We present a hemispherically averaged, seasonal model with land-sea and surface-air resolution that we use to compare the transient temperature response of different model components to changes in the radiative forcing. Horizontal heat transport coefficients were obtained by tuning the model to reproduce the present seasonal climate as closely as possible. Cross-sensitivity experiments with a step function solar constant increase and a variety of ocean parameter feedback processes essentially duplicate the results obtained in part 1 (Harvey and Schneider, 1985) with a globally averaged model. We intercompare the global and hemispheric model surface and deep-ocean transient response and show how both the global mean surface and mixed-layer only transient response of the hemispheric model can be reproduced by using the globally averaged model. For a time-dependent CO2 increase scenario we find that the mixed-layer warming lags moderately behind the atmosphere warming but that the atmosphere over land and atmosphere over sea warmings are rather tightly coupled. Implementation of a number of ocean parameter feedbacks to a time-dependent CO2 increase has a similar qualitative effect on the transient temperature response as when implemented for a step function solar constant increase but causes less variation in the transient response than uncertainties about the equilibrium climate sensitivity. During the first year following a small negative perturbation to the solar constant, in order to mimic a volcanic dust veil, we find that the air temperature response over land is about twice that of the ocean mixed-layer temperature. This suggests that one should distinguish between temperature responses over land and over ocean in any studies attempting to compare model-generated volcanic signals with empirical surface temperature data. For complex short- and long-term mixed forcing changes, such as volcanic and CO2, the lag between atmosphere and mixed-layer temperature changes does not become large until the long-term forcing change becomes dominant, although the land-sea lags associated with short-term perturbations may still be significant for regional climatic anomalies associated with these perturbations.