The Vendor and SPA will maintain eleven (11) real-time, long-term, and high-frequency water quality monitoring stations above and below SPA-managed reservoirs (Table 1) to achieve Objective 1. At each station, five water quality parameters (i.e., temperature, conductivity, dissolved oxygen, pH and turbidity) will be collected using multi-parameter sensors (YSI Sonde). Sensor maintenance will occur at 4 6-week intervals following current USGS operating procedures (Wagner et al. 2006). All time-series records collected will be analyzed and approved following USGS procedures (Wagner et al. 2006) using Aquarius Time-Series (Aquatic Informatics, Vancouver, Canada) which the Vender will purchase and maintain, providing SPA with remote access. The Vendor and SPA will maintain real-time data acquisition and photovoltaic power systems (where installed) at each site. The Vendor will maintain the communication and retrieval between all deployed sensors, dataloggers, computers, and servers where data will be stored. Continuous water quality datasets will be made publicly available through CUAHSI HydroShare, a collaborative platform for sharing hydrologic data, models, and code (Objective 1). It supports findability, accessibility, interoperability, and reusability (FAIR) data principles. Each dataset will include comprehensive metadata and documentation, detailing the source, collection methods, and any processing steps.
Findings from this project should demonstrate the importance of continued high-frequency sampling to further elucidate how reservoirs are impacted by riverine water and sediment deposition and how these impacts are affected by climatic and hydrologic variability and the impacts of watershed-scale disturbances (e.g., drought, flood, wildfire), point-source and non-point-source pollution, and alteration of the natural flow regime (Objective 2). The Vendor will propose specific research topics that incorporate sonde data collected under this project, in addition to other readily available data sets, to address these objectives and that are also of interest to the greater scientific community.
While collecting accurate and reliable suspended sediment data is critical to water resources planning, reservoir management, and research, the cost of installing, operating and maintaining the suspended sediment gages has caused a decline in the existing network. The use of in-stream turbidity and streamflow data to compute time-series suspended sediment concentrations and loads reduces the cost of continuous suspended sediment monitoring compared to current methods (Rasmussen et al. 2009). The Vendor and SPA will use turbidity and suspended sediment data to quantify sediment loading to, and transport from, SPA-managed reservoirs (Objective 3). To support this investigation, discrete samples will be collected quarterly and during high flow events to capture the hydrological variability and extremes and analyzed by the Vendor for total suspended sediments during each site visit following standard methods (Edwards et al. 1999, Davis 2005, Groten and Johnson 2018).
Collecting turbidity data has become less expensive and more reliable, however, technological limitations exist that impact the ability to compute suspended sediment concentration and loads and turbidity measurements are more expensive and less reliable to operate than other water quality parameters. Turbidity is measured by emitting a light and measuring the amount scattered in the water using a photodetector. The deployment of an onboard brush reduces fouling of the photodetector, but is prone to splaying, malfunction, or sediment laden high flow events. In addition, as technology changes, new turbidity sensors can report different values from old sensors in the same waterbody, even from sensors produced by the same manufacturer that report the same units. This difference in reporting is due to the number, wavelength, and angle of light sources and photodetectors can cause different readings in the same media and operating range (Rasmussen et al. 2009, Snazelle 2020, Foster et al. 2021). Additionally, conditions can exceed the range of the sensor, which is common following monsoon precipitation events within the study area (Reale et al. 2015, Van Horn and McGibbon 2024). Finally, the properties of the transported particles can result in errors including, for example, post-fire black carbon events, which have been documented upstream and downstream of multiple SPA reservoirs (Reale et al. 2015, Nichols et al. 2024, Van Horn and McGibbon 2024). During these events entrained ash and charcoal absorb transmitted light from the turbidity sensor, simulating conditions of no light scattering or 0 FNU rather than the actual value (Dahm et al. 2015). In contrast to the turbidity sensor, the conductivity, dissolved oxygen, and pH sensors use an electrode cell, optical luminesce, and ion selective electrode, respectively. These sensors have captured the range in observed values upstream and downstream of the reservoirs (Van Horn and McGibbon 2024), including post-fire pulses (Reale et al. 2015, Nichols et al. 2024). Van Horn and McGibbon (2024) developed a machine learning model to estimate sediment delivery to Conchas Reservoir using high-frequency streamflow, turbidity, and conductivity to simulate turbidity values prior to and following the Hermits Peak Calf Canyon wildfire. Thus, collecting multiple continuous water quality variables provides vital information to model sediment transport in streams and rivers and loading to reservoirs. The Vendor will utilize machine learning techniques to continue to explore the patterns in high-frequency streamflow, turbidity, and other water quality parameters to evaluate surrogate signals for turbidity at additional SPA reservoirs (Objective 4).
The major research findings will be documented in technical reports drafted by the Vendor and SPA for publication in peer reviewed journals (Objective 5). The Vendor will be responsible for all associated publishing fees. Annual reports using internally and externally collected water quality and quantity data throughout the year above, within, and below each SPA-managed reservoir, will be authored by the Vendor and SPA.
There are five main objectives:
1. Collect, review, and disseminate real-time and high-frequency water quality data upstream and downstream of SPA reservoirs.
2. Assess episodic, seasonal, and interannual trends in water quality and the influence on SPA reservoirs.
3. Leverage turbidity and streamflow records to calculate high-frequency suspended sediment concentrations and loads into and out of SPA reservoirs.
4. Utilize machine learning techniques to explore the patterns in high-frequency water quality data to understand the drivers of sediment loading for each reservoir.
5. Prepare scientific professional and technical reports for publication in peer reviewed journals.
Supplies and Materials:
SPA will provide equipment and materials for water quality data collection in Appendix A. The Vender is responsible for acquiring all remaining supplies and material to successfully complete the project.
Site Locations:
o Arkansas River above and below John Martin Reservoir, CO
o Purgatoire River above and below Trinidad Reservoir, CO
o Canadian River above Conchas Reservoir, NM
o Pecos River above and below Santa Rosa Reservoir, NM
o Rio Chama above and below Abiquiu Reservoir, NM
o Rio Grande above Cochiti Lake, NM
o Rio Grande downstream of confluence with Rio Jemez, NM
Applicants should have expert knowledge and work experience in selecting, installing, operating, maintaining, and overseeing multi-year and continuously deployed (e.g., year-round) high-frequency water quality sensor (i.e., temperature, conductivity, DO, pH, and turbidity) networks in large rivers, preferably in systems with high suspended sediment loads within the southwestern U.S. Previous experience evaluating discrete water quality and suspended sediment data with high-frequency water quality and streamflow data is preferred. The candidates should have experience reviewing and approving data from water quality sensor networks using Aquarius Time-Series (or equivalent software). The candidates should have experience publishing in peer reviewed journals assessing water quality using data from long-term and continuously deployed high-frequency water quality sensor networks, preferably in large rivers within the southwestern U.S.
