Major anions
The dominant sources of major ions for surface waters in the Slapy catchment were the application of synthetic fertilizers (for SO42–, NO3–, Cl–, and K+), liming of farmland (for Ca2+, Mg2+, and HCO3–) and its draining (for NO3–), atmospheric pollution/deposition (for SO42–), wastewaters (for NH4+, NO3–, and Cl–), and road de-icing (for Cl– and Na+) as discussed in detail by Kopáček et al. (2013a, b, 2014a, b, 2017). In short: The uneconomical use of synthetic fertilizers, excessive farmland drainage, and lack of environmental protection contributed to water pollution during the planned economy period until the late 1980s (Kopáček et al. 2017). As a result, concentrations of strong acid anions and BCs increased during this period (Fig. SI-6). In contrast, an economic recession immediately following the re-establishment of a market economy and continually improved legislation (regulation of emissions of S and N compounds and agricultural practice; Bičík et al. 2001; Kopáček et 2013b, 2014a,b) have resulted in dramatically reduced inputs of elements to farmland from fertilizers and to forests from atmospheric deposition; Figs. SI-11 and SI-12), and consequently, in decreasing (or levelling-off) trends in water pollution from diffuse sources since the 1990s. The long-term trends of major ions thus highlighted general relationships between the intensity of agriculture and land use on riverine element exports, well known from other large European and American catchments (e.g., Kaushal et al. 2005; Raymond et al. 2008; Aquilina et al. 2012). The exceptions from this pattern was Na+ due to continuously increasing use of NaCl for road de-icing from the early 1970s (Fig. SI-11F) that increased Na+ concentrations in surface waters, especially in the 2010s (Fig. SI-6G).
Seasonal variations in concentrations of major ions were relatively small. The SO42– and Cl– seasonality was similar, with maximum values in summer (Fig. SI-6B, C), due probably to lower dilution by discharge (Fig. 1A) and higher evaporation from the epilimnion. Seasonal cycles of HCO3– and NO3– exhibited inverse patterns (Figs. 3B, SI-6A) due to the HCO3– dilution by elevated discharges in spring, while NO3– concentrations increased during this period (see section Nitrate for details). Because the inverse seasonal variations in molar HCO3– and NO3– concentrations were of similar magnitude, the combined effect of their changes on the Ca2+ and Mg2+ seasonality was negligible. The seasonal variations of Ca2+ and Mg2+ ions were thus predominantly associated with their co-export with SO42– as its counter-cations.
Dissolved organic carbon
The highest DOC concentrations in the 1960s and their continuous decrease till 1992 (Fig. 4B) reflected decreasing surface water pollution with domestic and industrial wastewaters (Fig. SI-13). In the 1960s, ~36% of the Slapy catchment inhabitants were already attached to sanitary systems, but only 1–10% (3% on average) of these wastewaters were treated (Fig. SI-14A). The rest of the inhabitants used latrines, and biologically stabilized waste was hauled away from cesspools and used for soil fertilization in agriculture. By the mid-1990s, more than 80% of the catchment inhabitants were attached to sanitary systems, and the proportion of treated municipal wastewaters had increased to ~90% (Kopáček et al. 2013b). Consequently, the pollution of surface waters in the catchment with organic matter from municipal wastewaters steeply decreased in the 1990s (Fig. SI-14B).
Wood pulp and paper productions were other important DOC sources in the Slapy catchment. Most of the wood pulp production was based on the sulphite process and reached 72–171 Gg yr–1 in the 1960–1981 period (Kopáček et al. 2014a), with two major paper mills. Their wastewaters were directly discharged into the Vltava. Sulphite pulp effluents are heavily polluted with various types of wood extracts, including readily biodegradable compounds, as well as non-degradable substances such as lignosulfonates, and have concentrations of biological and chemical oxygen demands in the hundreds and thousands of mg L–1, respectively (Pitter 1990; Suriyanarayanan et al. 2005). Organic pollution associated with sulphite pulp production in the Slapy catchment can be estimated at 13 to 31 Gg yr–1 of biochemical oxygen demand (BOD5) in 1960–1981, assuming an average BOD5 production of ~180 g kg–1 of pulp (Pitter 1990). These wastewaters were not treated until the late 1960s. Their production, however, decreased in 1966 when one of the paper mills was shut down. Another decrease occurred in 1968 when wastewaters became partially treated using evaporators at the second paper mill. These measures decreased concentrations of brownish lignosulfonates in the surface water, and the low transparency that had been typical for the Slapy reservoir in the 1960s increased in the following decades (Fig. 1D). In 1986, the sulphite process was partly replaced by the sulphate process. In 1990, pulp effluents began to be efficiently recycled and treated, which led to a breakpoint in organic pollution of the Vltava above the reservoir cascade (Fig. SI-13) and a pronounced minimum in DOC concentrations in the Slapy reservoir over the next three years (Fig. 4B). Since this minimum, however, DOC concentrations have been steadily increasing.
The continuously increasing DOC leaching from diffuse sources in forested parts of the Slapy catchment (Hejzlar et al. 2003; Kopáček et al. 2013c) has reversed the decreasing DOC trend, associated with changes in surface water pollution by wastewaters from point sources, since the 1990s. The most likely reasons for the increasing DOC leaching were (1) climate, causing changes in temperature, precipitation, forest heath, and soil moisture (Hejzlar et al. 2003; Kopáček et al. 2018), and (2) decreasing atmospheric deposition of SO42– since the 1980s (Kopáček et al. 2013c), resulting in elevated DOC mobility due to increasing pH and decreasing ionic strength of soil water (e.g., Hruška et al. 2009; Evans et al. 2012).
Climate-related terrestrial exports of DOC from soils are usually associated with elevated intensities and frequencies of lateral water flows through the uppermost soil horizons (Bearup et al. 2014). Their probability increases during high precipitation and runoff events and is currently magnified by climate change due to more frequent torrential rains (Raymond et al. 2016). The climate-related, more frequent thaws, snowmelt, and elevated discharge during winter probably contributed to the significantly increasing DOC concentrations in the Vltava water during the winter and spring months of the 1993–2019 period (Fig. SI-8B). DOC leaching from forested parts of the Slapy catchment was further amplified after bark beetle-induced spruce mortality in forests, which has increased since the late 1990s due to increasing air temperature. The tree mortality changes soil chemical and microbial characteristics, resulting in elevated DOC leaching (Kopáček et al. 2018). In parallel, the reduced water transpiration after tree death leads to increased soil moisture, shallower groundwater, and greater water fluxes through shallow, organic-rich soils, further accelerating terrestrial DOC exports (Hubbard et al. 2013; Mikkelson et al. 2013). The DOC concentrations thus continued their increase even after 2002, when pulp production in the Slapy catchment completely ceased.
Phosphorus, silica, and chlorophyll
Concentrations of TP and chlorophyll a increased from 1960 to the mid 2000s (Fig. 3), and then started to decrease, with breakpoints in 2004 (Figs. 4C, SI-9A). Primary production in the Slapy reservoir is P-limited (Komárková and Vyhnálek 1998). Hence, the increase in TP concentrations was paralleled with enhanced primary production, here proxied by the concentrations of chlorophyll a. The elevated TP availability resulted from increasing P exports from diffuse agricultural sources and point sources, especially domestic wastewaters (Fig. SI-14C; Hejzlar et al. 2016).
Increased exports of P from diffuse sources are usually associated with high flows after intense precipitation, which intensify the erosion of soil particles with adsorbed phosphate, as well as leaching of dissolved P from the uppermost soil layers during periods of lateral water flows (Bowes et al. 2008). The risk of P export associated with erosion increased with the increasing proportion of arable land in the catchment (Kopáček et al. 2017) and intensified P-fertilization till the late 1980s (Fig. SI-11D). Since the early 1990s, the risk has been decreasing due to the reduced use of both synthetic fertilizers as well as manure and slurry (due to a sharp reduction in cattle and pig production after 1989, Kusková et al. 2008), the increasing conversion of arable fields to pastures (Kvítek et al. 2009), and an agri-environment scheme implemented in the Czech law after the accession of the Czech Republic to the European Union in 2004 (Doležal and Kvítek 2004; Vystavna et al. 2017).
The P export from point sources increased in the Slapy catchment from the 1960s to the 1990s due to the increasing number of inhabitants attached to sanitary systems, the increasing use of P-rich detergents and the low P-retaining efficiency of wastewater treatment plants (Hejzlar et al. 2016). Then, the P export from point sources began to decrease (Fig. SI-14C) due to legislation changes. These included (1) the recommended and voluntary reduced use of P in detergents since the mid-1990s, followed by a complete legislative ban in 2006; (2) regulations of wastewater treatment efficiency, limits for P concentration in effluents, and fees for the amount of P discharged being set for large wastewater treatment plants (> 3 Mg yr–1 of P) since 2005; and (3) by the reduced use of P in dishwasher detergents since 2014 (Hejzlar et al. 2016; Vystavna et al. 2017). These measures resulted in a continuing decrease in the Vltava TP concentrations, similarly to freshwaters in other European countries (Hukari et al. 2016), and consequently, in lower primary production in the Slapy reservoir compared to the 1990s (Fig. 3E).
The summer decrease in the epilimnetic TP concentrations (Fig. 3G) was caused by the uptake of this limiting nutrient by primary producers, and was accompanied by seasonally elevated chlorophyll a concentrations (Fig. 3E), low water transparency (Fig. 1D), elevated water saturation with dissolved O2 (Fig. 1F), decreased H2CO3* concentrations (Fig. 1H), depleted NH4+ (Fig. 3A), and increased pH (Fig. 1I). The elevated P uptake by phytoplankton resulted in a PP contribution to TP of 49% from May to September 1995–2019, while it only formed 29% of TP on an annual basis.
A substantial seasonal drop in concentrations of dissolved Si (Fig. 3H) was associated with the increasing dominance of diatoms in the reservoir phytoplankton. Diatoms, a group of algae with siliceous cell walls, can form dense populations under favourable conditions in early summer and deplete nutrients and Si to limiting levels (Wetzel 2001; Znachor et al. 2008). Due to the presence of siliceous frustules, diatom cells are substantially heavier than water and sink down despite a stable thermal water stratification in summer. Consequently, they are continuously replaced by other phytoplankton groups with no obligate Si requirement (Sommer et al. 2012), and thus the Si pool in water is regenerated, as evidenced in the enhanced Si concentrations later in the season (Fig. 3H).
Ongoing analyses of long-term data show widespread, systematic changes in the seasonal timing (phenology) of ecological events over recent decades (Winder and Schindler 2004). In the Slapy reservoir, phenological changes were most pronounced in spring (Fig. 5), which is in line with the commonly accepted notion that a shift in the spring biomass peak is the most sensitive indicator of phytoplankton response to climatic forcing (Winder and Schindler 2004).
Nitrogen forms
Concentrations of total N (TN) were dominated by NO3– and TON (76 and 21% on average, respectively), while NH4+ and NO2– contributed only negligibly (2 and 1%, respectively). Long-term trends and seasonal variations of TN thus predominantly reflected the NO3– dynamic (Fig. 3). The individual N forms, however, exhibited different time-series, reflecting different physical-chemical and microbial processes governing N transformations and uptake in the whole catchment-water system, as follows.
Ammonium
The NH4+ concentrations in the surface layer of the Slapy reservoir were highest in the early 1960s and probably originated from wastewaters and mineralization of the flooded organic matter. The untreated domestic wastewaters were a significant NH4+ source for surface waters, especially in the 1960s, with minimum sanitary systems attached to wastewater treatment plants (Fig. SI-14A). During the following five decades, the proportion of biologically treated domestic wastewaters in the catchment increased and intensive water aeration during biological treatment nitrified NH4+ and decreased its export from wastewater treatment plants to surface waters (Behrendt et al. 2000; Kopáček et al. 2013b).
Another factor contributing to decreasing NH4+ concentrations was the increasing volume of surface waters due to the construction of new reservoirs (Table SI-1), resulting in a tripled water residence time in the catchment between 1960 and 1991. More NH4+ thus was assimilated and nitrified prior to entering the Slapy reservoir.
Some of the NH4+ in the surface layer of the Slapy reservoir also temporally originated from mineralized vegetation and soil organic matter on the flooded bottom, especially in the early 1960s, entering the surface layer during spring and autumn turnovers (Procházková 1966). In addition, the water outlet from the Orlík reservoir always originated from the cold (~4 °C), dark, and anoxic hypolimnion. Under such conditions, the liberated NH4+ could not be effectively removed by assimilation and nitrification (Wetzel 2001) and thus added to the Slapy reservoir pool of NH4+.
Concentrations of NH4+ exhibited pronounced seasonal variations in the Slapy surface layer, with elevated winter values and pronounced spring maxima during the first three decades of this study (Fig. 3A). The major epilimnetic NH4+ sinks during the growing season were probably its assimilation (NH4+ is usually the primary source of inorganic N for freshwater phytoplankton) and nitrification, as in other circum-neutral waters (Procházková et al. 1970; Wetzel 2001).
Nitrate
Similarly to other large European catchments (e.g., Kronvang et al. 2008), NO3– leaching from agricultural land was the major source for the Slapy reservoir (Procházková 1966; Kopáček et al. 2013b). The increasing trend in NO3– concentrations from the 1960s to the 1980s (Fig. SI-7) predominantly reflected high net anthropogenic N inputs to farmland from synthetic fertilizers (Fig. SI-11C) and mineralization of soil organic N due to increasingly intensive drainage and deep tillage (Kopáček et al. 2013a). Point sources of N exhibited a similar increasing trend during this period due to the increasing population attached to sanitary systems (Fig. SI-14D), but this contribution was about an order of magnitude lower than that of diffuse sources (Kopáček et al. 2013b). Similarly, the subsequent decline in NO3– concentrations between the 1990s and the 2000s primarily reflected the de-intensification of agricultural production, reduced N fertilization rates, aging of drainage systems and the partial conversion of arable land to grassland (Bičík et al. 2001; Kopáček et al. 2013a), while increasing denitrification in wastewater treatment plants played a relatively minor role (Kopáček et al. 2013b). As observed elsewhere (e.g., Randall and Goss 2008; Kvítek et al. 2009), the reduced N inputs and decreasing soil aeration were the most effective measures rapidly reducing NO3– leaching from agricultural land also in the Slapy catchment.
The high seasonal variations in NO3– concentrations, with maximum values typically occurring in spring (Fig. 3B), were associated with elevated discharge and the preceding dormant season (Procházková and Brink 1991). Nitrate accumulated in soils during dry periods (especially in winter, with negligible N uptake by vegetation) because of low hydraulic conductivity in the soil profile was mobilized and leached in the following wet period (Meisinger and Delgado 2002; Randall and Goss 2008).
Nitrite
The NO2– concentrations in the Slapy surface layer exhibited pronounced seasonal variations (Fig. 3C). They started to increase in spring together with water warming, and this process accelerated in April–May when temperatures exceeded ~10 °C (Fig. 1C). The maximum NO2– concentrations corresponded with the highest transparency in the clear-water periods (i.e., during the lowest chlorophyll a concentrations; Figs. 1D and 3E) and the longest and most intensive sunshine in June. In the rest of summer, the NO2– concentrations continuously decreased with decreasing water transparency and finally reached their minima after water temperatures decreased to < 10 °C in November. In contrast to the terrestrial drivers dominating the NO3– dynamic, the NO2– variation was predominantly associated with the seasonal development of in-lake processes and probably resulted from nitrification and denitrification. Nitrification could have played an important role during elevated NH4+ concentrations at the beginning of the study. However, we expect that the role of denitrification in the NO2– production increased in subsequent years with increasing NO3– availability, as suggested by the parallel NO2– and NO3– long-term trends (Fig. SI-7). The NO2– production by both nitrification and denitrification can be expected to exhibit maxima in summer for the following reasons.
The nitrification rate is low at < 10 °C and very sensitive to small changes in temperature over a range from 10 to 17 °C, with the inhibitory effect of low temperature greater for NO3– formers, such as Nitrobacter, than for NO2– formers, such as Nitrosomonas (Randall and Buth 1984). The inhibitory effect of low temperature thus temporarily results in a NO2– build-up during temperature increases above or decreases back to ~10 °C. Moreover, the rate of microbial NO2– oxidation is more inhibited by light then its production from NH4+ (Olson 1981; Vrba 1990).
The NO2– production associated with NO3– reduction occurs during photochemical and microbial denitrification (Spokes and Liss 1999; Stief et al. 2016). Photochemical denitrification is a first-order reaction related to the NO3– concentration and also depends on the concentration of organic matter present in the system and the intensity of solar radiation (Spokes and Liss 1999; Porcal et al. 2014). At high (and compared to NO2– relatively stable) NO3– and DOC concentrations during the year (Fig. 3), the intensity of photochemical NO3– reduction thus primarily reflects seasonal variations in transparency and sunshine intensity, both of which are high in June. We also cannot exclude microbial denitrification, despite the high summer concentrations of dissolved O2 (Fig. 1E), due to the formation of local anoxic microsites that may occur inside algal aggregates, especially diatoms (Stief et al. 2016). This process would also result in the highest NO2– production during the summer months with the highest primary production and algal concentrations in the epilimnion.
Organic nitrogen
Concentrations of TON exhibited a small but significant (p < 0.001) increase between 1960 and 2019 (Fig. SI-9B). Data on PN and DON, available from 1995 to 2019, showed that DON increased (p < 0.001), while PN concentrations significantly (p < 0.05) decreased during this period, similarly to PC, as well as also PP and chlorophyll a from 2004 to 2019. The long-term increase in TON was thus associated with the DON representing 92% of TON during the 1995–2019 period.
The decrease in molar DOC:TON ratios from ~23 in the 1960s to ~12 during the 1992–2019 period was apparently caused by reduced inputs of DOC-rich wastewaters. The stable DOC:TON ratios since 1992 indicated that the long-term increase in concentrations of organic N was probably associated with increasing terrestrial DOC export from diffuse forest sources. In-lake N assimilation by algae, however, contributed to the elevated TON (and also PN and DON) concentrations during the growing period, resulting in seasonal variations similar to chlorophyll a (Fig. 3).
pH
The long-term trend in pH and its seasonal variations, with winter minima and maxima in July and August (Fig. 4D), integrated the effects of changing external acidity sources, in-lake H+ producing/consuming processes, and the P inputs (affecting the intensity of P-limited primary production), as well as changes in water buffering capacity. These processes had different effects on trends in annual pH minima and maxima (Fig. 6).
There were at least three reasons contributing to lower pH values during the 1960s–1980s than in the three following decades: (1) Acidic deposition and the intensive use of synthetic fertilizers [including (NH4)2SO4)] acidified soils and the receiving waters (e.g., Reuss and Johnson 1986). (2) Decomposing organic matter on the bottom of the flooded valleys of the Slapy and Orlík reservoirs and microbial mineralization of organic pollution from domestic and industrial wastewaters were more important CO2 sources (resulting in higher H2CO3* concentrations in the Slapy water) until the 1990s than in the following years, with a breakpoint in the H2CO3* trend in 1993 (Fig. SI-15). The elevated H2CO3* concentrations lowered water pH (Stumm and Morgan 1981), especially in winter, when CO2 assimilation by algae was limited. (3) The H+ release associated with NH4+ assimilation and nitrification (Reuss and Johnson 1986) was higher during the 1960s–1980s than in the following decades due to higher NH4+ concentrations (Fig. 3A).
The increasing pH trend between the 1960s and 1990s thus probably resulted from (1) the decreasing pollution of surface waters with DOC and NH4+ from point sources and the mineralization of flooded organic matter during aging of the Slapy and the Orlík reservoirs since the 1960s, (2) the decreasing use of (NH4)2SO4 for soil fertilizing and intensifying liming of farmland since the 1970s (Kopáček et al. 2017), and (3) the recovery of headwaters from atmospheric acidification since the mid-1980s (Kopáček et al. 2013c). Another process contributing to the pH increase during this period was an increasing proportion of NO3– assimilation. The average molar N:P ratio of the seston was higher (~20) than the water NH4+:TP ratio (from <1 to 8) and stable. A substantial contribution of N2 fixing cyanobacteria to the PN budget was rather unlikely due to their low abundance (Komárková and Vyhnálek 1998). Hence, the stable sestonic N content was mostly maintained by NO3– assimilation as in other ammonium-poor waters (Wetzel 2001). The microbial NO3-N reduction consumes an equivalent amount of H+ (Reuss and Johnson 1986). With the decreasing NH4+ availability (Fig. 3A) and increasing primary production (Fig. SI-9A), the role of NO3– as an additional N source for algae (Wetzel 2001) continuously increased and contributed to the in-lake H+ consumption and increasing pH, especially in summer.
The breakpoint in the trends of Slapy water pH occurred in 1992 (Fig. 4D), despite the continuing increase in the annual pH minima (Fig. 6A). The decreasing trend in pH since 1992 thus predominately resulted from the decreasing summer pH maxima (Fig. 6B). Processes contributing to decreasing pH between the 1990s and 2019 include at least four processes: (1) The intensive liming of farmland almost ceased in the early 1990s (Fig. SI-11G), and agricultural soils in the Slapy catchment have started to re-acidify (Kopáček et al. 2017). (2) The increasing leaching of DOC from forested areas in the Slapy catchment (Hejzlar et al. 2003; Kopáček et al. 2018) has supplied more organic acids to surface waters since the early 1990s. The surface pH is usually higher than that of the original soil water, leading to higher dissociation of organic acids and H+ liberation (Stumm and Morgan 1981). (3) The decreasing P loading of the Slapy reservoir since the mid-2000s (Fig. 4C) has reduced the P-limited primary production, resulting in a lower depletion of epilimnetic concentrations of dissolved CO2 by growing algae, and hence in lower summer pH maxima (e.g., Wetzel 2001). (4) The decreased primary production also reduced NO3– assimilation and the associated H+ consumption, which further contributed to the decreasing summer pH maxima.
A shift in the water buffering capacity was another important factor contributing to higher annual pH maxima in the 1990s, compared to the 1960s and 2010s (Fig. 6B). The lower pHs in the 1960s were closer to the value 6.35 at which the buffering capacity of the carbonate system in water reaches a maximum, while in the 1990s, pH values were closer to the 8.3 (Fig. 4C), at which the buffering capacity is minimum (Stumm and Morgan 1981). The ability of water to mitigate pH changes caused by H+ production/consumption thus decreased from the 1960s to the 1990s, allowing a more sensitive pH response to H+ sources and sinks. Then, the increasing water buffering capacity due to decreasing pH values contributed to the less pronounced summer pH maxima between the 1990s and 2010s.
These results show that long-term trends and changes in the seasonal variations of pH may reflect anthropogenic and climatic effects, reservoir aging, and changes in water eutrophication, not only in acid-sensitive regions (Psenner and Catalan 1994) but also in circum-neutral systems.