic functional diversity changes in synchrony with lake ecosystem state

Managing ecosystems to effectively preserve function and services requires reliable tools that can infer changes in the stability and dynamics of a system. Conceptually, functional diversity (FD) appears as a sensitive and viable monitoring metric stemming from suggestions that FD is a universally important measure of biodiversity and has a mechanistic influence on ecological processes. It is however unclear whether changes in FD consistently occur prior to state responses or vice versa, with no current work on the temporal relationship between FD and state to support a transition towards trait-based indicators. There is consequently a knowledge gap regarding when functioning changes relative to biodiversity change and where FD change falls in that sequence. We therefore examine the lagged relationship between planktonic FD and abundance-based metrics of system state (e.g. biomass) across five highly monitored lake communities using both correlation and cutting edge non-linear empirical dynamic modelling approaches. Overall, phytoplankton and zooplankton FD display synchrony with lake state but each lake is idiosyncratic in the strength of relationship. It is therefore unlikely that changes in plankton FD are identifiable before changes in more easily collected abundance metrics. These results highlight the power of empirical dynamic modelling in disentangling time lagged relationships in complex multivariate ecosystems, but suggest that FD cannot be generically viable as an early indicator. Individual lakes therefore require consideration of their specific context and any interpretation of FD across systems requires caution. However, FD still retains value as an alternative state measure or a trait representation of biodiversity when considered at the system level.