Predictive entomology: a causal framework for detecting and attributing insect population change

Widespread insect declines have raised concerns about ecosystem stability, yet emerging evidence shows that population trends are heterogeneous and shaped by interacting drivers rather than single stressors. The challenge is therefore to move beyond documenting patterns toward understanding the mechanisms that generate them and predicting how insects will respond to future environmental change. Environmental pressures operate across spatial and temporal scales to shape insect populations. Mechanisms determine the physiological, behavioural and ecological processes underlying population shifts, while species traits constrain sensitivity. Detection processes generate observable signals, and causal inference reveals how multiple drivers propagate through ecological networks. Here, we synthesise these domains to outline predictive entomology - an integrative framework linking drivers, mechanisms, detection and forecasting through explicit causal inference. Drawing on long-term datasets, mechanistic experiments and emerging sensing technologies, we illustrate how multiple drivers interact across scales and how causal tools can help separate confounding processes and detection artefacts from causal effects. Embedding mechanistic and causal understanding into monitoring and analytical pipelines provides the foundation for predicting ecological change, identifying emerging risks and guiding proactive conservation under global change.