Modelling thermal reaction norms for development and viability in Drosophila suzukii under constant, fluctuating and field conditions

Bréa Raynaud-Berton,Patricia Gibert,Christelle Suppo,Sylvain Pincebourde,Hervé Colinet

DOI: https://doi.org/10.1016/j.jtherbio.2024.103891

Abstract:Phenological models for insect pests often rely on knowledge of thermal reaction norms. These may differ in shape depending on developmental thermal conditions (e.g. constant vs. fluctuating) and other factors such as life-stages. Here, we conducted an extensive comparative study of the thermal reaction norms for development and viability in the invasive fly, Drosophila suzukii, under constant and fluctuating thermal regimes. Flies, were submitted to 15 different constant temperatures (CT) ranging from 8 to 35 °C. We compared responses under CT with patterns observed under 15 different fluctuating temperature (FT) regimes. We tested several equations for thermal performance curves and compared various models to obtain thermal limits and degree-day estimations. To validate the model's predictions, the phenology was monitored in two artificial field-like conditions and two natural conditions in outdoor cages during spring and winter. Thermal reaction norm for viability from egg to pupa was broader than that from egg to adult. FT conditions yielded a broader thermal breadth for viability than CT, with a performance extended towards the colder side, consistent with our field observations in winter. Models resulting from both CT and FT conditions made accurate predictions of degree-day as long as the temperature remained within the linear part of the developmental rate curve. Under cold artificial and natural winter conditions, a model based on FT data made more accurate predictions. Model based on CT failed to predict adult's emergence in winter. We also document the first record of development and adult emergence throughout winter in D. suzukii. Population dynamics models in D. suzukii are all based on summer phenotype and CT. Accounting for variations between seasonal phenotypes, stages, and thermal conditions (CT vs. FT) could improve the predictive power of the models.