1,721,008 research outputs found
Automatic input variable selection for analog methods using genetic algorithms
Full text linkAnalog methods (AMs) are statistical downscaling methods often used for precipitation prediction in different contexts, such as operational forecasting, past climate reconstruction of climate change impact studies. It usually relies on predictors describing the atmospheric circulation and the moisture content of the atmosphere to sample similar meteorological situations in the past and establish a probabilistic forecast for a target date. AMs can be based on outputs from numerical weather prediction models in the context of operational forecasting or outputs from climate models in climatic applications.AMs can be constituted of multiple predictors organized in different subsequent levels of analogy that refines the selection of similar situations. The development of such methods is usually a manual process where some predictors are assessed in different structures. As most AMs use multiple predictors, a comprehensive assessment of all combinations becomes quickly impossible. The selection of predictors in the application of the AM often builds on previous work and does not evolve much. However, the climate models providing the predictors evolve continuously and new variables might become relevant to be considered in AMs. Moreover, the best predictors might change from one region to another or for another predictand of interest. There is a need for a method to automatically explore potential variables for AMs and to extract the ones that are relevant for a predictand of interest.We propose using genetic algorithms (GAs) to proceed to an automatic selection of the predictor variables along with all other parameters of the AM. We even let the GAs automatically pick the best analogy criteria, i.e. the metric that quantifies the analogy between two situations. The first test consisted of letting the GAs select the single best variable to predict daily precipitation for each of 25 selected catchments in Switzerland. The results showed great consistency in terms of spatial patterns and the underlying meteorological processes. Then, different structures were assessed by varying the number of levels of analogy and the number of variables per level. Finally, multiple optimizations were conducted on the 25 catchments to identify the 12 variables that provide the best prediction when considered together
Analogue Methods and ERA5: Benefits and Pitfalls
Full text linkPerfect prognosis statistical downscaling relies on the statistical relationships established using observational data for predictands and predictors. Predictors are often retrieved from reanalyses, which are considered pseudo-observations. The impact of the choice of a reanalysis dataset on the performance of the downscaling method is usually overlooked, as global reanalyses are frequently assumed to be equivalent for the last few decades and data-rich regions such as Europe. However, it was recently shown that the reanalysis dataset can have a bigger impact on the method skill than the choice of predictor variables. Generally, reanalyses processed by more recent atmospheric models assimilate more data and perform best.
This work is aimed at assessing the extent of potential gains from the use of ERA5, following its release, compared to other global reanalyses. The assessment was carried out using six variants of analog methods, which are statistical downscaling techniques, to predict daily precipitation at 301 stations across Switzerland. ERA5 proved to be one of the best performing reanalyses across the different analog methods. Due to data availability, we recommend using 20CR for applications starting between 1851 and 1900, CERA-20C for those between 1900 and 1950, and ERA5 for applications after 1950.
However, ERA5 high spatial resolution (0.25°) turned out to be a trap for simple calibration techniques. The domains over which the predictor fields are compared need to be optimized, and high-resolution grids come along with numerous sub-optimal local solutions. An enhanced calibration procedure, thus, must be used. Besides the risk of poorly-calibrated domains, the high resolution also requires much higher computational time with no gain in skill, provided that the predictors considered are relevant at a synoptic scale. Although ERA5 should be the dataset of choice, its use at a lower resolution to predict daily precipitation should provide equivalent performance
- …
