Clock Drift and Sampling Time Difference

Temperature was recorded every 3 minutes at DS, and every 5 minutes at DN. We consider how the difference in sampling would affect the number of overturns detected, and the estimated dissipation. The temperature at DS is resampled every 5 minutes using linear interpolation, and subjected to the same overturn detection algorithm. We find that the resampling from 3 to 5 minutes slightly reduces the percentage of profiles containing at least one overturn (from $34$ to $30\%$ ), and also reduces the estimated dissipation from $1.2\times10^{-8} Wkg^{-1}$ to $1.02\times10^{-8}Wkg^{-1}$ .

According to the manufacturer, the instruments have a very slow temporal drift, typically less than 0.2 millidegree per month. We checked for differential sensor drift over time by repeating the bias corrections based on monthly averages. The results are similar to the 3 month (8 month) time averages, suggesting that differential drift errors are not an important factor. The sensor clock drifts were found to be less than $30 sec$ at any given sensor for DS, and less than $40 sec$ at DN, so we conclude that instrument clock drift has minimal effect on our 3 and 5 minutes averages. The difference of sampling times between the moorings therefore cannot explain the large differences found in the estimated dissipation between the two moorings.

Figure 5.1: Schematics of the overturn detection algorithm, showing a vertical density profile (black line), see section 5.1 for the definition of the points.

Figure 5.2: Distribution of dissipation events at mooring DS (red) and DN (blue), as a percentage of the detected overturns (top) and as a percentage of the total number of profiles (bottom).

Figure 5.3: Mooring DS a) The distribution of the number of overturns, before and after the temperature bias correction, as a function of the observed temperature difference across the depths of the overturn. b) The distribution of overturn size as a function of the temperature difference across the overturn. c) The distribution of the percentage of the total dissipation contributed by the overturns, as a function of the temperature difference across the overturn.

Figure 5.4: a) Vertical profile at DS on day 249, hour 11.5 (Figure 6.4) of measured absolute temperature after removing the long term bias. *'s indicate the locations of the instruments. Thin lines represent the measurement plus or minus the nominal accuracy of the instruments. b) same as a) but with inferred potential density ( $-1000 kgm^{-3}$ ).

Figure 5.5: Same at figure 5.3, but for mooring DN.

Figure 5.6: same as 5.4 but at DN for day 375, hour 11.5 (Figure 4.3).