Abstract 164

Abstract ID: 164

Forecast opportunity at subseasonal-to-seasonal timescales in the South Eastern Pacific

Lead Author: Boris Dewitte

Keywords: South Eastern Pacific, Intraseasonal Kelvin wave, S2S forecasts, ENSO, Coastal El Niño

Abstract: Eastern Boundary Upwelling Systems (EBUS) are amongst the most productive regions of the planet hosting a wealth of marine resources and concentrating economically important activities for the surrounding countries. Owing to their geographical setting they are in general under the direct influence of tropical variability through oceanic and atmospheric teleconnections, and they thus experience extreme climate events in the forms of heat or cold waves. There is currently an urgent need to predict such events sufficiently in advance for managing sensitive fisheries and marine ecosystems. As a basis for the design of an early warning system for the Central Chile region, we conduct a forecast skill assessment of sea level and sea surface temperature anomalies in the south eastern Pacific based on three resources, the S2S/C3S databases from the ECMWF platform, and a new research tool for performing climate reanalyses and seasonal-to-decadal climate predictions developed at CERFACS-CECI in the frame of the H2020-TRIATLAS project benefiting from 30 members. Considering the efficient equatorial oceanic teleconnection along the west coast of South America at a wide range of frequencies, we first evaluate skill for intraseasonal coastally-trapped Kelvin waves and SST anomalies, as well as for interannual SST variability associated with Eastern Pacific El Niño events, long-lasting La Niña events and coastal events (e.g. Chile El Niño). The systems differ the most in terms of the initialization method considering that the TRIATLAS database provide forecasts initialized using a so-called nudging method that consists in restoring the coupled model to SST observations to produce initial conditions. While this method “conserves” inherent biases of the coupled system at the initialization, it can yield improvement in forecast skill at seasonal timescales. We first illustrate the impact of the nudging method on the forecast skill showing that it is efficient in reducing both the shock at the initialization and the coupled model bias in accounting for the inter-event variability (or ENSO diversity). The analysis further documents the dependence of the sub-seasonal forecast skill in S2S as a function of the magnitude of the intraseasonal equatorial Kelvin wave activity (IEKW), showing that it is significantly improved for periods of high variance of the IEKW. We also document the relationship between the forecast skill at subseasonal timescale and that at interannual timescales in case studies. The mean state bias of the prediction systems is finally contrasted and its relationship with forecast skill is evaluated based on the ensemble forecast runs.

Lucas Glasner (CEAZA, La Serena, Chile)
Emilia Sanchez (CERFACS/CECI, Toulouse, France)
Cristian Martinez-Villalobos (Universidad Adolfo Ibañez, Santiago, Chile)
Orlando Astudillo (CEAZA, La Serena, Chile)