The objectives of this Limited Scope project were to evaluate the performance of instream sensors measuring solute concentrations in high-latitude streams and to assess the relevance of high-frequency records of stream chemistry for studying ecosystem responses to disturbance. High-latitude ecosystems are subject to disturbance by fire and permafrost thaw. Data describing material outputs from catchments could indicate the duration over which ecosystems are responding to disturbance, thereby indicating when military-owned lands may be most sensitive to training activities. This was a “high-risk” project because instream sensors have not been applied broadly in high-latitude streams. High-latitude streams are characterized by sharp changes in water temperature and by colored dissolved organic matter, both of which can interfere with the signals measured by optical sensors. Research was therefore designed to deploy and test the performance of sensors measuring dissolved organic matter and nitrate in boreal streams.

Technical Approach

Commercially available sensors measuring nitrate, fluorescent dissolved organic matter, temperature, turbidity, conductivity, and optical properties of organic matter were deployed in two streams draining the United States Army’s Yukon Training Area. Data were collected by sensors every 15 minutes, and autosamplers collected less frequent grab samples for calibration of sensor output. One stream drained lowland boreal forest that had not burned in more than ~75 years, and a second stream drained a catchment partially burned two years prior to the research.

Laboratory tests were performed to evaluate the response of sensor output to temperature, turbidity, and colored dissolved organic matter. High-frequency data were analyzed using wavelets, a spectral method for describing the temporal scales of variation. Indicators of impending regime shifts were also calculated to compare the stability of biogeochemical cycles in the recently burned and unburned catchments.


High-frequency data collected by sensors revealed temporal patterns that were not captured by lower-frequency monitoring programs. Importantly, large fluxes of nitrate and dissolved organic matter occurred during storms, and these would be missed by sampling programs conducted using grab samples. Second, diurnal fluctuations were apparent in both nitrate and dissolved organic matter, which may ultimately be useful in understanding the relative contributions of biology and hydrology to the biogeochemical cycles of catchments. Temporal patterns in all solutes measured by sensors were remarkably similar between the burned and unburned catchments, which might indicate resilience of lowland catchments to fire.

This Limited Scope project also indicates several technical considerations and improvements for implementing instream sensors as part of an environmental monitoring program. First, due to the brief duration of this project, sensors were installed on the streambed, and this introduced several limitations on data quality that would be eliminated by permanent mounting of sensors to existing infrastructure. Second, in boreal streams, which carry high concentrations of colored dissolved organic matter, fluorescent dissolved organic matter cannot provide an index of dissolved organic carbon concentration across the full range of concentrations observed, due to the changing chemical composition of organic matter at the scales of both individual storms and across seasons. Future research that probed the nature of composition vs. concentration relationships could yield reliable calibration relationships between fluorescent dissolved organic matter (fDOM) and dissolved organic carbon.


Use of instream sensors for monitoring water chemistry is feasible in high-latitude, boreal streams. Sensors could be installed across a network of sites to capture unpredictable changes, such as those resulting from fire and climate regimes. Alternatively, installation at a particular location downstream of training activities or infrastructure could provide near real-time monitoring of the effects of activities occurring within the catchment on water quality. A nitrate sensor would be particularly useful for capturing effects of training activities involving nitrate-based explosives on downstream water quality.