Environmental Correlates
I began by investigating changes in the percentage of stems and galls knocked over among years. Note that while the extreme weather event occurred in 2018, the affected gall collection occurred in 2019 and is referred to as such (i.e. the 2018 extreme event affected the 2018/19 generation). I began by using a linear mixed model (LMM) to test for differences in the ln-transformed percentage of stems knocked down across years, including year as a factor and population as a random effect. This model is designed to test for changes in the mean of a key environmental condition. However, extreme events can also change patterns of environmental variation among populations and communities. As such, I next used a Levene’s test to test for difference in variance among pairwise combinations of years. As above, I controlled for consistent population-level differences using a random effect. Together these two tests quantify the mean change and among-population variance change in stem knock down.