| 05 Aug 2026
Phytoplankton community, seasonality, and water chemistry modulate the lacustrine diurnal carbon engine and carbonate δ13C values
Abstract. Stable isotopic measurements of carbon in carbonate rocks (δ13Ccarb) have long been used to investigate paleoclimate and global carbon cycle dynamics. Recent work documented the diurnal carbon engine effect, which describes the impact of daily photosynthetic cyclicity on dissolved inorganic carbon (DIC) and, in turn, δ13Ccarb and δ13CDIC values, in low-latitude shallow marine environments. Because carbonate precipitation is temporally structured over the diurnal cycle, the resulting δ13Ccarb reflects a precipitation-weighted bias rather than a simple daily average. However, this system has not yet been explored in lakes. Here we present an adapted diurnal carbon engine model and constrain this effect for three end-member lake systems: Great Salt Lake (UT, USA), Green Lake (NY, USA), and Morrison Lake (MT, USA). We document how differing lake water chemistry and phytoplankton communities modulate local diurnal carbon engines. Because of geochemical and ecological differences, the impact of the diurnal carbon engine on δ13Ccarb values varies greatly by lake system. We estimated Δ¹³C offsets, which represent the daily variability expected resulting from the diurnal engine and precipitation weighting, of ~0.36 ‰ in Great Salt Lake and ~0.24 ‰ in Green Lake. We also modeled how seasonality impacts the diurnal carbon engine in Great Salt Lake, which resulted in ~0.10 ‰ of δ13Ccarb variability. Modeled diurnal variability is consistent with observed dynamics in Green Lake. Predicted δ13Ccarb values are consistent with those measured from most carbonate sedimentary facies in each lake. Lacustrine δ13Ccarb shifts may reflect local environmental and ecological conditions, rather than changes in long-term lake δ13CDIC values or the global carbon cycle, complicating lacustrine δ13Ccarb record interpretations.
Line 223: use (rather than exhibit)
Line 225: I think it would be more effective to highlight the different CCMs in alpha and beta cyanos here rather than introducing this below (e.g. line 243). Then Line 241 can focus on differences in carbon isotopes because the differences in C fixation are already established above.
Line 241: Its odd to highlight that these are distinct, including algae, but then not include any fractionation numbers for algae. I realize this can difficult due to the differences in CCMs etc. but as a reader it feels odd that algal numbers are not included.
Line 246: Its really just the two CCMs, its probably more about the environment, etc. “Diverse CCMs” is not really established unless (maybe) you also include algae. Or does this refer more to complexity in phenotype?
Lines 339-340: How are Green Lake and Morrison lake “end members”. Presenting Green Lake as a freshwater lake is likely a bit misleading. This seems more accurately described as systems that produce carbonate that vary in community composition. I would remove the “different end-member environments” unless this can be justified for all sites.
Line 342: Phytoplankton above the chemocline in Green Lake.
Line 397: As above, end members of what specifically. There are many lakes that differ in climate and seasonality - how are, for example, Green Lake and Morrison Lake the absolute end members of those while Great Salt Lake is an n of 1 rather than an end member. I like the framework of the 4 lakes and appreciate the idea for the modeling, but having data from 4 different or distinct lakes is not the same as these systems being endmembers.
Line 399: I’m still on this end member bit. How can ML be an end member when the chemistry of the lake is not included. Why include ML at all in this model without empirical data from the lake? Does that not compromise the model? Or are there other high elevation mountain lakes that might help bolster whether the estimates uses here are anywhere near actual conditions at ML?