题 目：On the distinct U.S. impacts and predictability of El Nino flavors
报 告 人: 张韬研究员
单 位: Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, and Physical Sciences Division, NOAA/Earth System Research Laboratory (ESRL)
地 点：3号楼 717会议室
Understanding and characterizing the different effects in spatial expressions of El Nino-related sea surface temperatures (SSTs) may be useful for enhancing overall US forecast skill relative to simply specifying a single (linear) impact pattern. However, the impacts of two types of El Nino on U.S. climate remain elusive, which could be due to the small sample size used in observations or the dependency of response in modeling approaches.
Empirical indication for distinct US impacts of different El Nino flavors during boreal winter is tested using a large ensemble of historical atmospheric model simulations forced by observed SST variability during 1980-2018. The ability to predict such flavors and their US impacts is further examined by diagnosing coupled model seasonal forecasts spanning the last three decades. The ensemble mean of atmospheric model simulations indicates that the forced signal of US surface air temperature during two types of El Nino, so-called Central Pacific (CP) and East Pacific (EP) events, show a considerable difference in the sign of anomalies over many US regions, with generally more widespread cold anomalies during EP events. And whereas the US precipitation patterns of EP and CP events are found to be more similar than their temperature patterns, EP events yield considerably wetter conditions over the contiguous US as a whole. The model confirmation of the empirical indications for colder/wetter US winters during EP El Nino compared to CP is shown to be robust across various atmospheric models, and a sensitivity is also discovered in initialized coupled model forecast patterns, at least at short lead times.
Concerning the predictability of these distinct US impacts of El Nino flavors during winter, analysis of coupled model seasonal forecasts exhibits considerable realism and skill at 3~4 months lead times. At long leads, the predictions studied herein fail to distinguish the El Nino flavors. Further analysis reveals that the forecast failure of El Nino flavors is mostly symptom of too few CP events compared to observations. We speculate that this forecast bias is linked to the climatological biases in the mean tropical Pacific SSTs, with the conjecture biases in ocean dynamics linked to a weakened zonal SST gradient inhibiting the generation of the CP El Nino events.