New insights on obtaining phytoplankton concentration and composition from in situ multispectral Chlorophyll fluorescence

A three-channel excitation (435 nm, 470 nm, and 532 nm) Chlorophyll fluorometer (695 nm emission) was calibrated and characterized to improve uncertainty in estimated in situ Chlorophyll concentrations. Protocols for reducing sensor-related uncertainties as well as environmental-related uncertainties were developed. Sensor calibrations were performed with thirteen monospecific cultures in the laboratory, grown under limiting and saturating irradiance, and sampled at different growth phases. Resulting uncertainties in the calibration slope induced by natural variations in the in vivo fluorescence per extracted Chlorophyll yield were quantified. Signal variations associated with the sensors (i.e., dark current configurations, drift, and stability) and the environment (i.e., temperature dependent dark currents and contamination by colored dissolved organic matter [ CDOM] fluorescence) yielded errors in estimating in situ Chlorophyll concentration exceeding 100%. Calibration protocols and concurrent observations of in situ temperature and CDOM fluorescence eliminate these uncertainties. Depending upon excitation channel, biomass calibration slopes varied between 6- and 10-fold between species and as a function of growth irradiance or growth phase. The largest source of slope variability was due to variations in accessory pigmentation, and thus the variance could be reduced among pigment-based taxonomic lines. Fluorescence ratios were statistically distinct among the pigment-based taxonomic groups, providing not only a means for approximating bulk taxonomic composition, but also for selecting the appropriate calibration slope to statistically improve the accuracy of in situ Chlorophyll concentration estimates. Application to 5 months of deployment in China Lake, Maine, USA reduced the error in estimating extracted Chlorophyll concentration from > 30% to < 6%.