Analysis and downscaling multi-model seasonal forecasts in Peru using self-organizing maps

Abstract

We present an application of self-organizing maps (SOMs) for analysing multi-model ensemble seasonal forecasts from the DEMETER project in the tropical area of Northern Peru. The SOM is an automatic data-mining clustering technique, which allows us to summarize a high-dimensional data space in terms of a set of reference vectors (cluster centroids). Moreover, it has outstanding analysis and visualization properties, because the reference vectors can be projected into a two-dimensional lattice, preserving their high-dimensional topology. In the first part of the paper, the SOM is applied to analyse both atmospheric patterns over Peru and local precipitation observations at two nearby stations. First, the method is applied to cluster the ERA40 reanalysis patterns on the area of study (Northern Peru), obtaining a two-dimensional lattice which represents the climatology. Then, each particular phenomenon or event (e.g. El Niño or La Niña) is shown to define a probability density function (PDF) on the lattice, which represents its characteristic 'location' within the climatology. On the other hand, the climatological lattice is also used to represent the local precipitation regime associated with a given station. For instance, we show that the precipitation regime is strongly associated with El Niño events for one station, whereas it is more uniform for the other. The second part of the paper is devoted to downscaling seasonal ensemble forecasts from the multi-model DEMETER ensemble to local stations. To this aim, the PDF generated on the lattice by the patterns predicted for a particular season is combined with the local precipitation lattice for a given station. Thus, a probabilistic or numeric local forecast is easily obtained from the resulting PDF. Moreover, a measure of predictability for the downscaled forecast can be computed in terms of the entropy of the ensemble PDF. We present some evidence that accurate local predictions for accumulated seasonal precipitation can be obtained some months in advance for strong El Niño episodes. Finally, we compare the multi-model ensemble with single-model ensembles, and show that the best results correspond to different models for different years; however, the best global performance over the whole period corresponds to the multi-model ensemble.

CHAC: a weather pattern classification system for regional climate downscaling of daily precipitation

“…Solman and Nuñez 1999;Gutiérrez et al 2004) have developed and used statistical downscaling methods to predict daily precipitation or temperature at local scale (weather station) in South America. The present study is a contribution to fill this gap.…”

“…In the following, we will compare the performance of two often-used techniques: k-means (Gutiérrez et al 2004) and the SOM (Self-Organizing Maps; Hewitson and Crane 2006). While other methods (analogue and k-nearest neighbours) have also been implemented, we only discuss these two clustering methods, which gave the best results.…”

“…The starting point of the statistical downscaling methodology is mainly based on the previous work by Gutiérrez et al (2004). This methodology makes use of a clustering technique to connect atmospheric fields with time series of observed daily precipitation or temperature at local stations.…”

“…Statistical techniques have proven to be as successful as dynamical techniques, being also far less computationally expensive and simpler to develop (e.g. Gutiérrez et al 2004;Hewitson and Crane 2006). However, their major drawbacks are their tendency to under estimate extreme precipitation events, and their incapacity in predicting non-stationary behaviours in the transfer function (especially under climate and land-use/land-cover changing conditions).…”

“…However, such an approach is not suitable for daily precipitation given that a linear relationship between precipitation and weather patterns is way too far from reality due to the stochastic nature of daily precipitation. For this reason, other SDS approaches have been taken such as classification and regression trees (CART; Hughes et al 1993) and clustering techniques (k-means, Gutiérrez et al 2004; neural networks, Hewitson and Crane 2006; analogues and nearest-neighbours; Mehrotra and Sharma 2005). Data clustering is a non-linear technique that covers fields of statistical analysis, machine learning, data mining, patter recognition, neural networks, and fuzzy theory (e.g.…”