%0 Conference Proceedings
%9 ACTI : Communications avec actes dans un congrès international
%A Correa, K.
%A Machu, Eric
%A Diouf, D.
%T Distribution and biomass of phytoplankton over the Senegalese shelf : a need for regional ocean colour algorithm [résumé]
%S International conference ICAWA 2016 : extended book of abstract : the AWA project : ecosystem approach to the management of fisheries and the marine environment in West African waters
%C Dakar
%D 2017
%E Brehmer, Patrice
%E Ba, B.
%E Kraus, G.
%L fdi:010072124
%G ENG
%I SRFC/CSRP ; IRD
%@ 978-2-9553602-0-5
%K SENEGAL ; ATLANTIQUE
%P 77
%U https://www.documentation.ird.fr/hor/fdi:010072124
%> https://horizon.documentation.ird.fr/exl-doc/pleins_textes/divers18-01/010072124.pdf
%W Horizon (IRD)
%X The acquisition and processing of 1179 images from the SeaWiFS (Sea-viewing Wide Field-of-view Sensor) sensor on board the Seastar satellite enabled us to monitor the chlorophyll a ( Chl-a) between December 2003 and November 2007 over the Senegalese plateau. The level 1 data from this sensor were processed using an algorithm combining a neuronal classification and Self Organizing Maps before retrieving chlorophyll-a concentrations with a variational method (S-NV) . The results obtained allowed us to distinguish a typical seasonality and a strong spatio-temporal variability of phytoplankton over the Senegalese plateau. The spatial distribution is marked by a maximum concentration along the coasts and especially south of the Cape Verde peninsula. Temporal variability is marked by a concentration that increases from winter to spring with a high spring concentration that exceeds 13 mg/m3 on average, a sharp decline in summer, and a new increase in autumn. The phytoplankton mean biomass estimated using the standard OC4V4 SeaWiFS algorithm was ~5 mg/m3 while the S-NV algorithm provided a mean concentration of ~4 mg/m3. Although estimates are comparable, S-NV allows to obtain a mean annual coverage over 50% while S-STD reached only 24%. Seasonal differences can be even stronger. Over the investigated period, years 2004 and 2007 show the highest concentrations in spring. S-STD give concentrations over 17mg/m3 whereas S-NV provide values of ~6mg/m3 for the same year/month. Results indicate that S-NV algorithm generates concentrations in better agreement with in situ observations than the standard algorithm.
%B ICAWA : International Conference AWA
%8 2016/12/13-15
%$ 034 ; 126 ; 021