Evaluation of stream temperature spatial variation using distributed temperature sensing

Abstract

Water temperature in rivers and streams is an important factor for aquatic ecosystem health. Measurement of stream temperature has traditionally been accomplished by point temperature measurements, continuous point temperature loggers, and more recently, airborne remote sensing techniques such as Forward-Looking Infrared Radar (FLIR) or Thermal Infrared Radiometry. While each of these measurement techniques has certain advantages, none allows for the combined spatial and temporal information provided by Distributed Temperature Sensing (DTS). DTS employs fiber optic signals to measure temperature and is a relatively new temperature measurement technology for hydrologic sensing applications.

Nine DTS stream temperature datasets were collected in the Middle Fork John Day River (MFJDR) as part of a basin-wide stream monitoring effort. The datasets encompassed five 1-3 kilometer long reaches, some monitored over three summers (2009-2011). In contrast to existing stream temperature measurement technologies, DTS can provide stream temperature data in both the spatial and temporal domains. Techniques and challenges of interpreting DTS stream temperature data were documented, and three applications of the technology to stream temperature monitoring were explored.

Cold water patches, potentially used by fish as thermal refugia during stream temperature maximums, were located using DTS. No identified cold patch exceeded 2.31°C cooler than ambient stream temperature. Tributary inflows provided some of the most temperature-differentiated cold patches. These findings provide a reference for the degree of thermal heterogeneity in the MFJDR system and beg the question of whether fish respond to small (<3°C) spatial temperature variations. Theoretical predictions of stream mixing potential (Richardson number and cavity flow mixing predictions) suggested that increasing stream thermal heterogeneity would require channel modification to decrease stream flow velocity in select areas.

The combined spatial and temporal coverage of a DTS stream temperature dataset on the Oxbow Conservation Area allowed diagnosis of a 2°C longitudinal stream temperature decrease observed in multiple Thermal Infrared Radiometry (TIR) and Forward-Looking Infrared Radiometry (FLIR) datasets collected on that reach. Advection velocity and channel depth, rather than groundwater or tributary inflows, were the main cause of the decrease, and the magnitude of the decrease peaked in the early afternoon, disappearing completely by evening. This finding suggests caution for interpretation of FLIR and TIR stream temperature datasets, which represent "snapshot" temperature measurements. For these datasets, knowledge of flow conditions (velocity and depth) may help avoid misinterpretation of temporally-transient temperature anomalies.

Diurnal slope periodicity was observed in linear-like spatial trends in four DTS datasets, and an analysis was made to examine this subtle spatially and temporally varying phenomenon. The phase of the diurnal slope variation differed between river reaches, suggesting that propagation of larger-scale thermal waves might be one driving mechanism. Temporally-constant offsets between slope magnitudes within reaches suggested some intra-reach differences in heat fluxes.