Josef Kˇreˇek , Jn Nov′kov′ , Lislv Pl′n , Ev Pˇzourkov′ , Evˇzen Stuhlík

a Department of Hydrology, Czech Technical University in Prague, Th′akurova 7, CZ-16629, Prague 6, Czech Republic

b Independent Botanist, Na Vrcholu 4, CZ-130 00, Prague, Czech Republic

c AON Impact Forecasting, V′aclavsk′e n′amˇestí 19, CZ-110 00, Prague, Czech Republic

d Institute of Hydrobiology, Biology Centre, Czech Academy of Sciences, Na S′adk′ach 7, CZ-370 05, ˇCesk′e Budˇejovice, Czech Republic

Keywords:Headwater catchment Forest functions Water resource recharge Acid atmospheric deposition Climate change

ABSTRACT Forest practices in headwater catchments are related to environmental and social contexts. The aim of this study is to analyse the changing protective functions of forests in the upper plain of the Jizera Mts.(North Bohemia, Czech Republic) since the second half of the 19th century. With time, forests have gradually changed from native stands to spruce plantations (Picea abies), introducing exotic conifers(more resistant to air pollution), and, recently, back to more diverse mixed stands. The priority in protective forest functions there shifted from flood protection to integrated control of water resources(quantity and quality). In the 1980s, forest - water interactions were degraded by consequences of extreme acid atmospheric deposition, forest die-back, and extensive clear-cut. In the Jizerka catchment,first signs of recovery were observed in the early 1990s, but, stream waters there are still affected by prolonged acidification.While reconstruction of stream water chemistry at Jizerka follows the drop of the acid deposition in some 5 years, the revival of stream biota takes 10-15 years. In 2071-2100, the projected climate change shows rising annual air temperatures by 3.0-4.6 °C, decreasing water yield by 65-123 mm, 60% drop in ‘minimum residual discharge’, and 20-30% rise in peak-flows. However, these projected environmental changes cannot substantially decline the high potential in water resource recharge, or, start reverse processes in recent recovery from acidification and radically affect the existence of planned mixed forests in the upper plain of the Jizera Mts.

1. Introduction

Forest functions are strictly related to environmental and social contexts, reflecting knowledge, technical capabilities, economic interests, levels of human pressure,and also human emotions(for example fear from forest wildlife and environment),Corbin(1998),Klvaˇc (2006), Woitsch (2006), Leggewie and Mauelshagen (2018).Three groups of forest functions have generally been considered(Krott,2005):(i)the productive function(focusing on the economic utilization of timber), (ii) the protective function (including physical,chemical,and biological effects within the scope of the natural environment), and (iii) the recreation function (focusing on the effects of the forest on human health and wellbeing).

In the history of the human civilisation,deforestation was one of the key processes in the decline of soil and water environments:already in 1530, a specific law of the Venetian Senate limited the harvest of forests to prevent possible landslides and water erosion phenomena in the Pre-Alpine territory(Pilli&Pase,2018).The poor state of mountain forests in Europe by the end of the 19th century was given to the context of catastrophic floods frequency(Magnein,1937), and initiated declarations of the soil improvement and the safe diversion of mountain waters. From that time, the steady growth in timber demand and the spread of new silvicultural practices resulted in the sustainable timber production and gradually reduced deforestation in Europe. Later, due to the increasing complexity of the social needs, the forest functions have been related with more recent environmental services and resulted in the multiresource forestry; particularly, in the headwater catchments, forestry practices should respect possible regulations of water yield,runoff timing,frequency of extreme events,and water quality including the biota (Willis, 2002). Agnoleti et al. (2002)claim the importance of rising science in ‘modern forestry’, to emphasize ecological and social sustainability.

After the UN (1992) statement, particularly, the vital role of forest in biodiversity protection and carbon storage gained attention. Land cover changes (deforestation mainly) in the past have contributed almost 20 ppm to today’s elevated atmosphere CO2concentrations (McGrath et al., 2015). In Europe, headwater catchments are mostly forested, thus, forestry strategies and practices are among the most important factors of their environmental services (FAO, 2008). At present, the share of untouched forest or forest with minimal human intervention is probably below 5% only(Schnitzler, 2014, in McGrath et al., 2015).

Considering water security in the future, the participation of water users in the management of headwaters should be practiced(Trottier & Slack, 2004), and a multi-functional view of forests respected due to the competitive interaction between forest services. Viviroli et al. (2007) considers the headwater catchments of mountain regions as ‘water towers’ with high potential by supplying lowland users. Priorities of forest protective functions in headwater ecosystem services are recently claimed by Kˇreˇcek et al.(2017). Their importance for human society is rising with the population pressure (UNEP, 2007) and the global climate change(IPCC,2015);the elevation-rising sensitivity to warming(Kohler&Maselli,2009)can even accelerate changes in mountain forests and affect the hydrological cycle of headwater catchments.FAO (2007)calls particularly for timely adaptive practices, because, decisions on the forest management might remain irreversible for decades or even centuries.

The aim of this paper is to evaluate impacts of changing environment and society on the forest functions in a headwater catchment of the Jizera Mts. (Czech Republic) during the period 1850-2100.

2. Material and methods

2.1. Study site

The study was performed in the Jizerka experimental catchment(50°48′21” - 50°48′59′′N, 15°19‘34“ - 15°20‘48“E) located in the upper plain of the Jizera Mountains (Northern Bohemia, Czech Republic, Fig.1), morphological characteristics are in Table 1. This catchment belongs to the Elbe river district 1-05-01-004 and to the‘high contributing potential to water resources’ (relative water yield RWY = 4.69 exceeds the critical value 1.25, according to Viviroli et al., 2007).

Climate characteristics of that unit are assigned from Tolasz et al.(2007):North temperate zone,K″oppen Dfc sub-arctic region,mean annual precipitation 1,400 mm,mean annual air temperature 4°C,and the average maximum snowpack of 120 cm (snow cover usually lasts from the beginning of November to the end of April).Sandy-loamy podsols developed on porphyritic granite show the depth between 0.7 and 1.2 m. The root system dominates in the topsoil occurring up to the depth of 15 cm.The topsoil is created by litter (Ol, depth of 0-2 cm), humus layer (Of+ Oh, 2-10 cm), and leached horizon (Ah,10-15 cm); mor is the most common humus(2-5 cm) there. In the runoff genesis, surface and fast subsurface forms (direct runoff) dominate, and, groundwater bodies occur in the shallow subsurface layer only;the soil-bedrock complex shows a relatively low buffering capacity to acidification (Kˇreˇcek &Hoˇrick′a, 2006).

According to the classification of Zlatník (1976), the climax forest there belongs to the belts of“Picea abies”and“Picea abies-Fagus sylvatica”. By the end of the 19th century, the native forests were harvested (timber for constructions, charcoal, and, namely, for the nearby Jizerka glass manufacture) and converted to even-aged plantations of Norway spruce (Picea abies). In the 1970s and 1980s,these stands were defoliated by the‘acid rain’impact,and in 1984-1988 harvested by clear-cutting. The subsequent reforestation (mainly Norway spruce and introduced exotic conifers) was complicated by the acid environment,a relatively large open space,and competition of the herbaceous vegetation (namely,Calamagrostissp.), Fig. 2.

Fig.1. The Jizerka experimental catchment.

Fig.2. Forest regrowth of the Jizerka catchment in 2008(20 years after the clear-cut).

Table 1 Geomorphology of the Jizerka catchment.

2.2. Field observations

The experimental catchment(Fig.1)was instrumented in 1982.The outlet is equipped by the sharp-crested V-notch weir with the automatic water pressure and temperature recorder ALA 4020(ALA, Buˇcovice, Czech Republic) operating in ten minutes.In situmonitoring of stream waters included temperature, pH and conductivity measured by the portable sensor WTW-350i (WTW,Weilheim, Germany). Standard meteorological observations (air temperature and humidity, global radiation, wind speed, atmospheric precipitation, soil moisture) were carried out in elevations of 875 and 975 m. Additionally, twelve passive fog collectors were installed along the catchment transect between elevations 862 and 994 m: At each collector, fog drip is generated by 400 m of Teflon line (diameter = 0.25 mm, surface area index SAI = 5) exposed at the height of 1.7 m above the ground. Detailed under-canopy observations were accomplished in the stand of mature spruce (area of 30×30 m,975 m a.s.l.,80-100 years old):the canopy throughfall sampled by ten bulk collectors (200 cm2area), randomly distributed,and stemflow collected at two average trees.Next three observed forest plots are located outside from the catchment;including a comparable mature spruce stand in lower elevation(745 m a.s.l.) and two beech stands (Fagus sylvatica, 160 and 40 years old) in elevations of 420 and 690 m.

Collected rain/fog waters were sampled in monthly intervals and stream waters weekly; chemical analyses were performed in the laboratory of the Hydrobiological Station Velký P′alenec(Charles University in Prague) by the standard techniques developed for soft water studies in mountain environment (Stuchlík et al., 2006), concerned on pH, conductivity, and acidifying substancesThe potential growth of algae in rain/fog collectors was reduced by the dark sampling bottles and installed shelters, also, the bacterial activity in our field is controlled by relatively cold mountain climate (and also by relatively high concentrations of nitrogen,low pH values and dissolved organic carbon), Cape et al. (2001). Benthic macroinvertebrates were sampled in the stream channel near the catchment outlet in May (after snowmelt), July/August (high summer period), and in September/October(relatively dry period)in 1994,2004-2005(by the Czech regional campaign, Horecký et al., 2013) and 2015. The kick-net sampling technique(Rosenberg&Resh,1993)with a hand net (500 μm) was used, the collected material sieved by a 300 μm net,and,counting of each taxa performed in the laboratory(by eye and under a binocular microscope at 12-16× power).

Detailed botanical investigations were carried out in vegetation seasons of 1991, 2002, 2005, 2008 and 2018. Phytosociological relev`es (4 × 4 m) were taken at each of the twelve-point transect along the main slope of the catchment,and eight-point transect of the fragmented mature spruce stand.For each botanical relev`e,the Ellenberg’s indicator of soil moisture was calculated as the weighted average of all recognized species (Ellenberg, 1974). To include the impact of all species abundance, the data were transformed from the Braun-Blanquet’s scale to a nine-point scale according to the approach of Van der Maarel(1979).The monitoring of forest stands corresponds to standard forest inventory techniques (K¨ohl et al., 2006); tree species, age, basal area, height,timber volume,horizontal canopy density,vitality of trees,and leaf area, were estimated. For the aerial extrapolation of ground observations, the free available satellite imagery of Landsat (4,5) and Sentinel (2) was employed (Pal′an et al., 2018).

Standard methods of historical research(Gunn,2011)were used to identify the relationship between landscape and the built environment in the Jizera Mts. region with particular attention to the experimental catchment area. This activity included both the documentary evidence of archives and related historical writings.

2.3. Theory and calculations

In hydrological years(1st November-31st October),the annual water budget of the investigated catchment can be considered by equation (1), Shaw (2011).

Where,ETis the evapotranspiration loss,Pprecipitation input,andRrunoff observed at the catchment outlet (water yield), [mm].

According to Arora(2001),components of potential evaporation(available energy) and precipitation determine annual runoff regime in a region. Grismer et al. (2002) reported about fifty empirical methods to estimate potential evapotranspiration; their worth depend namely on quality and extent of meteorological inputs. Based on the Penman - Monteith combination approach(Dingman, 2002), Allen et al. (1998) suggested the reference crop evapotranspiration RET(FAO-56)representing a hypothetical grass reference cover with specific characteristics (cool-season grass,height of 0.12 m,albedo 0.23,surface resistance 70 s m-1,not short of water) calculated in daily steps by the equation (2).

Where,RETis the reference evapotranspiration[mm day-1],Rnnet radiation [MJ m-2day-1],Tair temperature at 2 m height [oC]),uwind speed at 2 m height [m s-1],esatsaturation vapour pressure[kPa],eaactual vapour pressure [kPa], Δ slope vapour pressure curve[kPa°C-1],γ psychrometric constant [kPa°C-1].

In mountain coniferous forests, canopy interception is considered as the dominant component of the total evapotranspiration loss (FAO, 2008; Pal′an & Kˇreˇcek, 2018; Ringgaard et al., 2014.).Interception loss (I) from tall canopies has been estimated by observing the net precipitation(P′)reaching the soil surface under the canopy, and the gross precipitation in an adjacent ‘open field’(PG), equation (3):

Where,P′is the net precipitation,PGopen field (gross) precipitation,PFfog drip,PTthroughfall under the canopy, andPSstemflow(in spruce stands supposed to be negligible), [mm].

The analytical model of Gash is the storm-based approach using the mean evaporation and rainfall rates, and the detailed rainfall pattern(Gash et al.(1995).This model considers rainfall to occur as a series of discrete events, each comprising a period of wetting,when the rainfall(PG),is less than the threshold value necessary to saturate the canopy (PG’), a period of saturation, and a period of drying out after rainfall ceases.The forest structure is described in terms of a canopy storage capacity (S), which is defined as the amount of water left on the canopy in zero evaporation conditions when rainfall and throughfall have ceased, a free throughfall coefficient (p), which determines the amount of rain which falls directly to the forest floor without touching the canopy, and the proportion of rainfall diverted to stemflow (ps). The mean evaporation rate during rainfall (E) and the mean rainfall rate (R) are required to evaluate the loss from a saturated canopy until rainfall ceases.The separate components of the interception loss(I1,I2,I3,I4)are calculated as the interception ofmrelatively small storm insufficient to saturate the canopy (I1):

The loss(I2)by wetting the canopy,for n storms that saturate the canopy (PG＞PG‘):

Then, the total amount of intercepted rainwater (I) in a certain period is:

Atmospheric precipitation inputs are affected by both elevation and vegetative canopy (Gash et al., 1995; Pal′an & Kˇreˇcek, 2018).Thus, the hypsometric method was used to extrapolate point observations of gross precipitation and canopy throughfall,as well as the deposition of sulphate, nitrate and ammonia. Seasonal atmospheric loads of studied elements were estimated by equation(9):

Meteorological observations at Jizerka (1982-2016) were extrapolated with the data of two neighbouring climatological stations operated by the Czech Hydrometeorological Institute(Desn′a-Souˇs,P2DESN01,772 m a.s.l.,1951-2016,and,Jablonec n.N.,U2JANIOJ,495 m a.s.l.,rainfall registered since 1898),and trends in the time series detected according to WMO(2001).Concerning the future climate projection, simulated air temperatures (at 2 m height) and precipitation amounts were downloaded from the PRUDENCE archive (http://prudence.dmi.dk, Christensen &Christensen, 2007). These data are generated by the regional climate model RegCM driven for outputs of two global climate models HadAM3H A2 and HadAM2H B2 respect two emission scenarios (IPCC, 2013a, 2013b): a permanent growth of emissions(A2) and an effective emission control during this century (B2).Downloaded files of daily precipitation (precip.ICTP.A2.nc, precip.-ICTP.B2.nc) and mean daily temperatures (t2m.ICTP.A2.nc,t2m.ICTP.B2.nc) were transformed with Intel® Array Viewer Version 3.3 to standard MS Excel base;these data are available for periods 1961-1990 (control) and 2071-2100 (forecast). The elevation gradients (-0.6°C per 100 m) for temperature and(+50 mm per 100 m)for precipitation(Kˇreˇcek and Punˇcoch′aˇr 2012)were used to transform the downloaded data(corresponding with the average elevation of the 50 × 50 km square)to the catchment elevation. Daily catchment runoff was calculated by HBV conceptual model (Seibert, 2005) respecting periods of calibration(1991-2010),validation(2011-2016)and simulation(2071-2100).

The standard descriptive statistics and one-way ANOVA were applied to analyse the collected data sets and to identify relationships between the groups of data(Motulski&Searle,1998).Trends in the time series were detected by the CTPA programme (WMO,2001).

3. Results

3.1. Changes in the vegetative canopy

According to the 2016 inventory in the Jizerka catchment, four recognized canopy groups (mature spruce stands, middle age stands,young stands with grass,and areas with dominating grass)are introduced in Table 2;and,the canopy structure reconstructed back to 1880 and projected to 2100 is given in Fig.4.By the end of the 19th century,forests of the studied area were harvested(mainly providing the timber for local glassworks, Fig. 3) and converted to even-age plantations of Norway spruce (Rabˇstejnek,1969). In the period 1983-2016, these stands were damaged by the acid rain impact and harvested, the clear-cut and forest regeneration were documented by the supervised classification of multiband raster images(Landsat 4,5, Sentinel 2).

Nowadays, coniferous stands (P. abiesand introduced exotic species,P.pungensmainly)still dominate in the studied catchment with 90%; while deciduous species, mountain-ash (Sorbus aucuparia) and silver birch (Betula pendula) represent only 10%. It is evident, that the regrowth of forests was complicated by the extended clear-cut, acid environment and grass competition.Future forest projections respect the recommended adaptation strategies of IPCC (2015) and MZP (2017). Thus, by the end of the 21st century, the forest cover of the upper plain of the Jizera Mts.considers both existing stands of Norway spruce and deciduous species(Fagus sylvatica,Sorbus aucuparia,Betula pendula)replacing the recently introduced exotic confers.

3.2. Changing climate

3.2.1. Trends in air temperature and precipitation

Relatively high correlations between annual air temperatures and precipitation sums observed at Jizerka and neighbouring stations Desn′a - Souˇs and Jablonec n.N. (R = 0.96 and 0.83, respectively, by Rcrit= 0.33, n = 34, p = 0.05) allowed to extrapolate/interpolate data representing the studied catchment(Fig.5,Fig.6).

Both time series correspond to normal distribution and show an increasing tendency (2.2°C and 90 mm per 100 years), but, the statistically significant linear trend was found only for temperatures:correlation coefficient R=0.54(Rcrit=0.24,n=66,p=0.05),the two-tailed P value is below 0.0001 (extremely significant correlation), and not a significant departure from linearity (P value 0.09). The most intensive warming per a decade was identified in the winter season (months XII, I and II) 0.26°C (R = 0.36) and lowest in the autumn (IX, X and XI) 0.09°C (R = 0.22, not a significant trend at p = 0.05). Statistically significant was found the rising trend of annual extreme values: daily temperature maxima by 0.22°C (R = 0.37) and annual daily rainfall maxima by 4.5 mm(R = 0.25) per decade, Fig. 7 and Fig. 8.

3.2.2. Warming projections in 2071-2100

In the climate normal period 1961-1990, there was found a relatively strong correlation between the simulated and observed mean monthly temperatures (R = 0.86, Rcrit= 0.19, p = 0.05), but not significant for precipitation data (R = 0.11). In 2071-2100, the model shows an increase in the mean annual temperature of 2.2°C(B2 scenario)to 4.5°C(A2 scenario).Seasonal distribution of mean monthly temperatures (Fig. 9) shows higher increases in the high summer:maxima 4°C(B2)and 8.3°C(A2),in July-August.We can expect minor changes in annual amounts of precipitation, but,higher precipitation income during the winter(December-March)and lower in the summer (June-September), by 20-35%. These changes might lead to the shift in climax forest zones with the dominance of beech (Fagus sylvatica).

Generally, predictions of air temperatures and precipitation sums by global and regional climate models show a relatively high uncertainty,concerned particularly in the precipitation data,Prein and Gobiet (2017), Rajczak and Sch¨ar (2017). In mountain regions,these uncertainties accelerate namely with a lack of observation points(Kˇreˇcek and Punˇcoch′aˇr,2012)and complicated orography in mountain catchments (Ban et al., 2015). However, the climate projections in the Jizerka catchment correspond with outcomes of the recent world climate studies (IPCC, 2015).

Table 2 Vegetative canopy in the Jizerka catchment, 2016.

Fig. 3. Forest harvest in the Jizera Mts. by the end of the 19th century.Source: Museum of glass production, Kristi′anov, Czech Republic

Fig. 4. Changing canopy in the Jizerka catchment: 1880-2090.

3.3. Water budget

3.3.1. Application of the conceptual catchment model

The HBV hydrological model was calibrated for the period 1991-2010, and simulated daily runoff showed a good agreement with the observed data in the validation period of 2011-2016(Nash-Sutcliffe efficiency NSE = 0.82, Nash & Sutcliffe, 1970). In 2071-2100, water yield (mean annual outflow from the Jizerka catchment)decreases by 123 mm(i.e.14%,A2)and 65 mm(7%,B2)in comparison with the period 1991-2016,Table 3.By comparable precipitation inputs, the annual water loss by evapotranspiration increases from 434 mm (recent) to 557 mm (A2) or 499 mm (B2).

Seasonal runoff in 2071-2100 (Fig.10) has significantly shifted against 1991-2100: during four summer months (VI - IX), the runoff percentage (to the annual outflow) decreases from 47% to 29% (A2) and 33% (B2), while in the dormant season (X-IV), that percentage increases from 53% to 70%(A2) and 67%(B2). While in 1991-2016, the monthly runoff volume culminated by melting snowpack in April(14%of the annual flow),in 2071-2100,with the reduced snow pack volume and duration, that culmination is expected already in March.

Fig. 5. Annual precipitation at Jizerka: 1900-2016.

Fig. 6. Mean annual temperature at Jizerka: 1950-2016.

3.3.2. Evapotranspiration potential

Mean annual sums of reference evapotranspiration RET (FAO-56) calculated in daily steps (eq. (2), (Allen et al., 1998) are presented in Table 4.For the period 2071-2100,increases of RET by 128 mm(A2)and 65 mm(B2),correspond with results of the HBV catchment model:drop in water yield by 117 mm(A2)and 65 mm(B2).

Forest canopy can intensify evaporation loss by the interception process(eqs.(4)-(8)).In the warm season of 2016,interception loss observed in the mature stand of Norway spruce is given in Table 5.With climate change, in 2071-2100,can be expected rising values of wet canopy vaporization(E from 0.21 to 0.26 mm hour-1),but,on the other hand,with predicted forest cover(Fig.4),changing values of the canopy storage capacity S(from 1.7 to 1.4 mm),the threshold rainfall saturating the canopy P’ (from 2.8 to 2.0 mm), throughfall coefficient p (from 0.22 to 0.3), and stemflow coefficient psevidence at deciduous trees(from 0 to 0.24).Thus,the rising potential of evaporation can be still modified by changes in forest canopy:by identical rainfall patterns(Table 5),in 2071-2100,the interception loss can change from 271 mm (current state) to 292 mm (A2) and 246 mm (B2).

However, relatively higher evapotranspiration of spruce stands in the investigated catchment has been affected by an additional precipitation income by the canopy fog drip.For the recent canopy structure(Fig.4,Table 2),the annual income from fog drip is 81 mm(i.e., 10-12% of the annual runoff). In 2071-2100, by comparable atmospheric patterns and expected changes in forest stands(Fig.2),the fog drip income might decrease to 61 mm(i.e.75%of the recent fog drip).

Fig. 7. Maximum daily temperature at Jizerka: 1960-2016.

3.3.3. Moisture status

Perhumid environment of the Jizerka catchment is identified by the relatively high value of Lang’s moisture factor, ratio between mean annual precipitation and mean average air temperature(Shaw, 2011) LMF = 350. Correspondingly, the moisture status of the Jizerka catchment is characterized by high values of Ellenberg’s soil moisture index IF (Fig.11) and relatively small differences between potential and actual evapotranspiration. However, after clear-cut of mature spruce stands(1984-1990),the average IF value has increased from original 5.76 (identified before the forest harvest in 1982) to 6.45 (1991), reaching maximum 6.97 (2005), and,decreasing back to 6.33(2018)with forest regrowth.These changes are related to a temporary lower evapotranspiration after the forest harvest with the delay of a couple of years.

Fig. 8. Maximum daily rainfall at Jizerka: 1960-2016.

Fig. 9. Projected mean monthly temperatures at Jizerka (873 m a.s.l.): 2071-100.

Table 3 Monthly (Rm)and annual (Ra)runoff at the outflow of the Jizerka catchment in a ‘normal year’ of periods 1991-2016 and 2071-2100.

3.4. Floods and draughts

Frequency of peak flows in the Jizerka catchment is driven namely by probability of summer rainstorms causing flash floods.Annual maxima of daily rainfalls for the continuous observation period (1961-2019) show an extreme asymmetry (asymmetry coefficient Cs=1.92 by variation Cv=0.45).Using the Type 1 Gumbel extreme value distribution (Shaw, 2011), maxima of daily rainfall Pdmax(mm day-1)of return period 1,10,20,50,100 and 1000 years re given in Table 6.

Historical records in the Jizera Mts(Munzar&Ondr′aˇcek,2010).identified the maximum daily rainfall of 300 mm observed at Jizerka(and 345 mm at the nearby Nov′a Louka station)on the 29thJuly 1897. According to the frequency distribution (Table 5), this rainstorm (and related flood) might return in 600-700 years.During such an extreme event the role of forests (with retention capacity in an order of magnitude lower) is restricted only to a slope stabilisation and runoff retardation.

The set of annual peak flows observed at the Jizerka outlet(1982-2019) was extrapolated with the Log-Pearson Type III distribution (Shaw, 2011), Table 7. Discharge values with return periods of 10 and 100 years are Q10=2.5 and Q100=5.2 m3s-1km-2.With rising temperatures, according to the climate change scenarios A2 and B2,these values might increase up to Q10=3.3(3.0)and Q100= 6.8 (6.3) m3s-1km-2respectively; based on the hypothesis of intensifying rainstorm intensities by ca 7%peroK(Knist et al.,2018).

Frequencies of observed daily flows at the Jizerka outlet in 1982-2019 and modelled in 2071-2100 are given in Table 8. The 90% frequency discharge (value reached or exceeded during 330 days a year)is interpreted by the Water Act 254/2001 Coll(Tureˇcek,2002). as the ‘minimum residual discharge’ necessary to preserve the aquatic environment; in 2071-2100, there is the significant change in Q90%from 7.52 to 2.9 l s-1(decreasing to 39% of its original value).

Fig.10. Projected changes in mean monthly runoff Rm (mm per month). at the outlet of the Jizerka catchment.

Table 4 Mean annual reference evapotranspiration RET at Jizerka(873 m a.s.l.) by the climate change scenarios A2 and B2.

Table 5 Results of the Gash storm-based interception model applied in the mature spruce stand at the Jizerka catchment, 2016,I1, I2, I3, I4 - interception components (eqs.(4)-(8)), m, n - number of rainfalls (not enough, and saturating the canopy,respectively).

During 1982-2019, there were not registered any hydrological draught (manifesting an abnormally low streamflow according to Van Loon,2015).Significant summer drops(ca 50%)were identified in the soil moisture record (Fig. 12) at the top of the Jizerka catchment in years 2017-2019 with the lack of precipitation(reaching only 72%of the normal year),but not registered in lower parts of the catchment.

3.5. Acid atmospheric deposition

Fig.11. Ellenberg’s soil moisture indicator IF in the Jizerka catchment: 1991, 2005 and 2018.

Table 6 Frequency of maximum daily rainfall at Jizerka,1961-2010.

Table 7 Peak discharge QN(N is return period in years)at the Jizerka outlet,1982-2016 and 2071-2100.

Table 8 Daily discharge frequencies at the outlet of the Jizerka catchment in periods 1982-2019 and 2071-2100 (Qa is mean annual discharge).

Fig.12. Volumetric soil moisture at the top of the Jizerka catchment, 2010-2019.

In the period of observation(1982-2019),the ratio between the total load of nitrogen and sulphur in the Jizerka catchment has increased from 0.35(1982-1988)to 2.0(after 2011),and this trend is predicted also for the end of the 21st century (IPCC, 2013a,2013b).

3.6. Water quality and biota

In the Jizerka catchment,stream water acidification culminated in the 1980s with the peak in atmospheric deposition of sulphur and nitrogen (Fig.17), Table 9. In the 1990s, the recovery resulted from a synergy of the drop in SO2emissions (following the 1985 Sulphur protocol, Holen et al., 2013) and the decline of the forest canopy(reducing the surface area as well as the air mass turbulence above the canopy). Later, open field deposition of sulphur and nitrogen have been stabilised, and stream water chemistry became controlled mainly by the regrowth of forest stands (Fig. 4).

Considering water resources recharge, no health-based guideline value has been proposed for the pH of water,but,WHO(2004)suggest that optimum pH for drinking water quality between 6.5 and 8.5.Low pH values observed in the Jizerka stream indicate also low hardness(up to 10 mg of calcium and magnesium per litre)and particularly higher contents of aluminium (Table 9). The limit of aluminium in drinking-water is 0.1 mg l-1for large treatment facilities, and 0.2 mg l-1for small facilities (WHO, 2004). In 2071-2100, the expected drop in annual catchment outflow(Table 3)can cause higher mean annual concentrations of sulphate:3.4 mg l-1(A2 scenario) and 3.1 mg l-1(B2 scenario), nitrate:2.2 mg l-1(A2)and 2.0 mg l-1(B2)and aluminium:0.12 mg l-1(A2)and 0.11 mg l-1(B2).

The investigated stream at Jizerka was found fishless in the 1980s,and remained without any fish evidence also in 1990-2016.In 1994 (six years after the peak in acid load), the number of identified benthic taxa(36)there still indicated a‘strongly acidified environment’, and, in 2005, the number of taxa increased to 68 detecting a ‘moderately acidified environment’ (with pH between 5.0 and 6.3), Table 9.

Fig.13. The approximated bulk deposition of sulphur (S-SO4). and nitrogen (N-NO3 and N-NH4) at Jizerka,1850-2090.

Fig.14. The atmospheric load of sulphur (S-SO42-) estimated under the canopy. in the Jizerka experimental catchment,1880-2090.

Fig.15. The atmospheric load of nitrogen (N-NO) estimated under the canopy in the Jizerka experimental catchment,1880-2090.

Fig.16. The atmospheric load of nitrogen (N-NH) estimated under the canopy in the Jizerka experimental catchment,1880-2090.

Table 9 Timber harvest and selected parameters of stream water quality at the outlet of the Jizerka catchment,1982-2016.

3.7. Historical context of forest functions in the Jizera Mts

Since the first human interventions, priorities of forest management in the Jizera Mts. have gradually changed from early protection of forests (to secure the state border and big game hunting), to sustainable timber harvesting and recent multiresource forestry (Rabˇstejnek, 1969; Kreˇcmer & Kˇreˇcek, 1986).Particularly, with industrialisation in the 19th century, the occurrence of more frequent floods with intensive timber harvest in the Jizera Mts(Fig.3).was documented by Dlask in Kutnar(1941).This situation initiated the declaration of soil and water protection(Austro-Hungarian Acts 116 and 117/1884 Coll.). Particularly, the catastrophic flood of 1897 commenced there the complex of protection activities including construction of retention dams,torrent control works, and reforestation activities, based on Surell’s principle(1841,in Magnein,1937):“The presence of forest prevents the formation of torrents”.Anyway,by the relatively fast reforestation,the original forests of native tree species (Fagus sylvatica L., Picea abies, and Abies alba Mill.) were converted to the plantations of Norway spruce (Picea abies) with lower ecological stability,mentioned already by Gayer (1886).

In 1956, the Protected Landscape Region of the Jizera Mts. was claimed by the Czech government to preserve the unique natural elements of the region (Nature Protection Act 40/1956 Coll.); and,after the UN (1992) statement, the vital role of forest there was supported by Nature and Landscape Protection Act 114/1992 Coll.Later, with rising demand on drinking water supply, the Protected Headwater Area of the Jizera Mts. was declared by the Decree 40/1978 of the Czech government (according to the Water Act 138/1973, 254/2001 Coll.) to support the regional recharge of water resources (quantity and quality) by controlling the reduction of forests and drainage development.

However, the traditional forest benefits in protecting water quality were declined by the acid rain phenomena, when the acid atmospheric deposition (namely by the airborne emissions of sulphur and nitrogen) was rising with the canopy area and roughness (Kˇreˇcek & Hoˇrick′a, 2006). Acid rain affects the ecosystems of the Jizera Mts.(part of the European‘Black Triangle’)since the 1850s (Figs. 13 and 17); culminating in the 1980s when the widespread acidification of freshwaters with damage to drinking water quality and populations of fish and other aquatic organisms became critical (Table 9). First signs of recovery occurred in early 1990s with both, the 1985 Helsinki Protocol on the Reduction of Sulphur Emissions or their Transboundary Fluxes (with loads of sulphur decreased to approx. 40% of those in 1987, Holen et al.,2013, Fig. 13), and the extensive dieback and clear-cut of spruce plantations (Figs. 4 and 17).

By reforestation,in the 1990s,several exotic species(blue spruce-Picea Pungens,black spruce-Picea Mariana,etc.)were introduced to produce stands more resistant to air pollution (ˇSpul′ak et al.,2019). These exotic conifers show comparable performance(nutrient pools, topsoil chemistry, and growth) with stands of Norway spruce in the Jizera Mts.Nowadays,conversion projects to produce forests near the native composition are motivated by the need for higher ecosystem stability,the danger of biotic agents(e.g.high damage caused by cloven-hoofed game to stands ofP.pungens)and global climate change. While impacts of air pollution and acidification on forest functions was the most discussed topic in Central Europe in the 1980s and 1990s, nowadays, the global climate change and weather extremes are considered as the most important driving factor of mountain forests (Braun et al., 2015).

4. Discussion

While at the end of the 19th century, the protective role of forests in the Jizera Mts. concentrated on flood control, nowadays,water resources recharge and biodiversity became the focus. With the expected status of society,climate, and water resources by the end of the 21st century(FAO,2008;IPCC,2015),these priorities in headwater catchments might be still emphasized.

By the end of the 20th century, in the Jizera Mts., widespread acidification, forest dieback, and extensive clear-cut changed priorities in forestry for environmental reconstruction and assisted regeneration (Perrow & Davy, 2002). While emissions of sulphur are controlled effectively since the end of the 1980s(Protocol on the Reduction of Sulphur Emissions,Holen et al.,2013),nitrogen is still considered as a future problem. H°unov′a et al. (2020) reported a significant trend in reducing emissions of NOxin big industrial cities of the Czech Republic since 1994,but not in mountain regions.Since 1982,the ratio between the total load of nitrogen and sulphur in the Jizerka catchment has gradually increased from 0.35 to 2.0.Contrary to the statement of UNEP(2007)that the historical decline in temperate forests has been reversed between 1990 and 2005,forest ecosystems in the Jizera Mts. are still threatened by acidification.

First signs of stream water recovery from acidification were observed at Jizerka in the early 1990s as results of both the 1985 Helsinki Protocol on the Reduction of Sulphur Emissions (Holen et al., 2013) and reduced canopy area by forest clear-cut (Figs. 4 and 17, Table 9). Thirty years after, forest - water interactions in the Jizerka catchment are still affected by the prolonged acid atmospheric deposition and lack of environmental forestry practices.Without acid rain impact,streams in the Jizera Mts.would show a mean pH level close to 6.5 (Kˇreˇcek & Hoˇrick′a, 2006). According to Jenkins et al. (1990), besides the significant reduction in atmospheric emissions,stream water recovery requires also the control of forest canopy and treatment of forest soils by the addition of base cations; forests contribute to historical acidification of soils and surface waters by the nutrient demand. Thus, regrowth of coniferous stands can still slow down the recovery of surface waters in the Jizerka catchment (Kˇreˇcek et al., 2019). Invertebrate communities are considered the best indicator to assess the environmental impact of stream water acidification(Guerold et al.,2000),strongly correlated with water pH (Horecký et al., 2006). During the investigated period 1982-2019,the number of identified benthic taxa at Jizerka(Table 9)manifested the change from‘strongly acidified’to‘moderately acidified’ environment. While recovery of stream water chemistry there follows the drop of the atmospheric load in some five years,the revival of stream biota takes evidently a longer lag period (about ten to fifteen years).

The analysed instrumental time-series at Jizerka show a significantly rising trend in both annual mean air temperatures and annual maxima of daily temperatures (0.22°C per decade) with more intensive warming in the winter (December-February,0.26°C per decade), and in annual daily rainfall maxima (4.5 mm per decade), but elevated annual precipitation (9 mm per decade)was recognized as not statistically significant. In comparison with 1961-1990,the projected mean annual temperature for 2071-2100 rises by 4.6°C (A2 scenario) and 3.0°C (B2 scenario); mean monthly temperatures rise particularly in the high summer by 9°C(A2)and 4°C(B2).Generally,climate change predictions are linked to relatively high uncertainties (an IPCC outcome refers to the likelihood greater than 66% probable, Latif, 2011). In mountain regions,particularly,uncertainties in precipitation accelerate namely with the lack in observation points (Kˇreˇcek and Punˇcoch′aˇr, 2012,Prein&Gobiet,2017)and complicated orography(Ban et al.,2015).However, the climate projections in the Jizerka catchment correspond well with recent trends in instrumental data.The projected warming can lead to the shift in climax zones to those with dominating beech forests, more diverse mixed stands with higher ecological stability,and lower acceleration of the acid atmospheric load. However, the projection of future forests considers the survival of existing spruce stands and conversion of recently introduced exotic conifers to deciduous species. Such a scenario corresponds with the findings of Vacek and Balcar (2004) and ˇSpul′ak et al.(2019).The survival of native spruce(Picea abies)in the changing mountain environment can be still supported by planting its more resistant forms (Frýdl et al.,2011).

In 2071-2100, annual water yield of the Jizerka catchment is expected to decrease by 123 mm (i.e.14%, A2) and 65 mm (i.e. 7%,B2),Table 3,with the seasonal shift from April to March(according to changes in snow accumulation and melting). The increase in vaporizing intensity(Table 4)can be still modified by forest canopy:the annual canopy interception can change from recent 271 mm to 292 mm (A2) or 246 mm (B2), Table 5. Moreover, the annual fog drip income might decrease from the recent 81 mm(ca.10%of the annual runoff) to 61 mm (i.e. 75 % of the recent fog drip). Despite these changes in water balance, the water supply potential of this headwater catchment will not be significantly declined: the contemporary high ‘relative water yield’ RWY = 4.69 (long-term runoff ratio between the catchment cell and the whole lowland area,Viviroli et al.,2007)might still rise with expected more rapid drops in lowland mean annual discharge(reported by Trnka et al.,2016). The study of IPCC (2013a, 2013b) highlighted namely the problem of increasing draught episodes as a serious limiting factor of water supply facilities and biodiversity conservation. During 1982-2019,there were still not registered hydrological draughts at the outlet of the Jizerka catchment. Similarly, Hisdal et al. (2001)reported that it is still not possible to conclude when hydrological draughts in Europe at the end of the 19th century have become more severe or frequent. In 2071-2100, the 90% - frequency discharge Q90%(‘minimum residual flow’to preserve aquatic life)is considered to drop from 7.52 to 2.9 l s-1(i.e.to 39%of recent value),Table 8, while the peak flow of 100 year return period Q100could increase from original Q100=5.2 m3s-1to 6.8(6.3)m3s-1for A2(B2)scenarios, respectively, Table 7. This shift in peak flow frequencies can significantly affect the flood control design, for example, the design flow of Q100will shift to Q50(B2) or Q40(A2).

The expected drop in annual runoff can cause higher mean annual concentrations of sulphate: 3.4 mg l-1(A2) and 3.1 mg l-1(B2); nitrate: 2.2 mg l-1(A2) and 2.0 mg l-1(B2), and, aluminium:0.12 mg l-1(A2)and 0.11 mg l-1(B2),Table 9.These values might still rise by episodic runoff events (Kˇreˇcek & Hoˇrick′a, 2006), but, they cannot affect the drinking water standard or the biodiversity conservation. Generally, the decline of stream water quality is associated with lower stream flows and higher contents of nutrients(Strayer&Dudgeon,2010).On the other hand,climate warming can positively influence biogeochemical weathering associated with higher alkalinity of mountain waters (Catalan et al., 2014), and,thus, stimulate their chemical and biological recovery from acidification(Kop′aˇcek et al.,2019).In addition,rising water temperature might allow colonization of stream waters by less stenothermic species from lower parts of the watercourse, and, thus to avoid a sharp change of the original (pre-acidification) biodiversity of the aquatic environment. Stream biota reflects also the stream bed structure (Rosenberg & Resh, 1993), thus, a higher frequency of peak-flows(Table 7) might intensify changes in stream bed with a higher potential for colonization by alien species. The drop in minimum flows (Table 8) is still not causing an intermittent flow regime, but, decreasing Q90% can significantly affect stream water biota.

5. Conclusions

During the last two hundred years, forests in the Jizera Mts.gradually changed from native stands to spruce plantations,introducing exotic conifers resistant to the air pollution,and back to the preference of stands near the native composition.While at the end of the 19th century,the protective role of forests concentrated mainly on flood control, nowadays, protection of water resources and biodiversity became the focus.

The study in the Jizerka experimental catchment, 1982-2019,shows that the acid atmospheric deposition (and stream water quality) can be modified by forestry practices, and, the recent recovery from acidification there is the result of both the drop in sulphur emissions and the clear-cut of mature spruce stands.

The projected environmental changes in 2071-2100 will not substantially decline the high potential in water resource recharge,water quality and aquatic biodiversity,or,start reverse processes in recent recovery from acidification in the upper plain of the Jizera Mts.

Declaration of competing interest

None.

Acknowledgement

This research was supported by the Ministry of Education,Youth and Sports (Czech Republic, INTER-EXCELLENCE LTC 17006), European Cooperation in Science and Technology (COST Action CLIMO CA15226) , Czech Technical University in Prague (Czech Republic,Project SGS20/110lOHR1l2T/11), and Czech Science Foundation(Czech Republic, PROGRESS: 20-08294S).