Variability of Pacific tropical convergence zones in observations

The instantaneous South Pacific Convergence Zone (SPCZ) and east Pacific Intertropical Convergence Zone (ITCZ) are identified in long-term satellite observations using two automated statistical models. The statistical models are designed to emulate visual identification of convergence zones, balancing the complex definition of a convergence zone as an elongated envelope of convection including clouds of varying heights as well as clear sky, against the need for automatic detection in large amounts of data. Identification occurs in 3-hourly infrared (IR) images from geostationary satellites from 1980-2012. For the SPCZ, the study is limited to November through April but the east Pacific ITCZ is identified year round. Interannual variability, seasonal evolution, and intraseasonal variability of the SPCZ are quantified using 3-houlry SPCZ labels. The SPCZ is found to have two distinct parts: a tropical segment which is more active, particularly in December through February, having a mostly zonal orientation and a subtropical segment which is less active and has a tilted orientation. The El Niño Southern Oscillation (ENSO) influences the SPCZ on interannual time scales as the SPCZ shifts equatorward during El Niño and poleward during La Niña. On the intraseasonal time scale the SPCZ changes intensity and location according to various phases of the Madden Julian Oscillation (MJO). The SPCZ also has a distinct diurnal cycle in area, mean IR temperature, and cloud height, which changes throughout the season and is influenced by ENSO and the MJO. In the east Pacific, the ITCZ can take on several configurations. A statistical model is used to automatically assess the daily state of the east Pacific ITCZ based on the location of cloud bands: north of the equator (nITCZ), south of the equator (sITCZ), simultaneously north and south of the equator (dITCZ), and over the equator (eITCZ). A fifth state describes when no cloud bands exist in the east Pacific (aITCZ). Most of the year is dominated by the nITCZ state but in the boreal spring all states occur and variability is high. In March and April the dITCZ occurs, on average, 34% of the time, indicating that the double ITCZ occurs frequently in instantaneous data