PEREZ Gonzalo Luis
congresos y reuniones científicas
ACE Project 8 SORPASSO (CTOM Absorption, see how much information can we get?)
GONZALO L. PÉREZ
Conferencia; NOSASSO; 2019
Studying the data by Legs allows us to have a first glimpse of the CTOM absorption variability, however, this could be more related to oceanic and geographic characteristics. Therefore, I studied the absorption coefficients and spectral characteristics through the temporal series, and also following the analysis or SIM I related the variability with some important geographic points. We can see the temporal variability of absorption coefficients at several reference wavelengths. They are colored by Legs. Of course, we can also see the variability within legs, and higher values at different points during ACE. High values at all evaluated wavelengths were observed near Islands Kergulend and Heard, and also in South Georgia. Higher values were also observed near Mertz, Off. Ross Sea and Siple.If we compare this variability against other variables, like chlorophyll and salinity, it is clear that the increment in absorption coefficients near Mertz, Off. Ross Sea, Siple, and South Georgia, were concomitant with an increment of chlorophyll a. It seems not to be the case for the islands Kergulend and heard, but discrete samples of chlorophyll a by HPLC also showed an increment in this area. Salinity presented the lowest values at Mertz and a decreasing pattern near Off. Ross Sea. Additionally, If we analyzed the variability of spectral characteristics in the temporal series, we can see that near geographic point where higher values of absorption coefficients were detected, spectral slopes were in general lower than other samples. The slope ratio, instead, showed the same pattern than absorption coefficient. This is clearly observed against continuous values of chlorophyll a and salinity. The general opposite pattern between abs and slopes, in better observed in the following plot.To further explore the data set, we can also search for similarities between spectra and overlap these new groups with the chronological order. For this I normalized the spectra (between 250 to 500 nm by 5 nm interval) and using cluster analysis I have obtained 3 different groups.