Diagnosing coastal ocean CO2 interannual variability from a 40 year hydrographic time series station off the east coast of Australia

Abstract

Advancing our understanding of the ocean's role in modulating atmospheric carbon dioxide interannual variability is important but limited by the lack of temporal ocean CO2 observations throughout most of the world's oceans. In particular, the role of the coastal ocean for interannual variability of CO2 is unknown, but expected to be large given the dynamics of land to ocean CO2 exchanges. Two coastal hydrographic time series stations at Port Hacking off Sydney, Australia, have collected and measured hydrographic properties (temperature, salinity, nitrate, phosphate, and silicate) since 1942. Here I examine the potential to use empirical relationships derived from nearby discrete CO2 measurements to diagnose CO2 interannual variability at the Port Hacking time series station. I test the approach by predicting interannual variability of CO2 at the Bermuda Time Series Station (BATS) in the North Atlantic between 1988 and 2008 and show the method captures 78% of the observed interannual variability found at BATS. I further use oxygen as a tracer to diagnose the length scales of variability in the southwest Pacific to provide insight into the applicability of the method. After applying the approach at Port Hacking, significant CO2 interannual variability was found (up to ± 30 ppm) that was largely driven by local regional variability in nutrients (nitrate) and temperature with little link to larger modes of variability (SAM or ENSO). Oceanic CO2 was diagnosed to be almost always lower than atmospheric CO2 levels (ΔpCO2 ∼ 10–55 μatm), indicating a quasi‐permanent coastal ocean CO2 sink during winter off Sydney. Direct pCO2 measurements taken along the NSW coastline in April and August 2008 confirm the empirical predictions and a wider CO2 sink along Australia's east coast. If extrapolating the Port Hacking CO2 results to the New South Wales coastline, the annual coastal CO2 sink would be about −0.05 Tg C/yr. Both the empirical predictions and direct measurements of CO2 presented here along the Australian east coast support recent reviews suggesting continental shelves act as a significant CO2 sink, at least during autumn and winter.