Annual phytoplankton blooms are observed around most sub-Antarctic islands during austral spring and summer, but are absent in the surrounding high-nutrient low-chlorophyll (HNLC) ocean. The CROZEX study (2004/2005) tested the hypothesis that annual blooms occurring immediately north of the Crozet islands in the Polar Frontal Zone (PFZ) develop because of natural iron fertilisation, while to their south in HNLC waters, there is insufficient iron (Fe) to promote blooms. Size-fractionated nitrogen uptake (ρN) experiments using 15N-NO3-, NH4+ and urea addressed three major goals. Firstly, measurements of NO3- uptake (ρNO3-) aimed to establish whether ρNO3- responds to natural Fe fertilisation. Secondly, we compared regional ρNO3- rates, hypothesising that in Fe-fertilised regions, ρNO3- should exceed that in HNLC regions. Thirdly, by using satellite imagery, we extrapolated ρNO3- measurements made during a declining bloom to reconstruct seasonal ρNO3- by the spring bloom. Finally, we estimated the 'new' Fe demand required to support ρNO3-, comparing this with estimated Fe fluxes. Diatoms and colonial Phaeocystis dominated phytoplankton communities north of the islands, while to the south, Phaeocystis was absent. Total ρN was elevated north of the islands (∼400 μmol m-2 d-1) relative to south of the islands (∼250 μmol m-2 d-1). Nitrate uptake showed a clear response to Fe fertilisation, exhibiting a strong north (∼198 μmol m-2 d-1) to south (∼74 μmol m-2 d-1) gradient, while neither ρNH4+ nor ρurea showed such significant latitudinal gradients. The N-S integrated f-ratio gradient was 0.47-0.28 while specific N uptake (VNO3 d-1) rates were significantly higher in the Fe-fertilised region relative to those in the southern HNLC region. The potential for NH4+ inhibition of ρNO3- did not appear to be significant. High PON:chl-a ratios combined with relatively low 14C:15N uptake ratios suggested that most phytoplankton were relatively chlorotic and carbon stressed, with the exception of those growing actively within a cyclonic eddy where neither Fe nor light appeared to be limiting. Size-fractionated ρNO3- and f-ratios exhibited a complex response to NH4+ and Fe availability, with f-ratios in the >20-μm fraction being low (∼0.3) in the HNLC region, but significantly higher (∼0.7) in a localised diatom-dominated bloom in the northern Fe-fertilised region. In contrast, f-ratios in the <2-μm size class were similar everywhere (∼0.44), indicative of Fe-limitation for large-celled diatoms in the southern HNLC region. As a result of Fe-regulated ρNO3-, new production showed a N-S gradient of ∼24 to ∼15 mmol C m-2 d-1, very similar to carbon export determined from NO3- 'draw-down' and from 234Th measurements. The estimated DFe demand required to support seasonal ρNO3- in the northern region, based on conservative cellular Fe:N quotas, required surface (to 100 m) pre spring-bloom DFe concentrations of ∼0.75 nmol l-1. Our results support the hypothesis that phytoplankton blooms north of the islands are stimulated by natural Fe fertilisation, with a direct impact on ρNO3-, particularly for larger cells, resulting in higher new production rates relative to those from the Fe-limited HNLC region south of the Crozet islands. © 2007 Elsevier Ltd. All rights reserved.
