No abstract

By repeated microstructure profiling during PATCHEX, we observed 11 cycles of daytime stratification and turbulent decay within the surface mixed layer during a period of generally light winds. Absorption of solar radiation strongly stratified water near the surface, but also weakly stratified all depths in the remnant mixed layer below the diurnal thermocline. Advection due to relaxation of horizontal density gradients also routinely contributed to remnant layer restratification. Between 0.2 and 0.4 MPa stratificationincreased linearly with time from (N 2) • 5 x 10 -7 s -2 at the end of nighttime convection to (N 2) • 3.5 x 10 -6 s -2 by the end of the day. Throughout the day we observed high dissipation rates ((½) > 10 -6 W kg -1) in the near-surface zone, which extended down to 0.07-0.10 MPa. Within the remnant layer, dissipation remained nearly constant for the first half hour after the end of convective forcing and then decayed by a factor of 40 in 4 hours, reaching 5 x 10 -lø W kg -1; throughout the decay, overturning scales were limited by the growing stratification. Within the seasonal thermocline there was a daily cycle in ½, with nightly maxima of about 3 x 10 -8 W kg -1 and daytime minima around 2 x 10 -9 W kg -1. The mean turbulent heat flux within the seasonal thermocline was comparable in magnitude to the mean heat flux through the surface. velocity microstructure. Lombardo and Gregg [1989] (herein referred to as LG) used this time series to compare turbulence in the convective nighttime mixed layer with similarity scaling. We use it to study the processes controlling restratification and turbulent decay following the cessation of convective forcing.Section 2 is a brief description of the oceanic measurements taken and the instruments used, and section 3 discusses the meteorological observations and background oceanic conditions.In section 4 we examine the observed daily cycle in stratification and turbulence. Section 5 is a summary and discussion of our results. These results are used as a basis for modeling in Part 2 [Brainerd and Gregg, this issue]. INSTRUMENTATION AND METHODSObservations were taken from the R/V Thompson, starting next to the moored Floating Instrument Platform (FLIP) for about 3 days, then following a midwater float for 5 days until it was 19 km from FLIP, and finally returning to FLIP for the last 3 days. Neither of the major ship movements occurred during restratification.Microstructure measurements were made using the Advanced Microstructure Profiler (AMP), a tethered freefalling instrument that measures temperature, conductivity, pressure, and microscale temperature gradient and velocity fluctuations. Small-scaJe vertical shears of horizontal velocities were measured by two airfoil lift probes capable of resolving shears from centimeter to meter scales [Osborn and Crawford, 1980], sufficient to resolve dissipation spectra during PATCHEX. The rate of dissipation of turbulent kinetic energy, e, was determined by integrating the shear spectra [Shay and Gregg, 1986], assuming t...

Abstract. On the Tropical Ocean-Global Atmosphere-Coupled Ocean-Atmosphere Response Experiment pilot cruise, at 147øE, in the western Pacific Warm Pool, we profiled for 17 days at 0øN and for 5 days at 2øN. Winds were generally light and variable in direction, but rainfall was often quite intense. Contrary to what is seen in the central equatorial Pacific, we did not observe a deep diurnal cycle in dissipation extending below the mixed layer. Strong daytime restratification often prevented nightly convective deepening down to the seasonal thermocline, resulting in surface forcing remaining trapped in a shallow layer. The relaxation of horizontal density gradients into vertical appears to be an important process driving restratification. Turbulent fluxes in the bottom of the mixed layer were generally small. Following rainfall, we observed pools of fresh water that typically disappeared within a few hours, leaving the mixed layer nearly homogeneous in salinity; thus we did not observe a permanent barrier layer. Modeling such events using the Price-Weller-Pinkel model suggests a fresh pool will be mixed away on timescales of a few days, primarily by nighttime convection. The observed vertical structure can be accounted for by local vertical mixing processes. We made three-four drops per hour, totaling 1100 profiles on the equator and 275 at 2øN. At the equator we lost nearly a day of profiling due to the loss of an AMP (year day 52), and In this study we distinguish between mixed layers and mixing layers [Brainerd and Gregg, 1995]. By mixed layer we mean the depth zone that has been mixed down from the surface within the relatively recent past; in this case it corresponds to the layer above the top of the seasonal thermocline. By mixing layer we mean the zone in which there is active turbulence being directly forced by surface fluxes; at night it is generally the zone in which convection is active, while during the daytime it is the zone in which wind-and wave-generated turbulence is active. Observations Meteorological ConditionsWesterly bursts are most likely in winter, but in February 1990 the Southern Oscillation Index (SOI) dropped to its low-10,437

We find that daytime restratification of the remnant layer is important to modeling both the decay of convective turbulence during the day and convective deepening the following night. Penetrating solar insolation accounted for about 60% of the observed restratification of the remnant layer. Of the other processes supplying the remaining restratification, we believe lateral advection is the most important and sets a limit on the capabilities of one-dimensional mixed-layer models. In the morning, with the end of convective forcing, there was an initial period lasting nearly an hour, similar to the convective time scale, during which ½ in the remnant layer remained nearly constant. After that, turbulence in the remnant layer could be modeled in accordance with a balance between e and the storage term for turbulent kinetic energy (TKE). Energy storage is computed using q2 = (el/C)2/3, where I = 0.84Lo, the Ozmidov scale, matches observed overturning scales during most of the decay. Calculating Lo for the observed linearly increasing stratification gives a modified exponential form for the decay. A value of C = 0.04 gives the best fit. This decay lasted about 5 hours, until (e) reached • 5 x 10 -lø W kg -1. Then for the remainder of the restratification period, dissipation and production of TKE due to vertical shear appeared to be approaching a balance.