Premium
Using machine learning to correct for nonphotochemical quenching in high‐frequency, in vivo fluorometer data
Author(s) -
Lucius Mark A.,
Johnston Kenneth E.,
Eichler Lawrence W.,
Farrell Jeremy L.,
Moriarty Vincent W.,
Relyea Rick A.
Publication year - 2020
Publication title -
limnology and oceanography: methods
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.898
H-Index - 72
ISSN - 1541-5856
DOI - 10.1002/lom3.10378
Subject(s) - fluorometer , mean squared error , chlorophyll a , chlorophyll fluorescence , environmental science , chlorophyll , irradiance , chemistry , atmospheric sciences , mathematics , fluorescence , statistics , physics , biochemistry , organic chemistry , quantum mechanics
In vivo fluorometers use chlorophyll a fluorescence ( F chl ) as a proxy to monitor phytoplankton biomass. However, the fluorescence yield of F chl is affected by photoprotection processes triggered by increased irradiance (nonphotochemical quenching; NPQ), creating diurnal reductions in F chl that may be mistaken for phytoplankton biomass reductions. Published correction methods are mostly designed for pelagic oceans and are ill suited for inland waters or for high‐frequency data collection. A machine learning‐based method was developed to correct vertical profiler data from an oligotrophic lake. NPQ was estimated as a percent reduction in F chl by comparing daytime values to mean, unquenched values from the previous night. A random forest regression was trained on sensor data collected coincident with F chl ; including solar radiation, water temperature, depth, and dissolved oxygen saturation. The accuracy of the model was assessed using a grouped 10‐fold cross validation (mean absolute error [MAE]: 7.6%; root mean square error [RMSE]: 10.2%), which was then used to correct F chl profiles. The model also predicted NPQ and corrected unseen F chl profiles from a future period with excellent results (MAE: 9.0%; RMSE: 14.4%). F chl profiles were then correlated to laboratory results, allowing corrected profiles to be compared directly to collected samples. The correction reduced error (RMSE) due to NPQ from 0.67 μ g L −1 to 0.33 μ g L −1 when compared to uncorrected F chl data. These results suggest that the use of machine learning models may be an effective way to correct for NPQ and may have universal applicability.