Anomaly Nesting: A Methodology to Downscale Seasonal Climate Simulations from AGCMs

In this paper a methodology is proposed to downscale coarse-resolution atmospheric general circulation model (AGCM) seasonal simulations. Anomaly nesting involves replacing the climatology of the driving AGCM with observed (in this case the National Centers for Environmental Prediction reanalysis) climatology at the lateral boundaries of the nested regional climate model (the regional spectral model). In this study the methodology is tested over South America and the neighboring ocean basins. A comparison of the austral summer seasonal simulation with the conventional way of nesting, namely driving the regional model with full AGCM forcing, reveals that substantial gains in the deterministic skill are realized through anomaly nesting. It is also shown that the high-frequency variance (at 3–30- and 30–40-day time scales) is more realistic from the anomaly nesting procedure.