Cold temperatures can be detrimental to insect survival, and experienced winter temperature minima have been shown to have an impact on the northern range of bark beetles. For species like emerald ash borer, which is devastating ash populations in North America, the actual range will be determined by winter temperatures in their under-bark microclimate. Since monitoring under-bark temperature over a large area, such as Ontario, is not feasible, we need to be able to predict microclimate conditions from routinely collected weather data. We used a Newtonian convective cooling model, and compared the model predictions of under-bark temperature to actual under-bark temperatures of ash trees in 6 Ontario locations ranging from Sault Ste Marie in the north to London in the south. Each of the 6 locations contained 5 trees in an urban environment, and 5 trees in a woodlot environment. The model was applied from November 2008 to March 2009.

Results/Conclusions

We were able to predict the daily minimum under-bark temperature of ash trees with an average root mean squared error of 1.4 degrees Celsius. Not all estimates of daily minimum under-bark temperatures were conservative (i.e., some predicted under-bark temperatures were colder than those observed). We examined the model sensitivity to changes in the cooling constant (i.e., the rate of heat transfer between the tree and surrounding air), but found that a change in cooling constant of 8% only produced a change in average root mean squared error of 2%.  Therefore, since the model is relatively insensitive to realistic changes in the cooling constant, further improvements in prediction can only be achieved by including another explanatory variable (e.g. solar radiation).