maximum flood resuspension on October 10, 2002, range in size from the limit of resolution of the video system (approximately 30 ^m) to approximately 400 ^m. Large flocs also dominated the LISST data (see below).

Encouragingly, the floc size estimates from the VISTA and LISST were reasonably similar. Figure 10.3b shows a comparison between the D50 estimates from the LISST and VISTA deployed on the profiling rig during the October 10, 2002 anchor station. The LISST size distributions have been adjusted by excluding particles below the resolution of the VISTA and too rare (large) to have at least a 1% probability of being observed by the VISTA, before calculation of D50. LISST D50 estimates are generally between 80% and 200% of the VISTA D50 estimates. The LISST and VISTA D50 estimates are significantly correlated (r2 = 0.37) and the mean of the LISST estimates is significantly larger than the mean of the VISTA estimates (paired t-test, p < 0.05). Although it is possible that the LISST slightly overestimated D50, a more likely explanation is that the VISTA sampling procedure disrupted the largest, weakest flocs, and therefore resulted in a slight underestimation of D50 relative to the LISST. In either case, Figure 10.3b does not establish the absolute accuracy of either technique but lends considerable confidence to the use of both as reasonable estimators of in situ particle size.

Calibration of the transmissometer on the LISST to TSS sampled in situ was quite reasonable (Figure 10.3a), with r2 values for the three linear fits shown ranging between 0.82 and 0.85. The scatter in the data are due to the facts that the suspended sediment samples were taken over a 70 km reach of the upper Chesapeake Bay over all depths and over a 1-week period, and that there is a certain amount of unavoidable error associated with collection and analysis of suspended sediment samples themselves. Differences in, or departures from, a linear relationship between optical transmission and TSS are most often attributed to changes in particle size,60 with coarser particle sizes (and higher settling velocities) corresponding to a higher calibration slope (lower sensitivity). This was shown previously for the upper Chesapeake Bay by Sanford et al.,6 who pointed out corresponding seasonal and interannual variabilities in transmissometer calibrations. In 2002, these calibration differences were less apparent. In particular, the calibrations for the May and October 2002

0 0

Post a comment