Temperature and precipitation seasonal forecasts over the Mediterranean region: added value compared to simple forecasting methods

Filippo Calì Quaglia,Jost Von Hardenberg,Silvia Terzago

doi:10.1007/s00382-021-05895-6

Abstract

This study considers a set of state-of-the-art seasonal forecasting systems (ECMWF, MF, UKMO, CMCC, DWD and the corresponding multi-model ensemble) and quantifies their added value (if any) in predicting seasonal and monthly temperature and precipitation anomalies over the Mediterranean region compared to a simple forecasting method based on the ERA5 climatology (CTRL) or the persistence of the ERA5 anomaly (PERS). This analysis considers two starting dates, May 1st and November 1st and the forecasts at lead times up to 6 months for each year in the period 1993–2014. Both deterministic and probabilistic metrics are employed to derive comprehensive information on the forecast quality in terms of association, reliability/resolution, discrimination, accuracy and sharpness. We find that temperature anomalies are better reproduced than precipitation anomalies with varying spatial patterns across different forecast systems. The Multi-Model Ensemble (MME) shows the best agreement in terms of anomaly correlation with ERA5 precipitation, while PERS provides the best results in terms of anomaly correlation with ERA5 temperature. Individual forecast systems and MME outperform CTRL in terms of accuracy of tercile-based forecasts up to lead time 5 months and in terms of discrimination up to lead time 2 months. All seasonal forecast systems also outperform elementary forecasts based on persistence in terms of accuracy and sharpness.

Highlights

Seasonal forecasts of atmospheric variables like near-surface air temperature and precipitation are attractive for a variety of applications in different economic and socially relevant sectors, including hydropower and wind energy production (Torralba et al 2017; Clark et al 2017), management of water resources (Svensson et al 2015), fire risk, agriculture, transports (Palin et al 2016) and shipping, health (Lowe et al 2017), and in general, hazardous weather events which can cause serious economic damages (Morss et al 2008)
The main question which we addressed in this study is whether the most advanced seasonal forecast systems, or multi-model ensembles based on them, outperform elementary forecast approaches: (i) a climatological forecast (CTRL), which has been set as the benchmark; (ii) a persistence forecast (PERS) based on the persistence of ERA5 anomalies at lead times after the first month
Moving from the evaluation of the accuracy of tercilebased forecasts to the evaluation of the accuracy of the forecast distribution, in this second case we find that the improvement with respect to the climatological forecast is evident for lead time 0, limited up to lead time 2 and null for summer precipitation for selected models

Summary

Introduction

Seasonal forecasts of atmospheric variables like near-surface air temperature and precipitation are attractive for a variety of applications in different economic and socially relevant sectors, including hydropower and wind energy production (Torralba et al 2017; Clark et al 2017), management of water resources (Svensson et al 2015), fire risk, agriculture, transports (Palin et al 2016) and shipping, health (Lowe et al 2017), and in general, hazardous weather events which can cause serious economic damages (Morss et al 2008) In all these cases, a reliable indication of mean climate conditions a few months ahead can be associated with a well defined economic value (Bruno Soares et al 2018, and references therein). While the prediction of ENSO is quite good, issues in the lower stratosphere and at the tropopause are reported, which could influence the ability to extend forecast skill to the extra-tropics. Dunstone et al (2016) and Scaife et al (2014) focused their work on the predictability of the North Atlantic Oscillation, which profoundly influences North American and European winter climate, finding the UKMO and the HadGEM3-GC2 models to have skill in NAO prediction up to the following season

Results

Discussion

Conclusion