Comparison of cyanobacteria monitoring methods in a tropical reservoir by in vivo and in situ spectrofluorometry

International audienceA reliable and representative monitoring of phytoplankton abundance, especially cyanobacteria, is mandatory for choosing the best management or restoration strategy and for assessing its effectiveness in worldwide eutrophic lakes and reservoirs. In vivo and in situ fluorometry is expected to be faster than the traditional methods of phytoplankton counting and chlorophyll-a (chla) extraction and quantification. Contrary to water bodies in temperate climate, the technique still needs investigation in tropical fresh-waters. We assessed the performance of a spectrofluorometer, the FluoroProbe (FP) in Lake Pampulha (Minas Gerais, Brazil), a water body where cyanobacteria are the dominant group most of the time. FP chla concentrations were in average almost three times lower than spectrophotometric chla concentrations and underestimation was larger when Chroococales were the dominant species. The linear response of the probe was limited to 100 g chla L −1. The correlation between FP and spectrophotometric measurements was significant only for FP measurements below 100 g chla L −1 (r = 0.84, p < 0.001, n = 25). The correlations between FP measurements and biovolumes of phytoplankton (r = 0.64, p = 0.001, n = 23) and cyanobacteria (r = 0.60, p = 0.002, n = 23) were significant but weaker than reported in temperate water bodies. Oscillatoriales fingerprint was the most distant from FP default fingerprint for cyanobacteria, presenting a lower relative fluorescence at 610 nm. Applying an average experimental fingerprint obtained from four dominant cyanobacteria species in Lake Pampulha allowed us to improve FP measurements compared to spectrophotometric data, especially when a single species, Planktothrix isothrix, was highly dominant. Even so, high phytoplankton biomass in Lake Pampulha led to high optical density, which disturbs FP measurements. Thus, in highly eutrophic water bodies, additional laboratory analysis is necessary to validate in situ spectrofluorometer measurements, more than in oligotrophic and mesotrophic environments