Heat transports in the Indian Ocean estimated from TOPEX/POSEIDON altimetry and model simulations

Abstract

Estimates of the heat budget of the Indian Ocean computed using TOPEX/Poseidon (T/P) sea-level anomalies and the Miami Isopycnal Coordinate Ocean Model are compared to study the redistribution of heat in the Indian Ocean. In particular, the horizontal heat transport and heat storage are used because they typically change on time scales of months or years or longer, and are therefore a predictable element of the climate system. The results show that T/P-derived heat storage is weaker than that derived from the model but has similar spatial structure and temporal evolution. Complex principal component analysis shows that there are two main modes of heat content redistribution in the Indian Ocean. The most dominant mode has an annual signal peaking in the boreal summer, and depicts the response to strong southwest monsoon winds. This involves offshore propagation of heat in the northern Indian Ocean and southward propagation of heat across the equator. The other main mode of heat content redistribution in the Indian Ocean results from westward propagating equatorial Rossby waves. This process is prominent in the boreal fall to spring, and represents the dynamic readjustment of the Indian Ocean to near-equatorial wind forcing. This mode indirectly relates to the dipole mode index in the Indian Ocean. The minima of this time series coincide with the occurrence of the anomalous dipole structure in the equatorial Indian Ocean.

Introduction

The inter-hemispheric heating contrasts of land and ocean surfaces are at the heart of the monsoon phenomenon. The role of the ocean in the functioning of the monsoons is a subject of considerable relevance to initiatives exploring the ocean climate in the global context. The upper ocean heat supply feeds the necessary evaporation and atmospheric moisture transports during the monsoon season. The energetics of the combined hydrosphere–atmosphere system in the Indian Ocean sector and their role in the functioning of the monsoons have been subject to study, in a climatological sense, through observational data sets (Hastenrath and Greischar, 1993; Hsiung et al., 1989). Although the results differ in detail, there is agreement on the dominant seasonal characteristics of the Indian Ocean heat budget. During boreal winter, the hydrosphere of the southern tropical Indian Ocean exports heat that is carried to the north of the equator. In the northern portion of the basin the hydrosphere stores heat. During the boreal summer, the northern Indian Ocean cools drastically, with the net heat gain through the ocean surface being overcompensated by the larger southward heat transport across the equator, which in turn results in a large heat import to the hydrosphere of the southern Indian Ocean. This and the depletion of the oceanic heat content supply about a third of the energy required for the vigorous evaporation from the southern tropical Indian Ocean, which is further favored by the strong southeast trade winds peaking at this season of the year. The studies indicate the presence of a strong link between ocean dynamics and the atmospheric heat and moisture transfer that is critical to the strength of the monsoons.

A better understanding of the temporal variability and spatial redistribution of the upper ocean heat content is thus valuable to comprehend the monsoon dynamics (Webster et al., 1998; Meehl, 1994). Part of the uniqueness of the Indian Ocean is the fact that it is bounded to the north at a relatively low latitude (unlike any other ocean), so all heat absorbed in the northern Indian Ocean must escape southwards. There have been a few modeling efforts that have focused on the dynamical processes responsible for heat content redistribution in the Indian Ocean on a seasonal scale (e.g., McCreary et al., 1993; Wacogne and Pacanowski, 1995). More recently, Murtugudde and Busalacchi (1999) analyzed the heat budget terms and SST in the Indian Ocean. They found that during the boreal summer, apart from surface fluxes, entrainment cooling and horizontal advection play equally important roles in determining the SST in the Arabian Sea, Bay of Bengal and the southern tropical Indian Ocean. Lee and Marotzke (1998) and Garternicht and Schott (1997) provided a more detailed analysis of the mechanisms governing the meridional heat transport in the Indian Ocean. Garternicht and Schott (1997) find that a vertical overturning cell in the upper 500 m is the main contributor to the annual mean meridional heat flux across 5°S, whereas the horizontal gyre circulation, confined to the upper 500 m, dominates north of the equator. Lee and Marotzke (1998), through a dynamical decomposition of the overturning and heat transport, show that the time-varying Ekman flow plus its barotropic compensation can explain a large part of the seasonal variations in the overturning and heat transport. These studies also reveal the importance of the upper-ocean dynamics in the seasonal heat transports. The present study aims towards extending these recent findings about the meridional heat transports with an improved understanding of the spatio-temporal redistribution of the upper ocean heat content. The analysis intends to reveal heat propagation characteristics not only in the meridional direction but also in the zonal direction. The latter have assumed importance with the recent findings about an east–west dipole structure in the Indian Ocean and their impact on the intensity of the ensuing monsoons over east Africa and the Indian subcontinent (Webster et al., 1999; Murtugudde et al., 2000; Saji et al., 1999).

Satellite altimetry, with its abundant spatio-temporal coverage of the ocean provides an excellent opportunity to study the heat propagation in the ocean. These observations can be used for validating model results and also for improving models. Perigaud and Delecluse (1992), Perigaud and Delecluse (1993) analyzed Geosat altimeter data for comparison with numerical simulations from a reduced-gravity model forced by winds. After decomposition into complex empirical orthogonal functions (CEOF) they found that the low-frequency anomalies are largely described by the first two modes for observations, as well as simulations. Stammer et al. (1996) compared the sea-level response of the 0.25° global model of Semtner and Chervin (1992) to observations by the TOPEX/Poseidon (T/P) altimeter over ∼2 years. They found that, on scales >2°, the model sea-level variations associated with seasonal cycle in the Indian Ocean were similar to those in T/P observations. Several investigations (Yang et al., 1998; Basu et al., 2000; Subrahmanyam et al., 2002; Subrahmanyam and Robinson, 2002) have made more detailed comparisons between T/P observations and ocean model simulations using reduced gravity ocean models. They all show that dominant modes of the observed and simulated sea-level changes were related to important mesoscale and basin-scale variations.

Of particular interest to this study is the heat content redistribution over the Indian Ocean. The heat content anomalies of the ocean have been estimated from T/P sea-level anomalies (SLA), based on the assumption of a linear relation between sea-level change and the heat content of the water column. In regions of strong variability in the upper ocean heat content, such as the Indian Ocean, the derived heat content anomalies provide useful tools to study the spatial and temporal variability.

This study focuses on:

• testing the viability of using SLA data from T/P altimetry to produce a time-series of heat content anomalies over the Indian Ocean;

• assessing the hydrosphere heat budget and compare the heat budget terms derived from SLA observations with those derived from a numerical simulation of the Indian Ocean;

• using the results to comprehend the spatio-temporal variability in the oceanic heat transports in the Indian Ocean.