A0185
Title: Global and regional long-term climate forecasts: A heterogeneous future
Authors: Lola Gadea - University of Zaragoza (Spain) [presenting]
Jesus Gonzalo - Universidad Carlos III de Madrid (Spain)
Abstract: Climate is a long-term issue; therefore, climate forecasts should be long-run. Such forecasts are crucial for designing the mitigation policies required to fulfil one of the main objectives of the Paris Climate Agreement (PCA) and design adaptation policies that mitigate the adverse effects of climate change. They also serve as an indispensable instrument to assess climate risks and successfully steer the green transition. The aim is to propose a simple method to produce long-term temperature density forecasts from observational data using the realized quantile methodology introduced in an existing study, where unconditional quantiles are converted into time series objects. They complement the projections obtained by physical climate models, mainly focused on the mean temperature. These averages usually conceal wide spatial disparities, which, among other distributional characteristics, are captured by our density forecast. The proposed method consists of running an out-of-sample forecast model competition and combining the estimates of the resulting Pareto-superior models to eliminate the forecast model dependency. Furthermore, the approach considers climate change a non-uniform phenomenon; therefore, analyzing it from a regional perspective is crucial, offering different predictions for a heterogeneous future.
