Argonne National Laboratory

LPB Hydro-Climate Variability as Simulated by GCM Experiments: Role of Remote SST Forcing

TitleLPB Hydro-Climate Variability as Simulated by GCM Experiments: Role of Remote SST Forcing
Publication TypeJournal Article
Year of Publication2013
AuthorsCherchi, A, Carril, AF, Zamboni, L, Menendez, C
JournalClimate Dynamics
Date Published04/2012
Other NumbersANL/MCS-P3016-0712

A set of AMIP-type experiments is computed and analyzed to study springtime hydro-climate variability in the region of La Plata Basin (LPB). In particular, an ensemble of nine experiments with same interannually varying SST, as boundary forcing, and different initial conditions is used to investigate the relative role of the Pacific, Indian and Atlantic tropical oceans on modulating the local precipitation. The AMIP type ensemble results have been compared with a coupled model experiment (using the same atmospheric component). The comparison reveals that the model has a good performance in the simulation of precipitation over LPB and South America, with a slight overestimation of the seasonal mean and an underestimation of the variability. Nevertheless, an EOF analysis of South America precipitation shows that the model is able to realistically reproduce the dominant modes of variability in spring. Further, its principal component (PC1) when correlated with global SST and atmospheric fields identifies the pattern related to ENSO and the large-scale connections. Overall the teleconnection pattern in the tropical and Southern Pacific Ocean is well captured by the SST-forced ensemble, but it is absent or too weak in other oceanic areas. In the subtropical South Atlantic the correlation is more realistic in the coupled model experiment suggesting the importance of air-sea feedbacks for that region, even at lower than interannual timescales. When the composite analysis of SST and atmospheric fields is done only over the ensemble members having a PC1 in agreement with the observations, both in terms of sign and intensity, then the correspondence between model and data is much improved. The improvement relies on avoiding climate noise by averaging only over members that are statistically similar and it suggests a high level of uncertainty due to internal atmospheric variability. Some individual springs have been analyzed as well. In particular, 1982 represents a clean case with a clear wave train propagating from the central Pacific and merging with a secondary one from eastern tropical South Indian Ocean, and it corresponds to a strong El Nino. Another case, 2003, corresponds to a rainy spring for SESA but in this case the en-semble mean does not exhibit any teleconnection through the South Pacific and it is not able to reproduce the correct local precipitation pattern, suggesting that in this case regional effects are more important than remote forcing.