How does dynamical spatial variability impact 234Th-derived estimates of organic export?

In this study we first evaluate the small-scale spatial variability of particulate export, using a set of synoptic thorium-234 activity observations sampled within a one-degree radius. These data show significant variability of surface thorium activity on scales of the order of 100 km (∼270–550 dpm m−3). This patchiness of export potentially affects the robustness of point observations and our interpretation of them. Motivated by these observations we subsequently couple an explicit model of thorium-234 dynamics to a coupled physical–biogeochemical basin model capable of resolving these small-scales. The model supports the observations in displaying marked thorium variability on spatial scales of the order of 100 km and smaller, with highest values in the regions of large eddy kinetic energy and large primary productivity. The model is also used to quantify the impact of small-scale variability on export estimates. Our model shows that the primary source of error associated with the presence of small-scale spatial variability is related to the standard assumptions of steady state and non-steady state (>40% during bloom condition). The non-steady state method can misinterpret variations due to patchiness in thorium activity as temporal changes and lead to errors larger than those introduced by the simpler steady state approach. We show that the non-steady state approach could improve the flux estimates in some cases if the sampling was conducted in a Lagrangian framework. Undersampling the spatial variability results in further bias (>20%) that can be reduced when the sampling density is increased. Finally, errors due to the dynamical transport of thorium associated with small-scale structures are relatively low (<20%) except in regions of high eddy kinetic energy.

► Small-scale spatial variability of particulate export (20–50%) from two “high resolution” data sets: 234Thorium activity observations within one-degree radius and three sediment traps within 100 km radius.
► Coupled dynamical–biogeochemical–thorium model at small-scale.
► Impact of small-scale transport on thorium estimates of 20–50%.
► Estimation of errors due to undersampling and standard assumption (non-steady state, steady state etc) in conjunction with the presence of small-scale spatial variability.
► Highest variability and source of errors in region of high Eddy Kinetic Energy.