Methods are provided in Sherbo et al. 2023.
Sample Collection
Standard limnological sampling of each lake was undertaken monthly from May through September of 2018. Light profiles were collected using a flat plate quantum sensor (LICOR LI-192) at 0.5 to 1 m depth intervals until the reading was <1 % of surface irradiance (I0) and used to calculate light extinction coefficients (Kd). Water column profiles for water temperature and Chl a were taken using a YSI EXO2 or a RBR maestro sonde. In addition to manufacturer defined calibrations, values of Chl a from the sonde were corrected post hoc, using laboratory based fluorometric estimates of Chl a from epilimnetic water samples. Thermocline depths were calculated using R Lake Analyzer v1.11.4 (Winslow, 2018), and euphotic depths (Zeu) were calculated as the depth of 1% of surface irradiance. Incident photosynthetically active radiation (PAR) was collected at 30-minute intervals using a LICOR flat plate quantum sensor at the IISD-ELA meteorological station.
Duplicate epilimnetic samples were collected by raising and lowering 3.5 L opaque and insulated integrating sampling bottles (Fee, 1976) from the surface to 0.5 m above the thermocline and transported to the on-site analytical laboratory within 2 hours of collection. For lakes that were part of IISD-ELA’s long-term ecological research (LTER) program (L224, L239, L373, L442), samples for phytoplankton taxonomy were also collected from the metalimnion using the same depth-integrating sampling device. The metalimnion was operationally defined as the depth range between the thermocline (Ztherm) and Zeu (1 % of surface irradiance). One of the integrated samples from each strata (epilimnion, metalimnion) was homogenized by inverting several times and sub-sampling for nutrient chemistry and phytoplankton taxonomy, while the other (epilimnion only) was reserved for primary production assays during the July sampling period (see below).
Water Chemistry and Phytoplankton Taxonomy Protocols
Water chemistry analyses followed standard analytical protocols consistent with ELA’s LTER program (Stainton et al., 1977). Samples were homogenized then filtered through a GF/C (Whatman; nominal pore size 1.2 µm) filter within 4 hours of collection. Filters were immediately placed in a desiccator and retained for analysis of Chl a, and particulate carbon (C), N and P. The filtrate was retained for analysis of total dissolved P (TDP), total dissolved nitrogen (TDN), and DOC (Shimadzu TOC analyzer). Total N (TN) and P (TP) were calculated by summing dissolved and particulate fractions. Water samples for phytoplankton taxonomy were taken from raw lake water prior to filtering and were preserved with Lugol’s solution and stored in the dark at room temperature until analysis. Phytoplankton taxonomy and biovolume estimates (Findlay et al., 1994) were consistent with those used within ELA’s long-term phytoplankton taxonomy dataset.
Primary Production Assays
Primary production assays for each lake were undertaken using epilimnetic water samples collected during July, and were initiated within 2 hours of water sample collection following the methods of Davies et al. (2003). The method measures photosynthesis and respiration via changes in partial pressure of carbon dioxide (pCO2) within the headspace of water samples over the course of a 6-hour incubation. Within a dimly lit room, sample water was filtered through 82 µm mesh (to remove large zooplankton) into an aspirator bottle with a stir bar and continuously stirred. The filtered and homogenized sample was then subsampled into 9-14 serum bottles (160 mL; Wheaton) with 7-11 bottles optically transparent and 2-3 opaque bottles. After filling, the bottles were sealed with a serum stopper and a 30 cc headspace was introduced by injecting 30 cc of room air and displacing 30 mL of lake water. More precise measurements of aqueous and headspace volumes were undertaken by weighing the serum bottles before filling and after the headspace was introduced. The bottles were placed in temperature-controlled water bath at increasing distances from a light source, a 150 Watt high pressure sodium halide lamp placed next to one end of the water bath. UHP helium (15 cc), which was also used as the carrier gas for gas chromatography, was then injected into the headspace of each serum vial to ensure that repeated sampling of the headspace did not induce negative pressure. To equilibrate CO2 between the gas and aqueous phases before sample collection, the serum bottles were lightly agitated then allowed to incubate for 1 hour prior to the collection of the first sample (see below). Midway through the experiment light measurements were taken in the position of each bottle using a Biospherical Instruments (QSL-100) spherical quantum sensor. Room barometric pressure was recorded using a (Fisherbrand Traceable) barometer. The water bath was kept within 0.5 °C of lake epilimnetic temperature though the addition of ice throughout the incubation.
pCO2 within the headspace of each serum vial was assessed at 1h, 3h and 6h of incubation using a Varian GC 8-A micro gas chromatograph fitted with a COx column. At each sampling interval, 5 cc of headspace gas was sampled using a gas tight syringe and injected into a 7 cc vacutainer that was prefilled with ultra high purity (UHP) helium at room pressure. Immediately after the 6h pCO2 sampling interval, serum bottles were acidified with 200 µl of phosphoric acid to convert all dissolved inorganic carbon (DIC) to CO2. The bottles were then shaken and allowed to equilibrate for 1-2 hours within the temperature-controlled water bath before the final headspace sample was taken for the measurement of DIC.
Vacutainers with gas samples were stored under positive pressure for 1-3 weeks prior to analysis. Standard curves were constructed using known CO2 standards (50.36 ppm, 499.6 ppm, 5004 ppm; Linde Canada) and used to calculate the pCO2 within the headspace of each serum bottle at each timestep, after correction for pressure due to the injection of helium within the serum bottles and vacutainers, and changes in bottle pressure from repeated sampling of the serum bottles (Davies et al., 2003).
The pCO2 primary production method (Davies et al., 2003) relies on the relationships between CO2 (pCO2), DIC, pH and carbonate alkalinity (ALK). To determine DIC at each timestep, two of these five variables must be known. Aqueous and headspace CO2 concentrations at each timestep were calculated based on the headspace pCO2, water temperature, and pressure within the bottles using Henry’s law. ALK was calculated at the final timestep using the R package ‘Seacarb’ (Gattuso et al., 2021) based on the concentrations of CO2 (pre acidification) and DIC (post acidification) under the assumption that ALK remained constant throughout the incubation (Davies et al., 2003). DIC at each time step was then calculated based on CO2 and ALK using the R package ‘Seacarb’ (Gattuso et al., 2021), and changes in DIC (ΔDIC) over the incubation were calculated. The relationship between -ΔDIC and PAR was plotted and fit to the Jasby & Platt (1976) photosynthesis vs. irradiance model using Sigmaplot 15 (Systat Software Inc.). Maximal photosynthetic rates (Pmax), and the slope (α) of the light limited portion of the photosynthesis vs. irradiance (P-I) response curve were derived from the model (Table S1S2. Supplemental material). Respiration (R) was calculated by averaging the change in DIC within the 2-3 opaque bottles. While the method was sensitive, the loss of pressure within the incubation bottles, syringes, or vacutainers occasionally occurred and samples were discarded. For two lakes (L239, L442) such errors resulted in too few sampling points for P-I curves and associated parameters, and phytoplankton productivity, to be calculated.
Estimates of depth integrated net primary production through the euphotic zone were calculated using the equation:
∑_(Z_0)^(Z_(eu ))〖NPP〗_Z = [P_max^b × tanh〖(α^b × I_Z )/(P_max^b )〗- R^b ] ×b_Z [Eq 1.]
where NPPZ represents net primary production at depth Z, IZ represents PAR at depth Z, and bZ represents Chl a concentration at depth Z. The superscript b indicates that values were normalized to phytoplankton biomass (estimated using Chl a). While zooplankton were removed prior to in vitro incubations, estimates of R, NPP and GPP are also influenced by bacterial respiration. NPPZ was calculated at 0.5 m depth increments from the water surface (Z= 0) to the euphotic depth (Zeu). Sub-epilimnetic production is was calculated based on P-I relationships derived from phytoplankton in the epilimnion that domay not fully account for adaptation to low light conditions. To assess the implications of this approach on depth-integrated productivity, we first examined how metalimetic values of PbMAX and αb varied from epilimnetic values from the historical ELA dataset. We used the median, and the 25th and 75th percentile, differences in photosynthetic variables to evaluate effects on depth-integrated productivity in the lake with the highest water clarity (L224), where potential errors to depth-integrated productivity would be maximal. This analysis indicated that omitting direct estimates of metalimnetic photosynthetic parameters likely resulted in modest underestimates of depth-integrated GPP (-15.5 to +4.1 %, median = -3 %), and NPP by (-25.9 to -1.9 %, median = -8 %) within the highest water clarity lakes.
To remove effects of lake size and volume differences between lakes, values of NPPZ within each depth layer were not volume weighted. GPP was calculated at each depth interval by converting respiration (R) to a positive number (i.e. multiplying by -1) then adding to NPP. Depth integrated productivity was calculated by summing GPP and NPP across the depth layers from Z= 0 to Z= Zeu (Eq 1.). Mean daily PAR during July 2018 was calculated by averaging PAR values collected the IISD-ELA meteorological site at 0.5 hour intervals throughout the photoperiod. Iz was calculated for each lake using the estimate of mean daily PAR (450 µM photons m-2 s-1) and lake specific values of Kd.
Stainton, M. P., Capel, M. J., & Armstrong, F. (1977). The chemical analysis of freshwater (Misc. Special Publ. 25, Issue. F. R. B. Canada.
Davies, J. M., Hesslein, R. H., Kelly, C. A., & Hecky, R. E. (2003). PCO2 method for measuring photosynthesis and respiration in freshwater lakes [Article]. Journal of plankton research, 25(4), 385-395. https://doi.org/10.1093/plankt/25.4.385
Findlay, D. L., Hecky, R. E., Hendzel, L. L., Stainton, M. P., & Regehr, G. W. (1994). RELATIONSHIP BETWEEN N-2-FIXATION AND HETEROCYST ABUNDANCE AND ITS RELEVANCE TO THE NITROGEN BUDGET OF LAKE-227 [Article]. Canadian Journal of Fisheries and Aquatic Sciences, 51(10), 2254-2266. https://doi.org/10.1139/f94-229
Gattuso, J. P., J.M., E., Lavigne, H., & Orr, J. (2021). Seacarb: seawater carbonate chemistry. In R package version 3.3.0. http://CRAN.R-project.org/package=seacarb.
Jassby, A. D., & Platt, T. (1976). Mathematical formulation of the relationship between photosynthesis and light for phytoplankton. Limnology and Oceanography, 21(4), 540-547.
