Accounting for Horizontal Tracer Gradients in Biological Productivity Estimates From Semi‐Lagrangian Platforms

Abstract Marine net community production (NCP), a metric of ecosystem functionality, is often estimated as the residual term in a mass balance equation that aims to describe upper ocean variations in the time series of a chemical tracer. The advent of biogeochemical (BGC) Argo profiling floats equipped with nitrate, pH, and oxygen sensors has enabled such NCP estimation across vast ocean regions. Floats typically drift at 1,000 m depth between profiling from ∼2,000 m to the surface every 10 days, resulting in quasi‐Lagrangian time series that can reflect different upper ocean water masses over time. However, limited information about real‐time horizontal tracer gradients often leads to lateral processes being omitted during tracer budget closure, which can bias the residual‐term NCP estimates. To determine the potential magnitude of such biases, we developed a method to quantify and adjust for the impact of lateral float movement across horizontal tracer gradients using dissolved inorganic carbon (DIC) as our case study. We evaluated the method by extracting artificial float profiles from a depth‐resolved observation‐based DIC product to generate an artificial DIC time series. We then estimated NCP before and after accounting for horizontal gradient effects and compared the results to NCP estimates from an artificial DIC time series extracted at a fixed location along the float trajectory. Testing 10 biogeographical domains with moderate to substantial horizontal DIC gradients, our method significantly improved the precision (by ∼50 to ∼80%) and accuracy (by ∼10 to ∼100%) of regional NCP estimates. This method can be applied to other tracers with multi‐month‐long residence times.