DOI: 10.5593/sgem2017H/33/S12.050


R. Markova, S. Kohnova, J. Parajka, J. Poorova, J. Hall
Thursday 23 November 2017 by Libadmin2017

References: 17th International Multidisciplinary Scientific GeoConference SGEM 2017, www.sgemviennagreen.org, SGEM2017 Vienna GREEN Conference Proceedings, ISBN 978-619-7408-27-0 / ISSN 1314-2704, 27 - 29 November, 2017, Vol. 17, Issue 33, 399-406 pp; DOI: 10.5593/sgem2017H/33/S12.050


This study uses a pooling approach to identify catchments with similar flood characteristics in the Alpine-Carpathian region. The identification of similar regions is performed using the nonhierarchical k-means clustering algorithm and different sets of catchments characteristics (catchment area, forest cover, mean catchment elevation) and flood seasonality characteristic (concentration of flood around the mean date of flood occurrence). The optimal number of cluster is determined using the silhouette method. The analysis is conducted in the period in the period 1960-2010 using 555 gauging stations located in Slovakia (59) and Austria (496), with the length of observation ranging 30 to 51 years. The analysis is performed for two different combinations of catchments characteristics and flood seasonality characteristic. The first combination includes the mean catchment elevation, forest cover and concentration of flood around the mean date of flood occurrence (r). The second combination is the same as above and also includes the catchment area. Based on silhouette method, 2 clusters can be considered the optimal number of clusters for both combinations. Not including the catchment area into the analysis (first combination) resulted in both, a higher global average silhouette width and higher average silhouette with over each individual clusters. These larger silhouette width indicates higher confidence in the two clusters obtained. Cluster 1 is located in the lowland areas and is characterized by wide temporal spread of the flood timing (February to July). Cluster 2 is located in mountain areas and is characterized by summer floods (June and July), which have a high temporal concentration (strong seasonality).

Keywords: clustering algorithm, floods, silhouette method, Slovakia and Austria