Kd averages calculated by month from studies conducted in St. John, US Virgin Islands from 2014-2017 (NCEI Accession 0277386)

Acquisition Description:
Light was measured as the radiant energy between 400 and 700 nm wavelength (i.e., PAR, μmol quanta m-2 s-1) as Photosynthetic Photon Flux Density (PPFD). In situ light was measured using two logging meters fitted with a cosine-corrected PAR sensor and wiper (Compact LW, JFE Advantech Co., Ltd, Japan), that were deployed at ~ 19.1-m depth (height of the sensor) in Great Lameshur Bay (18° 18 ́ 37.04N, 63° 43 ́ 23.17W).

These instruments recorded downwelling PAR, and were deployed six times from 2014and 2017, from August to March and from March to August. The meters were operated in burst mode, during which they would wake up, clean the sensor with a wiper, and record a burst of multiple records before returning to sleep. The Compact LW meter is designed for oceanographic applications to 200 m depths, is fitted with a photodiode sensor, and has a stated accuracy of ± 4% (over 0–2000 μmol photons m2 s-1) and resolution of 0.1 μmol photons m2 s-1. Both meters were purchased new for this study, and were deployed individually and sequentially between field samplings with comparisons between consecutive deployments used to screen for calibration drift. One sensor was used for a combined duration of 16 months during, and the other sensor was used for 4 months, returned to the manufacturer for servicing (May 2016), and then used again for 3 months. In between deployments, sensors were inspected for abrasions that would affect calibration, and were carefully cleaned with vinegar.

Different configurations of the meter were employed to prolong battery life. In the first and second deployments (starting 21 August 2014 and 19 March 2015, respectively), a burst of 10 measurements was recorded at 0.033 Hz (i.e., every 30 s) every 1.5 h; the instrument failed during the third deployment (starting August 2015); in the fourth and fifth deployments (starting 16 March 2016 and 29 July 2016, respectively) a burst of 10 measurements was recorded at 0.033 Hz every 1.0 h; and in the sixth deployment(starting 23 February 2017) a burst of 30 measurements was recorded at 0.100 Hz (i.e.,every 10 s) every 2.0 h. The timing of bursts was not standardized to local time and, therefore, the number and timing of bursts bracketing noon (which were used to calculate transmission, described below) differed among deployments. The sampling frequency within each burst was sufficient to alleviate the bias resulting from wave-induced light flecking (Zheng et al. 2002). As a result of varying power demands of each sampling configuration, the meter did not always record for the full duration of each deployment.

Surface light was recorded with two cosine-corrected PAR sensor (S-LIA-M003, OnsetComputer Corporation) attached to loggers (Micro Station Data Logger H21-002, OnsetComputer Corporation) recording at 0.0033 Hz (i.e., every 5 minutes). The sensors were calibrated by the manufacturers, and were mounted ~ 4-m above sea level adjacent to Great Lameshur Bay (18° 19 ́ 6.61N, 64° 43 ́ 27.73W), and ~ 0.875 km from the sensor recording in situ light. The paired surface sensors were used to ensure data integrity should one sensor fail, and the paired deployments provided a means to detect erroneous records due to sensor drift or failure. The surface sensors were downloaded, reprogrammed, and cleaned in July of each year, and have been deployed for 11 y(from 2007). Here, surface light data for 2014-2017 are presented to provide temporal concordance with the submerged sensor.