Understanding trends in satellite-derived oceanic evaporation
Frank J. Kelly 1, Alberto M. Mestas-Nuñez1, Abderrahim Bentamy2, Kristina B. Katsaros3, Rachel T.
Pinker4, William M. Drennan3, Anthony Santorelli4 and James A. Carton4
1
Department of Physical and Environmental Sciences, Texas A&M University-Corpus Christi, 6300 Ocean
Drive, Corpus Christi, TX 78412, U.S.A, Alberto.Mestas@tamucc.edu
2
Institut Francais de Recherche pour l’Exploitation de la Mer, B.P. 70, 29280, Plouzane, France
3
Department of Applied Physics, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149, U.S.A.
4
Department of Atmospheric and Oceanic Science, University of Maryland, College Park, MD 20742, U.S.A.
Warming trends in sea surface temperature during the latter portion of the 20th century have raised
inquiries abut trends in oceanic evaporation. The expectation is that evaporation should increase in a
warming ocean. Satellite-derived evaporation datasets over the global ocean should contribute to the
discussion. In this study, we use an extended 15-year (1993-2007) version of the Institut Francais de
Recherche pour l’Exploitation de la Mer (IFREMER), satellite, air-sea flux dataset to estimate trends
and variability in oceanic evaporation over the global oceans. In general, the IFREMER dataset shows
a pronounced, positive trend in global evaporation. Trend patterns show that the increase in oceanic
evaporation is global in scope. Notable exceptions occur in the cold tongue regions of the Pacific and
Atlantic oceans where evaporation trends are flat to slightly negative. Portions of the high-latitude
Pacific and Indian oceans exhibit a pattern similar to the cold tongues.
The observed satellite evaporation trend is enhanced by a short period of rapidly increasing evaporation
between 1999 and 2004. Our analysis shows that the enhanced trend is mainly attributable to a rise in
sea surface temperature. However, increased variability in wind speed accentuates the trend. In
addition, a small, but perceivable jump in evaporation beyond June 2002 corresponds to a jump in sea
surface temperature caused by merging AMSR to AVHRR data. Comparisons with other evaporation
datasets show both similarities and differences. For example, some products document a positive
although less pronounced trend in global evaporation because they contain more regions with a
negative trend; regions with a positive trend are restricted to the trade winds and western boundary
currents. Discrepancies between datasets arise from using different input state variables within the
latent heat flux bulk algorithm and/or whether the product relies on model output. The in-depth
analysis of the IFREMER dataset and comparisons with other products gives us an uncertainty range in
evaporation trend estimates. Our study also provides some clues on how to improve the estimation of
air-sea turbulent fluxes by examining the nonlinear relationships between the input state variables.