An Evaluation of Underwater Epoxies to Permanently Install Temperature Sensors in Mountain Streams

—Stream temperature regimes are fundamentally important to understanding patterns and process in aquatic ecosystems and inexpensive digital sensors enable accurate and repeated measurements of temperature at a site. Most temperature measurements in mountain streams are made only during summer months because of logistical constraints associated with stream access and concerns that large annual floods will destroy sensor installations. We assessed six underwater epoxies to determine whether sensors could be attached to large rocks already in streams to provide durable installations and whether temperature measurements would be biased by heat conduction through the rocks. Only one of the six test epoxies bonded sensors firmly to rock surfaces in laboratory trials. In subsequent field trials, 9 of 11 sensors attached to rocks with this epoxy successfully weathered above-average floods in four Idaho and Nevada streams in 2010. Comparisons of daily maximas, minimas, and means between rock-mounted sensors and control sensors at 10 rocks suggested temperature measurements were not biased by attachment to rocks. We also assessed the effect of direct sunlight on sensors by removing solar shields from some sensors and noted rapid and statistically significant increases in daily means (+0.21 °C) and maximas (+0.54 °C), but not minimas (-0.01 °C). Use of underwater epoxy for permanent installation of temperature sensors in mountain streams is a viable technique if an appropriate epoxy is chosen, sensors are shielded from direct sunlight, and rocks large enough to withstand floods are used. Moreover, installations using epoxy are rapid (approximately 20 minutes) and firm attachments to rock surfaces over a range of stream temperatures (5 to 20 °C) are possible. Thermal regimes are fundamental to understanding aquatic ecology given the ectothermic physiologies of most aquatic organisms. Increasing evidence suggests that 1 Corresponding author: email: [email protected] significant departures from historical thermal conditions are underway in response to a warming climate (Hari et al. 2006; Isaak et al. 2010; Kaushal et al. 2010; Rieman and Isaak, In press). Although considerable amounts of stream temperature data are now routinely collected using inexpensive digital temperature sensors, most data in mountain streams are collected during the summer after annual snowmelt floods have abated and logistical access to mountain streams is easiest. This provides a narrow view of stream thermal regimes and misses ecologically relevant information about the date of spring onset, the annual accumulation of thermal units, and length of the growing season to name a few (Neuheimer and Taggart 2007, Olden and Naiman 2009). Moreover, it is inefficient to visit field sites twice each year (first to install a sensor, second to retrieve the sensor), especially when many modern temperature sensors have batteries and memory capacities that often last several years. Collection of annual stream temperature data has been limited in mountain streams because it requires labor intensive instrumentation of permanent sites capable of withstanding large annual floods associated with snowmelt runoff. Typically, steel cables or other obtrusive structures are used to hold temperature sensors in streams but oftentimes instrument loss rates remain high. An ideal installation method would require minimal effort and materials, be largely foolproof, and provide durable installations that withstand floods and associated bed-load movement. Moreover, it would be desirable if installations were relatively innocuous so that sensors were not stolen or vandalized during lengthy deployments. Here, we report on an assessment of underwater epoxies to directly attach temperature sensors to large rocks in mountain streams. Our study had two primary objectives: 1) determine whether sensors attached to rocks with epoxy could withstand Underwater Epoxy Evaluation 2 annual floods and 2) determine whether attachment to rocks biased temperature measurements from heat conduction through the rock. As a secondary objective, we also tested whether the lack of a solar shield affected temperature measurements.