Using machine learning to correct for nonphotochemical quenching in high‐frequency, in vivo fluorometer data

Abstract

In vivo fluorometers use chlorophyll a fluorescence (Fchl) as a proxy to monitor phytoplankton biomass. However, the fluorescence yield of Fchl is affected by photoprotection processes triggered by increased irradiance (nonphotochemical quenching; NPQ), creating diurnal reductions in Fchl 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 Fchl by comparing daytime values to mean, unquenched values from the previous night. A random forest regression was trained on sensor data collected coincident with Fchl; 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 Fchl profiles. The model also predicted NPQ and corrected unseen Fchl profiles from a future period with excellent results (MAE: 9.0%; RMSE: 14.4%). Fchl 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 μg L−1 to 0.33 μg L−1 when compared to uncorrected Fchl 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.

Fluorometer optical path interference via zooplankton phototaxis: Implications for high‐frequency data collection

“…Unfortunately, the coupling of elevated nighttime F chl (due to FIZ) with depressed daytime F chl (due to NPQ) may lead to exaggerated differences between day and nighttime values. In this study, F chl data from the exclusion experiment revealed fluorescence quenching in both open and exclusion treatments and its magnitude was congruent with other studies in Lake George (Lucius et al 2020). However, only the open water treatment showed the characteristic, large‐magnitude spikes associated with FIZ bias, indicating that these nighttime F chl increase where not associated with fluorescence quenching.…”

“…As F chl is a common proxy for phytoplankton biomass, NPQ correction methods have been developed to generate F chl data more representative of true phytoplankton biomass. Methods for correction often use nighttime data as a reference for comparison and correction of temporally adjacent daytime data in marine (Thomalla et al 2018; Xing et al 2018) and freshwater environments (Lucius et al 2020). However, nighttime data that are collected at high frequencies (e.g., once per minute) may be subjected to FIZ, leading to the use of biased coefficients when correcting for daytime NPQ.…”

Cyanobacteria species dominance and diversity in three Australian drinking water reservoirs

“…Optimization of fluorescence sensors through machine learning can assist in this field of research, including detection of anomalies to trigger management responses (Almuhtaram et al, 2021) and correction of light-induced fluorescence quenching (Lucius et al, 2020;Rousso et al, 2021). The application of machine learning techniques to predict cyanobacteria species dominance based on environmental conditions has been suggested as a way to improve calibration of fluorescence sensors and to provide species-specific information in real time to support improved cyano-HAB management (Rousso et al, 2020).…”

Toward a Holistic Understanding and Models of Nonphotochemical Quenching Effects on In Vivo Fluorometry of Chlorophyll a in Coastal Waters

“…daily, weekly, and interannually) in a cost-effective way (2,3). Reliability of hourly measurements, however, remains problematic (4)(5)(6)(7). Although Chla concentration is estimated by this method, it is in vivo fluorescence (IVF) that is actually measured.…”

“…IVF is the light reemitted by photosystems from excited to ground states. Chla-specific IVF (IVF B ) is variable and can be affected by phytoplankton taxonomy (8,9), cell size (10), nonphotochemical quenching (4,7,11,12,13), temperature (1,14) and nutrient status (15). Nonphotochemical quenching (NPQ) is a protection mechanism used by plants and algae to dissipate excess absorbed light energy as heat (i.e.…”