LONG-TERM DEVELOPMENT OF DISCHARGE AND NITRATE CONCENTRATIONS IN THE LITTLE CARPATHIANS HEADWATERS

One of the requirements imposed by the Water Framework Directive 2000/60/EC (WFD, 2000) is to analyze and predict long-term evolution of surface water quality parameters. During 28-years period (1991–2018), the concentrations of selected pollutants were monitored in the Little Carpathians headwater basins by the Institute of Hydrology, SAS. In this study we analyse the long-term development of runoff and nitrates nitrogen concentrations in the Parná River at Horné Orešany water gauge station during the period 1991–2018. Discharges in the Parná River decreased slightly, but the trend is not statistically significant. In the case of nitrate nitrogen concentrations the marked decrease occurred in this river basin from the value of 5.11 mg l during the years 1991–1995 to the value of 2.49 mg l in the years 2015–2018. The relation between discharge and nitrate concentration was used to derive exponential empirical relations for estimation of the nitrate nitrogen concentrations in the stream based on mean daily discharge during the sampling day. These equations can be used for indirect estimation of nitrate nitrogen concentrations when they were not measured directly.


Introduction
One of the negative consequences of the expected global warming will be the increase of stream water temperature and runoff decrease during summer season, which will result in stream water quality worsening. The intensification of the agricultural mass production and increase of the inorganic nitrate fertilization doses worldwide (also in Slovakia) after the Second World War influenced the increase of nitrate concentrations in the stream water. Nitrate loading from an agricultural basin is generally strongly related to the amount of fertilizers applied. On Fig. 1 there is presented the annual development of average consumption of industrial nitrogen fertilizers applied to 1 ha of agricultural soils in Slovakia during the period 1960-2015 (upper). The lower part of the figure shows the monthly nitrate concentrations in three streams with long monitoring (Pekárová and Miklánek, 2013). In the plot we can see the direct influence of the increased nitrate doses upon the nitrate concentrations in the rivers. The fertilizer consumption per hectare arable land in 2012 in other EU countries published by the World Bank (WB, 2020) is plotted in Fig. 2 for comparison. This figure equates to 151 kg of fertilizers consumed per ha arable land on average for the EU countries. Influence of agricultural activity on the nitrate content in drinking water sources in Slovakia was studied by Hyánková et al (1995). Pavelková (2000) assessed the water quality in drainage channels in the Eastern Slovakian Lowland in [1994][1995][1996]. She demonstrated the insufficient water quality, mainly due to nitrites nitrogen concentrations. The long-term development of nitrate in house wells in the villages of Michalovce and Sobrance District was analysed by Pavelková and Babinec (2017). The intensive daily monitoring of nitrate concentrations was managed by IH SAS in 1986 in order to make detailed analysis of nitrate development at the inflow to water reservoir Veľká Domaša. Five sampling sites were established in the Ondava basin, and some other also in the Little Carpathians and Strážov Highland. It was proved by the study that the lowering of fertilisation doses and of the agricultural production resulted in lowering of nitrate concentrations in surface streams (Pekárová and Velísková, 1998;Pekárová and Pekár, 1996;Pekárová and Miklánek, 2013;Velískova et al., 2012). The influence of the forested and arable basin on nitrate concentrations was studied at experimental microbasins of IH SAS near to Kúnovec in the Strážov highland in 1986-2006(Pekárová et al., 2008. The nitrate concentrations in the small forested Vydrica basin near to Bratislava were studied by Pekárová and Pekár (1998), Pekárová and Onderka (2005), and Sebíň et al. (2007). The application of fertilizers was at low level in Slovakia in 2012 compared to other EU countries (Fig. 2), but due to increasing fertilization in Slovakia it is necessary to continue in evaluation of the nitrate content in the streams. Annual development of average consumption of industrial nitrogen fertilizers applied to 1 ha of agricultural soils in Slovakia during the period 1960-2015 (upper) in kg per 1 ha of agricultural land. Monthly nitrate nitrogen concentrations of the Hron, Ondava, and Váh rivers (SHMI data, according to Pekárová and Miklánek, 2013  .

Material
The Little Carpathians lie in the Northeast of Bratislava between 17°-17°50' Eastern Longitude and 48°10'-48°40' Northern Latitude. The Little Carpathians are affected by air streams from the agriculturally polluted part of the Western Slovakia and Southern Moravia, which are densely populated areas. Both, the intensive industry and agriculture have a negative influence on water quality in this region. For example, the total volume of waste water discharged from Chemolak Smolenice was 477.3 thousand m 3 in 1991, with BOD5 of 12.5 t.y -1 (tons per year), COD of 38.1 t.y -1 and 150.3 t.y -1 of insoluble substances (Molnár et al., 1993). The intensive agriculture in the lowland area has great negative impact on high nitrates content in surface water. Most local settlements (incl. Trnava) are exclusively supplied by drinking water from local groundwater sources. Because these sources are irreplaceable, it was necessary to give attention to the water quality problem.
Surface and groundwater quality in the basin of Trnávka River has been studied by several authors in the past. Gálik (1990) was interested in the impact of human activities on groundwater quality. He found that chloride, sulphate, and especially nitrate concentrations increase permanently in the groundwater. Guliš et al. (1994) evaluated drinking water quality. Lehotský and Tóth (1992) assessed the water quality of 96 forest springs and wells on the ridge of the Little Carpathians. Slaninka et al. (2005) monitored precipitation, surface runoff, and spring water runoff, for the sake of mass balance in the Vydrica catchment up to the Spariska section. The analysis of surface water quality in this study was based on the data obtained by IH SAS at four sampling sites ( Fig. 3 (karst, limestone) during third period in laboratory of IH SAS in Bratislava. The selected sampling sites of IH SAS represent a headwater quality (forested area). The sampling sites Gidra: Píla, Parná: Majdán and Trnávka: Buková are identical with the water gauge of the Slovak Hydrometeorological Institute (SHMI). The temporal variability of the nitrate nitrogen concentrations was very similar in all 4 stations in the period 1991-1994 (Fig. 4). Therefore we decided to focus on variability of the pollution concentrations only in Parná River at sampling site Horné Orešany water gauge (upstream the small reservoir Horné Orešany) and the sampling program was completed by another sampling site below the water reservoir in Horné Orešany village.

Methods
The statistical analysis of daily discharge series included the testing of homogeneity, stationarity, autocorrelation, multiannual cyclicity and long-term trends in measured data series. For testing we have used the autocorrelation and spectral analysis with application of the AnClim software (Štepánek, 2010). At the trend analysis of the time series, the parametric and non-parametric tests can be used (Procházka et al., 2001). The parametric test considers the linear regression of the random variable xi on time. The parameters of the trend line are calculated by using standard method for estimation of the parameters of a simple linear regression model, i.e. by using least square method. To identi-fication of the long-term trends we used Mann-Kendall nonparametric trend test. The Mann-Kendall nonparametric test (M-K test) is one of the most widely used non-parametric tests for significant trends detection in time series. By M-K test, we want to test the null hypothesis H0 of no trend, i.e. the observations xi are randomly ordered in time, against the alternative hypothesis H1, where there is an increasing or decreasing monotonic trend. The maximum concentrations of nitrates in surface water usually occur during the high discharges period (see Fig. 4) and therefore the impact of discharge cannot be omitted when analysing the mean (monthly, annual) nitrate concentrations. In order to assess the mean nitrate concentrations we have to calculate the weighted mean with respect to discharge, it means to calculate the mean concentration of nitrates in the stream according to the relation (Pekárová et al., 1995;Pekárová and Pekár, 1996):

Discharge analysis
The water gauge of Parná: Horné Orešany (Number 5250, river km 26.8) is situated above the water reservoir Horné Orešany (Výleta et al., 2017). The average annual discharge of the Parná River at the station of Horné Orešany for the period of 1961-2018 was 0.351 m 3 s -1 . The minimum and maximum flow observed during this period was 0.025 m 3 s -1 and 7.653 m 3 s -1 , respectively. Other statistical characteristics can be found in Table 2. On Fig. 5a, we can observe the variability of dry and humid periods. Periods 1971Periods -1974Periods , 1989Periods -1991Periods , 2001Periods -2003Periods , and 2017Periods -2018 were extremely dry. The regularity of changing the dry and humid years was studied by the autocorrelation and spectral analyses of mean annual discharge time series (Fig. 5b). Autocorrelogram shows significant autocorrelation near the 14 th year. This period was proved by the spectral analysis as well. The spectral analysis identified also other multiannual cycles: 4.5-5.5 years; 3.6 years, and 2.4 years. With respect to this more or less regular variability we can expect the more humid years to come. The intraannual variability of mean, maximum, and minimum daily discharge did not change significantly (Fig. 5c). On the other hand, the discharges are decreasing from the long-term point of view. With respect to the multiannual cycles of low and high discharges, we recommend to estimate the long-term trend since low period around year 1971 to low period around year 2018, (see results in the Table 3). Flows decreased slightly during this period, but the trend is not statistically significant.

Nitrate concentrations in surface water
Nitrates are of significant seasonal variability. Higher nitrate values occur during snow-melting in spring (Mendel, Halmová, 1993;Pekárová et al., 1994;Hyánková et al, 1995;Pekárová, Pekár, 1996). Incorrect application of fertilizers during the vegetation period leads to a rapid increase of nitrate concentration in stream waters. The time variation of measured nitrate concentrations is shown in Fig. 4. In Table 4 there are summarised the estimated means of nitrate nitrogen in three different periods and calculated by two different approaches: 1. Arithmetic average of measured data, 2. Weighted average with respect to discharge according to Eq. (1).
The highest nitrate pollution was found at the Parná River, above Horné Orešany in the first period 1991-1994, where average nitrate concentrations reached 4.2 mg l -1 , and weighted average 5.11 mg l -1 . After 1993 nitrate concentration in Parná headwater decreased.

Hydrological regime and estimation of nitrate nitrogen concentrations in the stream
As already mentioned, the water quality has a strong seasonal dependency, especially for some matters. Nitrate concentrations reach their maximum values during the wet periods, mainly during the spring snow melting. In Parná River, the minimum concentrations occur in summer-autumn (June-November), while the maximum concentrations are measured between December and May (Fig. 6a). Rain water washes nitrates off from the surface into a river. Generally, in the rivers of the Little Carpathian headwater region, an increased discharge leads to an increase in the nitrate concentration. But, during big floods, the nitrates concentration starts to be diluted by increasing discharge. We have derived exponential empirical relations for estimation of the nitrate nitrogen concentrations in the Parná River stream (Fig. 6b)

Conclusions
In Slovakia, the application of industrial fertilizers decreased significantly due to decrease of the agricultural mass production 30 years ago after 1989. The decrease in production impacted positively the decrease of nitrogen content in surface waters in Slovakia. After 20 years, since 2010, the fertilization dozes have increased to values close to the original ones before 1989. It can be expected that the nitrate concentrations will increase as well in our streams. Moreover, due to increasing air temperature (global warming) the evaporation increases as well and river runoff is decreasing slowly. The most   P1,1961-1990 P1,1989-2018 Fig. 6. a) Monthly regime of the nitrate nitrogen concentrations;b)Relationships between N-NO3concentrations and daily discharge, in the Parná River endangered are the lowland streams in agricultural areas. The aim of this study was to assess the long-term development of the hydrological regime of the Parná River at Horné Orešany gauge during the period 1961-2018 and to assess the development of the nitrate nitrogen concentrations at this sampling site during the period 1991-2018. The analysis of the hydrological regime proved the changing of wet and humid periods in approximately 14-years cycles in the Parná River series. In the main the discharge is decreasing slightly. In the case of nitrate nitrogen concentrations the marked decrease occurred in this river basin from the value of 5.11 mg l -1 during the years 1991-1995 to the value of 2.49 mg l -1 in the years 2015-2018. It was shown that the nitrate concentrations in surface streams are increasing with increasing discharge in the Little Carpathians headwaters, but there exists a threshold value of discharge when the nitrates concentration starts to be diluted by increasing discharge. The relation between discharge and nitrate concentration was used to derive exponential empirical relations for estimation of the nitrate nitrogen concentrations in the stream (Fig. 6b) based on mean daily discharge during the sampling day. These equations can be used for indirect estimation of nitrate nitrogen concentrations when they were not measured directly.