Global Synthesis of POC Using Satellite Data calibrated with Transmissometer
and POC Data from JGOFS/WOCE

National Science Foundation Grant No. 9986762

Talk presented to the JGOFS-SMP Workshop, July 16-20, 2001, Woods Hole, MA.

The use of beam attenuation to predict POC: bottle or pump data?
W. Gardner, M.J. Richardson, A.V. Mishonov
Dept. Oceanography, Texas A&M University
College Station, TX, 77843

One of the goals of the JGOFS program is to reduce the uncertainties in carbon budgets for the ocean. Particulate organic carbon is a small portion of the total carbon in the ocean, but it is the fraction that can move independently of the surrounding water, i.e. it can sink. There are two primary ways to sample and measure particulate organic carbon (POC) from water samples in the ocean, but the results can differ by than a factor of 5-10. The oldest and most common method is to collect water samples (up to a few liters), filter the water and analyze the filter for POC. The other method uses in-situ pumps to filter thousands of liters of water to obtain large samples of particulate matter. A fraction of the filter is then analyzed for POC. The problem is that POC concentrations measured from bottle samples are often 5-10 times the concentrations measured using in-situ pumps. A mechanism which we could use to determine which of the above POC sampling methods is more accurate comes from DOC measurements (dissolved organic carbon). DOC is measured by filtering a small water sample (10-200 mls) and measuring the total organic carbon (TOC) in the water, not carbon on a filter. In areas where POC concentrations are high, one can take two small volumes of water, and analyze one filtered and one unfiltered for total organic carbon (TOC). The difference between the two samples should be POC. The bottle POC data yields somewhat higher values than the TOC method, but the in-situ pump POC values are much lower than either of the other methods. At low POC concentrations (<2 µM/l) bottle POC values may be elevated due to adsorption of DOC on filters, but at concentrations >2 µM/l the contribution of adsorption is negligible. There are few studies where any two of these methods have been tested simultaneously, but optical data (using beam transmissometers) have been collected simultaneously with the different POC sampling methods and POC/beam Cp ratios can be used to compare results from different sites. We will present some of these comparisons and discuss reasons why the in-situ pump values are most likely too low. The primary reason seems to be the difference in filtration pressures with in-situ pumps (~1 atm) versus bottle samples (~ 0.25 atm) and secondarily the nominal pore size of quartz filters used with in-situ pumps (~1 µm quartz) versus filters used with bottle samples (0.7 µm GFF). It is crucial to understand which of the above methods is correct when making carbon budgets, especially as we employ more optical methods to quantify carbon pools more rapidly.

NSF-SMP 9986726 Project web & CD / Updated: 12-Apr-2004 / 2004 A. Mishonov