Field comparison of five in situ turbidity sensors

Abstract

Five commercially available turbidity sensors were field tested by the U.S. Geological Survey Hydrologic Instrumentation Facility for accuracy and data comparability. The tested sensors were the Xylem EXO (EXO), the Hach Solitax sc (Solitax), the In Situ Aqua TROLL sensor installed onto a TROLL 600 sonde (TROLL 600), the Campbell Scientific OBS501 (OBS501), and the Observator ANALITE NEP–5000 (NEP–5000). The sensors were deployed at Pearl River at National Space Technology Laboratories Station, Mississippi (U.S. Geological Survey site 02492620), and were serviced weekly. In addition to the five in situ turbidity sensors, corresponding discrete samples were collected and analyzed during the evaluation on a calibrated Hach 2100N benchtop turbidimeter. The OBS501 malfunctioned early in the evaluation and eventually failed, resulting in few data from the sensor. During this study, the four remaining sensors (minus the OBS501) changed similarly throughout the field test; however, sensor data from the EXO consistently demonstrated lower results than the Solitax, TROLL 600, and NEP–5000, possibly because of the variation in raw signal processing among manufacturers. Results from a single factor analysis of variance test and a Tukey Honestly Significant Difference test verified the low bias observed in the EXO data and indicated there was a significant difference between the EXO data and data from the Solitax, TROLL 600, and NEP–5000 but an insignificant difference among the data when the Solitax, TROLL 600, and NEP–5000 were compared to each other.