Estimates of primary production by remote sensing in the Arctic Ocean: Assessment of accuracy with passive and active sensors

Abstract

The Arctic Ocean, including its regional shelf seas, is assumed to play an important role in the global carbon cycle. However, the true magnitude of annual production is unknown, as in situ data are sparse in time and space. Remote sensing technology has the potential to provide large scale estimates of phytoplankton biomass at much higher frequency and spatial coverage than shipboard observations in this remote region. Subsurface peaks in both biomass and primary production (PP), which are the characteristics of the Arctic, are shown to limit the reliability of ocean color based integrated PP (IPP) models in the Chukchi Sea. Here we report that the retrievals of IPP from remotely sensed ocean color data were accurate only when limited to 1.2 optical depths, which severely constrains the utility of ocean color remote sensing for the assessment of Arctic Ocean dynamics. Active sensors such as LIDAR, can, in combination with passive ocean color, dramatically improve our ability to estimate IPP for the Arctic. IPP retrievals were improved to within a factor of 2–3 of the measured values, when the vertical distribution of Chl a was determined to a resolution of 1 m using modeled LIDAR retrievals of the beam attenuation coefficient. This was far better than models using only passive ocean color. The instrument specifications of the current NASA spaceborne LIDAR (CALIOP) allow for the retrieval of K d at a depth resolution of 23 m. Even with this constraint, however, the accuracy of the modeled IPP was improved over passive ocean color retrievals to approximately a factor of 3. The Arctic is a perfect location to merge ocean color and LIDAR measurements as the polar orbit of CALIOP provides complete grid coverage of the area every 8 days, crossing the horizontal gradients in Chl a already known to exist from passive ocean color observations.