Non steady-state dynamics of stable oxygen isotopes for estimates of metabolic balance in large lakes

The ratio of gross photosynthesis to community respiration (P:R) is a vital characteristic of ecosystem function but is poorly defined in large lakes because traditional methods are impractical for measuring P and R on appropriate time and space scales. The new steady state in situ18O method has a great power to assess metabolic patterns in large lakes. However it could be confounded by seasonal and/or episodic temperature changes, which can violate the steady-state assumptions by altering the saturation level of dissolved O2 without actual gain or loss of O2 when air-water gas exchange is too slow to compensate for the altered solubility. We estimated P, R and P:R at three coastal sites of varied depth, productivity and physical exposure in the Laurentian Great Lakes by both the 18O method and experimental incubations. At two shallow sites with low and moderate exposure to open lake forces, the 18O and incubation methods returned similar estimates for P, R and P:R. The deepest site experienced large and frequent upwelling episodes and the 18O method was unreliable, with P and R estimates usually negative. We define an O2 saturation ratio (γ) that determines the applicability of the 18O method to any given system. Estimates obtained using observed O2 saturation above this threshold are invalid. The γ ratio increases with lake productivity. Higher productivity and/or shallower lakes appear good candidates for use of the 18O method, but applications in deeper, physically-driven lakes will likely need explicit attention to temperature-driven non-steady state effects.