Table 1. Chl Algorithm Performance for CI < -0.0005 sr-1 Using the NOMAD Data Seta
Algorithm RMS URMS Mean Ratio Median Ratio R2 (Linear) R2 (Log) N
OC4v6 34.9% 28.2% 1.11 1.08 0.73 0.85 50
CI 16.5% 16.2% 1.01 1.01 0.78 0.95 50
OClow 23.0% 22.3% 1.03 1.05 0.73 0.85 50
aOClow represents a local polynomial fit between the log-transformed band ratio and Chl for low concentrations only (CI < -0.0005; Figure 3a, red line), which shows improved performance than the globally tuned OC4v6. The regression equation is ChlOC_low = 10.y and y = 0.3906 - 1.5479Chi + 3.2125Chi2 - 3.1073Chi3. URMS is "unbiased" RMS (see text for details).