The use of field-deployable fluorescence sensors to better understand dissolved organic matter concentrations and composition has grown immensely in recent years. Applications of these sensors to critical monitoring efforts have also grown, encompassing post-fire monitoring, wastewater tracking, and use as a proxy for various contaminants. Despite the growth, it is well known that these sensors require corrections for temperature (Watras and others, 2011) and are subject to many light-field interferences caused by both scattering and absorbance due to dissolved and particulate substances (Downing and others, 2012; Lee and others, 2015; Booth and others, 2023). The most common fluorescence sensors used by the U.S. Geological Survey (USGS) include those targeting fluorescent dissolved organic matter (fDOM) and chlorophylls. Because fDOM sensors primarily measure fluorescence in the dissolved to colloidal phases, corrections to the interferences caused by particulates can be made relatively easily. By the end of 2024, the USGS had 69 fDOM sensors deployed within official water quality monitoring networks included on the USGS National Water Dashboard (https://dashboard.waterdata.usgs.gov/app/nwd/en/) and numerous others used in surveys and research applications across the Nation.

Although temperature corrections are widely applicable across sensor models, interference corrections can be model specific due to differences in design specifications across manufacturers and models (Booth and others, 2023). The corrections are also potentially subject to changes in manufacturing within a specific sensor model. Recently, USGS staff obtained information regarding possible changes in the manufacturing of its most widely-used fDOM sensor model, raising concerns about data consistency and quality in the USGS fDOM sensor networks.

Furthermore, changes in turbidity sensors since the corrections guidance was performed may also affect the performance of the corrections. The turbidity sensor used in the original experiments (Downing and others, 2012) was determined to have a signal output approximately 1.3 times higher than the output of the turbidity sensor currently used in an extensive field comparison study (Messner and others, 2023). With these changes, it is imperative that the corrections be reevaluated to maintain data consistency and continuity across the USGS.

In this study, we evaluated turbidity corrections for fDOM sensors over a range of serial numbers covering manufacturing dates 2015 through 2022 and turbidity serial numbers covering the range 2013 through 2022. The goal was to determine whether reported changes in the manufacturing process of the fDOM and turbidity sensors affected the correction approach developed by Downing and others (2012) such that additional guidance would be required to address this manufacturing change. To evaluate, we repeated a laboratory-based test similar to that performed by Downing and others (2012) in which a series of tank experiments with multiple sensors were deployed in a suspension of Elliot Silt Loam (ESL). High turbidities of the ESL suspension were maintained throughout the tank by turbulent recirculation using submersible pumps. Particulates were removed using a recirculated line equipped with a capsule filter (0.45 micron). Measurements were collected throughout the filtration until turbidities reached approximately 5 formazin nephelometric units (FNU; data available in Baxter and others, 2023). Each experimental run included a mixture of unique sensor combinations to account for variability imposed by the turbidity and temperature sensors. The fDOM correction factor was calculated for each combination of fDOM and turbidity sensors included in the test.

We observed no systematic change in fDOM correction coefficients across serial numbers representing manufacturing years 2015 through 2022. However, the results highlighted questions raised about the corrections for high-turbidity samples, as noted in USGS Techniques and Methods (Booth and others, 2023). Applying the inverse of the commonly-used fDOM ratio with a quadratic fit performed better than the exponential fits when correcting fDOM data for turbidity in the ESL laboratory filtration test and generated a simple scale factor correction equation. This approach also served as a better indicator of data quality than the exponential fit approach. Similar to fDOM, more rigorous quality assurance measures may be necessary to evaluate turbidity sensor calibrations and performance. Sensors exceeding a certain age may need to be replaced despite passing quality assurance checks during calibration. Further testing of the turbidity corrections for different sediment and water types is warranted to better understand the variations in the fits and correctable ranges of turbidity in different systems.