A TOVS Dataset for Study of the Tropical Atmosphere

Abstract

Temperature and moisture data from the TIROS Operational Vertical Sounder (TOVS) archives are examined for applicability to studies of the tropical atmosphere on time scales ranging from intraseasonal to interannual. Comparisons with monthly mean radiosonde data from island stations confirm the ability of TOVS to track short- and long-term atmospheric variability. Biases and rms errors are generally different for near-equatorial and subtropical stations. Inclusion of soundings derived under cloudy conditions increases negative temperature bias while improving or leaving unchanged the rms temperature error at all levels. The large volume of TOVS data (up to 16 000 soundings per day) is reduced to manageable form by the creation of a gridded product. A structure function analysis is performed to assist in the choice of gridding parameters. The objective analysis routines used are designed to handle data voids common in satellite data fields. A time-longitude diagram of 5-day mean precipitable water (PW) in the 1000?700 mb layer shows a strong interannual (El Niño) signal as well as time variability in the 30?60 day range in the western Pacific and Indian oceans. On the shorter time scales, maximum PW is generally coincident with axes of minimum 250 mb velocity potential.