Several autonomous, in situ sensors were deployed on a subsurface mooring in Placid Lake, Montana USA, a dimictic, freshwater lake, to measure the partial pressure of CO2 (pCO2), dissolved O2 (DO), temperature and several other variables. These high temporal resolution time-series were obtained over six different deployments from winter 1997 to summer 1999. Data from periodic profile measurements and local and regional meteorology were obtained to aid the interpretation of the in situ data. Emphasis is placed on the analysis of the short-term and seasonal biogeochemical variability during the 1997 and 1998 under-ice deployments and the 1999 deployment from ice cover to summer stratification.

Gas variability on diel or shorter time scales was small or undetectable during most of the ice-covered periods, only becoming significant prior to ice-out when runoff and light penetration increased, promoting convective currents, vertical mixing, and biological production. A surprising 7.6 d period oscillation, believed to be driven by a baroclinic seiche, dominated the short-term variability during the first year. Increasing pCO2 and decreasing DO, as a result of microbial metabolism, characterized the long-term variability.

Vertically distributed sensors within the water column during 1999 showed that convective currents led to lake turnover and complete mixing under ice. A one-dimensional (vertical) physical mixed-layer model for the upper ocean was adapted for use in freshwater. The model simulated the thermal structure of the lake exceptionally well in both ice-covered and ice-free conditions. Simple two-box biogeochemical models were developed for pCO2 and DO and were coupled to the mixed-layer depth output from the physical model. The biogeochemical models were used to quantify the relative importance of biology, air-water gas exchange, mixing, and heating and cooling on pCO2 and DO. Net community production dominated the average daily variability of pCO2 (78%) and DO (98%) under ice. After ice out, net community production, gas exchange, mixing, and heating and cooling each contributed 33, 30, 14, and 23%, respectively, to pCO2 variability in the surface mixed-layer. These same processes contributed 36, 45, 19 and 0% to the DO.