Project: Primary Production Model Intercomparison: Primary Production Algorithm Round Robin 4

The goal of our proposal is to evaluate algorithms which estimate primary
production (PP models) from ocean color measured by space-borne sensors. This
project aims to continue a project that is nearing completion, originally funded
under NASA's Carbon Cycle Science Program, the Primary Production Algorithm
Round Robin 3 (PPARR3). We found that greater variability between modeled
estimates of primary production occurred at low sea-surface temperature (SST) and
low or high chlorophyll concentrations. A comparison to in situ data from the
tropical Pacific revealed that the participating models consistently underestimated
the variance of depth-integrated primary production (PP). Model performance was
independent of traditional groupings of model structure: some of the simplest
(depth- and wavelength-integrated) models performed as well as some of the most
complex (depth- and wavelength-resolved.) The significant decrease in root-meansquare
error of participating models in the equatorial Pacific (by 62% on the
equator and 35% off the equator) between PPARR2 and PPARR3 is testimony of the
success of the PPARR exercises.
We propose to expand the PPARR3 exercise by comparing participating PP
models with in situ primary production data acquired during process studies and
time series sites associated with the Joint Global Ocean Flux Study (JGOFS), data
from the Southern Ocean, including a database from the Antarctic Peninsula, and a
compilation of coastal measurements. JGOFS process study and time series data
provide high quality measurements of biogeochemically active regions and
oligotrophic subtropical gyres respectively. The Southern Ocean presents a daunting
challenge for space-based primary production models, as common temperature
dependent functions fail at low temperatures in addition to the challenges of
determining the chlorophyll concentration at low sun angles. High biomass and the
presence of other optically active substances which absorb and reflect light make
primary production particularly difficult to model adequately in the coastal ocean.
By addressing these focus regions, we will evaluate model performance for different
environmental conditions. As in PPARR3 we anticipate that model developers will
use these comparisons to refine and reformulate their models.