This was a proposal to study the ocean’s surface mixed layer and stratified thermocline in the equatorial Indian Ocean and their effects on the initiation and evolution of the Madden-Julian Oscillation (MJO) using an array of three upper-ocean / surface-flux moorings plus two RAMA moorings operated by NOAA. Working with R.-C. Lien from the Applied Physics Laboratory at the University of Washington, our goals were to quantify and understand the dynamics of the preconditioning, evolution, and recovery of the upper ocean as it interacts with MJO events. The mooring array was deployed along 78E in the center of a radar sounding array consisting of two ships and two island stations, representing core observational components of the larger DYNAMO program. Each mooring was equipped with meteorological sensors, arrays of CTD sensors, moored microstructure sensors (cpods), and profiling current meters. Additional arrays of cpods were deployed on RAMA moorings at 0, 80E and 0, 90E.

An unexpected outcome of the complementary shipboard experiment (1059055) was the fish aggregations beneath the ship that contaminated Doppler sonar measurements of ocean currents and generated high turbulence levels via active swimming. This was diagnosed by the analysis of Pujiana etal (2015) using independent measurements from high-frequency shipboard echosounder and, importantly, from the moored cpods at the 0, 80E mooring 1 km away from the ship. Day/night discrepancies in e at the ship (Revelle) in the deep, marginally stable flow beneath the surface mixed layer were not replicated at the mooring (Figure 1). It was determined that fish, following the upwards nighttime migration of zooplankton into the near-surface waters illuminated by the ship’s lights and actively swimming against the mean current, generated this contamination. This did not happen at the mooring in the absence of nighttime lighting.

Consequently, we learned a lesson in ichthyogenic turbulence (Pujiana etal, 2015), finding the principal characteristics to be (i) low wave number roll-off of shear spectra in the inertial subrange relative to geophysical turbulence, (ii) Thorpe overturning scales that are small compared with the Ozmidov scale, and (iii) low mixing efficiency. These factors extend previous findings from very different biophysical regimes and support the general conclusion that the biological contribution to mixing the ocean via turbulence is negligible.

Perhaps most significantly, these new moored measurements of the ocean’s response to the intense surface winds and cooling by two successive MJO pulses showed persistent ocean currents and subsurface mixing after pulse passage, thereby reducing ocean heat energy available for later pulses by an amount significantly greater than via atmospheric surface cooling alone (Figure 2). This suggested that thermal mixing in the upper ocean from a particular pulse might affect the amplitude of the following pulse. We tested this hypothesis by comparing 18 pulse pairs, each separated by < 55 days, measured over a 33-year period (Figure 3). With this comparison, we found a significant tendency for weak (strong) pulses, associated with low (high) cooling rates, to be followed by stronger (weaker) pulses. We therefore proposed that the ocean introduces a memory effect into the MJO whereby each event is governed in part by the previous event (Moum etal, 2016).

The correlation shown in Figure 3 is consistent with a broader potential association of high SST and high ³RMM (where ³RMM is a measure of MJO intensity; Figure 4). The occurrence and intensity of tropical cyclones (TCs, including hurricanes and typhoons) is dynamically associated with warm SST through its effect on the moisture content of the atmospheric boundary layer. However, showing this statistically (important for projecting occurrence frequencies and intensities in a changing climate) has not been straightforward. Tropical SST has been shown to be correlated with the maximum cubed surface wind speed within hurricanes, a direct measure of the power in individual TCs. A strong relationship also exists between the frequency of TCs with SST, however indirect the SST / TC number relationship. These studies benefited from relatively long records (> 80 years) and direct measurements of cyclone winds and occurrences.

By comparison, our record of ³RMM is short and measurements are less direct. However, a relatively unsophisticated analysis using ³RMM of the 84 MJO pulses identified in our record compared to SST averaged locally at 0°, 80°E is suggestive. We sorted the 84 pulses into three ranges of pre-existing SST and then counted the number of pulses in six ranges of ³RMM (Figure 4). This shows a progressively greater propensity for large values of ³RMM to occur with progressively larger SST.

Last Modified: 01/11/2017
Modified by: James N Moum