Seasonal and interannual variability of ocean color and composition of phytoplankton communities in the North Atlantic, equatorial Pacific and South Pacific

https://doi.org/10.1016/j.dsr2.2003.07.018Get rights and content

Abstract

Monthly averaged level-3 SeaWiFS chlorophyll (Chl) concentration data from 1998 to 2001 are globally analyzed using Fourier's analysis to determine the main patterns of temporal variability in all parts of the world ocean. In most regions, seasonal variability dominates over interannual variability, and the timing of the yearly bloom generally can be explained by the local cycle of solar energy. The studied period was influenced by the late consequences of the very strong El Niño of 1997–1998. After this major event, the recovery to normal conditions followed different patterns at different locations. At the equator, Chl concentration was abnormally high in 1998, and then decreased, while away from the equator it was low in 1998 and subsequently increased when equatorial upwelled waters spread poleward. This resulted in opposed linear trends with time in these two zones. Other noticeable examples of interannual variability in the open ocean are blooms of Trichodesmium that develop episodically in austral summer in the south-western tropical Pacific, or abnormally high Chl concentration at 5°S in the Indian Ocean after a strong Madden–Julian oscillation. Field data collected quarterly from November 1999 to August 2001, owing to surface sampling from a ship of opportunity, are presented to document the succession of phytoplankton populations that underlie the seasonal cycles of Chl abundance. Indeed, the composition of the phytoplankton dictates the efficiency of the biological carbon pump in the various oceanic provinces. We focus on the North Atlantic, Caribbean Sea, Gulf of Panama, equatorial Pacific, South Pacific Subtropical Gyre and south-western tropical Pacific where field data have been collected. These data are quantitative inventories of pigments (measured by high performance liquid chromatography and spectrofluorometry) and picoplankton abundance (Prochlorococcus, Synechococcus, picoeucaryotes and bacteria). There is a contrast between temperate waters where nanoplankton (as revealed by pigments indexes) dominate during all the year, and tropical waters where picoplankton dominate. The larger microplankton, which make most of the world ocean export production to depth, rarely exceed 20% of the pigment biomass in the offshore waters sampled by these cruises. Most of the time, there are large differences in the phytoplankton composition between cruises made at the same season on two different years.

Introduction

Marine primary production can be assessed directly using flux measurements in the field, such as 14C fixation experiments. However, these measurements are expensive and time-consuming, and world ocean databases contain much more data of chlorophyll a (Chl a) concentration. The later can easily be measured, and it is a key variable in models of photosynthesis. Errors in Chl measurements are sometimes high, and often consist in biases caused by handling artifacts (for instance, too high filtration pressure or poor conditions for preservation of samples) or by the measurement concept itself (Chl b seen as pheophytin a in the fluorescence-acidification technique or in vivo Chl fluorescence taken as a proxy for the Chl a concentration). The other important variables that force primary production are light, temperature and nutrient concentrations, the variability of which are much better known and understood than that of Chl concentration. As a consequence, in the past we have learned more about variability of marine primary production by looking at the distribution of nutrients and coupled physical–biogeochemical models than from the numerous measurements of Chl a concentration made at sea (Sverdrup, 1955; Dutkiewicz et al., 2001).

The first series of satellite sea-color data, provided by the Coastal Zone Color Scanner (CZCS), showed that Chl a concentration could be estimated from space and over sampled, under clear sky conditions. Unfortunately, the CZCS was not programmed for global coverage, and the 1978–1989 data set has large gaps in many regions. Based on these data however, it was possible to identify ecological provinces and to describe their annual cycle of phytoplankton (Longhurst, 1998). One decade later, the SeaWiFS sea-color sensor was launched and is providing data that cover the global ocean at 9-km resolution (1 km for local area coverage mode). This global data set is not affected by cruise or method biases, unlike the in situ data sets. Furthermore, careful calibration of this instrument (Barnes et al., 2001; Eplee et al., 2001) may make it possible to detect long-term evolution of Chl a concentration with time, a point of interest in the context of climate change. Algorithms that convert the signal seen by the satellite into Chl a concentration are improving, and may provide estimates of new geophysical quantities in the future, such as other phytoplankton pigments or particulate organic carbon (Sathyendranath et al., 1994; Stramski et al., 1999; Loisel et al., 2001).

In this work, we used a simple and global analysis of the SeaWiFS data to describe the major patterns of seasonal and interannual variability over the period from January 1998 to December 2001. An additional objective was to estimate how variations in Chl concentration correspond to variations in composition of the phytoplankton. Indeed, the impact of primary production on marine geochemistry strongly depends on the species of phytoplankton that photosynthesize. Well-known examples are biocalcification by the coccolithophorids, which reduces the alkalinity of seawater and thus modifies the dissolved carbonate equilibrium (Robertson et al., 1994), or diazotrophy by the cyanobacteria Trichodesmium that increases the pool of reactive nitrogen through fixation of atmospheric N2 (Capone et al., 1997). Aside from these extremes, the fraction of marine primary production that sinks to depth (export production) depends on the phytoplankton species present: the larger species (diatoms, dinoflagellates) are responsible for massive export of carbon while most of the production by smaller species is rapidly recycled. Attempts to detect some phytoplankton species from space (Brown and Yoder, 1994; Subramaniam et al., 1999) are limited to surface bloom conditions, and knowledge of the distribution of phytoplankton groups is still obtained from oceanographic cruises.

Here, we used field data collected quarterly from a commercial shipping line that spans a wide range of latitude and oceanic conditions, as part of the GeP&CO (Geochemistry, Phytoplankton and Color of the Ocean) and GeP&SIMBAD programs. GeP&CO is a component of the French program PROOF (PROcessus Océaniques et Flux), supported by the Institut National des Sciences de l’Univers, the Institut de Recherche pour le Développement, the Centre National d’Etudes Spatiales and the Institut Français pour l’Exploration de la Mer. GeP&SIMBAD (GeP&CO and SIMBAD) is supported by the Centre National d’Etudes Spatiales.

Section snippets

Data and methods

This study is based on monthly averaged level-3 binned SeaWiFS Chl data issued by the third reprocessing. Data are first averaged on a 0.5° longitude×0.5° latitude grid, and each time series in all grid elements was analyzed using fast Fourier transform (FFT). The grid elements where the surface of the ocean was not seen by SeaWiFS during more than two consecutive months, or more than 12 months over the 1998–2001 period, were removed from the analysis. Missing months were interpolated using

Linear trend

Prior to the FFT analysis, a linear trend p×t, where p is the Chl per month increase, and t is time in months from January 15, 1998 to December 15, 2001, was subtracted from the SeaWiFS monthly Chl concentrations in each 0.5°×0.5° grid elements. The distribution of p in the world ocean is shown in Fig. 2. Rapid increases (greater than 0.004 mg m−3 month−1) in Chl concentration can be seen in some restricted areas. Among these are the Costa Rica Dome region, the plumes of Amazon and Rio de la Plata

Occurrence time of the seasonal chlorophyll maximum

The Chl concentration at the sea surface generally responds to the seasonal cycle of solar energy that strongly impacts the timing and intensity of vertical nutrient flux, and vertical stability. Hence, at 50°S, the peak of biomass occurs in November, December or January, i.e. at the season where light and vertical stability combine to trigger growth of phytoplankton (Fig. 4). At lower latitudes, the Chl maximum occurs earlier, as early as July (i.e. in austral winter) at about 20°S. The

Large-scale ecosystem observations by GeP&CO

Marine primary production can be estimated from Chl satellite data with acceptable accuracy (Morel, 1991; Behrenfeld and Falkowski, 1997). However, what is pertinent in global carbon geochemistry is new production. The fraction of primary production that corresponds to new production, i.e. the f ratio (Dugdale and Goering, 1967), is strongly dependent on the population of phytoplankton. The general consensus is that large diatoms grow on nitrate and rapidly export large amounts of carbon,

Conclusion

This analysis of SeaWiFS data was limited to places where time gaps, caused by clouds, did not exceed two consecutive months and amounted to less than 12 months for the entire 1998–2001 period. This constraint excluded high latitudes where the seasonal bloom of phytoplankton is known to force a major export flux of oceanic carbon to depth. The remaining low- and mid-latitudes represent, however, about 45 of the world ocean. The chlorophyll concentrations collected by SeaWiFS (third processing)

Acknowledgements

We thank Marine Consulting & Contracting and Ms. Alexandra Rickmers in Hamburg who respectively manage and own the container carrier Contship London, and kindly agreed to host scientific observers onboard of this ship. Scientific observers Philippe Gérard, Joël Orempuller and François Baurand ensured high-quality observations at sea. James Murray read the manuscript and improved the English language usage. The comments of two anonymous reviewers helped to improve the manuscript. Overall, thanks

References (43)

  • F Partensky et al.

    Vertical structure of picophytoplankton at different trophic sites of the tropical northeastern Atlantic Ocean

    Deep-Sea Research I

    (1996)
  • J.E Robertson et al.

    The impact of a coccolithophore bloom on oceanic carbon uptake in the northeast Atlantic during summer 1991

    Deep-Sea Research I

    (1994)
  • K Banse et al.

    Satellite observed winter blooms of phytoplankton in the Arabian Sea

    Marine Ecology Progress Series

    (1986)
  • R.A Barnes et al.

    Calibration of SeaWiFS. I. Direct techniques

    Applied Optics

    (2001)
  • M.J Behrenfeld et al.

    Photosynthetic rates derived from satellite-based chlorophyll concentration

    Limnology and Oceanography

    (1997)
  • R.R Bidigare et al.

    Evidence for phytoplankton succession and chromatic adaptation in the Sargasso Sea during spring 1985

    Marine Ecology Progress Series

    (1990)
  • C.W Brown et al.

    Coccolithophorid blooms in the global ocean

    Journal of Geophysical Research

    (1994)
  • D.G Capone et al.

    Trichodesmium, a globally significant marine cyanobacterium

    Science

    (1997)
  • E.J Carpenter et al.

    Nitrogen fixation and uptake of combined nitrogenous nutrients by Oscillatoria (Trichodesmium) thiebautii in the western Sargasso Sea

    Limnology and Oceanography

    (1975)
  • M.-E Carr

    Estimation of potential productivity in Eastern Boundary Currents using remote sensing

    Deep-Sea Research II

    (2002)
  • H Claustre

    The trophic status of various oceanic provinces as revealed by phytoplankton pigment signatures

    Limnology and Oceanography

    (1994)
  • Cited by (0)

    View full text