Temperatures Experienced by Emerald Ash Borer and Other Wood-boring Beetles in the Under-bark Microclimate

Abstract

Most studies of under-bark microclimate have been restricted to observations of a few coniferous trees in wooded southern latitudes. This limitation is worrying because of emerging wood-boring pests that specialize on deciduous trees in Canada, such as emerald ash borer (Agrilus planipennis). Using a large data set that includes 60 ash trees spread across both urban and woodlot sites in 6 different Ontario locations, I found that the under-bark microclimate of deciduous trees can provide wood-boring beetles with an environment in which temperatures which differ from air temperature. On average, daily minimum under-bark temperatures are significantly warmer than air temperatures in the winter months. At temperatures low enough to cause substantial cold-temperature mortality to emerald ash borer, the difference between under-bark and air temperature can be large. In addition, I observed that the difference between daily minimum under-bark and air temperature can vary, and consequently that assumptions of a constant level of difference between the two are not valid. In the spring season, I found that daily under-bark temperature maxima on the south side of the tree are significantly warmer than air temperature maxima. This difference lead to faster predicted development times for beetles in the southern under-bark microclimate of urban trees as compared to predictions based on air temperature, suggesting that city trees may impact overall population dynamics.

While it is clear that under-bark temperatures differ from air temperatures, and are important to predicting possible range and population growth of wood-boring insects, large scale measurements of microclimate conditions are not feasible. I tested the ability of a simple Newtonian cooling model to predict under-bark temperature extremes using weather station data. While the model did not predict daily under-bark temperature maxima accurately, predictions of minima were quite accurate (1.31˚C average root mean squared error), especially when compared to the errors from assuming under-bark temperature is the same as air temperature (3.20˚C average root mean squared error). I recommend use of the Newtonian cooling model to predict under-bark temperature minima of deciduous and coniferous trees.