Custom In-Situ Fluorescence Sensors


The three custom fluorometers designed for this study were deployed at the CRW DWTP intake, but only for a limited time, and late in the study period (fig. 28 and table 4). The majority of this period covered the prolonged recession to the summer base-flow period, through autumn, and into the January rainy season. Individually, the two sensors centered near Peak C designed to measure the FI tracked DOC concentration and were highly correlated to FDOM from the standard Cyclops-7 fluorometer, which is also focused on this Peak C region but has a wider band-pass (table 5).


While both of the Peak C custom sensors tracked DOC concentration, there was a change in the ratio of these two sensors, referred to here as FI_in-situ (fig. 28). Prior to this study, the FI has only been calculated using benchtop measurements of fluorescence; the ratio of emission at 470 to 520 nm at an excitation wavelength of 370 nm effectively provides information about the slope of the fluorescence response in this region of EEMs space (McKnight and others, 2001). As described above, the FI provides information about DOM source and composition. The two in-situ sensors developed for this study, however, have a broader band pass than benchtop instruments. Thus, while the ratio of these two sensors is expected to be correlated to the benchtop FI ratio, the absolute value of the FI_in-situ is expected to differ from what is commonly reported for benchtop measurements of surface‑water samples—a range of 1.2 to 1.8 (Cory and others, 2010). The FI_in-situ values for this study ranged from 0.9 to about 1.2 (fig. 28). 


Over the 10-month deployment of these sensors, the FI_in-situ values first decreased as streamflow and DOC concentration decreased and then increased in September through December 2011 when streamflow and DOC concentrations remained low and relatively stable (fig. 28). Because FI values are assumed to be independent of concentration, these trends suggest a change in carbon composition. The initial decrease in FI_in-situ suggests a trend towards a DOM pool increasingly dominated by terrestrial, high molecular-weight material. The increase during the later period of deployment in September–November coincided with the drawdown of Timothy Lake, and a period when benthic algal populations may have started to senesce at the end of the growing season. Concentrations of chlorophyll-a at the Oregon City monitor produced occasional spikes up to about 10 µg/L (fig. 28); these periodic peaks suggest moderate sloughing of benthic algae. With the onset of rain, FI_in-situ values increased, perhaps because of suspension of decaying algae in the river or some other factor. Although data collected for this study are not conclusive, these initial results show a dynamic response in the FI_in-situ to changes in DOM character and quantity, and show promise for future water-quality monitoring applications.


The design of the experimental Peak T custom sensor (ex 270/em 340 nm) required the use of a low-ultraviolet LED (table 5). The signal from this sensor was low, likely because of the low amount of fluorescence response from DOM in this region of the EEMs. Furthermore, the loss of signal from the sensor in mid-July was determined to be caused by loss of signal output from the lamp. Testing of LEDs in this region has shown that photon output from these deep ultraviolet LEDs are short-lived. As LED manufacturers continue to improve the efficiency of deep ultraviolet LEDs—the ability of the device to convert electrons to photons or the external quantum efficiency—these deep ultraviolet LED-based field sensors are expected to improve concurrently. Testing of these low-ultraviolet fluorometers in waters that are known to have higher fluorescence in this region (wastewater impacted rivers, for example; Hudson and others, 2008; Goldman and others, 2012), may also prove to be a more suitable application for these sensors. 


First posted February 11, 2013