Influences on shallow ground temperatures in high flux thermal systems

• Soil temperatures in Yellowstone NP were measured before and after precipitation events.
• The range, mean, and skewness of measured temperatures decreased, but the correlation structure remained unchanged.
• We conclude geostatistical inferences based on shallow temperature patterns are robust to moderate perturbation.

Ground temperature measurements are a useful indication of subsurface processes and heat flux, particularly in volcanic and hydrothermal systems, but obtaining reliable data at sufficient resolution can be difficult. Investigators commonly use temperature measurements at 1 m depths to minimize land surface boundary impacts; however, these measurements are time-consuming and invasive, limiting the number of points that can be surveyed. Alternatively, shallow ground temperature measurements (≤ 25 cm depth) offer a rapid and minimally-invasive way to collect a large number of observations in a target area. Although this method has obvious appeal, changing atmospheric conditions can impact the observed temperatures, and thus may reasonably be expected to influence interpretations arising from the data. Here we examine the impact of precipitation and changing air temperature on shallow ground temperatures in the vicinity of a group of hot springs located in Yellowstone National Park, Wyoming. We find that the mean, the range, and the skewness of the observed temperatures were decreased by changing atmospheric conditions; however, the model variogram representing data taken after several days of moderate precipitation adequately described the spatial correlation of data taken before precipitation. We therefore conclude that the ability to differentiate between high- and low-flux areas may be somewhat reduced by moderate precipitation and changing atmospheric conditions, but that interpretations made on the basis of characteristics of the inferred variograms are likely to be robust to such perturbations in high heat flux environments.