Weather-induced changes in moth activity bias measurement of long-term population dynamics from light trap samples

Interpretation of light trap catches of moths is complicated by daily variation in weather that alters flight activity and numbers caught. Light trap efficiency is also modified by wind and fog, and daily weather may effect absolute abundance (numbers actually present). However, actograph experiments and other sampling methods suggest that changes in daily activity are large by comparison to changes in absolute abundance. Daily variation in weather (other than wind and fog) is therefore a form of sampling error in absolute abundance estimates. We investigated the extent of this sampling bias in 26 years of population dynamics from 133 moth species. In a subset of 20 noctuid and geometrid species, daily numbers caught were positively correlated with temperature in 14 species, and negatively correlated with rainfall in 11 species. The strength of correlations varied between species, making it difficult to standardize catches to constant conditions. We overcame this by establishing how weather variation changed with time and duration of the flight period. Species flying later in the summer and for shorter periods experienced more variable temperatures, making sampling error greater for these species. Of the 133 moth species, those with shorter flight periods had greater population variability and more showed significant temporal density dependence. However, these effects were weak, which is encouraging because it suggests that population analyses of light trap data largely reflect factors other than sampling error.