1. Unusual blooms in the Equatorial Pacific in 1998 : A model point of view Menkes C. E., Gorgues T., Alvain S., Aumont O., Bopp L., Moulin C., Dandonneau Y., Cravatte S., Vialard J. (LODYC, LSCE, Paris) SeaWifs, Ryan et al., 2002

2. The Strong 1998 La Nina return produced very strong and unusual blooms (Chavez et al., 1999, Ryan et al., 2002). What type of species were dominant in bloom 2 ? Why ? What was the exact role of Tropical Instability Waves Did the major bloom really propagate eastward ? To address these question we use a coupled dynamical- biogeochemical model and a new method for deriving phytoplankton species from SeaWifs (Alvain et al., 2004).

3. Coupled dynamical-biogeochemical Model OPA Dynamics 1992-2000 Sponge layer 30S and 30N 7-day ERS-TAO stress/1-day NCEP fluxes Model OPA vs TAO currents. Correct structure but bias at 110°W (Lengaigne et al., 2003, Vialard et al., 2002) Coupled on-line to: PISCES (Aumont et al., 2002) Retroaction light-chlorophyll

4. The model allows to distinguish between diatoms or pico-nano planktons. To do so in reality, we use the algorithms of Alvain et al. (2004) that allows to extract from SeaWifs reflectances at different wavelengths, five dominant groups of species: Green Algae Prochlorococcus Cyanobacteria Diatoms Coccolith. bloom Ex: 10 days synthesis in Feb 1998

5. The model allows to distinguish between diatoms or pico-nano planktons. To do so in reality, we use the algorithms of Alvain et al. (2004) briefly summarized: 20 pigments were measured daily (5 observation per day) during 12 GeP&CO cruises from France to New Caledonia between November 1999 and July 2002. Then SeaWiFS daily Level 3 binned (V4) products were used to derive normalized water-leaving radiances (nLw) along Gep&Co cruises. GeP&CO cruises

6. 5/ 11 SPECTRAL VARIABILITY OF nLw From a large set of SeaWiFS Chl-a and nLw ( ), we fitted an “ideal” nLw mod that depends only on [Chl a] : nLw swfs ( ) + Chl-a swfs  nLw mod (Chl-a swfs, ) To evidence the second order spectral variability (SV) of nLw, we divided nLw by this nLw mod : SV ( ) = nLw obs ( ) / nLw mod (Chl-a swfs, ) wavelength SV wavelength nLw Twenty homogeneous classes of ‘Spectral Signature’ were identified

7. 9/ 11 SV CLASSES & PHYTOPLANKTON GROUPS Mean Spectral Signature for the 5 groups Five phytoplankton groups can be identified from our 20 Sp.Sign. Classes by using Chemtax (Mackey et al., 1997) to go from pigments to groups. - Haptophytes (classes 1-4) -Prochlorococcus (classes 5-14) -Cyanobacteria (classes 15-18) -Diatoms (classe 19) -Coccolithophorids bloom (classe 20) wavelength SV Haptophytes Prochlorococcus Cyanobacteria Diatoms Coccolithophorids

8. DATA(1998-2000): MODEL (1998-2000): SST AVHRR SST-1C SeawifsModel Chl The patterns of SST and Chl are reasonable but upwelling too weak

9. EQ. SST CHLORO SEA LEV. ANO. DATA MODEL AUG-SEP TIW=BREAK TIW JUL JUN-JUL LA NINA MAY EL NINO 1998 1999 1998 1999

10. The dominant ( more than 70%) group during the 1998 blooms are diatoms in both model and SeaWifs at the equator. Only in 1998 are diatoms dominant at the equator DATAMODEL Again, the eastward propagation in 1998 is not obvious and the diatom blooms (july and sep-aug) seem separated.

11. 2°N SST CHLORO SEA LEV. ANO. DATA MODEL At 2°N (and south), the signal takes the form of TIWs In the model, TIWs are weaker than observed for the first bloom and stronger than observed at the end of 1998 TIW

12. The blooms of dominant diatoms correspond to strong outbursts of Iron in the surface layers In the model, the b) Aug-Sep bloom is not associated to EUC surfacing but to enhanced vertical diffusion and upwelling of Fe due to the trades. b b b a )bloom is modulated by an irruption of TIW signal that propagates to the west a a a The first a) Jun-July bloom corresponds to a patch of eastward current where the eastward EUC surfaces near 120°W (Picaut et al., 2002, Izumo et al., 2002)and strong westward current is to the east

13. U at EQ Fe at EQCHL at EQ 23°C 0m 50m 100m 200m Model resembles the data with TIW undulations Strong upwelling at 130°W: Iron input and bloom EUC surfaces. SSTTMI CHLORO MODELSEAWIFS BLOOM 1: end of June 130°W Strong shear of zonal currents  generation of very strong TIWs east of 130°W, near 110°W

14. U at EQFe at EQCHL at EQ 23°C SSTTMI CHLORO MODELSEAWIFS BLOOM 2: end of August Model chl is too Strong west of 130°W Model resembles the data with TIW undulations and Separation near 130°W This separation corresponds to the westward moving TIW, now near 130°W that advects chl-poor and warm waters from the north.

15. The coupled OPA-PISCES model reproduces some of the major features of the 1998 unusual blooms. These blooms are dominated by diatoms and 1998 is the only period when this is true in both model and observations. The first bloom in June-July is associated with strong upwelling near 130°W linked to EUC surfacing with north-south undulations due to instability waves generated by strong current shear. This shear generates near 110°W an extremely strong TIW signal that advects nutrient-poor and warm waters from the north and that reduces the bloom to the west as it propagates. Then a second bloom appears to the east of 120°W, confined to the eastern Pacific and is not obviously propagating eastward either in the data or in the model.