The water-quality sonde

A station is a tube on a piling: a dock's, a bridge's, or one driven for the purpose. At North Inlet–Winyah Bay in South Carolina, the tube is 4-inch pipe strapped to the piling with 19 mm nylon banding and held away from it by a 2 × 6 inch PVC board, so the water flows around it. The tube stands on the creek bed, and a stop bolt through it 30 cm above its base holds the sonde, so the probes stay 30 cm above the bed at any tide.

Four slots 2 × 12 inches (5 × 30 cm), cut around the tube just above the bolt, let the water flow over the probes. Pairs of holes 2.5 inches (6.4 cm) across, drilled every foot up the rest of the tube, let it flush as the tide rises and falls. The figure shows a PVC tube; at North Inlet the lower part is now copper, which slows the growth of organisms on it. Where a creek is deep, a sonde may instead be fixed at a set depth: Winyah Bay's surface sonde is about 0.2 m below the water at mean low water.

The reserves use YSI's EXO sondes, approved for the program in 2013 to replace the older 6600 model; North Inlet–Winyah Bay switched between 2015 and 2017. The EXO2 shown here is 7.62 cm across and 70.52 cm long with its guard, and weighs 3.60 kg. Its base, the bulkhead, has six ports for probes around a central port for the wiper; an empty port is sealed with a plug. A slotted guard screws on over the probes to protect them.

Four D cells in the sonde's body power it for about 90 days at one reading every 15 minutes. The sonde stores each reading in its own memory, and a cable from the connector at its top runs up the tube to the transmitter on the piling.

Conductivity and temperature share a probe: four nickel electrodes measure how well the water conducts electricity, and a thermistor measures its temperature. Conductivity rises with temperature, so the sonde reports specific conductance, the conductivity corrected to 25 °C. Salinity is not measured; it is calculated from conductivity and temperature using the Practical Salinity Scale, and reported in practical salinity units.

Dissolved oxygen is measured with light. Blue light makes a dye in the probe's cap glow, and oxygen in the water shortens the glow. The sonde times the glow, against a red reference, to get the oxygen as a percentage of saturation. The concentration in mg/L is calculated from that, the temperature and the salinity.

pH comes from a glass electrode and a reference electrode in one probe; the voltage between them follows the water's acidity. Turbidity is the near-infrared light (860 nm) that particles in the water scatter to a detector at 90°, in formazin nephelometric units (FNU). Where a reserve installs one, a chlorophyll probe shines blue light (470 nm) to make the chlorophyll in phytoplankton fluoresce, and measures the glow. It gives an estimate of chlorophyll, calibrated against a dye.

Depth comes from a strain gauge behind two small holes in the bulkhead. It senses the weight of the water and of the air above it, about 1 cm of depth for each hectopascal of air pressure, so the reserves subtract the air's share using the pressure measured at their weather station. The depth rises and falls with the tide; it is the depth of water above the sensor, not of the creek.

YSI's stated accuracies, as listed by North Inlet–Winyah Bay (where two are given, the greater of the two applies): temperature ±0.2 °C; specific conductance ±1% of the reading or 0.002 mS/cm; salinity ±2% or 0.2; oxygen saturation ±1% of the reading or 1% saturation, and concentration ±1% or 0.1 mg/L (up to 20 mg/L); depth ±0.004 m; pH ±0.1 within 10 °C of the calibration's temperature; turbidity ±2% or 0.3 FNU. Chlorophyll has no stated accuracy, since it depends on the method.

At each 15-minute mark, the wiper turns once across the faces of the probes and back to clean them, and then the sonde reads each probe. In the setting used for moorings, it filters the reading for up to about 40 seconds, rejecting outliers. So each value is the water at that moment, not an average over the 15 minutes: a squall, a front or a pulse of fresh water that passes between readings is missed or caught only in part.

Anything left in an estuary soon grows a coat of algae, barnacles and silt, and the readings drift as it does. At North Inlet–Winyah Bay, each sonde stays out for two weeks, a period set by fouling and battery life. Each station has two sondes, so a freshly calibrated one goes in the same day the other comes out. The surfaces of the probes that don't measure are covered in copper tape to slow the growth.

Before it goes out, a sonde is calibrated in the laboratory: pH with standards of 7 and 10, turbidity with 0 and 124 NTU, specific conductance with a 10 mS/cm standard, and dissolved oxygen and depth by the air pressure at the time, with oxygen in air-saturated water. When it comes back, it measures fresh standards again before cleaning, to find out how far it drifted while it was out.

Every hour, the transmitter on the piling sends the last four readings to a GOES weather satellite, and from there they reach the reserves' Centralized Data Management Office. The office checks them automatically, flagging readings that are missing or outside the sensor's range, and they become provisional data. After each deployment, the full record from the sonde's memory goes through the same checks. The reserve then reviews it and flags what it finds, and the office makes a final check before the data become authoritative, published at nerrsdata.org.

NOAA's National Data Buoy Center passes the provisional readings on, and that is what Bluegraph shows: the water's temperature, specific conductance, salinity, dissolved oxygen as saturation and as concentration, pH, turbidity, chlorophyll where measured, and the sensor's depth, each a reading 15 minutes apart. You can find the stations on the stations list, among the estuaries.