Detection of Rossby Waves in Ocean Colour Data Paolo Cipollini, Peter G. Challenor, David Cromwell, Graham D. Quartly and Stefano Raffaglio 1 James Rennell Division for Ocean Circulation and Climate - Southampton Oceanography Centre European Way, Southampton SO14 3ZH, United Kingdom Phone: +44-23-80596404 Fax: +44-23-80596400 e-mail: cipo@soc.soton.ac.uk 1 Now at Roke Manor Research Ltd, Romsey, United Kingdom ABSTRACT This paper deals with some observations of the signature of mid-latitude planetary waves (Rossby waves) in global chlorophyll-a data from the Ocean Colour and Temperature Scanner (OCTS) and the Sea-viewing Wide Field-of-view Sensor (SeaWiFS). By producing longitude-time plots of the merged OCTS and SeaWiFS monthly composites we detect, at some latitudes, westward propagating signals whose characteristics are consistent with those expected for long- wavelength baroclinic Rossby waves. The propagating signals are superimposed on the much stronger annual phytoplankton cycle, however their detection may be enhanced by band-pass filtering the longitude-time plots. The propagation speed depends on latitude, with speed increasing equatorward, as expected for Rossby waves. We show an initial comparison of the results with those obtained from sea surface height data. Finally, we discuss possible explanations for the phenomenon INTRODUCTION Oceanic planetary waves or Rossby waves are of great importance for ocean circulation and climate. They transport momentum and information across the main oceanic basins, affect currents and delay the effects of climatic anomalies such as El Niño. Owing to their dimensions, and in particular their surface amplitude of just a few centimeters coupled with a wavelength of some hundreds of kilometers, they cannot be directly observed from ships. Recent advances in satellite oceanography, and in particular the accuracy of the retrieval of Sea Surface Height (SSH) by satellite-borne radar altimeters such as the TOPEX/POSEIDON and those on board ERS-1 and ERS-2, have allowed a detailed study of these waves and proven them to be almost ubiquitous and to travel faster than previously expected [1]. The signature of Rossby waves has also been observed in the global Sea Surface Temperature (SST) dataset. A comprehensive review of the techniques used to extract the information about Rossby waves from satellite data, as well as of the main results, is in [2]. Having observed Rossby wave propagation in global satellite datasets of SSH and SST, it is a natural step to examine global datasets of ocean colour (chlorophyll-a contained in phytoplankton) which are becoming increasingly available. This paper presents some results of observations of Rossby waves in the global ocean colour record built from NASDA’s Ocean Colour and Temperature Scanner (OCTS) and NASA’s Sea-viewing Wide Field-of-view Sensor (SeaWiFS) datasets. DATA AND METHODS In virtue of the early launch (August 1996) of OCTS on board the Japanese satellite ADEOS (whose operation was unfortunately terminated on 30 June 1997 because of a major failure of the solar array) and the subsequent (August 1997) launch of SeaWiFS, the scientific community has access to an almost 4-year long time series of ocean colour data. This is now being extended by SeaWiFS itself and the recently launched Moderate resolution Imaging Spectrometer (MODIS) For the purposes of this study we used OCTS Global Area Coverage (GAC) level 3 Binned Map data (monthly chlorophyll-a composites) which are currently available from NASDA-EORC. The data have been processed with version 4 chlorophyll algorithm and cover the period November 1996 to June 1997 (8 months in total). For SeaWiFS we used GAC level 3 data from NASA-GSFC DAAC. The data have been reprocessed by GSFC with version 2 chlorophyll algorithm. The present study includes 20 months of SeaWiFS data, from October 1997 to May 1999. The original OCTS and SeaWiFS data are on a 0.0879° x 0.0879° grid, which is far too detailed for observing large- scale features like Rossby waves. So the data have been rebinned onto a 0.5° x 0.5° grid, still more than enough to detect large-scale propagating signals. This additional binning reduces the noise and the effect of potential remnant cloud contamination on the data. Finally OCTS and SeaWiFS datasets have been merged into a single dataset covering 31 months from November 1996 to May 1999, with a three- month gap in summer 1997.