Developmental Differentiations of Major Maize Stemborers Due to Global Warming in Temperate and Tropical Climates.

While many insects are in decline due to global warming, the effect of rising temperatures on crop insect pests is uncertain. A capacity to understand future changes in crop pest populations remains critical to ensure food security. Using temperature-dependent mathematical models of the development of four maize stemborers in temperate and tropical regions, we evaluated the potential impacts of different climate change scenarios on development time. While recognizing the limitations of the temperature-dependent development rate approach, we found that global warming could either be beneficial or detrimental to pest development, depending on the optimal temperature for the development of the species and scenarios of climate change. Expected responses range from null development to 1.5 times faster development than expected today. These results suggest that in the medium term, the studied species could benefit from global warming with an accelerated development, while in the long term, their development could either be delayed or accelerated, which may impact their dynamics with implications on maize cultivation.

Modeling Temperature-Dependent Development Rate in Insects and Implications of Experimental Design.

Characterizing the temperature-dependent development rate requires empirical data acquired by rearing individuals at different temperatures. Many mathematical models can be fitted to empirical data, making model comparison a mandatory step, yet model selection practices widely vary. We present guidelines for model selection using statistical criteria and the assessment of biological relevance of fits, exemplified throughout a Lepidoptera pest dataset. We also used in silico experiments to explore how experimental design and species attributes impact estimation accuracy of biological traits. Our results suggested that the uncertainty in model predictions was mostly determined by the rearing effort and the variance in development times of individuals. We found that a higher number of tested temperatures instead of a higher sample size per temperature may lead to more accurate estimations of model parameters. Our simulations suggested that an inappropriate model choice can lead to biased estimated values of biological traits (defined as attributes of temperature dependent development rate, i.e., optimal temperature for development and critical thresholds), highlighting the need for standardized model selection methods. Therefore, our results have direct implications for future studies on the temperature-dependent development rate of insects.