Cross-flow filtration of dissolved and colloidal nitrogen and phosphorus in seawater: Results from an intercomparison study

A cross-flow filtration (CFF) intercomparison study was conducted to evaluate the effectiveness of a variety of CFF systems in fractionating and recovering greater than or equal to 1-kD (kilodalton) molecular weight seawater organic matter. Inorganic nutrients and total and organic nitrogen and phosphorus results are presented for CFF-processed, 0.2-mu m-filtered seawater representative of both shallow coastal (Woods Hole) and deep open-ocean (Hawaii) environments. Concentrations of NO3+ + NO2-, NH4+, PO43- and Si(OH)(4) all showed evidence of contamination or scavenging within individual CFF systems at one time or another. When adequate precautions were taken, however, nutrients displayed predicted conservative behavior in all systems. Organic N was generally observed to be less of a potential contaminant in CFF systems than was organic C. Contamination by inorganic and organic P was relatively common. Due to the low natural abundances of P in these and other natural seawater samples, even slight P contamination may confound interpretation of colloidal P recoveries. Those CFF systems that recovered significant quantities of colloidal (greater than or equal to 1-kD molecular weight) organic N generally recovered significant quantities of colloidal organic C as well. Recoveries of colloidal organic N ranged from 0 to similar to 50% of total seawater organic N. High variability in colloidal organic N recovery (similar to +/- 20%) was characteristic of identical or nearly identical CFF systems. High variability was also observed in the recovery of colloidal organic P (range = 20-80% of total seawater organic P) from coastal Woods Hole seawater. Using open-ocean Hawaii seawater, in which no total seawater organic P was detected, CFF systems that recovered significant quantities of colloidal organic C and N showed recovery of an apparent colloidal organic P fraction. This finding suggests that scavenging of P into the colloidal fraction may occur during CFF processing. C/N elemental ratios of CFF-processed organic fractions showed several differences between Woods Hole and Hawaii seawater. In general, C/N values for coastal Woods Hole seawater organic matter were relatively similar in the various fractions compared to the unfractionated, starling seawater. In contrast, the high-molecular-weight (greater than or equal to 1 kD) fraction of open-ocean Hawaii seawater had highly elevated C/N values relative to both unfractionated seawater and low-molecular-weight (less than or equal to 1 kD) organic matter. It was not possible to adequately assess C/P ratios at this time due to potential artifact-associated ambiguities in organic P data. These results emphasize the need for greater controls and calibration both among and within different makes of CFF systems. Excessive variability in both quantitative recoveries and qualitative characterization of organic matter isolated by similar or identical systems suggests that the performance of CFF systems may be sensitive to subtle differences in clean-up protocols or operating parameters, or to differences in manufacturing quality control.