Assessing the error in photosynthetic properties determined by fast repetition rate fluorometry

Fast repetition rate (FRR) fluorometry is an optical technique for estimating photosynthetic properties of phytoplankton from measurements of variable fluorescence yield. I determined the minimum error in such estimates contributed by inherent instrument biases, improper measurement protocols, and the type of optimization algorithm used to infer photosynthetic variability from changes in variable fluorescence. Many of these errors were nonrandom in origin and would not be reduced with repeated sampling or averaging. Characterization of a commercial FRR fluorometer (FRRF) showed undocumented hardware biases with magnitudes roughly equivalent to those addressed by the current characterization approach. Robust optimization algorithms were less likely to misidentify these biases as representing actual photosynthetic variability. In general, robust algorithms improved the accuracy, precision, and distribution of error in analyses of both simulated and actual variable fluorescence measurements. When reanalyzed with robust algorithms, in situ data from an FRRF indicated different photosynthetic behavior than did the analysis tool used originally. Methods to improve and standardize the collection and analysis of FRR variable fluorescence data are essential for evaluating the strengths and limitations of this powerful, but involved, technique. Although the error minimization procedures described were developed primarily to minimize errors and artifacts with FRR fluorometry, several are generally applicable to any fluorescence yield technique in which a physiological model is used to estimate photosynthetic parameters from variable fluorescence measurements.