Search

The increase in hydrological extremes during the last two decades has had a significant impact on natural and social environments. These hydrological extremes depend greatly on changes in the meteorological parameters. The task of this research was to evaluate the impact of meteorological factors (snow water equivalent and heavy rainfall) on the formation of spring floods in the basins of the Nemunas, Lielupė and Venta rivers. Five Lithuanian rivers (Venta, Šešuvis, Mūša, Merkys and Žeimena) from these basins were analysed in detail. These rivers fall within the Western, Central and Southeastern hydrological regions of Lithuania. Long‐time‐series data for daily discharge, precipitation and thickness of snow cover from 12 meteorological and five hydrological gauging stations were used. The evaluation of the relation between these factors was carried out for two periods: 1961–1987 and 1988–2014. The relation between the maximum discharge of the spring flood, the maximum snow water equivalent before the flood and the precipitation amount 10 days before the flood was analysed by multiple regression analysis. The high correlation co‐efficients between the observed and predicted maximum discharges of the spring flood for the created regression models fluctuated between 0.63 and 0.86. The verification of the selected regression model was performed with Hydrologiska Byråns Vattenbalansavdelning (HBV) software, which also showed a high correlation co‐efficient (0.71). The applied methodology of this research could be used for better perception of flood‐formation consequences in different hydrological regions of Europe. The multiple regression analysis of the maximum discharge of spring flood (Qmax, m3/s) conducted with the maximum snow water equivalent before the flood (SWEmax, mm) and the precipitation amount 10 days before the flood (P10, mm) was carried out for five rivers in Lithuania. The reliability of the regression models shows high correlation co‐efficients (0.63–0.86) between the observed and predicted Qmax. To verify the regression models, hydrological modelling with Hydrologiska Byråns Vattenbalansavdelning (HBV) software was performed. The analysis revealed close correlation co‐efficients (0.71–0.78) related to the observed Qmax (Obs), the modelled Qmax of the HBV (Mod), the predicted Qmax of the regression model of the reference period (RegRef), the predicted Qmax of the regression model of 1961–1987 (Reg1), the predicted Qmax of the regression model of 1988–2014 (Reg2), and their range for the verification period (1986–2005). [ABSTRACT FROM AUTHOR]

The climate change impact on water resources is observed in all the Baltic States. These processes became more evident in the last decades. Although the territory of the Baltic States (Lithuania, Latvia, Estonia) is not large (175 000 km²), the climatic differences are quite considerable. We performed a regionalization of the territory of the Baltic States depending on the conditions of river runoff formation which can be defned according to percentages of the river feeding sources (precipitation, snowmelt, groundwater). Long-term series of temperature (40 stations), precipitation (59 stations) and river discharge (77 stations) were used to compose ten regional series. This paper addresses: (1) variability in long-term regional series of temperature, precipitation and river discharge over a long period (1922-2007); (2) changes in regional series, comparing the periods 1991-2007 and 1931-1960 with the reference period (1961-1990), and (3) the impact of temperature and precipitation changes on regional river discharge. [ABSTRACT FROM AUTHOR]

Recent advances in deep learning, especially the long short-term memory (LSTM) networks, provide some useful insights on how to tackle time series prediction problems, not to mention the development of a time series model itself for prediction. Runoff forecasting is a time series prediction problem with a series of past runoff data (water level and discharge series data) as inputs and a fixed-length series of future runoff as output. Most previous work paid attention to the sufficiency of input data and the structural complexity of deep learning, while less effort has been put into the consideration of data quantity or the processing of original input data—such as time series decomposition, which can better capture the trend of runoff—or unleashing the effective potential of deep learning. Mutual information and seasonal trend decomposition are two useful time series methods in handling data quantity analysis and original data processing. Based on a former study, we proposed a deep learning model combined with time series analysis methods for daily runoff prediction in the middle Yangtze River and analyzed its feasibility and usability with frequently used counterpart models. Furthermore, this research also explored the data quality that affect the performance of the deep learning model. With the application of the time series method, we can effectively get some information about the data quality and data amount that we adopted in the deep learning model. The comparison experiment resulted in two different sites, implying that the proposed model improved the precision of runoff prediction and is much easier and more effective for practical application. In short, time series analysis methods can exert great potential of deep learning in daily runoff prediction and may unleash great potential of artificial intelligence in hydrology research. [ABSTRACT FROM AUTHOR]