Hydrochemical Resilience of Mountain Forest Catchments to Bark Beetle Disturbance: A Central European Study

Abstract

Over the last decade, bark beetle outbreaks have significantly impacted forests in Central Europe, causing extensive loss of forest cover. We evaluated the impact of partial deforestation in three mountain forest catchments in the Jeseníky Mountains, comparing them with the unaffected Červík catchment (Beskydy Mountains) and the severely affected Pekelský stream catchment (Czech-Moravian Highlands). Atmospheric deposition in the catchments was similar, with total element input driven primarily by precipitation volumes rather than ion concentrations. We did not observe the hypothesized increase in DOC and nitrogen export, although nitrate outflow was slightly higher than atmospheric input in two cases. Significant export of calcium, magnesium, and bicarbonates was driven mainly by the geology of the individual catchments. The limited impact of bark beetle outbreaks on DOC dynamics can be attributed to the relatively low proportion of clear-cut areas and the rapid development of ground vegetation on impacted sites.

1. Introduction

Forest catchments play a key role in the hydrological cycle [1,2] and serve as a crucial source of high-quality drinking water for the human population in many regions [3,4]. The chemical composition of surface waters draining from these catchments is a sensitive indicator of overall ecosystem health [5] and its responses to external pressures [6,7]. Traditionally, the hydrochemistry of forest catchments has been understood as the outcome of a complex interplay of several factors, including the geological bedrock determining the baseline mineralization, soil processes, climatic conditions, forest management, and long-term inputs of atmospheric deposition.

Climate change, manifested mainly through more frequent and intense droughts [8,9], has weakened forest vitality [10,11], thereby creating conditions for massive outbreaks of the bark beetle (Ips typographus) [12]. Especially in Central Europe, where Norway spruce (Picea abies) was planted as the major productive species in even-aged monocultures, these outbreaks have led to devastating disturbances over large areas since 2018 [13,14]. In the Czech Republic, salvage logging peaked in the years 2018–2022, when the total amount of wood felled due to bark beetle infestation was nearly double the normal logging volume [15,16]. The total area of salvage clear-cuts created in 2018–2023 exceeded 200,000 ha [17]. Significant loss of forest canopy cover was also reported from other Central European countries, e.g., 501,000 ha in Germany during the period 2018–2020 [18]. Unprecedented damage by bark beetle was reported by Hallas et al. [19] for Austria and by Carr et al. [20] for Poland, which led to justified concerns about the impact on ecosystem services, specifically regarding the capacity of forests to fulfill their ecological functions.

The loss of forest cover over large areas represents an abrupt change in site conditions and poses significant environmental risks connected with changes in micro-climatic conditions [21], soil erosion [22,23], increased decomposition of organic matter [24], nutrient leaching [25] and carbon loss [26], and shifts in bacterial [27], fungal [28], vascular plants [29] and insect populations [30]. The consequences of large-scale disturbances on the hydrological cycle [31,32], water quality and elemental cycling are not yet fully understood. Deforestation and soil surface disruption can accelerate the mineralization of organic matter [33], intensify nitrification, and lead to increased nutrient leaching [34,35], especially of nitrogen [36], into surface waters. It remains an open question to what extent these new drivers override or modify traditional controlling factors, such as geology or the hydrological regime, and how catchments with different levels of disturbance respond.

In 2020, three small research catchments were established in the Jeseníky Mountains in the northeastern part of the Czech Republic (Central Europe) to study changes in the hydrological regime of forests affected by bark beetles [37]. In addition to detailed monitoring of meteorological and hydrological parameters [38], the chemical composition of atmospheric deposition and runoff water was measured to evaluate changes in the elemental balance of disturbed ecosystems. Given that the study in the Jeseníky Mountains is relatively short-term and data before the start of salvage logging are missing, we compare the results with two other long-term monitored forested catchments: (i) the Pekelský stream in the Bohemian-Moravian Highlands, selected as a comparative disturbed site with different geological bedrock (paragneiss), and (ii) the Červík stream in the Beskydy Mountains, serving as an undisturbed reference site with intact forest cover.

The study covers the period 2022–2024. The objectives were to: (1) compare the atmospheric input and runoff output of elements in catchments with varying degrees of disturbance; (2) determine whether the recent bark beetle outbreak caused significant nutrient leaching; and (3) evaluate the relative importance of disturbance versus geological bedrock in controlling stream water chemistry.

2. Materials and Methods

2.1. Study Area

In 2020, experimental monitoring was established in the Hrubý Jeseník Mts. within the catchments of three right-hand tributaries of the Černá Opava River: the Slučí (SL), Sokolí (SO) and Suchý (SU) streams (hereafter referred to collectively as the ‘3S’ catchments) (Figure 1). The aim was to assess the impact of the advancing bark beetle outbreak on hydrological conditions in mountain catchments.

The management strategies regarding the bark beetle outbreak differ among the catchments. While typical forest protection measures, involving salvage logging of infested stands, were applied in the SO catchment, the SU catchment includes a nature reserve (48.8 ha) where bark beetle control is prohibited and dead trees are left to decay. Part of the SL catchment falls within a Natura 2000 site (Birds Directive); here, felling of infested trees is permitted, but high stumps (~2 m) must be retained along the stream. All catchments are fully forested, with elevation ranging from 630 to 1216 m a.s.l. The parent material consists of igneous and metamorphic rocks (meta-granitoids, blastomylonites and quartzites), with fluvial sediments in the lower reaches. The most common soil types are Cambisols, Podzols and, at higher elevations, Leptosols. All the catchments belong to the cold climatic region C6-C7 according to Quitt’s classification [39].

The Pekelský stream (ZE) is a right-hand tributary of the Švihov (Želivka) drinking water reservoir in the Bohemian-Moravian Highlands and has been monitored since the 1970s [40]. The catchment covers an area of 124 ha, comprising 122 ha of forest land and 2 ha of agricultural (arable) land. Forest management in the catchment is typical productive even-age forestry in stands dominated by Norway spruce. The elevation ranges from 374 to 470 m a.s.l. The soil cover consists of deep Cambisols and Stagnosols of various subtypes, with moder humus forms developed on loams, slope debris, eluvium, and isolated Neogene clays and gravels. The geological bedrock consists of Moldanubian rocks, specifically medium- to coarse-grained biotite paragneisses. The area belongs to the moderately warm climatic region MW7 according to Quitt’s classification [39].

The Červík (CE) stream is located in the Moravian-Silesian Beskydy Mts. and has been monitored for more than 70 years [41]. The catchment has an area of 185 ha and is fully forested. Productive Hercynian mixed forests (Norway spruce, European beech and silver fir) are managed less intensively as the catchment is part of a protected landscape area. The elevation ranges from 696 to 958 m a.s.l. The Červík stream flows into the Šance reservoir; the experimental catchment is situated in the upper headwaters. The bedrock is formed by sandstone and shale (flysch), overlain by Cambisols approximately 1 m deep. The lower part of the catchment belongs to the moderately warm region MW1 and the upper part to the cold climatic region C7 according to Quitt’s classification [39]. Basic characteristics of each catchment are shown in

Table 1. Catchment characteristics.

Stream	Slučí (SL)	Sokolí (SO)	Suchý (SU)	Červík (CE)	Pekelský (ZE)
Catchment area (km2)	3.98	3.99	2.05	1.85	1.24
Stream length (m)	3436	3632	1680	5426	1671
Minimum catchment elevation (m a.s.l.)	649	614	646	505	374
Maximum catchment elevation (m a.s.l.)	1202	1215	1058	958	470
Mean catchment elevation (m a.s.l.)	914	919	892	696	445
Forest area (km2) *	3.98	3.99	2.05	1.85	1.22
Forest cover (%) *	100	100	100	100	98
Mean annual precipitation (mm) **	984	947	866	1071	676
Mean runoff (L·s−1·km−2) **	9.96	9.60	7.16	18.18	6.35

* Relates to areas classified under Czech forestry law (Act No. 289/1995 Coll.) as “land designated for forestry purposes”. ** For the period from June 2022 to May 2024.

.

2.2. Identification of Clear-Cut Areas

To quantify the area and development of clear-cuts in each catchment, georeferenced aerial imagery provided by the Czech Office for Surveying, Mapping and Cadastre (ČÚZK) was used. Orthophotos were acquired at three-year intervals until 2012, and subsequently at two-year intervals; these data are publicly available for download or via WMS services. The areas of newly created clearings were manually vectorized using ArcMap 10.5 software for the period 2006–2024.

2.3. Sampling of Precipitation and Stream Water; Analytical Methods

Precipitation for chemical analyses was sampled in an open area (bulk deposition = wet + dry deposition) according to ICP Forests methods [42] using three funnels with a collection area of 500 cm²; water was collected in polyethylene vessels placed in soil pits to shield them from temperature fluctuations and sunlight. Samples were collected three times per month and subsequently pooled into volume-weighted composite samples on a monthly basis for chemical analysis.

Stream water samples were collected in polyethylene bottles on the same days as atmospheric deposition, at the respective catchment outlets (discharge measurement profiles). They were also pooled into monthly samples for chemical analysis.

The following parameters were analyzed: pH (potentiometrically) and alkalinity (titrimetrically) using a pH meter (798 MPT Titrino, Metrohm, Herisau, Switzerland); conductivity (conductimetrically) using a conductometer (Orion 3 Star, Thermo Scientific, Waltham, MA, USA); total nitrogen (TN) and dissolved organic carbon (DOC) by elemental analysis (TOC/TN analyzer Formacs, Skalar, Breda, The Netherlands); anions (S-SO₄²⁻, N-NO₃⁻, F⁻, Cl⁻) by ion chromatography (EcoIC, Metrohm); cations (Al, Ca, Cu, Fe, K, Mg, Mn, Na, Zn) by optical emission spectrometry (ICP OES 5110 VDV, Agilent Technologies, Santa Clara, CA, USA) and N-NH₄⁺ and P-PO₄³⁻ by automated spectrometry (SANPlus, Skalar). The concentrations were expressed in mg·L⁻¹; for statistical analysis, these concentrations were converted into meq·L⁻¹.

The study is based on a two-year dataset from June 2022 to May 2024, which allows for comparison across all monitored catchments.

2.4. Hydrological Data

As previously mentioned, hydrological monitoring in the Hrubý Jeseník Mts. began in 2020. In contrast, the Červík and Pekelský stream catchments are among the longest-monitored small forested catchments in the Czech Republic, with records dating back to 1954 and 1976, respectively.

Precipitation in the Hrubý Jeseník catchments was monitored by three automatic meteorological stations (one in each catchment) and nine totalizers (three in each catchment). The meteorological stations measured air temperature, air humidity, global solar radiation, and precipitation continuously at 10 min intervals. Precipitation was measured using a Pronamic Pro (Skjern, DK) rain gauge with a 200 cm² collection area and 0.2 mm resolution. Data were stored on a Greybox N2N datalogger (EMS Brno, Czech Republic), powered by a solar panel and a battery system. Since the rain gauges were unheated, precipitation measurements during the winter season relied on totalizers, which were measured manually every three months, starting at the beginning of the hydrological year (1 November). The areal precipitation for each catchment was calculated using the Horton method [43].

For the 3S catchments, water levels were measured using LMP307 hydrostatic submersible pressure sensors (FIEDLER AMS, Czech Republic) installed in natural, unmodified stream beds [37]. Discharge was calculated from continuous water level measurements using rating curves derived specifically for the stream cross-sections.

In the Červík and Pekelský stream catchments, precipitation was monitored using the same instrumentation as in the Hrubý Jeseník Mts. (rain gauges and totalizers), but discharge was measured using standardized open rectangular flumes, with water levels converted to discharge using established rating curves. Water levels were monitored using PSH-30 float gauges and TSH22 pressure and temperature sensors (both FIEDLER AMS), with data managed via Fiedler M4016 control units.

2.5. Elemental Budgets of the Catchments

Annual elemental budgets for the individual catchments were determined based on a two-year monitoring period (June 2022–May 2024). The budget was defined as the difference between elemental input via atmospheric deposition and output via runoff. Input was quantified based on the chemical analysis of bulk deposition and monthly precipitation volumes in each catchment. Output was quantified by multiplying the total monthly discharge volume (measured continuously) by the element concentration determined in the corresponding monthly composite water sample. The mean annual values were derived by averaging the data from the two monitored years, with the net balance calculated as the difference between annual input and annual output. The resulting values are expressed in kg·ha⁻¹·yr⁻¹.

2.6. Statistical Methods

All statistical analyses were performed using the open-source software R, version 4.5.1 [44].

2.6.1. Compositional Data Analysis (CoDa)

In numerous studies, hydrochemical data are treated as compositional data, represented by a matrix of non-negative relative values constrained by a constant sum [45,46]. The application of standard statistical methods to such data can lead to misleading inferences and spurious correlations [47,48]. To analyze these data correctly, transformations such as log-ratios are employed to map the compositions into (standard) Euclidean space, enabling the use of conventional multivariate techniques. Alternatively, methods that operate directly within the simplex geometry of compositional data can be applied.

The ILR-ION plot was used as a visualization tool derived from Compositional Data Analysis (CoDA) [49], specifically using isometric log-ratio (ilr) transformations to represent major ion composition in water samples. This approach mitigates the limitations of traditional plots like the Piper diagram, such as data compression and overlap at plot boundaries, by expanding edge data and clustering central data for clearer interpretation [50,51]. Four coordinates convert the compositional data from the simplex to real space [46]:

$$z_1 = \sqrt{{\displaystyle \frac{2}{3}}}ln{\displaystyle \frac{\sqrt{\left[{\mathrm{C}\mathrm{a}}^{2+}\right]\left[{\mathrm{M}\mathrm{g}}^{2+}\right]}}{\left[{\mathrm{N}\mathrm{a}}^{+}+{\mathrm{K}}^{+}\right]}}$$

(1)

$$z_2={\displaystyle \frac{1}{\sqrt{2}}}ln{\displaystyle \frac{\left[{\mathrm{C}\mathrm{a}}^{2+}\right]}{\left[{\mathrm{M}\mathrm{g}}^{2+}\right]}}$$

(2)

$$z_3=\sqrt{{\displaystyle \frac{2}{3}}}ln{\displaystyle \frac{\sqrt{\left[{\mathrm{C}\mathrm{l}}^{-}\right]\left[{\mathrm{S}\mathrm{O}}_4^{2-}\right]}}{\left[{\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}\right]}}$$

(3)

$$z_4={\displaystyle \frac{1}{\sqrt{2}}}ln{\displaystyle \frac{\left[{\mathrm{C}\mathrm{l}}^{-}\right]}{\left[{\mathrm{S}\mathrm{O}}_4^{2-}\right]}}$$

(4)

2.6.2. Principal Component Analysis (PCA) on Isometric Log-Ratio (Ilr) Transformed Data

Principal Component Analysis (PCA) was applied to the isometric log-ratio transformed data.

The initial matrix containing eight compositional variables (Ca²⁺, Mg²⁺, Na⁺, K⁺, Cl⁻, SO₄²⁻, HCO₃⁻, NO₃⁻) was transformed into seven orthogonal ilr variables. These transformed variables were then merged with standardized supplementary variables (EC, pH, DOC, TN), and the combined dataset was subjected to PCA. The ILR transformation and PCA were performed in the R environment [44], utilizing the ‘compositions’ package [52]. Samples were visualized in a biplot based on their PC scores, and their grouping was highlighted using 95% confidence ellipses.

3. Results

3.1. Spatiotemporal Dynamics of Clear-Cuts

The intensity and timing of disturbances in the respective catchments were quantified by analyzing the area of new clear-cuts (ha) within defined time intervals, determined by the availability of aerial imagery (

Table 2. Quantification of disturbance: Area of clear-cuts (ha) in the monitored catchments from 2003 to 2024.

Period	SL	SO	SU	ZE
2003–2006	1.33	1.32	0.00	8.40
2006–2009	0.22	0.00	0.82	4.08
2009–2012	1.17	1.37	0.66	5.54
2012–2014	0.20	0.00	0.00	1.06
2014–2016	1.03	0.48	0.02	3.73
2016–2018	3.88	9.22	2.49	6.28
2018–2020	8.73	18.81	5.81	21.86
2020–2022	15.81	7.73	7.46	18.00
2022–2024	0.24	6.08	0.00	7.06

). Data for the reference catchment CE are not presented, as no significant logging occurred there during the study period. Between 2003 and 2016, logging activity in all monitored catchments was minimal, with clear-cut areas generally not exceeding 1.5 ha per interval. A marked increase in logging, indicating the onset of the bark beetle outbreak, was recorded in the period 2016–2018.

This escalation continued in the subsequent intervals (2018–2020 and 2020–2022), representing the peak of the disturbances. The intensity of disturbance differed substantially among the catchments. The highest and most sustained disturbance occurred in the ZE catchment, where 39.9 ha (32.1% of the catchment area) was logged during the critical four-year period 2018–2022. Catchments in the Jeseníky Mountains were also significantly affected, mainly at lower elevations (up to 800 m a.s.l.); however, the relative extent of the disturbance was lower. In SO, the clear-cut area in 2018–2022 reached 26.5 ha (6.7% of the area), while in SL it was 24.5 ha (6.2%) and in SU 13.3 ha (6.5%). During the period of hydrochemical data collection (2022–2024), logging activity nearly ceased in SL and SU, while in SO and ZE it persisted (6.1 ha and 7.1 ha, respectively).

3.2. Precipitation and Stream Water Chemistry

The chemistry of precipitation is summarized in

Table 3. Chemical composition of precipitation in the monitored localities. Values are in mg·L⁻¹, except for pH (unitless) and electrical conductivity (μS·cm⁻¹).

Locality *		pH	NH4+	Ca2+	K+	Mg2+	Na+	Cl−	NO3−	SO42−	HCO3−	DOC	TN	EC
SL + SO + SU	Mean	5.56	0.34	0.28	0.29	0.06	0.17	0.34	0.91	0.78	1.15	3.02	0.67	10.06
	STD	0.38	0.32	0.18	0.19	0.02	0.11	0.20	0.53	0.32	1.13	1.04	0.34	2.94
CE	Mean	5.40	0.51	0.31	0.49	0.09	0.01	0.35	2.32	0.79	0.87	2.20	1.13	14.77
	STD	0.48	0.45	0.17	0.45	0.05	0.02	0.17	1.98	0.38	1.50	0.86	0.78	7.10
ZE	Mean	5.67	0.47	0.35	0.34	0.11	0.21	0.41	1.83	0.94	0.41	2.65	1.16	20.05
	STD	0.51	0.27	0.20	0.49	0.10	0.11	0.19	1.35	0.57	1.01	1.49	1.49	23.38

* SL—Slučí stream catchment, SO—Sokolí stream catchment, SU—Suchý stream catchment, CE—Červík stream catchment, ZE—Pekelský stream catchment. DOC—dissolved organic carbon, TN—total nitrogen, EC—electrical conductivity.

. In the Jeseníky Mts., precipitation was monitored at a single site due to the close proximity of the individual catchments. Although there were no pronounced differences in the chemical composition of precipitation among the three regions, the Jeseníky Mts. exhibited the lowest concentrations of potassium (K⁺) and magnesium (Mg²⁺) ions (0.29 mg·L⁻¹ and 0.06 mg·L⁻¹, respectively). Calcium concentrations (Ca²⁺) were also slightly lower (0.28 mg·L⁻¹) compared to CE (0.31 mg·L⁻¹) and ZE (0.35 mg·L⁻¹). Concentrations of nitrate (NO₃⁻), total nitrogen (TN), and ammonium (NH₄⁺) were lowest in the Jeseníky Mts. and highest in the CE catchment. In contrast, dissolved organic carbon (DOC) was highest in the Jeseníky Mts (3.02 mg·L⁻¹, compared to 2.20 mg·L⁻¹ in CE and 2.65 mg·L⁻¹ in ZE).

Table 4. Chemical composition of runoff in the individual streams. Values are in mg·L⁻¹, except for pH (unitless) and electrical conductivity (μS·cm⁻¹).

Locality *		pH	NH4+	Ca2+	K+	Mg2+	Na+	Cl−	NO3−	SO42−	HCO3−	DOC	TN	EC
SL	Mean	7.17	0.01	7.16	0.78	0.90	1.63	1.15	2.82	9.60	14.71	1.51	0.73	61.84
	STD	0.11	0.00	0.65	0.07	0.06	0.07	0.06	0.33	0.47	2.43	0.34	0.15	4.40
SO	Mean	7.32	0.01	9.99	0.80	1.11	1.90	1.14	3.52	11.77	21.94	1.78	0.87	79.48
	STD	0.16	0.01	1.22	0.06	0.12	0.11	0.07	0.63	0.90	3.24	0.52	0.16	6.85
SU	Mean	7.30	0.01	10.05	0.70	1.60	2.17	1.45	2.32	16.22	20.34	2.02	0.62	87.04
	STD	0.14	0.01	1.16	0.05	0.17	0.12	0.10	0.63	1.07	3.14	0.66	0.18	7.32
CE	Mean	7.01	0.02	4.43	0.99	2.09	2.38	0.94	1.32	11.27	14.04	2.83	0.42	60.83
	STD	0.21	0.05	0.81	0.18	0.39	0.64	0.05	0.23	0.69	5.38	1.01	0.24	10.20
ZE	Mean	7.76	0.01	11.69	1.66	3.64	8.06	4.29	3.80	22.94	38.47	5.93	0.94	144.70
	STD	0.31	0.00	1.25	0.16	0.26	0.48	0.38	1.80	5.30	4.77	1.22	0.36	13.00

* SL—Slučí stream catchment, SO—Sokolí stream catchment, SU—Suchý stream catchment, CE—Červík stream catchment, ZE—Pekelský stream catchment. DOC—dissolved organic carbon, TN—total nitrogen, EC—electrical conductivity.

summarizes the runoff chemistry in the monitored streams. The pH of runoff (7.0–7.8) was uniformly higher than the pH of precipitation water (5.4–5.7) in all catchments. The electrical conductivity (EC) of runoff was several times higher than that of precipitation; values in the Jeseníky Mts. (61.84 μS·cm⁻¹–87.04 μS·cm⁻¹) were slightly higher than in CE (60.83 μS·cm⁻¹) but lower than in ZE (144.7 μS·cm⁻¹). This reflects the concentration of base cations and sulfates, which were distinctly higher in ZE than in the other catchments. Ammonium concentrations were very low in all catchments (0.01–0.02 mg·L⁻¹), whereas nitrate concentrations were consistently higher than in precipitation. The highest nitrate levels were observed in ZE (3.80 mg·L⁻¹), followed by intermediate levels in the Jeseníky Mts. catchments (2.32–3.52 mg·L⁻¹) and the lowest in CE (1.32 mg·L⁻¹).

Differences in runoff chemistry among the studied catchments are clearly illustrated in Figure 2. The Slučí stream (SL) exhibits significantly lower concentrations of calcium, magnesium, and sulfates compared to the other Jeseníky Mts. catchments. The elevated values of most parameters in the ZE catchment, including pH and electrical conductivity, are also notable.

3.3. Retention and Export of Elements

To assess the function of the individual catchments as sources or sinks of elements, net annual budgets (output via runoff minus input via precipitation) were calculated (Figure 3,

Table 5. Net annual elemental budget (kg·ha⁻¹·yr⁻¹): Negative values (light orange) indicate net retention (sink), while positive values (light blue) indicate net release (source) of substances from the catchment.

	SL	SO	SU	CE	ZE
N-NH4+	−2.54	−2.54	−2.55	−3.25	−2.27
N-NO3−	0.24	0.61	−0.56	−3.48	−0.63
TN	−3.89	−3.59	−4.77	−8.25	−5.89
Ca2+	20.15	26.67	19.03	19.14	21.33
Mg2+	2.33	2.76	2.94	9.69	6.76
K+	−0.08	−0.17	−0.95	0.40	1.72
Na+	2.26	4.35	3.50	11.27	15.02
Cl−	−0.66	0.68	0.59	1.41	6.37
S-SO42−	8.03	9.39	9.60	17.96	14.18
HCO3−	34.85	52.30	31.42	54.78	70.68
DOC	−22.01	−21.75	−22.55	−7.12	−4.24

). All catchments functioned as net sources of weathering products, especially calcium (Ca²⁺) with a mean export of over 21 kg·ha⁻¹·yr⁻¹ and bicarbonates (HCO₃⁻), reflecting the dominance of chemical weathering processes over atmospheric deposition. The ZE catchment exhibited the highest release of HCO₃⁻ (70.68 kg·ha⁻¹·yr⁻¹) and Na⁺ (15.02 kg·ha⁻¹·yr⁻¹). This confirms that the gneissic crystalline bedrock (paragneiss) with a rich feldspar content undergoes intense weathering, which releases large amounts of Na⁺ and Ca²⁺. The Jeseníky catchments (SL, SO, SU) showed the lowest net release of HCO₃⁻ (mean 39 kg·ha⁻¹·yr⁻¹), which is consistent with the low absolute concentrations of HCO₃⁻ in these waters and their lower buffering capacity compared to CE and ZE.

All catchments acted as significant sinks for total nitrogen, with CE showing the highest retention (−8.25 kg·ha⁻¹·yr⁻¹), indicating intense biological consumption or accumulation of nitrogen in soils. All catchments were also net sources of chloride Cl⁻ (except SL), with the ZE catchment being the largest source (6.37 kg·ha⁻¹·yr⁻¹).

All catchments acted as sinks for DOC, with the ZE showing the lowest retention (−4.24 kg·ha⁻¹·yr⁻¹).

3.4. Hydrochemical Characteristics of Catchments

The hydrochemical characteristics of the catchments, based on the analysis of isometric log-ratios of major ions, are presented in the Ilr-Ion plot (Figure 4). The upper left panel shows the balance between Ca²⁺ and Mg²⁺ (x-axis) and the balance between Cl⁻ and SO₄²⁻ (y-axis). Data for all monitored streams plot on the positive side of the x-axis, indicating that Ca²⁺ concentrations are generally higher than those of Mg²⁺. Concurrently, all samples exhibit negative y-coordinates, indicating higher concentrations of SO₄²⁻ compared to Cl⁻.

Despite these common features, significant differences can be observed among the catchments. The CE catchment is clearly distinct, primarily due to a very low Ca²⁺ to Mg²⁺ ratio. The ZE catchment is characterized by the highest ratio of chlorides to sulfates. The values from the three Jeseníky Mts. catchments also differ, although SL and SO are clustered closer to each other than to SU.

The upper right panel (Figure 4B) displays the balance between Cl⁻/SO₄²⁻ and HCO₃⁻ (x-axis) versus Cl⁻ and SO₄²⁻ (y-axis). The negative coordinate on the x-axis indicates that HCO₃⁻ is the dominant anion. Bicarbonates in natural surface waters originate mainly from the weathering of geological bedrock.

The lower left panel (Figure 4C), where the y-axis represents the balance between alkaline earth metals (Ca²⁺, Mg²⁺) and the alkali metals (Na⁺ + K⁺), allows for the classification of waters based on their compositional type. While the SU, SL, and SO catchments belong to the Ca-dominant type (Ca²⁺ cations predominate), the ZE and CE catchments are classified as a mixed cation type, falling between the Ca, Mg, and Na+K categories. The mean values of isometric log-ratio coordinates z₁, z₂, z₃, and z₄ for the studied catchments are presented in

Table 6. The mean values of isometric log-ratio coordinates z₁, z₂, z₃, and z₄.

Catchment *	z1	z2	z3	z4
CE	0.35	0.18	−1.26	−0.76
SL	0.46	1.11	−1.31	−0.48
SO	0.59	1.20	−1.58	−0.66
SU	0.67	0.94	−1.29	−0.72
ZE	0.05	0.47	−1.24	−0.1

* CE—Červík stream catchment, SL—Slučí stream catchment, SO—Sokolí stream catchment, SU—Suchý stream catchment, ZE—Pekelský stream catchment.

.

The chemical composition of the three Jeseníky catchments corresponds to the classic calcium-bicarbonate type. This composition is primarily influenced by silicate weathering of the bedrock [53], which predominantly includes silicate metamorphic rocks (phyllites, schists, metagranitoids) and, locally, metabasites (amphibolites). Positive z₁ values (ranging from 0.48 to 0.67) indicate that the weathering of minerals supplying divalent cations is the main source of dissolved ions, a feature typical of environments containing metabasites and metagranitoids.

Furthermore, high positive z₂ values (0.94 to 1.11) demonstrate that the weathering of Ca-silicates (plagioclase, amphibole) is the dominant cation release mechanism. The z₃ values confirm that water composition is primarily governed by natural weathering processes. Negative z₃ values (ranging from −1.13 to −0.84) reflect the dominance of weathering-derived bicarbonate over the input of strong acid anions (Cl⁻ and SO₄²⁻). Negative z₄ values (ranging from −1.49 to −1.29) indicate that sulfate (SO₄²⁻) prevails over chloride (Cl⁻). This suggests a minor but discernible influence of atmospheric sulfur deposition or sulfide weathering in the bedrock.

The SO catchment exhibits the greatest dominance of weathering-derived HCO₃⁻. In contrast, the SU catchment displays the relatively highest proportion of Cl⁻/SO₄²⁻. Notably, this catchment also has the lowest z₂ value (0.94), indicating a relatively higher magnesium content in relation to calcium. This distinction is strongly associated with the presence of metabasites (greenschist/amphibolite) in the bedrock, making the SU catchment unique among the three.

The CE catchment serves as a reference for waters whose chemistry is controlled by flysch sedimentary bedrock. The waters are of the calcium-bicarbonate type; however, the relatively low z₂ value (0.18) indicates that Ca²⁺, Mg²⁺ concentrations are comparable. This higher relative proportion of magnesium is characteristic of the weathering of dolomitic cements or clay minerals within the flysch strata. The negative z₃ value (−0.80) confirms that the water composition is dominated by weathering-derived bicarbonate (HCO₃⁻), with minimal influence from acid anion input. Nevertheless, the negative z₄ value (−1.54) points to a strong dominance of sulfate (SO₄²⁻) over chloride (Cl⁻) among external anions. This suggests that, despite the generally pristine character of the catchment, the chemical balance is affected by regional atmospheric sulfur deposition and/or the oxidation of sulfides (e.g., pyrite) within the claystone components of the flysch.

The ZE catchment, characterized by a geological setting dominated by paragneiss rich in biotite and feldspars, exhibits a distinct hydrochemical signature of highly weathered crystalline bedrock. The water is of a mixed bicarbonate type with a high proportion of alkali metals. The low mean z₁ value (0.05) indicates that divalent cations (Ca²⁺, Mg²⁺) are almost equally represented in the composition relative to monovalent cations (Na⁺, K⁺). The high concentration of bicarbonates (38.47 mg·L⁻¹) reflects intense chemical weathering of the bedrock, providing the water with a high buffering capacity against acidification. The z₃ value (−0.78) demonstrates that, despite the high mean concentration of sulfates (22.94 mg·L⁻¹), bicarbonate remains the dominant anion. However, the high absolute concentration of sulfates and the z₄ value (−0.96) indicate a strong regional influence of both sulfates and chlorides.

3.5. Principal Component Analysis (PCA)

Principal Component Analysis (PCA) was performed on a data matrix consisting of seven ilr-transformed compositional variables (ilr1 to ilr7) representing the ratios of major ions, and four standardized supplementary variables (EC, pH, DOC, TN). The first three principal components (PCs) cumulatively explained 92.63% of the total variance in the data. Consequently, the first three components were retained for interpretation and visualization, providing a robust description of the main patterns in the dataset.

The correlations between the Principal Components (PC1, PC2, PC3), the ilr-transformed compositional variables and the standardized supplementary variables are shown in

Table 7. Pearson’s correlation coefficients (r) and corresponding p-values calculated between the Principal Components (PC1, PC2, PC3) and the ilr-transformed compositional variables (ilr1 to ilr7) and the standardized supplementary variables (EC, pH, DOC, TN).

	PC1		PC2		PC3
	r	p-Value	r	p-Value	r	p-Value	Correlation
ilr1	−0.331	<0.001	−0.754	<0.001	−0.172	0.060	Strong +
ilr2	−0.090	0.326	0.662	<0.001	−0.315	<0.001	Moderate +
ilr3	0.613	<0.001	0.662	<0.001	−0.254	0.005	Weak +
ilr4	−0.343	<0.001	0.521	<0.001	−0.381	<0.001	Weak −
ilr5	−0.318	<0.001	−0.404	<0.001	0.546	<0.001	Moderate −
ilr6	0.157	0.086	0.764	<0.001	−0.412	<0.001	Strong −
ilr7	0.729	<0.001	0.093	0.313	−0.030	0.746	Non significant
EC	0.960	<0.001	0.017	0.852	−0.008	0.929
pH	0.810	<0.001	0.095	0.303	0.559	<0.001
DOC	0.856	<0.001	0.302	<0.001	−0.378	<0.001
TN	0.555	<0.001	−0.805	<0.001	−0.174	0.058

. The correlations between the ilr vectors and the original, untransformed variables are presented in

Table 8. Pearson’s correlation coefficients (r) and p-values quantifying the relationship between the PCA input variables (ilr vectors) and the original, untransformed variables.

	ilr1		ilr2		ilr3		ilr4
	r	p-Value	r	p-Value	r	p-Value	r	p-Value	Correlation
HCO3−	−0.544	<0.001	−0.162	0.077	0.584	<0.001	−0.349	<0.001	Strong +
NO3−	0.520	<0.001	−0.512	<0.001	−0.153	0.096	−0.608	<0.001	Moderate +
SO42−	−0.208	0.022	0.229	0.012	0.525	<0.001	−0.473	<0.001	Weak +
Na+	−0.421	<0.001	0.097	0.290	0.810	<0.001	−0.128	0.164	Weak −
K+	−0.446	<0.001	0.067	0.466	0.848	<0.001	0.127	0.167	Moderate −
Ca2+	−0.141	0.123	−0.339	<0.001	0.040	0.664	−0.800	<0.001	Strong −
Mg2+	−0.550	<0.001	0.347	<0.001	0.897	<0.001	0.019	0.837	Non significant
Cl−	−0.285	0.002	0.032	0.726	0.691	<0.001	−0.273	0.003
	ilr5		ilr6		ilr7
	r	p-Value	r	p-Value	r	p-Value
HCO3−	−0.161	0.080	0.129	0.160	0.622	<0.001
NO3−	0.073	0.430	−0.459	<0.001	0.330	<0.001
SO42−	−0.280	0.002	0.272	0.003	0.664	<0.001
Na+	−0.534	<0.001	0.374	<0.001	0.776	<0.001
K+	−0.665	<0.001	0.437	<0.001	0.639	<0.001
Ca2+	0.380	<0.001	−0.317	<0.001	0.449	<0.001
Mg2+	−0.657	<0.001	0.639	<0.001	0.623	<0.001
Cl−	−0.420	<0.001	0.231	0.011	0.864	<0.001

.

The first principal component (PC1) accounts for a dominant share of the variance (59.44%) and is interpreted as the main gradient of total mineralization and dissolved substance content. The strongest and most statistically significant positive correlations with PC1 were observed for electrical conductivity (EC; r = 0.960, p < 0.001), dissolved organic carbon (DOC; r = 0.856, p < 0.001), and pH (r = 0.810, p < 0.001). Among the ilr variables, ilr3 (representing total ionic enrichment excluding Ca²⁺) and ilr7 (representing the influence of salts, especially NaCl), also showed strong positive correlations. Consequently, waters with high positive PC1 scores are characterized by high total mineralization, elevated pH, and high organic matter content. Conversely, negative PC1 scores indicate dilute, nutrient-poor, and more acidic waters.

The second principal component (PC2) accounts for an additional 21.53% of the variance and is interpreted as a contrast between the influence of nitrogenous substances and weathering and leaching processes. Its negative pole is defined by two key, strongly correlated indicators: total nitrogen (TN; r = −0.805, p < 0.001) and ilr1 (r = −0.754, p < 0.001), which represents the ratio of nitrates to base cations and buffering capacity. Conversely, several variables are positively correlated, most notably ilr6 (r = 0.764, p < 0.001), representing the ratio of base cations to nitrates, and dissolved organic carbon (DOC; r = 0.302, p < 0.001). A negative score on PC2, therefore, represents a strong influence of nitrogen, both in terms of its total amount and its relative dominance, while a positive score is associated with water chemistry dominated by the products of bedrock weathering and organic matter leaching from soil horizons.

The third principal component (PC3) accounts for an additional 11.96% of the variance and is interpreted as an acidity gradient, contrasting waters with higher pH against those acidified by organic matter decomposition. The positive pole of this axis shows a strong, significant correlation with pH (r = 0.559, p < 0.001). The negative pole is defined by a significant negative correlation with dissolved organic carbon (DOC; r = −0.378, p < 0.001) and several ilr variables associated with acid leaching (ilr6, ilr4). A positive score on PC3 thus indicates waters that are less acidic, while a negative score indicates waters where acidity is strongly influenced by organic acids.

The projection of the individual catchments into the space defined by the first two components (Figure 5) revealed a clustering into three distinct hydrochemical types. The ZE catchment formed a separate, clearly demarcated group with high positive values on the PC1 axis, which confirms its character as a catchment with highly mineralized and organic-rich waters. Its most striking feature is the broad scatter of values along the PC2 axis, ranging from strongly positive to strongly negative. This high variability suggests that the catchment functions as a dynamic and potentially unstable system. It likely alternates between periods when water chemistry is dominated by weathering products (positive PC2 scores, e.g., during baseflow) and episodes of nitrogen leaching (negative PC2 scores), which are typical of disturbed ecosystems following rainfall events.

At the opposite end of the PC1 gradient were the CE and SL catchments, which formed a second distinct cluster. Their strong negative scores on PC1 characterize them as catchments with dilute, nutrient-poor, and more acidic waters. However, these two catchments differ along the PC2 axis. The CE catchment (Beskydy) shows a predominantly positive score on PC2, which corresponds to its character as the least disturbed catchment, where water chemistry is controlled primarily by natural weathering and leaching processes. In contrast, the SL catchment (Jeseníky) is shifted toward lower PC2 values, indicating a slightly stronger relative influence of nitrogen.

The SO and SU catchments are located in the center of the ordination diagram with PC1 values close to zero, characterizing them as transitional catchments. However, a key feature is their significant scatter along the PC2 axis, ranging from positive to strongly negative values. This confirms that in these moderately mineralized catchments, the primary driver of internal variability is the dynamics of nitrogen, where conditions dominated by natural leaching processes alternate with periods under the strong influence of nitrogen.

The biplot of PC1 vs. PC3 (Figure 6) provides a complementary view, separating the catchments based on mineralization and the nature of acidity. This visualization highlights a key distinction within the group of mountain catchments. The CE catchment is not only the most dilute (low PC1 scores) but is also characterized by negative scores on PC3, identifying it as a system where acidity is governed by organic matter. In contrast, the Jeseníky catchments (SL, SO, SU), while also dilute, exhibit positive scores on PC3, indicating that their chemistry is less influenced by organic acids. The ZE catchment, due to its high buffering capacity, generally maintains a relatively high pH (positive PC3 scores) despite its high DOC content.

4. Discussion

4.1. Atmospheric Deposition

The input of elements via atmospheric deposition can substantially influence the chemistry of surface waters in catchments. The strong impact on forest ecosystems and stream water quality resulting from high loads of sulfates and nitrates during the late 20th century in Central and Northern Europe is well documented [54,55]. Our data confirm the currently relatively low input of these acidifying compounds in all three studied regions, which is consistent with trends in air pollution observed over the last three decades [56,57]. The positive effect of decreased pollution on stream water chemistry has been observed throughout Europe [58,59]. The total input of sulfates, nitrates and ammonium is driven primarily by precipitation volumes rather than by the differences in ion concentrations across the respective regions.

It should be acknowledged that our data represent bulk deposition, which includes wet deposition and the part of dry deposition accumulated on the collectors. However, actual dry deposition in forested areas is higher because of the high surface roughness of tree crowns, especially in coniferous species [60,61]. Throughfall deposition (collected beneath the tree crown) corrected by the appropriate canopy exchange model [62] would be more precise, although even this procedure can underestimate the input of some nitrogenous compounds in mountain areas [63]. On the other hand, we were looking for enhanced export of organic compounds from catchments. If we cannot detect this even with slightly underestimated inputs, the result should be reliable. It is also important to note that our monitoring relied on monthly composite samples. While this approach provides robust estimates of mean annual fluxes, it inevitably smooths out short-term hydrological and biogeochemical pulses, such as nitrate or DOC flushing during storm events or snowmelt.

4.2. Disturbance Effect on Runoff Chemistry

We hypothesized that the intensity of the recent bark beetle outbreak would be reflected in the chemical composition of stream water. Although this study is limited to only two years of hydrological monitoring in the Jeseníky Mountains, we expected an effect on the dynamics of dissolved organic carbon (DOC) and nitrate [64,65] compared to the undisturbed CE catchment. While there are differences in ion balance among the studied catchments, we observed net retention of DOC and total nitrogen. This capability aligns with our definition of resilience: the capacity of the forest ecosystem to retain nutrients, particularly nitrogen, and prevent nitrate leaching into surface waters despite the significant loss of canopy cover. This pattern suggests that the ecosystem is still capable of retaining organic carbon released in clear-cuts and that nitrogen saturation has not yet been reached. We hypothesize that this resilience is likely supported by the rapid development of ground vegetation (Calamagrostis villosa, Rubus idaeus) on clear-cuts, which likely acts as a temporary nitrogen sink, compensating for the loss of uptake by mature trees [66].

However, the high variability of nitrate and total nitrogen concentrations in the ZE catchment, as revealed by the PCA and boxplots, shows that this retention is less stable than in the undisturbed CE catchment. This is even more evident if we compare current nitrate concentrations in the runoff water of Pekelský stream (4.27 mg·L⁻¹) with those reported by Lochman [67] or Vícha et al. [68] for the period before the bark beetle outbreak (around 2.00 mg·L⁻¹). Episodic leaching events, likely driven by hydrological flushes or droughts [69,70], are highly probable in the Pekelský stream, even if the annual budget remains negative. In contrast, the higher-elevation mountain catchments (SO, SL) exhibited slight nitrate leaching, despite receiving lower inputs, although the overall nitrogen balance remained negative. The comparison between ZE on one hand and SO, SL, and SU on the other indicates that environmental conditions favorable for vegetation growth (warmer climate, lower altitude) can enhance the resilience of the nitrogen cycle following disturbance [71].

Other studies also document a short-term increase in the concentrations of Ca²⁺, Mg²⁺, and Na⁺ cations after logging, followed by a relatively rapid return to pre-logging levels. For example, Oda et al. [72] reported a short-term increase in Ca²⁺, Mg²⁺, and Na⁺ concentrations in an experiment with paired catchments in the period up to one year after clear-cutting. Similar observations were made by Swank [73]. Webster et al. [74] recorded an initial increase in Ca²⁺, Mg²⁺, and Na⁺ concentrations in waters from clear-cut, selective, and shelterwood cut catchments compared to control catchments, followed by a return to pre-harvest conditions approximately 4–6 years after harvesting. This is attributed to reduced plant uptake following logging, increased decomposition and mineralization rates of litter, and increased runoff generation due to limited evapotranspiration [75,76]. In our specific case, given the net retention of DOC and nitrogen, we do not attribute the relatively high export of base cations to the recent clear-cutting. Instead, this process is likely determined by local geology and the release of sulfates and associated cations—a legacy of the significant acidification period that affected these sites in the second half of the 20th century [77,78].

Mikkelson et al. [76] suggest that the effects of bark beetle outbreaks on water quality cannot be generalized, as they are difficult to predict and potentially regionally distinct. Schmidt et al. [33] found that several stream water parameters, including DOC and nitrate, responded in a catchment-specific manner and with varying lag times following forest disturbance. In the present study, the lack of a clear effect of the outbreak on DOC and nitrate release can be attributed to the early post-disturbance phase captured by our monitoring, as well as the relatively small extent of the affected area in the Jeseníky catchments compared to other studies [79]. Longer-term monitoring would be required to assess whether this resilience persists or if delayed nutrient leaching occurs as the pre-disturbance soil pool is exhausted.

4.3. The Dominant Role of Geology in Individual Catchments

The geological bedrock remains the primary determinant of the hydrochemical baseline, particularly for major base cations (Ca²⁺, Mg²⁺, Na⁺) [80]. The distinct separation of the ZE catchment from the Jeseníky (SL, SO, SU) and Beskydy (CE) catchments in the PCA underscores the importance of the mineralogical composition of the bedrock [81]. The paragneiss underlying ZE is generally expected to be more susceptible to weathering than the phyllites and quartzites of the Jeseníky catchments and the flysch of the Beskydy Mountains [82,83]. This is evidenced by the significantly higher concentrations of base cations and alkalinity in ZE, despite it being the most disturbed catchment. These results suggest that while deforestation may increase runoff and leaching potential, the reservoir of available elements in the soil and bedrock dictates the magnitude of these fluxes [84].

A notable feature demonstrating the sensitivity of stream chemistry to local lithology is the magnesium (Mg) anomaly in the SU catchment. The presence of metabasite intercalations (rich in Mg-silicates) resulted in elevated Mg concentrations compared to the nearby SL and SO catchments, despite similar climatic conditions and disturbance levels [85]. This confirms that even in small headwater catchments, fine-scale geological heterogeneity plays a crucial role that cannot be ignored when assessing disturbance effects.

5. Conclusions

The results obtained do not indicate a clear influence of the bark beetle outbreak in the Jeseníky Mts. on the chemistry of outflow from the forest catchments. The DOC export is significantly lower than the atmospheric input and also lower than export from the undisturbed catchment of the Červík stream (CE) in the Moravian-Silesian Beskydy Mts. or the disturbed Pekelský stream (ZE) in the lower-elevation Czech-Moravian Highlands. Nitrogen export was higher than atmospheric input in two catchments (SL and SO), but only by 0.24 and 0.61 kg·ha⁻¹·yr⁻¹, respectively. Nitrate concentrations in runoff were relatively low (2.82 and 3.52 mg·L⁻¹), and the net total nitrogen export was negative. We conclude that in this early post-disturbance phase, the relatively limited extent of clear-cuts and the rapid development of ground vegetation prevented more distinct changes in the elemental balance of the Jeseníky Mts. catchments.

In contrast, stream water chemistry is primarily driven by geological conditions. Compositional Data Analysis (CoDa) revealed distinct hydrochemical signatures governed by the mineralogical composition of the bedrock, distinguishing the Jeseníky Mts. catchments from both the Pekelský and Červík catchments.

Our findings confirm that geological heterogeneity, such as the specific mineral composition of metamorphic rocks, determines the baseline mineralization and buffering capacity of individual catchments. Consequently, local lithology acts as the dominant factor that can override or mask the hydrochemical signals of forest disturbances. The relatively high export of base cations and bicarbonates suggests active weathering and raises the question of the extent to which soil Ca and Mg reserves are being further depleted. Assessing the balance between weathering rates and elemental export is therefore crucial for predicting the future health of these forest ecosystems. The findings are particularly relevant for temperate mountain forests in Central Europe.