In order to more fully understand the processes involved in the production, accumulation and export of dissolved organic matter (DOM) in the Equatorial Pacific Ocean, we (Archer et al. (1997) developed and employed a three-dimensional upper ocean general circulation model. This model avoids the biases associated with the simpler 1-D (Feely et al., 1995) and 2-D (Peltzer & Hayward, 1996; Hansell et al., 1997) models. The measurement of dissolved organic carbon (DOC) and its spatial and temporal variation in the Equatorial Pacific Ocean are described elsewhere in this web-site.
A new modelling approach was developed in order to estimate the timescale of semilabile DOC turnover in equatorial Pacific surface waters. The approach is to balance the unknown timescale for DOC turnover against the relatively well known surface to deep water exchange timescale generated by the flow field of a general circulation model. We assume a priori, and verify the assumption with model results, that the ~80 µM C concentration of total organic carbon (TOC) measured in oligotrophic surface waters is close to a biochemical steady state where production is balanced by consumption. this approximation ignores the seasonal cycle, which may dominate DOM dynamics in subtropical waters (Carlson et al., 1994) but which we assume is of secondary importance to the dynamics of DOM near the equator. Deep waters are assumed to have no semilabile DOM.
The simplest kinectic rate laws for semilabile DOM cycling (that allow for the existence of a steady state) are a constant (zero-order) production confined to the euphotic zone combined with a first-order (in DOC) consumption throughout the water column. Of course, both processes are biologically mediated and certain to be spatially variable. Previous modelling studies have accomodated this by directly linking the rate of DOC production and accumulation to the production of POC (Bacastow & Maeir-Reimer, 1991; Najjar et al., 1992) thereby coupling the DOC production rate to the supply of nutrients in the euphotic zone by ocean circulation. To the extent that our simplification of the problem to spatially uniform kinetics reproduces the observed distribution of DOC in the equatorial Pacific, our model had the advantage of clarity and simple intuitive application to the real ocean, independent of other model parameters.
The effect of fluid flow was simulated using weekly flow fields from a high resolution primitive equation model of the equatorial Pacific ocean (Toggweiler & Carson, 1995). The circulation was driven by monthly mean climatological winds and monthly mean surface heat fluxes as specified by Philander et al. (1987). Grid spacing was 0.5° in latitude and 2° in longitude at the equator, and 2° latitude at the model domain limits of 36°S and 49°N. The advection mthod for DOC is based on the normal fluxes of piece wise constant tracer field across cell boundaries (simple upwinding). The routine was run with a "monotonized centered" limiter to prevent the formation of spurious maxima and minima in the tracer field. The model was initialized with zero concentration of semilabile DOC and run for 2 years, with the first year intended for spin-up (over a seasonal cycle and with a duration which is long relative to the response time of DOC) and the second year for presentation and analysis. The final state from this run was used as an initial condition for a 2 year simulation using various time constants for the production of DOC.
DOC concentrations at the model sea surface are shown in the mpeg-videos below. Three clips are presented corresponding to the three production time-scales of 30, 60 and 120 days. In these videos, the upper panel shows concentration (green = 60 µM C grading to red = 80 µM C) while the lower panel shows production rates (red = fastest; blue = slowest). In addition, DOC concentrations for vertical sections of the model DOC field corresponding to the 1992 transects along 140°W for the three production time-scales. In these videos, a single panel is shown (South is on the left, North is to the right), and DOC concentration ranges from blue (40 µM C) to red (80 µM C). The horizontal scale is 5° of latitude per division and 100 m per vertical division. Each video animation is an assemblage of 48 images (12 x 4 weekly fields) comprising one annual cycle.
All of the model runs have minimum sea surface DOC concentrations at the equator, caused by the exposure of low-DOC thermocline waters to the sea surface in the equatorial divergence, as observed in the data. Poleward of the equator, sea surface DOC increases because these waters have been at the sea surface for a long time relative to the time-scale of DOC production. The minimum DOC concentration at the equator and the meridonal width of the minimum are both functions of the timescale for DOC production: faster timescales generating a shallower and more narrow equatorial concentration minimum.
When sea surface concentrations from the model are compared across the equator with the data, the shape and magnitude of the observed equatorial minimum appears to be best fit by the model using a timescale of 60 days, and bracketed by 30 and 120 days. Most of the net production of DOC is centered right at the equator, where the divergence of surface currents brings cold dense nutrient rich water to the sea surface. The advective rates of DOC export from the equatorial zone calculated using time constants between 30 and 120 days average 4-8 mmol C per m²-day.
We are grateful to Robbie Toggweiler and Steve Carson for access to their model flow fields and extensive consultation. We also acknowledge helpful discussion with Jorge Sarmiento, Paul Quay and Parker MacCready.