Climate Dynamics | 2021
Correspondence relationship between ENSO teleconnection and anomaly correlation for GCM seasonal precipitation forecasts
Abstract
Precipitation forecasts generated by global climate models (GCMs) are valuable for environmental modelling and management. At the seasonal timescale, GCMs’ forecast skills are generally attributed to low-frequency climate signals, particularly to the internal climate variability mode of El Niño-Southern Oscillation (ENSO). This study proposes a diagnostic method targeting the correlation coefficients for both Niño3.4 index and GCM forecasts in relating them to global precipitation. The similarities versus differences are identified. For the Climate Forecast System version 2 (CFSv2) forecasts, regions with significantly positive anomaly correlation corresponding to significant ENSO teleconnection are highlighted, and regions with significantly positive anomaly correlation in the absence of significant ENSO teleconnection are revealed. As December–January–February (DJF) is the peak season of ENSO, it is estimated that for every 100 grid points with significant ENSO teleconnection, ENSO teleconnection determines significantly positive anomaly correlation for 37 of them; the ratio is respectively 34/100 in March–April–May (MAM), 30/100 in June–July–August (JJA), and 38/100 in September–October–November (SON). It is found that anomaly correlation benefits substantially from ENSO teleconnection over Southern Africa, Central America/Mexico, Amazon and Central North America in DJF, Southeast Asia, North-East Brazil and Central Asia in MAM, Southeast Asia and West Coast South America in JJA, and Southeast Asia, Australia and New Zealand in SON. Limited contributions of ENSO teleconnection to anomaly correlation are observed for North Europe in DJF, West Cost South America and East North America in MAM, North-East Brazil and Central America/Mexico in JJA, and Amazon in SON. Overall, the proposed method contributes to disentangling the sources of GCM forecast skill of seasonal precipitation and fostering better seasonal forecast benchmarking and diagnosis.