The files within this directory consist of 2 types of gif images: 1) sonar images with files names that end in 'a.gif' eg. dec0102a.gif (denoting Dec 01,1992 0200 hours GMT); and 2) wave spectra with file names ending in 'b.gif' eg. dec0102b.gif (denoting Dec 01,1992 0200 hours GMT). The data are for the period: October 23,1992 to March 5,1993. 1. False-Color Sonar Images The observations shown in these images were taken between October of 1992 and March of 1993, during the Intensive Operating Period (IOP) of the COARE component of TOGA. The instrument package collecting the data was the Extended Life Sonar Instrument (ELSI). ELSI was deployed as part of the moored array component of the COARE Intensive Flux Array(IFA), at 1deg 28.3'S, 155 deg 43.0'E in 2079m of water. The mooring was designed to put the top of the instrument package at 25m below the surface in a zero-current environment: in fact, the mean depth over the period of the deployment was 24m, with occasional excursions to a few meters deeper. The instrument was attached to the tight subsurface mooring line at the bottom, and the net buoyancy of 300kg was concentrated at the top of the package; thus the righting moment was large. A vane was attached to one side of the package to keep "wobble" to a minimum. The instrument heading was typically steady to within about 5 degrees, and tilts were steady to within a degree or two. The operation of ELSI was implemented by a preprogrammed micro controller; and timing was based on a crystal oscillator. Most of the time was spent in "sleep" mode, with the instrument powered down. Every hour, power was automatically turned on, and readings of temperature, salinity and pressure were recorded. Once every 25 hours a complete acoustic data set was collected, consisting of 2400 "pings" at a frequency of 2Hz. Each "ping" consisted of (in sequence): a short pulse (.272ms) from all of the upward- looking sonars; a long pulse (4.35ms) from the upward-looking sonars; a long pulse (also 4.35ms) from the sidescans; and a 48ms listening period for the hydrophone. The long pulses from both the upward- looking sonars and the sidescans were coded transmissions, consisting of 12 repetitions of a 4-bit Barker code. Incoherent Doppler processing of the returns from these transmissions give estimates of velocity; the returns from the short pings are used for fine scale resolution of bubble clouds beneath the ocean surface, and for estimates of surface height variation due to gravity waves. The images in this data set are the fine-scale images of bubble clouds and other scatterers seen by the 55 kHz upward looking conical beam sonar. The images are 25 hours apart, except for a period of higher frequency sampling at the very beginning of the deployment. The strongest scatterers are bubble clouds due to air-sea interactions (mainly wind, but sometimes rain), but at times biological scatterers can be seen in the images, both discrete large scatterers and plankton layers. Environmental conditions during the experiment indicated that the region is characterized by low wind speeds, and intense but intermittent rainfall. The most dramatic event was the Westerly wind burst which occurred between December 20 (Day 50) and January 3 (Day 64). 2. Wave height spectra. The short pulse returns from upward-looking sonars can be analyzed to give the surface height variation due to gravity waves. As the instrument was on a tight mooring, and since tilt angles were steady to within a degree or two, the time between transmission and the reception of a surface return gives an instantaneous estimate of the range to the surface, with a precision of 0.2m. The location of the surface return in the time series was estimated using a slope detection algorithm. The resulting 20 minute time series of the range to the surface sampled at 2Hz was then processed to give an estimate of the non- directional waveheight spectrum. First the time series was de-trended (and mean removed at the same time): this procedure was necessary because slow variations in subsurface currents during the 20 minute data collection occasionally resulted in changes in the instrument depth. Windowed 128 point sub-series were Fourier transformed to the frequency domain. Typically 30 of such series, 50% overlapped, were averaged together to give an estimate of the waveheight spectrum with a frequency resolution of 0.0156Hz. The -3dB beamwidth of the sonar was 5 degrees, giving a 4.4m diameter beam "footprint" at the surface for the typical instrument depth. We cut off the high frequency spectrum at 0.5Hz, corresponding to a wavelength of 6.25m, as it was not possible for the sonar to detect shorter waves because of the finite "footprint".