Ian W. Bishop Bridget R. Deemer Theodore A. Kennedy Robert A. Payn Robert O. Hall, Jr. Charles B. Yackulic 20260302 comma-separated values Flagstaff, AZ U.S. Geological Survey https://doi.org/10.5066/P13UHLAB Ian W. Bishop Bridget R. Deemer Theodore A. Kennedy Robert A. Payn Robert O. Hall, Jr. Charles B. Yackulic 20260401 journal manuscript Limnology and Oceanography: Methods Wiley These data were compiled to model ecosystem metabolism in the Glen Canyon Dam tailwater (e.g., the Colorado River below Glen Canyon Dam). The primary objective of our study was to develop a new ecosystem metabolism model which we refer to as the variable flow two-station (VFTS) model. The VFTS modifies the classic two-station metabolism approach to account for unsteady flow, thus allowing for more accurate metabolism estimation in tailwater environments. In pursuit of this broader objective, we leveraged ongoing continuous water quality measurements being collected at two sites in the lower half of the Glen Canyon Dam tailwater (Colorado River; -8 mile to Lees Ferry) to assess our new model's performance under various input data constraints (e.g., travel time accuracy). We also assessed our new model's performance relative to both a highly complex two-station model and the community standard one-station modeling approach, the latter of which we believe is inappropriate for use in tailwater environments. These data include sub-daily environmental inputs required for our new model, which were collected in the lower half of the Glen Canyon Dam tailwater between March 2008 and April 2014 by the U.S. Geological Survey Grand Canyon Monitoring and Research Center (GCMRC). These data can be used to examine trends and drivers in Glen Canyon tailwater ecosystem metabolism and to explore the effects of input (travel time) accuracy and model choice on metabolism estimation in lotic habitats subject to subdaily flow variation. The purpose of these data is to develop a two-station metabolism model that is transferable to other river ecosystems in proximity to large river discontinuities, especially river segments downstream of dams which are often subject to hydropeaking or other subdaily flow fluctuations. Additionally, these data were generated to estimate ecosystem metabolism (GPP and ER) in the Glen Canyon Dam tailwater. These data could be used by future researchers to incorporate ecosystem metabolism into broader ecosystem analyses in the Colorado River downstream of Glen Canyon Dam, or to further assess the effects of experimental flows on downstream ecology and recreation. These data could also be used in broader assessments of ecosystem metabolism across freshwater ecosystems. 20080301 20140228 ground condition Complete None planned -111.59840 -111.49887 36.89060 36.83680 Glen Canyon Dam tailwater ISO 19115 Topic Category biota inlandWaters None river metabolism GPP ER k600 hydropeaking two-station USGS Thesaurus ecological models ecology USGS Metadata Identifier USGS:6887d457d4be024722b4aae2 None Colorado River Glen Canyon National Recreation Area tailwater Lees Ferry Glen Canyon No access constraints. Please see 'Distribution Information' for details. These data are marked with a Creative Common CC0 1.0 Universal License. These data are in the public domain and do not have any use constraints. Users are advised to read the dataset's metadata thoroughly to understand appropriate use and data limitations. Ian W Bishop U.S. Geological Survey Research Ecologist mailing and physical

2255 North Gemini Drive

Flagstaff AZ 86001 N/A ibishop@usgs.gov This research was funded by the U.S. Bureau of Reclamation through the Glen Canyon Dam Adaptive Management Program. A. P. Appling, R. O. Hall, C. B. Yackulic, M. Arroita 2018 publication JGR Biogeosciences Wiley Online Library https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2017JG004140 Herncin E. Garcia and Louis I. Gordon 1992 publication Limnology and Oceanography Wiley Online Library https://aslopubs.onlinelibrary.wiley.com/doi/abs/10.4319/lo.1992.37.6.1307 All raw dissolved oxygen, temperature, and specific conductivity values retrieved from YSI sonde deployments are presented as provided by the instrument. We did not round these values any further. Data users should note that the reported accuracies are +/- 0.1 mg/L for dissolved oxygen, 0.01 degrees Celsius for the temperature, and +/- 0.5% of the reading for specific conductivity. More information regarding quality assurance in these values can be found in the Process Step documentation, including sonde calibration methods used. We manually examined numeric variables in each dataset using the R functions summary() hist() to confirm that values for each fall into expected ranges. Data set is considered complete for the information presented, as described in the abstract. No formal positional accuracy tests were conducted. No formal positional accuracy tests were conducted. National Renewable Energy Laboratory (NREL) 2018 tabular digital data https://nsrdb.nrel.gov/ Digital 2008 2014 Sept 2025 NSRDB barometric pressure Subdaily barometric pressure data for DO saturation calculation U.S. Geological Survey 20240601 tabular digital data Flagstaff, AZ U.S. Geological Survey https://gcmrc.cr.usgs.gov:8443/discharge_qw_sediment/station/GCDAMP/09380000 Digital 2008 2014 Sept 2025 USGS GCMRC water quality portal Reach discharge data Michael D. Yard, Glenn E. Bennett, Steve N. Mietz, Lewis G. Coggins Jr., Lawrence E. Stevens, Susan Hueftle, Dean W. Blinn Unknown publication https://www.sciencedirect.com/science/article/abs/pii/S0304380004004375 Digital 2008 2014 Sept 2025 Glen Canyon solar insolation estimates Solar insolation time series for Glen Canyon study reach Dissolved oxygen (DO), specific conductivity, and water temperature were collected at 15-min intervals (96 measurements per day) via Yellow Springs Instruments (YSI) 6-Series sondes equipped with luminescent-based “ROX” dissolved oxygen sensors. Sondes were deployed at two stations starting in March 2008. This table only presents data collected between March 2008 and Feb 2014, matching the timespan analyzed by Payn et al. (2017). The downstream sonde was deployed mid-river, hanging 1-2m below the surface from a buoy near the Lees Ferry dock, 24.2-25.0 river kilometers (rkm) downstream from Glen Canyon Dam. Note that this station was moved downstream twice during the study period. The upstream sonde (12.9 rkm downstream of dam) was placed 1-2 m below the water surface, attached to a large boulder jutting out into the river current, a site that appeared to be subject to strong cross-sectional mixing. 2025 For each VFTS model run, several input data streams are required. Upstream and downstream dissolved oxygen and downstream water temperature were acquired from the Glen_Canyon_raw_sonde_data dataset. Upstream and downstream dissolved oxygen concentrations at saturation were calculated based on Garcia and Gordon (1992) using sonde water temperature and altitude-corrected barometric pressure retrieved from the National Solar Radiation Database (2018) for the nearest grid cell to the Lees Ferry dock over our study period. Reach depth was calculated assuming hydraulic continuity and a simplified rectangular channel as a function of downstream discharge, reach velocity and the assigned average reach width of 129m, as assumed in Payn et al. (2017). We generated a base estimate of solar insolation in our study reach during our study period using the Yard et al. (2005) incident light model, which produced reach-average, mid-river estimates of photosynthetic photon flux density (μmol quantum m−2 s−1) at 15-min resolution. For each downstream timepoint, these light estimates were summed over its associated travel time and divided by the sum of all light estimates for that day. Note that these estimates account for topographical/canyon shading but do not account for either cloud cover or water column light attenuation. Three travel time datasets were used, one for each VFTS model run. The first derives from an agent-based particle tracker simulation conducted by Payn et al. (2017) and was used in the VFTS-PT model run. The second set of travel time estimates was generated using the empirical travel times found in the Glen_Canyon_Conductivity-based_empirical_travel_times dataset. We used these empirical travel times to calculate travel time for all downriver timepoints using the following equation: ln(travel time)=ln(discharge) + dQt/dt, where dQt/dt was calculated using the change in discharge over the 30 minutes centered on time t. This latter term was added to account for the memory of past discharge that causes hysteresis in the relationship between reach-average travel time and flow. The third travel time set was used in model run VFTS-MT, where we applied a uniform/invariant travel time value based on simplified and averaged reach geometry (depth, width, length) and velocity) over the study period. In all three cases, the travel time variable is used to generate a lag variable, which denotes the number of time steps between a water parcel's observation at upstream and downstream sondes. The lag variable was used to properly pair an upstream observation to each downstream observation. For the One-station run, data inputs are the same, except that no upstream or travel time variables are required. 2025 Daily metabolism (GPP and ER) and gas-exchange (K600 or k600) estimates were acquired using both the newly presented variable flow two-station (VFTS) model and the streamMetabolizer process model “b_Kn_oipi_tr_plrckm.stan”. For the former, daily k600 (meters per day) were estimated but converted to K600 (per day) to compare with streamMetabolizer output. The VFTS models all used a normally distributed prior distribution for k600 with a mean of 3.48 and a standard deviation of 0.552, per Payn et al. (2017). DO_r2 values for VFTS output were determined the same way as in Appling et al. (2018). For the One-station model, the following argument values were assigned: K600_daily_meanlog_meanlog=log(3.46), K600_daily_meanlog_sdlog=0.16, and K600_daily_sdlog_sigma=0.05. For all model runs, 3000 warm-up iterations were run followed by 1000 saved iterations for each of 4 chains. 2025 A set of empirical travel times were calculated for the Glen Canyon study reach by detecting unambiguous specific conductivity peaks in the upriver sonde data and locating them in the downriver sonde in subsequent hours, with travel time being defined as the time it took each detected peak to travel between the two stations. Traces for both sondes were plotted and travel times were visually determined/measured. While the ions driving electrical conductivity are not entirely conservative in these waters, we assumed the dominant control on the timing between the corresponding upriver and downriver peaks was hydrologic transit time. We measured travel times across a broad range of discharge values (117-716 cubic meters per second). 2025 Quality Assurance (QA) - The YSI water quality instruments (models 6600 and 6920) were serviced at intervals of approximately 35-45 days. During servicing of the water-quality instruments, protocols described in Wagner and others (2006) were followed: a field meter was installed concurrent readings with the deployed water quality monitor(s) were taken, the probes were cleaned (wiper pads replaced when necessary) and then redeployed to check for biological fouling. Dissolved oxygen (DO) was calibrated on site (for both the deployed water quality instrument and field meter) in a shaded, air-saturated water bath at river temperature for at least 30 minutes. All dissolved oxygen probes used in the study area were equipped with mechanical wipers to reduce biological fouling. Quality Control (QC) - Temperature drift of the sensors was periodically checked by comparing readings of several instruments recording simultaneously in buckets of water in the laboratory. Dissolved oxygen data were not adjusted based on fouling and electronic drift, but instrument readings were compared between instruments during visits to assure that temperature readings from the sondes generally remained within 0.15 degrees C of each other and DO agreed within 0.1 mg L-1. QA/QC was ongoing, but the last QA/QC procedures were completed in early 2020, well after the time span associated with this data release. For more information, see https://www.usgs.gov/products/data-and-tools/data-management/manage-quality. Days with abnormal DO traces (e.g., anomalous departures in daily DO periodicity indicative of algal biofouling) were manually identified from data visualizations and were removed from this dataset before metabolism estimation. Daily estimates with poor model fits are included in the model output summary tables (“VFTS_and_One-station_model_output” and “CMH_model_output“), but these tables include a goodness of fit metric (coefficient of determination) which allows for filtering out of low-quality estimates. Lastly, several days included at least one travel time observation that exceeded 10 hours; these days were removed to conservatively retain estimates that were fully associated with that day’s metabolic activity, rather than estimates that represent metabolic dynamics taking place over the combined day of and previous day in the time series." 2025 Glen_Canyon_raw_sonde_data This data table represents the raw sonde measurements collected at upstream and downstream study sites. Briefly, YSI sondes were deployed near-continuously between March 2008 and Feb 2014 to measure dissolved oxygen, temperature and specific conductivity at two sites (upstream and downstream) in the Glen Canyon Dam tailwater. The downstream site position moved twice over the study period, which is apparent in the "sonde_location" column. These data served as the basis for key inputs required to estimate GPP and ER in this tailwater reach of the Colorado River. Note that sonde deployments lasted aproximately 30-45 days at a time, and sensor drift for each deployment was not measured. Producer Defined datetime Collection datetime for water quality measurements listed here. date format: YYYY-MM-DD HH:MM:SS; "America/Phoenix" timezone. Producer Defined 2008-03-01 18:15:00 2014-02-28 17:00:00 site Study site locality. "upstream" refers loosely to the colloquial -8 mile; "downstream" refers loosely to Lees Ferry dock. Producer Defined downstream Downstream site Producer defined upstream Upstream site Producer defined sonde_unit A letter assigned to each individual YSI 6000 series sonde setup. This field contains the letter of the deployed sonde. Producer Defined B Sonde B Producer defined C Sonde C Producer defined A Sonde A Producer defined D Sonde D Producer defined sonde_location River kilometers downstream of Glen Canyon Dam. Note that the downstream site was moved twice during the study period. Producer Defined 12.88 25.03 river kilometers dissolved_oxygen The instantaneous dissolved oxygen concentration measured. Producer Defined NA No Data Producer defined 2.51 10.98 mg/L temperature The instantaneous water temperature measured. Producer Defined NA No Data Producer defined 6.95 16.2 degrees C specific_conductivity The instantaneous water specific conductivity measured. Producer Defined NA No Data Producer defined 462 977 µS/cm decimal_days_since_deployment_start Number of days since deployment began/sonde was calibrated. Producer Defined 0.0 77.46 days VTFS_model_inputs This data table contains data inputs for 4 specific model runs, 3 using our new VFTS model and one using the streamMetabolizer (one-station) model. These data contain the necessary inputs to run the models, specifically: dissolved oxygen, disolved oxygen saturation, temperature, river depth, incident light, and also travel time for 96 daily time steps between sondes for the two-station approach. These data were used to generate daily metabolism estimates for our study reach and time period. Producer Defined model_run Name of model run. Three VFTS model runs are included, each used a different travel time series. "PT" = particle tracker"; "CQ" = discharge-based; "MT" = mean travel, a uniform value was used for all dates and times. Producer Defined OS one-station run using R package streamMetabolizer Producer defined VFTS-1 VFTS-1 = Variable Flow Two Station model run 1 (VFTS-PT) Producer defined VFTS-2 VFTS-2 = Variable Flow Two Station model run 2 (VFTS-CQ) Producer defined VFTS-3 VFTS-3 = Variable Flow Two Station model run 3 (VFTS-MT) Producer defined datetime The datetime for conditions at the downstream site. Producer Defined 2008-03-10T00:00:00Z 2014-02-28T23:45:00Z downstream_DO The instantaneous dissolved oxygen concentration measured downstream. Producer Defined 4.38 10.82 mg/L downstream_DO_sat The instantaneous dissolved oxygen saturation concentration measured downstream. Corrected to river elevation barometric pressure. Producer Defined 7.8 9.63 mg/L light This field represents the sum of light reaching the modelled river segment between each pair of upstream and downstream timepoints, scaled by the sum of light reaching that segment over the course of the day. The initial time series of incident light estimate for the reach was acquired through use of Yard et al. (2005) incident light model. Producer Defined 0.0 2025.76 unitless downstream_temp The instantaneous water temperature measured downstream. Producer Defined 6.95 15.3 degrees C reach_depth The estimated reach average river depth. Producer Defined 2.1 10.6 meters upstream_DO The instantaneous dissolved oxygen concentration measured upstream. "NA" for one station model. Producer Defined NA No Data Producer defined 3.98 10.98 mg/L upstream_DO_sat The instantaneous dissolved oxygen saturation concentration measured upstream. Corrected to river elevation barometric pressure, "NA" for one station model. Producer Defined NA No Data Producer defined 7.81 9.58 mg/L travel_time The estimated time a parcel arriving downstream took to travel there from upstream site. "NA" for one station model. Producer Defined NA No Data Producer defined 0.15 0.42 decimal days lag The number of time steps (15 minute intervals) between paired upstream and downstream measurements. "NA" for one station model. Producer Defined NA No Data Producer defined 14 40 VFTS_model_output This data table represents the summarized output of each model run reported in the associated manuscript/study. The output includes mean daily estimates of GPP, ER and K600, 95% credible interval bounds (2.5%-97.5%) for each parameter, coefficients of determination (our chosen model fit metric) for each day for each run, and Rhat metrics to convey the extent of MCMC chain convergence for each parameter. These data were used to analyze the role of model choice and travel time estimation method in ecosystem metabolism modeling of a desert tailwater ecosystem. Producer Defined model_run Name of model run. Three VFTS model runs are included, each used a different travel time series. "PT" = particle tracker"; "CQ" = discharge-based; "MT" = mean travel, a uniform value was used for all dates and times. Producer Defined OS one-station run using R package streamMetabolizer Producer defined VFTS-1 VFTS-1 = Variable Flow Two Station model run 1 (VFTS-PT) Producer defined VFTS-2 VFTS-2 = Variable Flow Two Station model run 2 (VFTS-CQ) Producer defined VFTS-3 VFTS-3 = Variable Flow Two Station model run 3 (VFTS-MT) Producer defined date Date of associated attribute values. date format: YYYY-MM-DD. Producer Defined 2008-03-11 2014-02-27 GPP_mn Posterior mean value for GPP parameter Producer Defined NA No Data Producer defined -6.69 30.46 g O2/m2/day GPP_low Lower bound (2.5th quantile) of credible interval for GPP parameter Producer Defined NA No Data Producer defined -8.5 28.45 g O2/m2/day GPP_hi Upper bound (97.5th quantile) of credible interval for GPP parameter Producer Defined NA No Data Producer defined -4.88 32.55 g O2/m2/day ER_mn Posterior mean value for ER parameter Producer Defined NA No Data Producer defined -64.84 6.99 ER_low Lower bound (2.5th quantile) of credible interval for ER parameter Producer Defined NA No Data Producer defined -73.64 5.99 ER_hi Upper bound (97.5th quantile) of credible interval for ER parameter Producer Defined NA No Data Producer defined -56.43 7.96 g O2/m2/day K600_mn Posterior mean value for k600 parameter scaled by depth Producer Defined NA No Data Producer defined 0.06 10.2 per day K600_low Lower bound (2.5th quantile) of credible interval for k600 parameter scaled by depth Producer Defined NA No Data Producer defined 0.01 8.53 per day K600_hi Upper bound (97.5th quantile) of credible interval for k600 parameter scaled by depth Producer Defined NA No Data Producer defined 0.15 11.73 per day GPP_Rhat Measure of MCMC chain convergence for GPP for a given model fit Producer Defined NA No Data Producer defined 0.999 1.041 unitless ER_Rhat Measure of MCMC chain convergence for ER for a given model fit Producer Defined NA No Data Producer defined 0.999 1.066 unitless k600_Rhat Measure of MCMC chain convergence for k600 for a given model fit Producer Defined NA No Data Producer defined 0.999 1.001 unitless K600_Rhat Measure of MCMC chain convergence for k600 for a given model fit Producer Defined NA No Data Producer defined 1.0 1.067 unitless DO_r2 Goodness of fit metric relating predicted and observed downstream DO values for a given day Producer Defined NA No Data Producer defined -0.079 1.0 unitless sigma Residual error for each day's best fit Producer Defined NA No Data Producer defined 0.082 0.12 g O2/m2/day Glen_Canyon_conductivity-based_empirical_travel_times This data table represents a set of empirically determined travel times that spans a gradient of concurrent discharge levels, which were collected in the Glen Canyon tailwater since 2008. These travel times were used to fit a travel time/discharge relationship to help estimate relatively simple and reasonably accurate travel times for two-station metabolism estimation. Producer Defined datetime Collection datetime for other attributes listed here. date format: YYYY-MM-DD HH:MM:SS; "America/Phoenix" timezone. Producer Defined 2008-03-02 21:15:00 2022-12-31 22:30:00 travel_time Observed travel time for traveling specific conductivity peak Producer Defined 3.42 15.25 hours discharge Measured discharge at the USGS gage Colorado River at Lees Ferry, AZ (09380000) Producer Defined 117.51 716.42 cubic meters per second change_in_discharge Finite estimate of change in discharge over 30 minute time period centered on a given time step Producer Defined -11.327 15.574 cubic meters per second per 15 minutes CMH_model_output This data table represents the summarized output of the CMH model, as estimated by the lead author of Payn et al. (2017) and used for comparative analysis in the associated manuscript/study. The output includes mean daily estimates of GPP, ER and k600, 95% credible interval bounds (2.5%-97.5%) for each parameter, coefficients of determination (our chosen model fit metric) for each day for each run, and Rhat metrics to convey the extent of MCMC chain convergence for each parameter. These data were used as high quality daily metabolism estimates against which VFTS and one-station model output performance could be assessed. Producer Defined date Date of associated attribute values. date format: YYYY-MM-DD. Producer Defined 2008-03-11 2014-02-27 GPP_mn Posterior mean value for GPP parameter Producer Defined 0.69 16.68 g O2/m2/day GPP_low Lower bound (2.5th quantile) of credible interval for GPP parameter Producer Defined 0.44 16.49 g O2/m2/day GPP_hi Upper bound (97.5th quantile) of credible interval for GPP parameter Producer Defined 0.93 16.9 g O2/m2/day ER_mn Posterior mean value for ER parameter Producer Defined -35.5 -0.0 g O2/m2/day ER_low Lower bound (2.5th quantile) of credible interval for ER parameter Producer Defined -36.97 -0.13 g O2/m2/day ER_hi Upper bound (97.5th quantile) of credible interval for ER parameter Producer Defined -33.76 -0.0 g O2/m2/day k600_mn Posterior mean value for k600 parameter Producer Defined -0.02 0.46 m/day k600_low Lower bound (2.5th quantile) of credible interval for k600 parameter Producer Defined -0.03 0.43 m/day k600_hi Upper bound (97.5th quantile) of credible interval for k600 parameter Producer Defined -0.0 0.49 m/day DO_r2 Goodness of fit metric relating predicted and observed downstream DO values for a given day Producer Defined 0.1 1.0 unitless U.S. Geological Survey - ScienceBase mailing address

Denver Federal Center

Building 810

Mail Stop 302

Denver CO 80225 1-888-275-8747 sciencebase@usgs.gov Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data for other purposes, nor on all computer systems, nor shall the act of distribution constitute any such warranty. Digital Data https://doi.org/10.5066/P13UHLAB None 20260312 Ian W Bishop USGS - SOUTHWEST REGION Research Ecologist mailing and physical

Flagstaff Science Campus,Bldgs.4And5

Flagstaff AZ 86001 N/A ibishop@usgs.gov FGDC Biological Data Profile of the Content Standard for Digital Geospatial Metadata FGDC-STD-001.1-1999