Abstract
The nature and magnitude of measurement uncertainties (precision and accuracy) associated with two approaches for measuring absorption by turbid waters (b(532 nm) ranging from 0.20 m−1 to 22.89 m−1) are investigated here: (a) point source integrating cavity absorption meters (PSICAM), and (b) reflective tube absorption meters (AC-9 and AC-s – both WET Labs Inc., USA). Absolute measurement precision at 440 nm was quantified using standard deviations of triplicate measurements for the PSICAM and de-trended, bin averaged time series for the AC-9/s, giving comparable levels (< 0.006 m−1) for both instruments. Using data collected from a wide range of UK coastal waters, PSICAM accuracy was assessed by comparing both total non-water absorption and absorption by coloured dissolved organic material (CDOM) measured on discrete samples by two independent PSICAMs. AC-9/s performance was tested by comparing total non-water absorption measured in situ by an AC-9 and an AC-s mounted on the same frame. Results showed that the PSICAM outperforms AC-9/s instruments with regards to accuracy, with average spread in the PSICAM total absorption data of 0.006 m−1 (RMSE) compared to 0.028 m−1 for the AC-9/s devices. Despite application of a state of the art scattering correction method, the AC-9/s instruments still tend to overestimate absorption compared to PSICAM data by on average 0.014 m−1 RMSE (AC-s) and 0.043 m−1 RMSE (AC-9). This remaining discrepancy can be largely attributed to residual limitations in the correction of AC-9/s data for scattering effects and limitations in the quality of AC-9/s calibration measurements.
DOI Link to Publication
