Synthesis of primary production in the Arctic Ocean: I. Surface waters, 1954-2007

The spatial and seasonal magnitude and variability of primary production in the Arctic Ocean (AO) is quantified with a pan-arctic approach. We synthesize estimates of primary production (PP), focusing on surface waters (0-5 m), using complementary methods that emphasize different spatial and temporal scales. These methods include (1) in situ observations of C-14 uptake mostly and possibly some O-2 production reported in units of carbon (in situ PP), (2) remotely sensed primary production (sat-PP), and (3) an empirical algorithm giving net PP as a function of in situ chlorophyll a (in situ Chl-PP). The work presented herein examines historical data for PP collected in surface waters only, as they form the majority of the values of a larger ensemble of PP data collected over >50 years (ARCSS-PP) by many national and international efforts. This extended set of surface and vertically-resolved data will provide pan-Arctic validation of remotely sensed chlorophyll a and PP, an extremely valuable tool in this environment which is so difficult to sample. To this day, PP data in the AO are scarce and have uneven temporal and spatial coverage which, when added to the AO's regional heterogeneity, its strong seasonal changes, and limited access, have made and continue to make obtaining a comprehensive picture of PP in the AO difficult. Daily surface in situ PP averaged 70 and 21 mg C m(-3) d(-1) for spring and summer, respectively, for the ca. 50 year period across the AO. Average daily estimates of in situ PP in surface waters on a pan-Arctic basis were several fold higher with respect to remotely sensed PP (sat-PP) and in situ chlorophyll-derived PP (Chl-PP) in the spring period, likely due to differences in data availability and coverage. Summer daily averages for surface in situ PP and sat-PP were similar and twice as high as in situ Chl-PP. Differences among annual estimates of surface in situ PP, in situ Chl-PP and sat-PP across the Arctic Ocean are presented and discussed. The mode of all three differences was 0 mg C m(-3) y(-1) and the median difference was within +/- 400 mg C m(-3) y(-1). Agreement between annual surface sat-PP and in situ PP was within +/- 1000 mg C m(-3) y(-1) for half of the data. We hope that the entire ARCSS-PP data set, especially if combined with any archiving by the latest IPY effort, will be employed to develop an improved, panarctic-specific PP remote sensing algorithm and to test hypotheses regarding the controls of PP in the AO. Furthermore, the ARCSS-PP is openly available to be used by AO modelers to calibrate, assimilate, validate and compare community ecosystem and biogeochemical numerical models and, when combined with field work, will allow for improved understanding, detection and prediction of long-term biological patterns. (C) 2012 Elsevier Ltd. All rights reserved.