1. Impact of river sources of P, Si and Fe on coastal biogeochemistry da Cunha 1, L.C., Le Quéré 1, C., Buitenhuis 1, E.T., Giraud 1, X. & Ludwig 2, W. 1 Max-Planck-Institut for Biogeochemistry, POSTFACH 10 01 64, Jena 07701, Germany (lcotrim@bgc- jena.mpg.de) 2 CEFREM UMR 5110, Université de Perpignan 52 Av. P. Alduy, Perpignan 66860 CEDEX, France. HIGHLIGHTS: We assess the role of river sources of P, Si, and Fe on coastal biogeochemistry and air-sea fluxes of CO 2 using a global ocean biogeochemistry model and observations. River inputs of P, Si, and organic carbon are taken from a global database. Gross dissolved river Fe inputs are ~32 Gmol Fe yr-1. We consider that between 99-80% of this dissolved iron is lost in estuaries, and test Fe input of 0.3-6.4 Gmol Fe yr-1. Results show that increased nutrient availability in coastal oceans leads to an increase in the coastal zone of 5-12% in export production 4.8-7.8% in primary production 20-28% in primary production of diatoms, and 1-5% in f-ratio. The increased efficiency of the biological pump leads to a reduction of the sea-to-air CO 2 flux by 21-53% in the coastal region, and the global oxygen-deficient areas extended from 8.1 to 9.6 10 6 km 2 because of increasing biological activity. P is the most limiting nutrient in the coastal zone of our model whereas Fe has the largest impact on the relative abundance of diatoms and consequently on carbon export and CO 2 fluxes. Based on elemental budgets from observations and on model results, nutrient inputs from rivers can sustain much of coastal export production: 15- 26% for dissolved inorganic P, 80-100% for dissolved silica, and 24-500% for dissolved iron. 1. FROM RIVERS TO THE COASTAL OCEAN Rivers transport terrestrial organic carbon and nutrients, both of natural and anthropogenic origin. Their material balance in the coastal zone determines to a large extent its role as either a source or sink of CO 2 relative to the atmosphere. We use different river- nutrient loading scenarios in a global ocean biogeochemistry model, PISCES (Table 1). 2. PISCES BIOGEOCHEMISTRY MODEL DESIGN The PISCES biogeochemistry model includes the representation of diatoms and nanophytoplankton, meso- and microzooplankton, and co-limitation by light and by P, Si and Fe [Aumont et al., 2003; Bopp et al., 2002]. It is embedded in the OPA-ORCA global circulation model [Madec et al., 1998; Madec and Imbard, 1996]. The horizontal resolution of OPA is on average 2° by 1.5° with a higher resolution at the equator (0.5°). The model has 30 vertical layers among which 10 are located in the top hundred metres of the ocean. Results show that increased nutrient availability in coastal oceans leads to an increase in the coastal zone of 5-12% in export production and 4.8-8.8% in primary production (Fig. 3). When only PO 4 and Si are added (simulations NUT and 12BDIP) we observe an enhancement in phytoplankton production, and coastal surface waters become impoverished in dissolved iron. When riverine iron is added (simulations NUTFE and 99PFE), there is a “fertilizing” effect on diatoms, thus increasing the export of organic matter and rising the e-ratio values. The increase in coastal primary and export production lead to an increase in Si and P uptake. Diatom coastal primary production increase up to 21%, and nanophytoplankton production increase up to 3.5%, creating large areas impoverished in PO 4, mainly in high productive and upwelling zones (Fig. 2). 4. NUTRIENT AVAILABILITY AND PRODUCTION Fig. 3 – Changes in modelled coastal primary and export production according to riverine DIP and dFe inputs The increased efficiency of the biological pump leads to a reduction of the sea-to-air CO 2 flux by 21-53% in the coastal region (Fig. 4). Fig. 4 – Difference between model scenarios coastal sea-to-air CO 2 flux (year 20). Global oxygen-deficient areas (O 2 < 25 µM) extend from 8.1 to 9.6 10 6 km 2 because of increasing biological activity. The ”extended” oxygen-deficient areas correspond to the Eastern Tropical Pacific, Arabian Sea, Bay of Bengal, and the Gulf of Guinea. The last two ecosystems are under influence of large river nutrient inputs, the Brahmaputra and Ganges rivers in the Bay of Bengal, and the Volta, Niger and Congo rivers in the Gulf of Guinea. Table 2 is a summary of the changes in coastal oceans after adding river nutrient fluxes to the PISCES model. Coastal export and primary production are enhanced by increasing PO 4 and iron. E-ratio value is higher when riverine POC is added, indicating a partial “trapping” of terrestrial organic matter. Phosphorus is the most limiting nutrient in the coastal zone. Combined river and resuspension P can sustain at maximum 70% of the coastal export production, while riverine Si and Fe inputs represent 80-100% (Si), and 24-500% of coastal export production. Fig. 2 – Difference between model scenarios and the reference run in surface PO 4 concentration (µM) after 20 years. NUT - NORIVER 12BDIP - NORIVER 99PFE - NORIVER Mineralisation of riverine organic matter (POC and DOC) is an additional source of PO 4 (C2SIPFE). When PO 4 inputs increase combined to riverine iron (simulations C2SIPFE and 12BFE), the “fertilizing” effect is improved, and we observe the highest increase in coastal primary (6.7-8.7%) and export production (12%). An increase of 5% in the e-ratio in simulation C2SIPFE is probably due to the export additional POC from rivers, and more available Si, favouring diatom production. Riverine PO 4 and Fe effects in coastal primary and export production are plotted in figure 3. Table 2 – PISCES simulations and their changes in coastal biogeochemistry ACKNOWLEDGMENTS: We thank W. Ludwig for making their data available to us, O. Aumont for letting us use the PISCES model, and K. Rodgers for his help with the OPA simulations. BIBLIOGRAPHY: Smith, S. V.; Swaney, D.; Talaue-McManus, L.; Bartley, J. D.; Sandhei, P. T.; McLaughlin, C. J.; Dupra, V. C.; Crossland, C. J.; Buddemeier, R. W.; Maxwell, B. A.; Wulff, F. Bioscience 2003, 53, 235-245. Bopp, L.; Le Quéré, C.; Heimann, M.; Manning, A. C.; Monfray, P. Global Biogeochemical Cycles 2002, 16, doi:10.1029/2001GB1001. Aumont, O.; Maier-Reimer, E.; Blain, B.; Monfray, P. Global Biogeochemical Cycles 2003, 17. Madec, G.; Delecluze, P.; Imbard, M.; Levy, C.; IPSL: Paris, 1998, pp 91. Madec, G.; Imbard, M. Clim. Dyn. 1996, 12, 381-388. Ludwig, W.; Probst, J.; Kempe, S. Global Biogeochemical Cycles 1996, 10, 23. Ludwig, W.; Probst, J. L. American Journal of Science 1998, 298, 265. Martin, J.-M.; Whitfield, M. In Trace metals in sea water; Wong, C. S., Boyle, E., Bruland, K. W., Burton, J. D., Goldberg, E. D., Eds.; Plenum: New York, 1983, pp 265-296. Martin, J.-M.; Meybeck, M. Marine Chemistry 1979, 7, 173-206. Chester, R. Marine Geochemistry; Unwin Hyman: London, 1990. Treguer, P.; Nelson, D. M.; Van Bennekom, A. J.; DeMaster, D. J.; Leynaert, A.; Queguiner, B. Science 1995, 268, 375. Fig. 1 – Riverine inputs of nutrients to the coastal ocean in mol s -1. a) dissolved iron, NET input considering 80% of loss to the particulate phase in estuaries; b) dissolved inorganic phosphorus, calculated accoring to [Smith et al., 2003] a) b)