Analysis of tropospheric waves with periods of 1 to 30 days using the tropical pacific profiler network

Tropospheric wind and reflectivity measurements over the tropical Pacific have been used to analyze the dominant atmospheric wave periods in the 1 to 30 day range. Wind profiling radar have been operating at a number of sites in the tropical Pacific including Biak (Indonesia), Ponape (Federated States of Micronesia), Christmas Island (Republic of Kiribati), and Piura (Peru) since at least the early 1990s. Data from these sites from 1991 to 1997 have been averaged into 6 hourly horizontal and vertical winds, and reflectivity, covering heights from 2 km to 20 km above sea level at 0.5 km intervals.

The mean horizontal and vertical winds are described over 1991 to 1997. A frequency analysis was carried out on this data using Fourier and wavelet techniques. The analysis was carried out in three period bands; 1-3 day, 3 -5 day, and 5-30 day. These bands correspond to the periods most commonly associated with mixed Rossby gravity waves, inertiogravity waves, and equatorial Rossby waves. A wavelet analysis and sliding Fourier analysis is used to show how the dominate wave periods in each of these bands change over the 1991 to 1997 analysis period. Power spectra of the zonal, meridional, and vertical winds are calculated as a function of season showing the atmospheric waves present in each season. Summer and autumn (Northern Hemisphere) show a dominant wave period of close to 4 days over Christmas Island and Piura. During the winter (Northern Hemisphere) a wave period of close to 15 days is observed. Using multiple wind profiler sites, the change in atmospheric wave phase as a function of height, time, and location are determined. From these measurements, the direction of phase and energy propagation is determined.

Relatively few studies have used wind profiler derived winds to examine atmospheric waves in the 1 to 30 day range. The wind profiler derived 6 hourly winds were compared with winds from the NCEP/NCAR reanalysis, and showed that many short period features were only captured in the wind profiler derived winds. This analysis complements studies that have used other forms of data including outgoing longwave radiation, atmospheric model derived winds and winds derived using alternative remote sensing techniques.