Abstract
AbstractMultimodel combinations are a well‐established methodology in weather and climate prediction and their benefits have been widely discussed in the literature. Typical approaches involve combining the output of different numerical weather prediction (NWP) models using constant weighting factors, either uniformly distributed or determined through a prior skill assessment. This strategy, however, can lead to suboptimal levels of skill, as the performance of NWP models can vary with time (e.g., seasonally varying skill, changes in the forecasting system). Moreover, standard combination methods are not designed to incorporate predictions derived from sources other than NWP systems (e.g., climatological or time‐series forecasts). New algorithms developed within the machine learning community provide the opportunity for “online prediction” (also referred to as “sequential learning”). These methods consider a set of weighted predictors or “experts” to produce subsequent predictions in which the combination or “mixture” is updated at each step to optimize a loss or skill function. The predictors are highly flexible and can combine both NWP and statistically derived forecasts transparently. A set of these online prediction methods is tested and compared with standard multimodel combination techniques to assess their usefulness. The methods are general and can be applied to any model‐derived predictand. A set of weather‐sensitive European country‐aggregate energy variables (electricity demand and wind power) is selected for demonstration purposes. Results show that these innovative methods exhibit significant skill improvements (i.e., between 5 and 15% improvement in the probabilistic skill) with respect to standard multimodel combination techniques for lead weeks up to 5. The incorporation of statistically derived predictors (based on historical climate data) alongside NWP forecasts is also shown to contribute significant skill improvements in many cases.
Highlights
Subseasonal-to-seasonal predictions, which range from a few weeks to a few months ahead, fill in the gap between weather prediction and seasonal forecasting
Given the small size of the ensemble but the frequent launch dates, a lagged ensemble is used to obtain a dataset with a structure and size comparable to European Centre for Medium-range Weather Forecasting (ECMWF) prior to the multimodel combination: the daily starts of National Centers for Environmental Prediction (NCEP) were subsampled biweekly to match the ECMWF starts (Mondays and Thursdays) and each of those starts was combined with the two preceding ones to generate a larger number of members (12 instead of 4)
To illustrate the aggregation methods and their evaluation, we first consider the case of United Kingdom (UK) electricity demand
Summary
Subseasonal-to-seasonal predictions (hereafter s2s), which range from a few weeks to a few months ahead, fill in the gap between weather prediction and seasonal forecasting. Within the realm of machine learning, a family of algorithms has been developed to perform “online prediction with expert advice” (Cesa-Bianchi and Lugosi, 2006), known as sequential learning or sequential aggregation rules These methods consider a set of predictors which, after initially being weighted uniformly, produce subsequent predictions in which the combination or “mixture” is updated continually over time to optimize a loss or skill function. Strobach and Bel (2015, 2016) first applied a similar method in climate forecasting, in the context of decadal prediction of atmospheric variables such as 2-m temperature They found that the exponentiated gradient algorithm was able to outperform the individual models and other standard benchmarks such as linear combinations and the climatology. Within the context of the EU-Horizon 2020 S2S4E project (Subseasonal-to-seasonal Forecasting for Energy1), these novel methods are applied to forecasts of national-average electricity demand and wind-power generation across a set of European countries, though the methodology can readily be applied to other s2s forecast properties
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
