Oceanic influences on the seasonal cycle in evaporation over the Indian Ocean

The annual mean and seasonal cycle in latent heating over the Indian Ocean are investigated using a simple, analytical ocean model and a 3D, numerical , ocean model coupled to a prescribed atmosphere, which permits interaction through sea surface temperature (SST). The role of oceanic divergence in d etermining the seasonal cycle in evaporation rate is reexamined from the vi ewpoint that the amount of rainfall over India during the southwest monsoon is a function of the amount of water evaporated over the "monsoon streamtu be'' as well as orographically induced convective activity. Analysis of Comprehensive Ocean-Atmosphere Dataset (COADS) shows that nearl y 90% of the water vapor available to precipitate over India during the sou thwest monsoon results from the annual mean evaporation field. The seasonal change in direction of airflow, which opens up a pathway from the southern Indian Ocean to the Arabian Sea, rather than the change in evaporation rat e is key to explaining the climatological cycle, though the change in laten t heating due to seasonal variations is similar to that needed to account f or observed interannual-to-interdecadal variability in monsoon rainfall. Th e simple model shows that net oceanic heat advection is not required to sus tain vigorous evaporation over the southern tropical Indian Ocean; its impo rtance lies in ensuring that the maximum evaporation occurs during boreal s ummer. Also shown with the simple model is that evaporation over the Arabia n Sea cannot increase sufficiently to make up for the loss of water vapor a ccumulated over the southern Indian Ocean should there be a change in circu lation such that the Southern Ocean is no longer part of the monsoon stream tube. Analytical, periodic solutions of the linearized heat balance equation for the simple model are presented under the assumption that the residual of ne t surface heat flux minus rate of change of heat content (DIV) is considere d to be an external periodic forcing independent of SST to first order. The se solutions, expressed as functions of the amplitude and phase of DIV, lie in two regimes. The first regime is characterized by increases (decreases) in the amplitude of DIV resulting in an increase (decrease) in the amplitu de of the solution. In contrast, in the second regime, the amplitude of the solution decreases (increases) as the amplitude of DIV increases (decrease s). It is noteworthy that the regime boundaries for SST and latent heating do not necessarily coincide. For the present climate, as determined from CO ADS, the southern Indian Ocean's annual harmonics of latent heating and SST lie in the second regime near the border, and so their tendencies are sens itive to the nature of the perturbation to the harmonic in DIV. The souther n Indian Ocean's semiannual harmonic of latent heating lies in the first re gime, and so its tendency is robust to the nature of the perturbation to th e harmonic in DIV; that of SST lies in the second regime near the border. Contrasting runs of the 3D numerical model, in which the Indonesian through flow differs by less than 4 X 10(6) m(3) s(-1) in the annual mean and less than +/-2 X 10(6) m(3) s(-1) in seasonal variability, provides new estimate s for its potential role in the Indian Ocean heat balance. Net surface heat flux differences of over 20 W m(-2) are found along the length and breadth of the southwest monsoon streamtube: particularly noteworthy regions are o ver the Somali jet and to the east of Madagascar. These changes can be expl ained in part by the changes in oceanic meridional transport generated by t he throughflow as well as by its heat input. Spatial resolution and upper o cean physics are sufficient for the throughflow to retain its zonal jet cha racter across the Indian Ocean and so inhibit meridional overturning. Signi ficantly, its presence reduces the amount of heat imported into the Souther n Ocean from the Arabian Sea during boreal summer, so making SSTs in the Ar abian Sea higher.