Water Surface Elevation Data Derived from Submersible Pressure Transducers Deployed in the Susitna, Yukon, and Nushagak Rivers, Alaska, Summers 2020-2021

Four unvented submersible pressure transducers (PTs; Onset HOBO water level data loggers; Onset, Bourne, Massachusetts) were deployed into the channel of the Yukon River in a reach above and below the Yukon River at Stevens Village, Alaska (15453500) streamgage (July - October 2020). A 5th PT was left in the gage house to correct for barometric pressure fluctuations. At the time of retrieval in October, each of the PTs was above the current river level and each PT position as well as water surface adjacent to it were surveyed with Global Navigation Satellite System (GNSS) equipment. A base station was set up for over 30 minutes at each PT location while the rover was used to survey the PT and water surface. A single base could not be set up to survey each location because the distance between locations exceeded radio range for the base and rovers. An OPUS-RS https://geodesy.noaa.gov/OPUS/ position solution was obtained for the base location at each PT, then this elevation in NAVD88 was used to correct the rover surveyed elevations for the PT and WS elevation. The PTs were programmed to record a depth measurement every 15 minutes during the deployment period. The depth record produced by the transducer was corrected to NAVD88 water surface elevations using the surveyed PT elevation in October as the elevation for zero depth of the transducer record. The record collected by the transducers were plotted alongside each other and the corrected gage height record from the streamgage to determine stability of each deployment. All transducers showed similar trends to the streamgage record and therefore were assumed to not have been disturbed during deployment. Based on this assumption, the records were able to be converted to NAVD88 for the entire period without survey information from the deployment in July.

Five unvented submersible pressure transducers (PTs; Onset HOBO water level data loggers; Onset, Bourne, Massachusetts) were deployed on the Susitna River in a reach up and downstream of the Susitna River at Sunshine, Alaska (15292780) streamgage (July - October 2020). Only 3 PTs were recovered. The PTs were programmed to record a depth measurement every 15 minutes during the deployment period. Instead of an independent PT for recording barometric pressure fluctuations, the records from the in-stream PTs were corrected using the barometric pressure record from the Talkeetna airport (PATK) during the period of deployment. Water surface elevations were surveyed using GNSS during the October recovery of the PTs and records were corrected to NAVD88 datum by determining the difference between depth and true elevation, then using that value as an offset to correct the entire period of depth record from the transducer. Again, comparison of PT records against one another and the corrected gage height record from the streamgage indicated that the PTs were stable during the period of deployment. On the Susitna, a single base set up was used to survey the water surface elevations at the location of the 3 recovered PTs, thus only a single OPUS solution from https://geodesy.noaa.gov/OPUS/ was obtained to correct the data into NAVD88 elevations.

Four unvented submersible pressure transducers (PTs; Onset HOBO water level data loggers; Onset, Bourne, Massachusetts) were deployed on the Nushagak River in a reach between the communities of Koliganek and New Stuyahok (May - September 2021). One PT was placed on shore near the location of submerged PT2 to collect a barometric pressure record to correct for atmospheric variability. At the time of deployment and again when the PTs were recovered, the water surface was surveyed with Global Navigation Satellite System (GNSS) equipment. A base station was set up for a minimum of 30 minutes at each PT location while the rover was used to survey the water surface. During the September recovery the water surface and an arbitrary control point were surveyed twice at each location to verify the quality of shots. A single base set up could not be used to survey all locations because the distance between locations exceeded radio range between the base and rover. An OPUS-RS position solution was obtained for the base location at each PT via the Trimble Access GNSS Processor (https://gnss-processor.trimbleaccess.com/gnssprocessor), which extends the CORS network by including National Park Service operated ground stations. Due to the remote location along the Nushagak River, a warning was generated noting that the survey locations were outside of polygons created by the reference CORS stations, thus solutions required extrapolation and should be used with caution. The maximum distance outside of the polygon was 17.5 km for the PT3 site. The OPUS-RS solutions were used as computed and the base elevation in NAVD88 was used to correct the raw water surface elevations surveyed on both days.

Nushagak PTs were programmed to record a depth measurement every 30 minutes during the deployment period. The depth record produced by each PT was converted to NAVD88 water surface (WS) elevations using the NAVD88 surveyed water surface from May as the zero depth of the water level record. The September surveyed NAVD8 WS elevation were used as a check for PT movement during deployment. The records collected by the PTs were plotted alongside each other and notes were made on the condition of the PT and anchor set-up when the PTs were retrieved. All evidence indicates stability at each PT deployment with records showing almost identical water level trends and the surveyed starting and ending water surface elevations within the reported precision of the OPUS-RS solutions. Upon download it was noticed that PT3's record had the same time and date for each data point. When a time increment was forced onto the data record, the plot looked like that from PT1 and PT2, thus the data was time corrected and considered valid. Only half the number of data points were collected by PT3 indicating it recorded data on an hourly interval rather than 30 minutes.