Night Moths Enter the Risk-Assessment Arena With a New Virtual Model

When pesticides are sprayed across European farmland, regulators can draw on a wealth of tools to assess the risks to honey bees and butterflies. But for the moths that fly the same fields after dark, no comparable framework has existed. That gap matters, because nocturnal moths are among the most important and least studied pollinators in agricultural landscapes, visiting night-blooming flowers and forming a vital link in food webs that sustain birds, bats and parasitoid wasps. A new study published in the Food and Ecological Systems Modelling Journal now offers the first comprehensive agent-based model designed to simulate how a typical night-flying moth, the common yellow underwing (Noctua pronuba), lives, moves and dies in a pesticide-treated landscape.

The work describes a formal model intended for integration into ALMaSS, the Animal Landscape and Man Simulation System, a long-established platform that simulates farms, fields, weather and wildlife at one-metre resolution across landscapes of typically ten by ten kilometres. Every moth in the new model exists as an individual digital agent with its own age, size, location and accumulating pesticide burden. Because each insect responds to its own simulated surroundings, the model can capture the kinds of chance events and spatial quirks that simpler, population-average assessments miss entirely. The choice of species was deliberate: N. pronuba is the type species of the enormous moth family Noctuidae, is widespread across its native Palearctic range, has become invasive in North America, and unusually for an insect, enjoys a rich historical literature on its biology, much of it from a doctoral study published in 1956.

A core technical achievement of the paper is the standardisation of how temperature drives development. The authors fitted Lobry–Rosso–Flandrois thermal performance curves, a mathematical form borrowed from microbial growth modelling, to decades-old laboratory data for eggs, larvae and pupae. Each curve is defined by four parameters: minimum, optimum and maximum developmental temperatures, and the growth rate at the optimum. For eggs, the fitted optimum came out at roughly 26 degrees Celsius with development failing above 29 degrees; larvae peaked near 29 degrees but died reliably at 27.5 degrees in the older survival experiments; pupae tolerated warmer conditions up to a 35-degree ceiling. Because the fitted curves estimated minimum development temperatures poorly, the team truncated them conservatively, assuming no development below 6 degrees for eggs and larvae and 7 degrees for pupae, values consistent with direct observation.

The larval data reveal an ecological subtlety with real consequences for pesticide exposure. Early laboratory work showed that first-instar caterpillars are active during the day, second instars are intermediate, and instars three through seven become strictly nocturnal, avoiding even moonlight. The final instar burrows beneath the soil surface, making it effectively invisible to any spray drifting overhead. In the model, larvae of instars one to six are therefore assumed to be above ground only between dusk and dawn when temperatures permit, while seventh-instar larvae remain protected below the surface. Since overspray is most likely during daylight hours when machinery operates, this behavioural rhythm substantially shapes which life stages encounter chemicals and when.

Adult behaviour is reconstructed from a patchwork of studies. Flight-mill experiments show N. pronuba can fly more than eleven kilometres, placing it at the top of the dispersal range measured among noctuids, and mark-release-recapture studies report mean movements of 867 metres between recaptures. The model gives each adult a daily dispersal distance drawn from a distribution between 100 metres and the 11-kilometre maximum. Energy budgets were scaled from sphinx moth physiology: a feeding weight of about 0.48 grams and a 21 percent weight loss over a maximum flight translate into an assumed daily nectar intake of roughly 86 microlitres. Foraging itself follows a deliberately neutral stochastic design, with moths moving toward high nectar concentrations along directed random walks, a placeholder until real foraging strategies can be characterised.

Reproduction is governed by photoperiod. Historical work identified a critical day length of around fifteen hours for triggering ovarian activation, suggesting females aestivate through high summer before breeding. The model assigns each female her own threshold drawn from a narrow Gaussian distribution, then allots her a total egg complement sampled from a skewed beta distribution averaging 1,323 eggs, reflecting laboratory figures of roughly 1,000 to 1,500 eggs laid in three to five batches. Egg numbers are split into batches using a broken-stick model constrained to match observed cluster sizes. Oviposition sites are chosen among patches of herbaceous vegetation, with attractiveness scaled to patch area, reflecting evidence that nocturnal, scent-oriented moths favour large habitat patches in ways diurnal butterflies do not.

Mortality comes from many directions, and the model treats them explicitly. Parasitoid wasps are represented as their own agents with functional responses combining search efficiency and handling time, following an approach already validated for a white butterfly model. Above certain densities, disease transmission and larval cannibalism impose additional spatially dependent mortality, while temperature-dependent survival curves capture the sharp thermal limits documented in the laboratory: at 25 degrees only 7 percent of larvae survived to maturity, and none at 27.5 degrees. Soil moisture adds another constraint, with larvae and pupae surviving best at 4 to 8 percent moisture and dying at extremes, represented by a symmetrical super-Gaussian function. Management is included bluntly: all soil cultivation kills non-adult stages, and complete vegetation removal is likewise assumed lethal, assumptions the authors flag for calibration against field data.

Pesticides enter the simulation through three separate pathways. Dietary exposure accumulates when adults drink contaminated nectar or larvae eat contaminated leaves. Contact exposure occurs when a moth crosses a treated surface, with uptake calculated from the pesticide load per square metre, the body area in contact and an absorption parameter. Overspray applies only when a spray event or drift coincides with a moth present in the field and not sheltered in soil. All three routes feed a single body-burden value per individual, which decays daily. In this first iteration, toxicology is deliberately simple: once the burden exceeds a threshold, the moth faces a daily probability of death. The authors acknowledge that sublethal effects, which in real moths can impair navigation, reduce fecundity and even harm offspring, remain outside the current framework.

The team is unusually candid about what the model cannot do. It excludes evolutionary responses such as resistance development, pheromone communication and social information, fine-scale microhabitat refugia within the one-metre grid, and the daily rhythms of nectar production that some flowers show. Extreme weather events, which can cause catastrophic mortality beyond anything captured in smooth laboratory-derived curves, are not represented. Nor does the model track interactions with competing herbivores or the indirect effects of pesticides eliminating preferred host plants. The authors position the tool accordingly: it is designed for comparing relative pesticide risks across landscape scenarios, identifying which life stages are most vulnerable, and testing whether spatial refugia could sustain populations, not for forecasting exact moth numbers or extinction probabilities.

Even so, the implications are considerable. Because the model is modular, parameterised primarily from N. pronuba but adaptable to other noctuids with similar life histories, it could become a template for assessing an entire guild of night-flying pollinators that current regulatory schemes ignore. Developmental patterns proved strikingly consistent across noctuid species, from pest species such as Helicoverpa armigera to wild relatives, suggesting the framework can be retuned with modest new data. A striking correlation strengthens that case: across noctuid species, fecundity and flight distance are strongly and positively related, hinting at a predictable axis of life-history variation the model can exploit. For the first time, regulators weighing a spray programme will be able to ask not only what happens to bees by day, but what unfolds among the moths that take over the same fields at night, and to do so with a tool whose assumptions, strengths and blind spots are stated in full.

Subject of Research: Agent-based modelling of Noctua pronuba for pesticide risk assessment in European agricultural landscapes

Article Title: The Formal Model for Noctua pronuba ( Lepidoptera , Noctuidae ) representing a typical agricultural-landscape night-flying moth species for pesticide risk assessment

Article References: Topping, C. J., & Pettersson, L. B. (2025). The Formal Model for Noctua pronuba (Lepidoptera, Noctuidae) representing a typical agricultural-landscape night-flying moth species for pesticide risk assessment. Food and Ecological Systems Modelling Journal, 6, Article e165232. https://doi.org/10.3897/fmj.6.165232

Image Credits: AI Generated

DOI: 10.3897/fmj.6.165232

Keywords: Noctua pronuba, ALMaSS, agent-based model, pesticide risk assessment, nocturnal pollinators, thermal performance curves, agricultural landscape, Lepidoptera, ecotoxicology, larval development, dispersal, Noctuidae