A data analysis system to extend the coverage capacity of meteorological stations for flood forecasting
Forecast Data provided by meteorological stations (MS) are crucial for Flood Forecasting Systems (FFS). These data are mainly related to temperature and precipitation. However, having enough MS to produce paramount of such data is challenging due to the high cost of their set up as well as their maintenance. As a consequence, it is almost impossible to get flood predictions in some regions due to the lack of meteorological forecast data. One solution to overcome such a drawback is to envision extending the data validity of a given area to another one. That is, we aim at using a MS of region A for estimating data we may have in region B if ever it had its own MS. In this respect, we propose an extension of MS forecast capacity by introducing a data analysis system based on a linear correlation technique. The system uses data collected from sensors networks installed on a given area not covered by a MS with data from a reference area that has a MS. Afterwards, it checks whether there is a linear correlation between the data of the two zones. In the affirmative case, a correlation function is deduced between the two areas and will be used for estimating data of the area without a MS. The results obtained from empiric experiments show the feasibility of our approach and its benefits.
Auteur(s) : Joel Tanzouak, Ndiouma Bame, Blaise Yenke, Idrissa SARR
Pages : 92-97
Année de publication : 2018
Revue : Conference on Geoinformatics and Data Analysis (ICGDA '18).
Type : Article
Mise en ligne par : SARR Idrissa