One of the known literature formulas for estimating Chl Selleckchem BMS354825 a obtained for the Baltic Sea environment is the one given by Siegel et al. (1994). It uses the green-to-red reflectance

ratio (but at wavelengths slightly shifted compared to the wavelengths already analysed in this work) and takes the following form: Chl a = 31.05(Rrs (510)/Rrs(670))− 2.115. If we used the modelled reflectance spectra obtained in this work, the equivalent formula would take the form Chl a = 32.3(Rrs(510)/Rrs(670))− 1.24 (n = 82; r2 = 0.7; X = 1.54). As can be seen in Figure 10a, these two last formulas would agree only in the ranges of the relatively low values of the Rrs(510)/Rrs(670) ratio (which corresponds to Chl a concentrations

of the order of 10 mg m− 3 and higher). For high values of that green-to-red reflectance ratio, the latter formula would predict Chl a values several times higher than the one given by Siegel et al. (1994). Roxadustat The other formula known from the literature is the one from the paper by Darecki et al. (2005). It uses the green-to-orange ratio of Rrs(550)/Rrs(590) and after simple transformation takes the form Chl a = 5.47 (Rrs(550)/Rrs(590))− 4.681. Based on the modelling results obtained in the present work, the equivalent formula using the same reflectance ratio would be Chl a = 30 (Rrs(550)/Rrs(590))− 3.33 (n = 82; r2 = 0.76;

tuclazepam X = 1.48). Figure 10b shows that these last two formulas would exhibit distinct differences. Both formulas are relatively steep functions of the green-to-orange reflectance ratio but for the same values of this, the predicted ranges of Chl a would differ by about one order of magnitude. However, in view of the results of the latter comparison, it has to be emphasised that the 590 nm reflectance band taken for that additional test lies relatively far from the modelling input data on the light absorption coefficient an(λ) (we recall that the nearest an input data bands were at 555 and 650 nm). As a consequence, the modelled values of Rrs at 590 nm band should be treated with a relatively low level of confidence. Nevertheless, the last two additional quantitative comparisons of the relationships between Chl a and different colour ratios should warn the potential user that all the results of the simplified modelling performed here, and in effect, all the semi-empirical (reflectance-based) formulas presented in this work, should be treated as qualitative rather than quantitative. Finally, let us comment on the comparison of all the statistical parameters obtained here for different variants of both empirical (see Table 1 and Table 2) and semi-empirical formulas (see Table 3 and Table 4).