Sensor Description
Multiparameter sondes (EurekaTM Manta 35) were equipped with sensors to measure temperature, pH, specific conductivity, optical dissolved oxygen (DO), turbidity, colored dissolved organic matter (CDOM), and optical brighteners (OB). An average value for the barometric pressure of sonde location was recorded for all of the time points based on the altitude of the location, which was entered in the EurekaTM Manta software. Measurements of each parameter were recorded every 15 minutes. Just before recording measurements, an automatic wiper wiped the ends of all of the sensors to prevent biofouling. Sondes were mounted inside PVC tubes with holes to allow for flowing water to pass over the sensors, anchored to the stream bottom, and the PVC tubes were secured to a bridge or rock on the stream bank with cables (Atlanta), anchored to the stream bottom on top of concrete blocks, and the blocks were secured to a tree on the stream bank with cables (Boston), or attached directly to the stream bank (Portland). Each sonde was powered by six 1.5V C-size alkaline batteries.
Sonde maintenance was completed approximately every two weeks during deployment (up to four weeks during warm weather when battery drain was slower). During maintenance, each sonde was removed from the stream then cleaned and rinsed with tap water. KimwipesTM and deionized water were used to gently clean the ends of the sensors. Data were downloaded from the sonde, and then the calibration of the different sensors was checked.
Deployment
Atlanta
Sondes were deployed from 16 November 2021 to 11 January 2023 at three stream sites in the metropolitan Atlanta, Georgia. Sondes were co-located with established U.S. Geological Survey (USGS) gaging stations so data could be correlated with discharge data (publicly available at https://dashboard.waterdata.usgs.gov). The three sites (SR02, USGS site 02203900; SR12 , USGS site 02203655; and SR24, USGS site 02203603) were selected because their watersheds are nested within the South River Watershed.
Boston
Sondes were deployed from 23 November 2021 to 20 December 2022 at three stream sites in the greater Boston, Massachusetts area. Sondes were co-located with established U.S. Geological Survey (USGS) gaging stations so data could be correlated with discharge data (publicly available at https://dashboard.waterdata.usgs.gov). The three sites (Alewife Brook, USGS site 01103025; Beaver Brook, USGS site 01104501; and Stony Brook, USGS site 01104370) were selected because their watersheds are similar in size (21.7, 28.8, and 26.3 km2, respectively) yet have a range of land cover (59.9, 41.4, 7.3% impervious cover, respectively). The sonde location in Alewife Brook is also downstream of six combined sewer overflow (CSO) outfalls.
Portland
Sondes were deployed from 23 July 2023 to 02 August 2024 at three stream sites in the greater Portland, Oregon area. Sondes were co-located with established U.S. Geological Survey (USGS) gaging stations so data could be correlated with discharge data (publicly available at https://dashboard.waterdata.usgs.gov). The three sites (Johnson Creek at Milwaukie, USGS site 14211550; Johnson Creek at Sycamore, USGS site 14211500; and Johnson Creek at Regner Rd, USGS site 14211400) chosen represent a downstream longitudinal urbanization gradient across the Johnson Creek Watershed, with more rural and agricultural defined areas in the headwaters (above Johnson Creek at Regner Rd) and more heavily urbanized areas towards the mouth (Johnson Creek at Milwaukie).
Calibration
Temperature readings did not require calibration. The optical sensors for CDOM and OB were factory-calibrated, but were immersed in standard solutions of 10 ppb and 100 ppb p-Toluenesulfonic acid (PTSA) solution to check for drift. For the other sensors, those giving readings more than +/- 5% different than the standard values were recalibrated in the field or lab according to the manufacturer’s instructions, and at least quarterly.
To calibrate pH, the sensor was tripled-rinsed with pH buffer solution and then immersed in the buffer using the calibration cup. The pH was calibrated in the EurekaTM Manta software using buffer solutions with pH 4.0, 7.0, and 10.0. To calibrate the specific conductivity sensor, the sensor was triple-rinsed with a conductivity standard solution and then immersed in the solution using the calibration cup. Specific conductivity was calibrated in the EurekaTM Manta software using a 1413 µS/cm standard solution. The optical DO sensor was calibrated using air-saturated water. A 500-ml bottle of deionized water was shaken vigorously for about one minute, and then the cap was removed and the bottle allowed to sit for about two minutes. The DO sensor was then triple-rinsed with and immersed in the oxygenated water using the calibration cup. DO was then calibrated using the EurekaTM Manta software. To calibrate the turbidity sensor, the sensor was triple-rinsed with and then immersed in deionized water (0 NTU) and a 100 NTU turbidity standard solution (Atlanta and Boston) or both 0 NTU and 1000 NTU standard solutions (Portland). Turbidity was then calibrated using the EurekaTM Manta software.
Data QA/QC
Sensor data went through several stages of quality control. First, field calibration checks were used to identify calibration drift. Calibration drift greater than 5% was corrected either assuming linear drift or using an offset correction, depending on visual inspection of the data.
Following drift correction, data were checked for reasonable value ranges, and values outside these ranges were removed. This included negative values for specific conductivity, turbidity, pH, CDOM, OB, and dissolved oxygen which should always be positive.
Outliers were flagged, but not removed from the dataset. These were defined as any values more than three standard deviations from a 6-hour moving average for each parameter.
Because chromophoric dissolved organic matter values can be affected by temperature, we calculated a standardized CDOM (at 20 degrees C) using the following equation:
CDOM20 = CDOM/(1-0.00273*(Water Temperature-20))
CDOM readings can also be affected by turbidity, especially above 15 NTU (Downing et al. 2012). Unfortunately, the turbidity sensors on our sondes were often unreliable. Where other turbidity data were available (i.e., from adjacent USGS gages), these were used to correct CDOM values, using literature values (Downing et al. 2012).
Data were flagged using the following system of flags:
C = data are corrected for calibration drift
N = sensor was out of the water for maintenance
R = parameter value has been removed because it was out of range (e.g., an impossible negative value)
F = the stream was frozen
M = data has been removed because of sensor malfunction
O = value is greater than 3 standard deviations from a 6-hour average value for this parameter
L = low stream flow may affect values
Downing, B. D., B. A. Pellerin, B. A. Bergamaschi, J. F. Saraceno, and T. E. C. Kraus. 2012. Seeing the light: The effects of particles, dissolved materials, and temperature on in situ measurements of DOM fluorescence in rivers and streams. Limnology and Oceanography: Methods 10: 767–775. doi:10.4319/lom.2012.10.767