Features and causes of catastrophic floods in the Nemunas River basin

The Nemunas River basin falls within the territories of five different countries – Belarus, Lithuania, Russia, Poland and Latvia. In general, the beginning of spring floods highly depends on rapid rise of air temperature, heavy precipitation and sudden snow melting in the analysed basin. In this paper, the conditions of formation and consequences of two catastrophic floods in 1958 and 1979 in the Nemunas River basin were studied regarding the hydrometeorological parameters (maximum snow water equivalent before the beginning of flood and precipitation amount during the flood) as well as runoff coefficients for each selected catastrophic flood. Differences between the main drivers and evolution of these floods were analysed. Spatial distribution of maximum snow water equivalent and precipitation, as well as runoff coefficient in different parts of the river basin, were identified as having the most significant impact on the formation of the studied catastrophic floods. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/). doi: 10.2166/nh.2019.147 om http://iwaponline.com/hr/article-pdf/51/2/308/682255/nh0510308.pdf 021 Vytautas Akstinas Diana Meilutytė-Lukauskienė (corresponding author) Jūratė Kriaučiūnienė Diana Šarauskienė Laboratory of Hydrology, Lithuanian Energy Institute, Breslaujos st. 3, LT-44403 Kaunas, Lithuania E-mail: diana.meilutyte-lukauskiene@lei.lt


INTRODUCTION
River basin are available (Kolupaila ). Some information on water resources dynamics in the Nemunas River basin, which influenced extreme events, can be found in Korneev et al. (). Therefore, the aim of this research is to analyse the main drivers and conditions of the formation of two (1958 and 1979) catastrophic floods in the Nemunas River basin and to find the most unfavourable combination of hydrological and meteorological factors which may cause catastrophic floods in river catchments of Eastern Europe.

STUDY AREA AND DATA
The Nemunas River is the 14th longest river in Europe and The evaluation of impact of meteorological and hydrological parameters on formation of the catastrophic floods was carried out using long-term series of daily discharge data from 12 WGS (Table 1), as well as data of daily precipitation (P, mm), monthly air temperature (T, C) and decadal (i.e., measured every tenth day) snow water equivalent (SWE, mm) from 58 MS.

METHODS
The beginning of spring flood in the Nemunas River basin mainly depends on climatic conditions (air temperature, precipitation, snow melting), whereas the end of flood may be influenced by many different elements (such as size, form and slopes of the basin, snow reserve in the basin, density of river network, etc.). In Figure 2, a scheme of flood formation in the Nemunas River is displayed, where the beginning of spring flood with abrupt increase of discharges, its course and culmination is presented. Sudden increase of Then, intensive snow melting causes abrupt increase of the water level in the river and these conditions give rise to the spring flood in the river basin.
A general scheme of the research methodology is presented in Figure 3. The first step in this research was probability distribution analysis in order to find out the floods of rare probability in the selected river basin. After selection of catastrophic floods, assessment of the main hydrological (water level during the flood (h flood , cm), daily discharge during the flood (Q flood , m 3 s À1 )) and meteorological (maximum snow water equivalent before the flood (SWE max , mm) precipitation amount during the flood (P flood , mm) and air temperature (T, C)) characteristics was carried out. In the following step, the number of inves-

SWAT (Soil and Water Assessment Tool) Baseflow
Filter (BF) program (https://swat.tamu.edu/software/) was used to separate a part of the groundwater feeding and surface runoff from the data of historical observations (i.e., hydrograph). This software provides an opportunity to inves- Calculations of hydrometeorological variables. In the Nemunas River basin, V res consists of water from SWE max together with P flood , which determine V flood . Volume of the selected flood was calculated by using the equation: where Q surface i is daily discharge during the flood without groundwater feeding (estimated from output of SWAT BF) (m 3 s À1 ), t is daytime (s), i is from 1 to n, n is flood duration expressed by days (the time period from the beginning until the end of spring flood, i.e., from the first day of sudden increase of hydrograph until the last day of sharp decrease in hydrograph after maximum discharge of the spring flood).

The volume of water resources V res in the Nemunas
River basin was calculated as: where H SWEmax is average height of maximum snow water equivalent (calculated from area of the whole basin) before the beginning of the spring flood (mm), which was selected from decadal (i.e., measured every tenth day) data of snow water equivalent, H P flood is average height of precipitation amount (calculated from area of the whole basin) during the spring flood (mm), which was calculated by assessing the time period from abrupt rise of the river discharge until the maximum peak of catastrophic flood, S basin is area of selected basin (km 2 ). Average heights of SWE max and P flood were estimated from the isoline maps.
In analysis of surface runoff processes, the runoff coefficient, an important input parameter in hydrologic modelling, characterised as the ratio of runoff volume and rainfall volume, is widely used. In the present study, the runoff coefficient ŋ is defined as a portion of accumulated water resources that directly becomes a part of the volume of catastrophic flood. A runoff coefficient was calculated for each WGS catchment (ŋ WGS ) and showed the ratio between V WGSflood and accumulated water resources from SWE max and P flood (V WGSres ):   During this flood, the Belorussian city of Grodno suffered     and the monthly amount of precipitation from December (1957 and1978) to April (1958 and1979)  SWE max (Figure 9(a)) in 1958 and smaller input of water from P flood (Figure 10

SUPPLEMENTARY MATERIAL
The Supplementary Material for this paper is available online at https://dx.doi.org/10.2166/nh.2019.147.