Variation of phytoplankton absorption coefficients in the northern South China Sea during spring and autumn

Abstract. We examined the temporal and spatial variabilities of phytoplankton absorption coefficients (α_phλ)) and their relationships with physical processes in the northern South China Sea from two cruise surveys during spring (May 2001) and late autumn (November 2002). A large river plume induced by heavy precipitation in May stimulated a phytoplankton bloom on the inner shelf, causing significant changes in the surface water in α_ph values and B/R ratios (α_ph(440)/α_ph(675)). This was consistent with the observed one order of magnitude elevation of chlorophyll α and a shift from a pico/nano dominated phytoplankton community to one dominated by micro-algae. At the seasonal level, enhanced vertical mixing due to strengthened northeast monsoon in November has been observed to result in higher surface α_ph(675) (0.002–0.006 m⁻¹ higher) and less pronounced subsurface maximum on the outer shelf/slope in November as compared that in May. Measurements of α_ph and B/R ratios from three transects in November revealed a highest surface α_ph(675) immediately outside the mouth of the Pearl River Estuary, whereas lower α_ph(675) and higher B/R ratios were featured in the outer shelf/slope waters, demonstrating the respective influence of the Pearl River plume and the oligotrophic nature of South China Sea water. The difference in spectral shapes of phytoplankton absorption (measured by B/R ratios and bathochromic shifts) on these three transects infers that picoprocaryotes are the major component of the phytoplankton community on the outer shelf/slope rather than on the inner shelf. In addition, a regional tuning of the phytoplankton absorption spectral model (Carder et al., 1999) demonstrated a greater spatial variation than seasonal variation in the lead parameter a0(λ). These results suggest that phytoplankton absorption properties in a coastal region such as the northern South China Sea are complex and region-based parameterization is mandatory in order for remote sensing algorithms.