Search

In this paper, the usefulness of artificial neural networks (ANNs) as a suitable tool for the study of the medium and long-term climatic variability is examined. A method for classifying the inherent variability of climate data, as represented by the rainfall regime, is investigated. The rainfall recorded at a climatological station in Cyprus over a long time period has been used in this paper as the input for various ANN and cluster analysis models. The analysed rainfall data cover the time span 1917-1995. Using these values, two different procedures were followed for structuring the input vectors for training the ANN models: (a) each 1-year subset consisting of the 12 monthly elements, and (b) each 2-year subset consisting of the 24 monthly elements. Several ANN models with a varying number of output nodes have been trained, using an unsupervised learning paradigm, namely, the Kohonen's self-organizing feature maps algorithm. For both the 1- and 2-year subsets, 16 classes were empirically considered as the optimum for computing the prototype classes of weather variability for this meteorological parameter. The classification established by using the ANN methodology is subsequently compared with the classification generated by using cluster analysis, based on the agglomerative hierarchical clustering algorithm. To validate the classification results, the rainfall distributions for the more recent years 1996, and 1998 were utilized. The respective I- and 2-year distributions for these years were assigned to particular classes for both the ANN and cluster analysis procedures. Compared with cluster analysis, the ANN models were more capable of detecting even minor characteristics in the rainfall waveshapes investigated, and they also performed a more realistic categorization of the available data. It is suggested that the proposed ANN methodology can be applied to more climatological parameters, and with longer cycles. Copyright © 2001 Royal Meteorological Society. 21 11 1401 1414 Cited By :59

Weather charts depicting the spatial distribution of various meteorological parameters constitute an indispensable pictorial tool for meteorologists, in diagnosing and forecasting synoptic conditions and the associated weather. The purpose of the present research is to investigate whether training artificial neural networks can be employed in the objective identification of synoptic patterns on weather charts. In order to achieve this, the daily analyses at 0000UTC for 1996 were employed. The respective data consist of the grid-point values of the geopotential height of the 500 hPa isobaric level in the atmosphere. A uniform grid-point spacing of 2.5° × 2.5° is used and the geographical area covered by the investigation lies between 25°N and 65°N and between 20°W and 50°E, covering Europe, the Middle East and the Northern African Coast. An unsupervised learning self-organizing feature map algorithm, namely the Kohonen's algorithm, was employed. The input consists of the grid-point data described above and the output is the synoptic class which each day belongs to. The results referred to in this study employ the generation of 15 and 20 synoptic classes (more classes have been investigated but the results are not reported here). The results indicate that the present technique produced a satisfactory classification of the synoptic patterns over the geographical region mentioned above. Also, it is revealed that the classification performed in this study exhibits a strong seasonal relationship. © Birkhäuser Verlag, Basel, 2007. 164 6-7 1347 1364 Cited By :25

Weather radars are used to measure the electromagnetic radiation backscattered by cloud raindrops. Clouds that backscatter more electromagnetic radiation consist of larger droplets of rain and therefore they produce more rain. The idea is to predict rainfall rate by using weather radar instead of rain-gauges measuring rainfall on the ground. In an experiment during two days in June and August 1997 over the Italian-Swiss Alps, data from a weather radar and surrounding rain-gauges were collected at the same time. The neural SOM and the statistical KNN classifier were implemented for the classification task using the radar data as input and the rain-gauge measurements as output. The rainfall rate on the ground was predicted based on the radar reflections with an average error rate of 23%. The results in this work show that the prediction of rainfall rate based on weather radar measurements is possible. 2 1393 1396 Sponsors: IEEE Neural Networks Society, INNS Conference code: 64098 Cited By :3

The aim of this work was to develop a system based on modular neural networks and multi-feature texture analysis that will facilitate the automated interpretation of cloud images. This will speed up the interpretation process and provide continuity in the application of satellite imagery for weather forecasting. A series of infrared satellite images from the Geostationary satellite METEOSAT7 were employed in this research. Nine different texture feature sets (a total of 55 features) were extracted from the segmented cloud images using the following algorithms: first order statistics, spatial gray level dependence matrices, gray level difference statistics, neighborhood gray tone difference matrix, statistical feature matrix, Laws texture energy measures, fractals, and Fourier power spectrum The neural network SOFM classifier and the statistical KNN classifier were used for the classification of the cloud images. Furthermore, the classification results of the different feature sets were combined improving the classification yield to 91%. 1 497 500 Sponsors: IEEE Conference code: 58800 Cited By :1

Forcasting the minimum temperature(Tmin) is one of the most important operational practices carried out by Meteorological Services worldwide. The great interest in developing methods for more accurate predictions has led the invention of several such methods of which the most commonly used can be considered to fall into three groups: physical, statistical-empirical, and numerical-statistical. In this study, an attempt is made to tackle the problem of minimum temperature forecasting using artificial neural networks. 112 114 Conference code: 15998 Cited By :9