Geoscience Reference
In-Depth Information
changes in the fluorescent properties of freshwater COM (>1 kDa) isolated from two estu-
aries, added to low-fluorescing ultrafiltration permeates to recreate a salinity gradient, and
then exposed to estuarine and marine bacteria. They found different responses in freshwater
COM fluorescence as a function of salinity from San Francisco Bay, which exhibited no
predictable changes in fluorescence intensity when compared to Chesapeake Bay, in which
humic fluorescence decreased as salinity increased. This result appeared to be related to
the different riverine sources of the COM used in these experiments (i.e., Sacramento vs.
Susquehanna River, Boyd and Osburn,
2004
).
COM fluorescence properties, as investigated by excitation-emission matrix (EEM)
spectra, show differences for natural water DOM based on source. As an example for dis-
cussion,
Figure 7.2
compares separately surface water DOM from a freshwater riverine
site in the Chesapeake Bay (
Figure 7.2A-C
) and from a coastal marine site in the Middle
Atlantic Bight (
Figure 7.2D-F
) with their respective high molecular weight (HMW)
retentate (>1 kDa) and low molecular weight (LMW) permeate DOM (<1 kDa) fractions
(Osburn, unpublished data). To produce these separations, a 1000 Da molecular weight
cutoff (MWCO) tangential flow (or cross-flow) ultrafiltration (TFF or CFUF) system was
used after initial 0.2 μm filtration (Boyd and Osburn,
2004
). To describe similarities and
differences in EEM results here and elsewhere in this chapter, we use the convention of
peak assignments developed by Coble (
1998
;
Table 7.1
) and we focus on terrestrial humic,
terrestrial fulvic, marine humic, and protein peaks (A, C, M, and T, respectively).
For both water samples, and for each fluorescence peak, the concentrated retentate had
the highest fluorescence (
Figure 7.2
). However, when normalized for the concentration fac-
tor, HMW fluorescence actually was less than the starting DOM fluorescence yet remained
higher than for the LMW permeate fluorescence (
Table 7.2
). The differences among the
three fractions were greater for the river sample than for the marine sample. Further, EEM
fluorescence varied for different peaks, and recovery was greater for the marine sample
than for the river sample. Recoveries are comparable with others reported (Wells,
2002
;
Boyd and Osburn,
2004
; Liu et al.,
2007
). Despite the lower recovery for the river DOM
(which could be due to sorption onto the filter' Mopper et al.,
1996
; Liu et al.,
2007
),
the EEMs are similar in appearance among the surface DOM, retentate, and permeate
fractions. In contrast, the marine surface DOM sample showed distinct A and T peaks
(
Figure 7.2D
). A peak shoulder is noticeable at the intermediate (Int) region between the M
and C peak regions, as described by Boyd et al. (
2010a
). After ultrafiltration, the Int peak
is notably lacking in the marine permeate (
Figure 7.2E
) while the T peak is less noticeable
in the marine retentate, potentially masked by the A peak fluorescence (
Figure 7.2F
). This
result suggests that protein (T peak) fluorescence occurs primarily in the LMW fraction
of marine DOM whereas the Int peak, intermediate between M and C, is primarily in the
HMW fraction consistent with work by Liu et al. (
2007
). Further, the Int peak is prominent
in the river CDOM, permeate, and retentate fractions (
Figure 7.2A-C
).
Because fluorescence intensity changed on molecular weight separation, peak ratios
could be employed to understand how DOM sources are partitioned between size classes,
based on their fluorescence (
Table 7.2
; Coble,
1998
; Parlanti et al.,
2000
). As defined in
