A halo-thermal, reduced-gravity model with four active layers is used to investigate how interannual rainfall anomalies affect sea surface height (SSH) variability in the Indian Ocean. The model is forced by monthly varying winds observed over the period 1980--2000 in two experiments that differ by their rainfall forcing, Run FSU and Run Arkin, forced by climatological and interannually varying rainfall, respectively. Compared to the large impact of wind on SSH (about 30 cm), the impact of rain is much smaller. Its maximum (found in the southeastern Indian Ocean during the rainfall deficits of 1994 and 1997) is only 2 cm. Because rainfall significantly affects model salinity and temperature, the deficits make the layers of Run Arkin colder and saltier than in Run FSU, causing a -5 cm change in sea level. Baroclinic adjustments also occur such that the top (bottom) two layers are thicker (thinner), increasing sea level by 3 cm and hence significantly reducing the SSH change due to steric effects alone. SSH variability in either Run Arkin or Run FSU compares very well with TOPEX data. Although the impact of rainfall on SSH is negligible, salinity variations significantly affect dynamic-height calculations of SSH. In the model, the neglect of salinity variations leads to an error of 5 to 10 cm along the eastern boundary, in the Bay of Bengal, and in the interior ocean south of 8¿S. This error is validated by the difference between TOPEX data and SSH derived from observed temperature profiles.