Abstract
The impact of anomalous tropical forcing on the anomalous tropical and extratropical circulation is examined by comparing the response of a global steady state linear primitive equation model with observations for nine (1978/79–1986/87) December–January–February (DJF) seasons. Outgoing longwave radiation (OLR) anomalies are used as proxies for tropical latent heat release. Root mean square amplitude and anomaly pattern correlation (APC) are chosen to be the verification tools.
It is shown that with OLR as the only forcing, the linear model produces a fairly realistic level of tropical interannual variability in the wind (both rotational and divergent) and geopotential height fields. The nine DJF's mean skill score, as measured by APC, is highest in the tropics for u, v and Φ between 600 to 900 mb where the moan APC is 0.5–0.6. There is skill in the simulation of both the rotational and divergent components of the circulation. In fact, the APC for the velocity potential at upper levels over the global domain is higher than that of the stream function. In the nine DJF's mean there is a small (APC 0.1–0.3), but significant skill in the Northern Hemisphere high latitudes in u, v and Φ at low levels. Since the explained variance is small, it leads us to conclude that in the mean, using our linear model, there is no demonstrable impact of much practical value on the extratropical circulation due to anomalous tropical heating.
On a year-by-year basis, however, the years with exceptionally large OLR perturbations in the tropics are also the years with the best model simulations (high APC over the globe as a whole and particularly in the tropics. In the ENSO DJF's of 1982/83 and 1986/87, the APC's are considerably above average in all regions including the high latitudes. The explained variance in the u and Φ fields at low levels is about 56% in the tropics and 30% in high latitudes. In thew years, the impact of the tropical heating anomalies on the remote high latitude circulation is therefore not only significant but also of practical value. The inclusion of the Hadley circulation in the model's basic state enhances the amplitude of the high latitude response, however, without improving the APC.
The model simulates the Southern Oscillation index (as measured by the Darwin–Tahiti 1000 mb height anomaly) with a correlation up to 0.9, as long as them is heating (cooling) in the lowest 300 mb in the prescribed vertical heating profile.
