Reconstructed long-term time series of phytoplankton primary production of a large shallow temperate lake: the basis to assess the carbon balance and its climate sensitivity

Aiming at building the carbon budget for further climate change impact research in the large and shallow northern temperate Lake Vrtsjarv, the present paper focuses on reconstructing the full phytoplankton primary production (PP) data series for the lake for the period of 1982-2009 covered by disconnected measurements, and testing the uncertainties involved both in the PP measurements and bio-optical modelling. During this 28-year period, in situ PP was measured in Vrtsjarv in 18 years with C-14-assimilation technique. We reconstructed the full time series using a semi-empirical PP simulation model based on continuously measured PAR irradiance and interpolated values of monthly measured chlorophyll a (C (chl)). The modelling results, which proved highly reliable during the calibration phase, correlated rather weakly with the annual PP estimates for the 18 years, which were based on 2-h incubations at midday, 1-2 times per month. Being based on continuous irradiance data, the modelled PP can be considered more reliable than the sparse measurements, especially for short to medium term studies. We demonstrate that in the long-term, the bio-optical method can be biased if changes in water colour or water level alter the light climate causing adaptive responses in the cellular chlorophyll content of light-limited phytoplankton.