Figure 13. Plots of measured radiance versus model reflectance factor for 3 sols (51, 210 and 302) and 3 filters (L6, R1 and R5), showing the related one-term fit model (solid line) and the two-term fit model (dashed line). The values of the slopes and offset are reported for both models in the lower right corner of each box. The white clean spot was not used for the computation of the fits.
The slightly greater flatness of the two-term fits that came out of the slope computation also resulted from the relative difference between the slopes in the two models. We quantify this relative difference\([(\text{slope}_{\text{two}\_\text{term}}-\text{slope}_{\text{one}\_\text{term}})/\text{slope}_{\text{one}\_\text{term}}]\)in Figure 14a for four filters (L6, L3, R2, R6). The data points at very low values on sol 9 and 23 were due to a low resolution (26 mm and 34 mm, respectively). If we exclude these two observations and those after sol 314, that were strongly perturbed by the dust event, the differences were negative and slightly increasing in absolute value with time. The difference in filter L6 (442 nm, purple circles) had a net decrease from -0.09 to -0.12 (-35%), while R6 (1022 nm, black circles) went from -0.17 to -0.20 (-13%). All the other filters followed a similar trend, with rates between -15% and -28%. In addition, the relative differences in slopes did not follow any linear dependency on wavelength. The plot of Figure 14b shows the averages of the relative differences for each filter over the sol range 100-180 (across the martian aphelion, highlighted by the two vertical solid lines in Figure 14a). All the points are included between -0.2 and -0.07, with L6 being the highest and R6 the lowest. The spectrum of the relative difference is characterized by a decrease from -0.09 in L6 to -0.15 in L4 (442 nm to 605 nm), followed by a rough slight increase up to -0.14 in R5 (978 nm) and a drop to -0.19 in R6.