The Benguela ecosystem in the south-eastern Atlantic Ocean extends from southern Angola (~17��S) to the southern tip of Africa near Cape Town (34��S), and is bounded at both extremities by the warm water of the Angola Current in the north and the Agulhas Current in the south (Shannon and Nelson 1996, Shillington 1998). The system is dominated by coastal upwelling, driven by longshore wind flow, which results in cold temperatures and high productivity (Bakun 1996). Sea surface temperature (SST) and surface chlorophyll a concentration (Chl) are two major parameters that are easily derived from space-borne sensors, and they can provide valuable information on physical and biological phenomena in the upper ocean. A detailed knowl- edge of the seasonal distribution and variability of SST and Chl is therefore useful for both climate and fish- eries research in the Benguela ecosystem. Reliable information about the mesoscale spatial and seasonal variability of SST and Chl has only been possible since the advent of satellite remote sensing. Knowledge of these variables provides insight into the dynamics of coastal upwelling and the extent of phytoplankton production. This coupling is of prime importance in the highly productive Benguela ecosystem, and in determining the reasons for its fisheries fluctuating (Hutchings and Field 1997). The primary goal of this study is to present an atlas of the spatial and temporal climatology of SST and Chl in the Benguela, and to describe the main patterns of seasonal variability. Chl climatology for this region has been lacking during the past decade, and despite improvements in quality, global SST climatologies (Reynolds and Smith 1995, Smith and Reynolds 1998) are of too coarse a spatial resolution (1 degree of lati- tude and longitude) for effective studies of coastal pro- cesses. The use of both SST and Chl allows a more comprehensive observation of the surface environment. The area covered by these climatological series is ad- jacent to three African countries, Angola, Namibia and South Africa, extending from 12��to 40��S and from 6��to 29��E, encompassing the whole Benguela up- welling region, as well as the southern coast of South Africa from Cape Town to Port Elizabeth (Fig.1). DATA AND METHODS SST and Chl from the global area coverage (GAC) data- sets (both at 4.5 km resolution) were made available by Afr. J. mar. Sci. 25: 363���372 2003 363 CLIMATOLOGY AND VARIABILITY OF SEA SURFACE TEMPERATURE AND SURFACE CHLOROPHYLL IN THE BENGUELA AND AGULHAS ECOSYSTEMS AS OBSERVED BY SATELLITE IMAGERY H. DEMARCQ*���, R. G. BARLOW��� and F. A. SHILLINGTON��� A climatology of satellite-derived sea surface semperature (SST) and surface chlorophyll a concentration (Chl), and their associated variability at time-scales from weeks to years, was constructed for the Benguela and Agulhas ecosystems. Global area coverage data at 4.5 km spatial resolution from both AVHRR and SeaWiFS sensors were used to assemble the climatology, from weekly and 5-day time-series respectively. The SST data series spanned 18 years (1982���1999), the Chl data the period September 1997���April 2002. The dominant pattern in the annual SST is the cold upwelled water on the western continental shelf of South Africa and Namibia. SST was high at the Angola-Benguela Front (15���17��S) and on the Agulhas Bank, the northern and southern extremities of the upwelling region. Monthly SST variability was high in both regions, except for the L��deritz and Cape Columbine-Cape Peninsula upwelling zones, where variability was low. The western Agulhas Bank exhibited a clear seasonal pattern of warm surface water in summer and cool surface water in winter, with an amplitude of 2.5��C. A band of high Chl (5���10 mg m-3) was apparent close to the coast from the Angola-Benguela Front to Cape Town, but there was a well-defined relative minimum at L��deritz. On the South Coast, highest Chl (3 mg m-3) was between Cape Agulhas and Port Elizabeth, in the form of a plume moving offshore. In contrast to SST, Chl variability was mainly at intra- monthly intervals, although variability was particularly high north of the Angola-Benguela Front in summer, and at various upwelling sites in winter. Key words chlorophyll, climatology, satellite imagery, temperature, variability * IRD, Institut de Recherches pour le D��veloppement, France. E-mail: demarcq@mcm.wcape.gov.za ��� Marine & Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa ��� Department of Oceanography, University of Cape Town, Rondebosch 7701, South Africa Manuscript received April 2003 accepted June 2003

the National Oceanic and Atmospheric Administration (NOAA) and the National Atmospheric and Space Administration (NASA) respectively. The original data were respectively acquired from the AVHRR and Sea- WiFS sensors, on board the NOAA and OrbView-2 satellites. The 4.5 km resolution matched the large-scale extent of the study, and is good enough to describe the spatial and temporal variability of the observed surface mesoscale patterns. The SST dataset was constructed and used in the European-Union funded ENVIFISH project (Shillington and Nykjaer 2002), where it was processed into weekly and monthly composites. SST data were processed using the method de- scribed in Kidwell (1995). The final product consisted of weekly averages of SST, from 1982 to 1999. In order to use the best data possible, and to keep the same overlap with the Chl dataset, this 18-year climatology was preferable to the 5-year period, as used for Chl. The clearer variability of SST, which is a less persistent variable than Chl, also confirmed the appropriateness of this choice. The SeaWiFS GAC database extended from the first available data of the SeaWiFS mission (September 1997) to April 2002. The original ���chlorophyll��� GAC level 2 data product was used to build the Chl com- posites directly from Version 3 of the SeaWiFS algo- rithms (O���Reilly et al. 2000) using the SeaDAS software package (Baith et al. 2001). A daily spatial compositing of the standard chlorophyll a product was performed by merging the data from different orbits (between 364 African Journal of Marine Science 25 2003 Fig. 1: (a) Annual average of SST and (b) mean monthly SD for SST computed from monthly composites for the period 1982���1999 (c) annual average Chl and (d) total CV of Chl computed from 5-day weighted composites from September 1997 to April 2002, and (e) mean monthly CV of Chl computed from monthly composites