Le et al. (2009) developed the four-band algorithm based on the three-band algorithm, and a fourth band at the NIR region is added to decrease the effect of pure water and non-algal particulates on Chl-a estimations. In Table 4, we can clearly see that the four-band algorithm performed the best of the four algorithms for water type II. For water type I, the bands at 665 nm, 709 nm and 865 nm were used, while the bands at 681 nm, 709 nm and 754 nm were applied for water types II and III in the four-band algorithm. However, the use of the four-band algorithm also has the requirement that the SCH 727965 of CDOM and non-algae particulates at the first band (λ1) are near those at the second band (λ2). From Fig. 4, it can be observed that, for all water types, the variation in the absorption coefficients for CDOM after 665 nm was minimal; however, for water type I, the absorption of non-algae particulates at 665 nm was close to the value at 709 nm. As a result, water type I partially meets this requirement. At the same time, for water types II and III, the absorption coefficient of non-algae particulates at 681 nm was near the coefficient at 709 nm. Thus, water types II and III partially satisfy this requirement. Another requirement is that the absorption of CDOM and non-algae particulates at the third band (λ3) is also close to the absorption at the fourth band (λ4). From the absorption spectral trend of non-algae particulates for water type I, shown in Fig. 4, it can be observed that the ad(865) (not shown in Fig. 4) was smaller than at 800 nm. This absorption spectrum indicates that the ad(800) is significantly lower than at 700 nm. As a result, the ad(865) is clearly smaller than at 709 nm. Therefore, water type I did not satisfy the second requirement. Based on the non-algae particulate absorption spectrum for water types II and III, we can see that the difference between 754 nm and 709 nm was small and that water types II and III partially meet the second requirement.