Population dynamics are controlled by life history and moderated by the environment. These factors determine the timing and magnitude of population abundance and are used to predict crop pest populations. Forecasts of pest populations typically focus on heat accumulation during the growing season when the pests are active or emerging from the soil if they have a diapause period. Weather conditions before and during diapause can impact population dynamics as well but tend to be understudied. This study used 16 years of Colorado Potato Beetle ( Leptinotarsa decimlineata ) abundance data combined with landscape and daily weather data to understand the importance of seasonal weather and landscape on predicting population abundance. Boosted-tree multiclass models predicted abundance at each life stage and across all models. Cumulative degree days, a measure of developmental time, were overwhelmingly important. To understand how the top non-temporal variables could interact with time, we used general additive mixed models to examine interactions between time and the top six ranked variables for each life stage, then generated predicted abundances across the growing season for the 10 th , 50 th , and 90 th percentiles of each variable, keeping all other variables constant. This revealed that while Colorado Potato Beetle abundance was most strongly affected by heat accumulation, other weather factors like precipitation and air saturation, as well as soil temperature during diapause, can also influence abundance trends. The only landscape variable consistently ranked in the top six was potato acreage.
Cohen et al. (Mon,) studied this question.