This data package includes one data file (GHG_2015_2023.csv), QAQC R script (GHG_qaqc_2023_2023.R), maintenance log (GHG_maintenance_2023_2023.csv), visual inspection R script (GHG_inspection_2015_2023.Rmd), sampling sites (site_descriptions.csv), and additional scripts (GHG_functions_for_L1.R). The QAQC script includes functions to process data, remove outliers, and apply the maintenance log to clean the data files included in the package (as described below); the visual inspection script provides code for plotting the data files; and the additional scripts include functions used in QAQC.
SAMPLE COLLECTION AND EQUIPMENT
Samples for greenhouse gas (methane, carbon dioxide) analysis were collected from Falling Creek Reservoir and Beaverdam Reservoir between 2015 - 2023. Most sampling occurred between the hours of 9:00 and 15:00; however, some sampling occurred outside of these hours, including some overnight sampling. For more information about nighttime sampling, see Doubek et al. (2018). Exact sampling times were included in the DateTime column starting in 2018. After 2018, exact times, when recorded during sample collection, are indicated in the DateTime column with a Flag_DateTime value of 0. If exact times were not recorded during sample collection, time was standardized to 12:00 and Flag_DateTime was set to 1. Prior to 2018, sample timestamps for samples collected between approximately 9:00 and 15:00 were standardized to noon and Flag_DateTime was set to 1.
Samples were collected from several routine depths at the deepest holes of both reservoirs (Site = 50 in both reservoirs) including: 0.1, 1.6, 3.8, 5.0, 6.2, 8.0, and 9.0 m in Falling Creek Reservoir and 0.1, 3.0, 6.0, 9.0, and 11.0 m for Beaverdam Reservoir. These sampling depths were chosen to approximately match outtake valves in Falling Creek Reservoir and Beaverdam Reservoir. Samples were collected with a 4-L Van Dorn sampler (Wildlife Supply, Yulee, FL, USA) and transferred into two replicate 20 mL glass vials without exposure to air and crimp sealed. These routine depth samples were collected approximately fortnightly from March-April, weekly from May-October, and monthly in November-February.
Surface samples from Falling Creek Reservoir were also collected at a gauged weir on the primary inflow (Tunnel Branch) to Falling Creek Reservoir (Site = 100), at the wetland stream (Falling Creek) adjacent to the reservoir (Site = 200), and at the reservoir outflow at the dam (Site = 1.1). In Beaverdam Reservoir, additional surface samples were collected from the several sites at the outflow below the dam (Site = 1.1, 1.2, 1.3, 1.4) and at the Mid-Reservoir Outflow Pipe (Site = 40). These samples were collected approximately fortnightly from May-October and monthly in November-April.
In summer 2019, additional surface samples were collected approximately monthly along the stream and reservoir continuum for both Falling Creek Reservoir and Beaverdam Reservoir as part of the Reservoir Continuum project (Falling Creek Reservoir: Creek Sites = 99, 100, 101, 102, 200; Reservoir Sites = 20, 30, 45, 50; Reservoir Outflow = 1 and Beaverdam Reservoir: Creek Sites = 100, 200; Reservoir Sites = 20, 30, 45, 50; Mid-Reservoir Outflow Pipe = 40). Surface samples were collected via grab samples for all stream and up-stream reservoir locations. All samples were stored on ice and analyzed within 24 hours of collection following McClure et al. (2018).
All sites are described with geographic coordinates in the site_descriptions.csv data file.
SAMPLE ANALYSIS AND EQUIPMENT
Dissolved methane and carbon dioxide concentrations were measured in each water sample on a Gas Chromatograph (GC) with a Flame Ionization Detector (FID) and Thermal Conductivity Detector (TCD) following McClure et al. (2018). Briefly, immediately prior to analysis, a 2-mL headspace was created with Helium (He) by displacing 2-mL of sample water. The headspace was equilibrated by shaking each 20-mL vial at 300 rpm for 15 minutes. The 2-mL headspace was then injected into the GC. The oven temperature was set to 35 degrees C with a carrier gas (He) flow rate of 15 mL/min. CH4 had a retention time of 1.3 min on the FID and CO2 had a retention time of 2.8 min on the TCD. Dissolved concentrations of CH4 and CO2 in water were calculated using the observed head-space concentrations and Henry's Law (McClure et al. 2018). From 2015-2018, samples were analyzed on a GC SRI, Model 8010. From 2019-onwards, samples were analyzed on a Shimadzu Nexis GC-2030.
DATA FLAGS
Replicate samples (Rep 1 and 2) were often collected for each Date, Site, and Depth. If a second rep was not collected, the data was flagged as '1' (sample was not collected). If the analyzed sample had a value below the minimum detection limit (MDL) the sample was flagged as '2' (sample below MDL). A running MDL was calculated for each GC separately (2015-2018, GC SRI, Model 8010; 2019-onwards, Shimadzu Nexis GC-2030) from routinely measured air concentrations throughout the year (2015-2021) or a reference gas tank (2022-2023) and calculated using a consistent temperature (2015-2018 20 deg C; 2019-2023 40 deg C) and pressure (30 psi). MDL for each year is presented; 2015 CH4 = 0.001 umol/L, CO2 = 2.13 umol/L; 2016 CH4 = 0.006 umol/L, CO2 = 4.45 umol/L; 2017 CH4 = 0.005 umol/L, CO2 = 12.70 umol/L; 2018 CH4 = 0.005 umol/L, CO2 = 21.84 umol/L; 2019 CH4 = 0.011 umol/L, CO2 = 5.08 umol/L; 2020 CH4 = 0.013 umol/L, CO2 = 6.17 umol/L; 2021 CH4 = 0.021 umol/L, CO2 = 5.74 umol/L; 2022 CH4 = 0.0107 umol/L, CO2 = 3.848 umol/L; 2023 CH4 = 0.0092 umol/L, CO2 = 7.9904 umol/L.
Replicates of each sample were assessed after sample analysis using calculated percent difference between replicate samples as well as the limit of quantification (LOQ = 3*MDL). If the difference between replicate samples was < LOQ, then the samples were not flagged ('0', samples are good). If the difference between samples was >LOQ and the percent difference between replicate samples was <30%, then the samples were not flagged ('0', samples are good). If the difference between samples was >LOQ and the percent difference between replicates was >30% but <50%, then the samples were flagged as '3' (difference between reps is above LOQ and above 30% difference and below 50% difference). If the difference between samples was >LOQ and the percent difference between replicates was >50%, then the samples were flagged '4' (difference between reps is above the LOQ and above 50% different). For all flags, replicates were retained in the dataset.
DATA PACKAGE UPDATES
We note several updates to the current data package following multiple revisions of earlier versions of this dataset. First, we now set any negative greenhouse gas values (methane or carbon dioxide) to zero, following best practices. Any samples that had a negative value have been flagged '3'. In addition, we standardized the calculation of MDLs for both methane and carbon dioxide to include a running MDL for all air samples analyzed during each respective year. Importantly, all MDLs are now calculated using a standard temperature (2015-2018 = 20 deg C; 2019-2023 = 40 deg C) and pressure (30 PSI).
WHOLE-ECOSYSTEM EXPERIMENTS
Multiple whole-ecosystem experiments have been conducted at Falling Creek Reservoir, including intermittent operation of hypolimnetic oxygenation (HOx) and pulsed epilimnetic mixing (EM) engineering systems. We encourage you to contact the lead author of the data package for more information.
REFERENCES
Doubek J.P., Campbell K.L., Doubek K.M., Hamre K.D., Lofton M.E., McClure R.P., Ward N.K., Carey C.C. (2018). The effects of hypolimnetic anoxia on the diel vertical migration of freshwater crustacean zooplankton. Ecosphere, 9(7):02332. https://doi.org/10.1002/ecs2.2332
McClure R.P., Hamre K.D., Niederlehner B.R., Munger Z.W., Chen S., Lofton M.E., Schreiber M.E., Carey C.C. (2018). Metalimnetic oxygen minima alter the vertical profiles of carbon dioxide and methane in a managed freshwater reservoir. Science of the Total Environment, 636. pp. 610-620. https://doi.org/10.1016/j.scitotenv.2018.04.255