Ambient temperature is frequently used to model the development of many herbivorous insects. Estimating development rates of cryptic saproxylic species in wood will be biased by the presence of differences between internal log temperatures, where larvae feed, and measured ambient conditions outside of logs. We tested for differences between ambient temperatures at the log surface and internal log temperatures as a function of habitat type (recent clearfell area and pine forest stands), log direction (north- and south-facing), depth (5–20 cm) and season (summer and winter). Internal log temperatures varied from log surface temperatures in all cases and were influenced by depth, depth × habitat, and depth × log direction interactions. A simple constant correction factor could not explain the relationship between log surface and internal log temperatures. Mean internal log temperatures were between 0.88 and 3.68 °C warmer in clearfells compared to forest stands irrespective of log direction or depth. During summer the average daily minimum temperatures were 2.3–4.1 °C warmer in logs compared to the log surface under a conifer canopy and up to 9.8 °C warmer in adjacent recently clearfelled areas. During winter the internal log minimum temperatures were 2.0–2.8 °C warmer under a conifer canopy and 2.6–5.2 °C warmer in recently clearfelled areas compared to log surface temperatures. Differences between log surface and internal minimum temperatures increased consistently with depth into the log. Accurate models that translate observed ambient temperatures to internal log temperatures that are representative of the location of feeding larvae are necessary to model saproxylic insect development at the landscape scale. Results from this study demonstrate that these translation models will have to consider depth, log orientation, and season.