Recent development in automatic monitoring systems affected in exponential growth of collected data. Simple methods of storage and spatial analysis of hydrometeorological phenomena are not efficient when dealing with big datasets. This paper presents results of research on databases structure for collection of high temporal density data (10 minutes) from several sensors installed on monitoring stations. Number of stations (from 5 to 50) was fitted to conditions of medium size county in Poland. Case study was performed for Żywiec County located in Southern Poland. Analysis of different relational database structures for typical scenarios of data request is presented. Optimisation criteria include i.a. request time and database size. At the databases design phase two main factors should be taken into account. First one is time of request, related to crisis management when fast access to data must be ensured. Second aspect is database size, related to planning and research phase when wide scope of data should be accessible and time in not so important. Some methods of database efficiency improvement, not related directly to database structure, are also presented. Two databases were used for comparison (MySQL and PostgreSQL) and their efficiency for different scenarios was tested and described.
Keywords: database, hydrometeorological monitoring, crisis management