Contrasting Feedback Attribution of Atmospheric and Surface Temperature Anomalies Related to Eastern Pacific ENSO and Central Pacific ENSO

Different feedback processes cause distinct spatial patterns of atmospheric and surface temperature anomalies in boreal winter between two types of El Nino–Southern Oscillation (ENSO): Eastern-Pacific (EP) type and Central-Pacific (CP) type. The decomposition of feedback processes is achieved by a climate feedback–response analysis method (CFRAM), which decomposes local temperature anomalies into partial temperature anomalies due to oceanic dynamics/heat storage, clouds, atmospheric dynamics, water vapor, ozone, and surface albedo. The sum of these diagnosed partial temperature anomalies is equal to the observed temperature anomaly of each type of ENSO. For both CP and EP types of ENSO, the oceanic dynamics/heat storage term and the water vapor term are positive, amplifying the surface temperature anomalies in the tropical Pacific, whereas atmospheric dynamics and cloud feedbacks are negative. Ozone and surface albedo terms are small. Their feedback attributions are opposite to local temperature anomalies in the tropical Pacific. The different feedback attributions of these two types of ENSO imply that the difference between the EP ENSO and the CP ENSO is mainly caused by oceanic dynamics, atmospheric energy transport and cloud forcing, especially longwave cloud forcing. Compared with the CP type, the EP ENSO has stronger ocean dynamical positive forcing in the entire air column over the tropical and subtropical Pacific. The cloud longwave feedback for the EP type can induce more intensive tropical warming and cooling over the entire extratropics in both hemispheres. In contrast, when the EP type of ENSO happens, cloud shortwave feedback drives a weaker tropospheric cooling over the southern extratropics.