The third primary production algorithm round robin (PPARR3) compares output from 24 models that estimate depthintegrated primary production from satellite measurements of ocean color, as well as seven general circulation models (GCMs) coupled with ecosystem or biogeochemical models. Here we compare the global primary production fields corresponding to eight months of 1998 and 1999 as estimated from common input fields of photosynthetically-available radiation (PAR), sea-surface temperature (SST), mixed-layer depth, and chlorophyll concentration. We also quantify the sensitivity of the ocean-color-based models to perturbations in their input variables. The pair-wise correlation between ocean-color models was used to cluster them into groups or related output, which reflect the regions and environmental conditions under which they respond differently. The groups do not follow model complexity with regards to wavelength or depth dependence, though they are related to the manner in which temperature is used to parameterize photosynthesis. Global average PP varies by a factor of two between models. The models diverged the most for the Southern Ocean, SST under 10 C, and chlorophyll concentration exceeding 1 mg Chl m À3 . Based on the conditions under which the model results diverge most, we conclude that current ocean-color-based models are challenged by high-nutrient low-chlorophyll conditions, and extreme temperatures or chlorophyll concentrations. The GCM-based models predict comparable primary production to those based on ocean color: they estimate higher values in the Southern Ocean, at low SST, and in the equatorial band, while they estimate lower values in eutrophic regions (probably because the area of high chlorophyll concentrations is smaller in the GCMs). Further progress in primary production modeling requires improved understanding of the effect of temperature on photosynthesis and better parameterization of the maximum photosynthetic rate. r

Depth-integrated primary productivity (PP) estimates obtained from satellite ocean color-based models (SatPPMs) and those generated from biogeochemical ocean general circulation models (BCGCMs) represent a key resource for biogeochemical and ecological studies at global as well as regional scales. Calibration and validation of these PP models are not straightforward, however, and comparative studies show large differences between model estimates. The goal of this paper is to compare PP estimates obtained from 30 different models (21 SatPPMs and 9 BOGCMs) to a tropical Pacific PP database consisting of similar to 1000 C-14 measurements spanning more than a decade (1983-1996). Primary findings include: skill varied significantly between models, but performance was not a function of model complexity or type (i.e. SatPPM vs. BOGCM); nearly all models underestimated the observed variance of PR specifically yielding too few low PP (< 0.2 g Cm-2 d(-1)) values; more than half of the total root-mean-squared model-data differences associated with the satellite-based PP models might be accounted for by uncertainties in the input variables and/or the PP data; and the tropical Pacific database captures a broad scale shift from low biomassnormalized productivity in the 1980s to higher biomass-normalized productivity in the 1990s, which was not successfully captured by any of the models. This latter result suggests that interdecadal and global changes will be a significant challenge for both SatPPMs and BOGCMs. Finally, average root-mean-squared differences between in situ PP data on the equator at 140 degrees W and PP estimates from the satellite-based productivity models were 58% lower than analogous values computed in a previous PP model comparison 6 years ago. The success of these types of comparison exercises is illustrated by the continual modification and improvement of the participating models and the resulting increase in model skill. (C) 2008 Elsevier BY. All rights reserved

[1] The performance of 36 models (22 ocean color models and 14 biogeochemical ocean circulation models (BOGCMs)) that estimate depth-integrated marine net primary productivity (NPP) was assessed by comparing their output to in situ 14 C data at the Bermuda Atlantic Time series Study (BATS) and the Hawaii Ocean Time series (HOT) over nearly two decades. Specifically, skill was assessed based on the models' ability to estimate the observed mean, variability, and trends of NPP. At both sites, more than 90% of the models underestimated mean NPP, with the average bias of the BOGCMs being nearly twice that of the ocean color models. However, the difference in overall skill GLOBAL BIOGEOCHEMICAL CYCLES, VOL. 24, GB3020, doi:10.1029/2009GB003655, 2010 GB30201 of 21 between the best BOGCM and the best ocean color model at each site was not significant. Between 1989 and 2007, in situ NPP at BATS and HOT increased by an average of nearly 2% per year and was positively correlated to the North Pacific Gyre Oscillation index. The majority of ocean color models produced in situ NPP trends that were closer to the observed trends when chlorophyll-a was derived from high-performance liquid chromatography (HPLC), rather than fluorometric or SeaWiFS data. However, this was a function of time such that average trend magnitude was more accurately estimated over longer time periods. Among BOGCMs, only two individual models successfully produced an increasing NPP trend (one model at each site). We caution against the use of models to assess multiannual changes in NPP over short time periods. Ocean color model estimates of NPP trends could improve if more high quality HPLC chlorophyll-a time series were available.Citation: Saba, V. S., et al. (2010), Challenges of modeling depth-integrated marine primary productivity over multiple decades: A case study at BATS and HOT, Global Biogeochem. Cycles, 24, GB3020,

Nearly half of the earth’s photosynthetically fixed carbon derives from the oceans. To determine global and region specific rates, we rely on models that estimate marine net primary productivity (NPP) thus it is essential that these models are evaluated to determine their accuracy. Here we assessed the skill of 21 ocean color models by comparing their estimates of depth-integrated NPP to 1156 in situ 14C measurements encompassing ten marine regions including the Sargasso Sea, pelagic North Atlantic, coastal Northeast Atlantic, Black Sea, Mediterranean Sea, Arabian Sea, sub-tropical North Pacific, Ross Sea, West Antarctic Peninsula, and the Antarctic Polar Frontal Zone. Average model skill, as determined by root-mean square difference calculations, was lowest in the Black and Mediterranean Seas, highest in the pelagic North Atlantic and the Antarctic Polar Frontal Zone, and intermediate in the other six regions. The maximum fraction of model skill that may be attributable to uncertainties in both the input variables and in situ NPP measurements was nearly 72%. On average, the simplest depth/wavelength integrated models performed no worse than the more complex depth/wavelength resolved models. Ocean color models were not highly challenged in extreme conditions of surface chlorophyll-a and sea surface temperature, nor in high-nitrate low-chlorophyll waters. Water column depth was the primary influence on ocean color model performance such that average skill was significantly higher at depths greater than 250m, suggesting that ocean color models are more challenged in Case-2 waters (coastal) than in Case-1 (pelagic) waters. Given that in situ chlorophyll-a data was used as input data, algorithm improvement is required to eliminate the poor performance of ocean color NPP models in Case-2 waters that are close to coastlines. Finally, ocean color chlorophyll-a algorithms are challenged by optically complex Case-2 waters, thus using satellite-derived chlorophyll-a to estimate NPP in coastal areas would likely further reduce the skill of ocean color models

Nearly half of the earth's photosynthetically fixed carbon derives from the oceans. To determine global and region specific rates, we rely on models that estimate marine net primary productivity (NPP) thus it is essential that these models are evaluated to determine their accuracy. Here we assessed the skill of 21 ocean color models by comparing their estimates of depth-integrated NPP to 1156 in situ ¹⁴C measurements encompassing ten marine regions including the Sargasso Sea, pelagic North Atlantic, coastal Northeast Atlantic, Black Sea, Mediterranean Sea, Arabian Sea, subtropical North Pacific, Ross Sea, West Antarctic Peninsula, and the Antarctic Polar Frontal Zone. Average model skill, as determined by root-mean square difference calculations, was lowest in the Black and Mediterranean Seas, highest in the pelagic North Atlantic and the Antarctic Polar Frontal Zone, and intermediate in the other six regions. The maximum fraction of model skill that may be attributable to uncertainties in both the input variables and in situ NPP measurements, was nearly 72%. Contrary to prior studies, ocean color models were not highly challenged in extreme conditions of surface chlorophyll-<i>a</i> and sea surface temperature, nor in high-nitrate low-chlorophyll waters. On average, the simplest depth/wavelength integrated models performed no worse than the more complex depth/wavelength resolved models. Water column depth (distance to coastlines) was the primary influence on ocean color model performance such that average skill was significantly higher at depths greater than 250 m, suggesting that ocean color models are more challenged in Case-2 waters (coastal) than in Case-1 (pelagic) waters. Given that in situ chlorophyll-<i>a</i> data was used as input data, algorithm improvement is required to eliminate the poor performance of ocean color models in Case-2 waters that are close to coastlines. Finally, ocean color chlorophyll-<i>a</i> algorithms are challenged by optically complex Case-2 waters, thus using satellite-derived chlorophyll-<i>a</i> to estimate NPP in coastal areas would likely further reduce the skill of ocean color models