The Strengbach catchment, located in the municipality of Aubure (Haut-Rhin, France) in the Vosges Mountains, has an area of 0.8 km² and an elevation between 850 and 1150 m. This catchment is mainly covered by acid brown soil and ochreous podzolic soil (Fig. 1) on Ca-poor granitic bedrock (Pierret et al., 2018). The forest covers 90 % of the catchment with 80 % planted with a monospecific spruce plantation (Picea abies) and the rest with beech (Fagus sylvatica) (Fichter et al., 1997; Pierret et al., 2018).

Meteorological, hydrological and geochemical variables have been monitored since 1986 in the Strengbach catchment by the Observatoire hydro-geochimique de l'environnement (OHGE; https://ohge.unistra.fr, last access: 18 September 2025). The creation of the observatory (with the installation of the first equipment) and the first studies were performed in order to understand the impact of acid rain on forested ecosystems (Probst et al., 1990, 1992, 1995; Pierret et al., 2019). Many studies are still being conducted on environmental issues concerning forest mountainous ecosystems and particularly about forest decline and nutrient deficiency (Fichter et al., 1997; Poszwa et al., 2003; Cenki-Tok et al., 2009; Brioschi et al., 2013; Beaulieu et al., 2020; Oursin et al., 2023).

The climate is oceanic mountainous with an annual mean temperature of 6.17 °C, and a mean precipitation of about 101 mm per month during the time period 1987–2024. The annual precipitation (over a calendar year) has varied between 814 mm yr⁻¹ (2023) and 1568 mm yr⁻¹ (2007) with an average of 1214 mm yr⁻¹. Outlet water flows vary between 464 and 1132 with an average of 732 mm yr⁻¹. Historically, rainfall was uniformly distributed throughout the year. But climate change consequences were observed at the local scale in the OHGE, with wetter summers and winters, and drier autumns when comparing before and after 2006 (Strohmenger et al., 2022). The OHGE databank is a key tool to understand environmental change consequences (Pierret et al., 2018; Strohmenger et al., 2022).

The plot (VP plot, Fig. 1) is a monospecific spruce (Picea abies) plot, planted in the 1900s, which has been declining for several decades due to different reasons: nutrient-poor soils amplified by acid rain until the 1980s (Probst et al., 1990, 1992, 1995), several recent violent storms, successive bark beetle attacks as well as intense drought episodes. The site, where trees were still alive, was instrumented from 2015 to 2023. Soil solutions were collected with 40×40 cm zero-tension lysimetric plates installed at four depths (5, 10, 30, 60 cm) in 1990 and sampled every six weeks during the period 2015–2023. Throughfall (fraction of precipitation that reaches the ground after passing through the canopy) samples were collected with two gutters (2.0×0.2 m) every two weeks (Pierret et al., 2018). The minimum volume of water needed for analyses is 25 mL, volumes of soil solution samples collected are given in Supplement (Table S1).

The studied soil profile (Fig. 1) is a Cambisol (Alocrisol in the French taxonomy) on granitic bedrock. We can distinguish two litter horizons. The upper litter (OL) which is composed of old needles, herbaceous roots and pieces of wood, is 5 cm thick. The lower humic horizon (OH-OF) is 6 cm thick and is black-colored with a smooth texture. The dark brown/grey horizon A of 5–6 cm is unequally distributed on the profile with wavy transitions. It is followed by a transition layer of 5 cm. At 17 cm depth, a brown/red mineral horizon B is found with a lumpy texture. This horizon is homogeneous and very sandy with centimetric gravel.

Soil solutions were collected in clean polyethylene (HDPE) containers sampled from the field and filtered with cellulose acetate filters 0.45 µm pore diameter membrane Merck-Millipore on a PTFE filtration unit. Conductivity and pH were measured with HQ440d Hach multimeter. The dissolved organic carbon (DOC) was determined using a carbon analyser (Shimadzu TOC-VCPH) with an uncertainty of 2 % to 3.5 %. The accuracy of the analysis was assessed by regular measurement of PERADE-17 river water CRM (Environment and Climate Change Canada matrix reference materials). K⁺, Na⁺, Ca²⁺, Mg²⁺, NH4+ and NO3- concentrations were determined by ion chromatography on a Thermo Scientific Dionex ICS-5000+. Al and Si concentrations were measured by ICP-AES on a Thermo Scientific iCAP 6000 SERIES. PERADE-17 and SLRS-6 water CRMs (certified by Environment and Climate Change Canada) were used as standards and the analytical uncertainty for ions concentrations is ± 2 %.

To characterise the dissolved organic matter UV-fluorescence of soil solution was measured. Fluorescence spectra were obtained on a Hitachi F-2500 spectrofluorometer equipped with a Xenon lamp, using FL Solution 2.0 software and a 1 cm × 1 cm (3.5 mL) quartz cuvette. Emission spectra were collected for three excitation wavelengths (λex= 254, 310 and 370 nm) with a step of 1 nm and slits of 2.5 nm. The humification index HIX was calculated on emission spectra for λex= 254 nm as the ratio of the sum of the fluorescence intensities (F) between 435 and 480 nm over the sum of the fluorescence intensities between 300 and 345 nm (Zsolnay et al., 1999), (Eq. 1). HIX quantifies aromatic molecules in solutions and is also qualitative of the complexity of aromatics (Zsolnay et al., 1999; Serène et al., 2025). The index BIX was calculated as the ratio of fluorescence intensity emitted at 380 nm over 430 nm for λex= 310 nm (Huguet et al., 2009) (Eq. 2). BIX is a proxy of biological activity in aquatic environment (Parlanti et al., 2000; Huguet et al., 2009). The index FI was calculated as the ratio of fluorescence intensities at 450 and 500 nm for λex= 370 nm (Eq. 3) (McKnight et al., 2001). FI indicates the origin of DOM and distinguishes two endmembers: terrestrially and microbially derived DOM. 1HIX=∑F(435-480nm)∑F(300-345nm),λex=254nm2BIX=F(380nm)F(430nm),λex=310nm3FI=F(450nm)F(500nm),λex=370nm

Drought events can create variations in soil water content. To evaluate the changes in water availability in the soil profile induced by rainfall decrease, an in-house daily water balance calculation code was applied to a soil profile under spruces. The BILHYDAY code, whose structure is inspired by the model BILJOU developed by Granier et al. (1999), takes meteorological data as inputs with stand parameters (leaf area index, soil porosity and residual water content). Based on a conceptual approach, the model operates on a daily time step to describe the hydrological processes of the soil–tree–atmosphere continuum. Potential evapotranspiration is estimated following the Penman formulation. Precipitation feeds a canopy reservoir that simulates interception up to a threshold capacity, while transpiration is represented as a function of potential evapotranspiration, soil water storage, and leaf area index. Transpiration is distributed among soil layers according to the root density profile. Soil evaporation is estimated from the incident energy (induced from direct radiation) and an extinction coefficient. Five soil layers are represented as cascading reservoirs that drain sequentially from top to bottom according to their field capacity. The soil is divided into 5 layers of 5-5-20-30-40 cm thickness (from the surface). This discretization of the vertical soil profile is specifically based on the installed instrumentation. It includes a top layer (0 to 5 cm), which is particularly sensitive to exchanges with the atmosphere, and a bottom layer (60 to 100 cm) that drains toward the deeper soil zone, with three intermediate layers (5 to 10 cm, 10 to 30 cm, and 30 to 60 cm) corresponding to the layers at which soil solution samples are collected. The uppermost layer is supplied by non-intercepted precipitation and is also subject to evaporation losses. Root water uptake associated with transpiration is distributed among the different soil layers according to the root density profile. The BILHYDAY model is used as a diagnostic tool to provide complementary hydrological information supporting the interpretation of geochemical measurements of soil solutions. The model structure and parameterisation are based on previous developments and applications, notably those described by Biron (1994), during which the different components of the soil water balance were assessed. In the present study, this existing parameterisation was retained for soil properties, and the leaf area index (LAI) was updated to account for changes in forest stand development. Such an approach has recently been used for the soil hydrodynamic characterization of two other plots in the Strengbach catchment; BILHYDAY was run in combination with a physically based hydrological model coupled with an inverse parameter estimation module (Belfort et al., 2025). To address potential concerns regarding the realism of the simulated soil moisture dynamics, additional soil water content measurements collected at locations close to the soil solution sampling site are provided in Supplement (Fig. S2). Although these measurements are not strictly co-located with the simulated soil profile, they exhibit temporal dynamics consistent with the model outputs, thereby supporting the coherence of the simulated soil moisture values used in this study. The relative extractable soil water was calculated similarly to Granier et al. (1999) and can be used to assess water stress. Although the model does not simulate photosynthesis by integrating a dynamic stomatal conductance, transpiration is nevertheless adjusted according to soil water availability, with regulation occurring below a threshold of 0.4, as proposed by Bréda et al. (2006).

To find outlier points, we used a robust method based on the calculation of the first and the third quartiles (Q1 and Q3) and the interquartile range (IQR = Q3-Q1). It consists of determining an interval with the lower bound (Q1-3× IQR) and the upper bound (Q3+3× IQR), whereby points outside this interval are considered outliers. The value of 3 was chosen to identify only extreme outliers (Schwertman et al., 2004; Grunsky, 2010). A box plot representation allows outliers to be visualized as points outside this interval (Supplement Fig. S3).

Non-metric multidimensional scaling (NMDS) was performed to identify atypical chemical signatures within the long-term monitoring dataset. NMDS is effective for highlighting atypical observations in long-term monitoring and can reveal the impact of extreme events on soil when many parameters are measured (Knight et al., 2024), or the consequences of extreme meteorological events on hydrochemistry (Setia et al., 2021). The analysis was conducted using the metaMDS function in the `vegan' library in R (Oksanen et al., 2001) on the dataset of chemical measurements (pH, conductivity, Na⁺, K⁺, Mg²⁺, Ca²⁺, Al, Si, NH4+, NO3-, DOC) for the period 2015–2023. To evaluate depth-specific anomalies, the data were partitioned by sampling depth (5, 10, 30, and 60 cm). Plotting the coordinates of different soil solution samples highlighted atypical points. Finally, the scores assigned to the chemical parameters were analyzed to determine which variables primarily drove the observed anomalies and distinguished the 2022 drought samples from the long-term baseline.

On the OHGE site, considering the hydrological years over the recorded period from 1987–2023 (year running Y from 1 October to 31 September of year Y+1), 2021/2022 (including the 2022 drought) was the third driest year on record (after 1990/1991 and 1995/1996) with a cumulated precipitation amount of 970 mm (-20 % compared to the average value over the period), the fourth warmest year with an average annual temperature of 7.47 °C (+20 % compared to the average for the period – just after 2019/2020, 2023/2024 and 2022/2023), and with the second highest number of days with an average daily temperature above 18 °C, i.e. 39 d compared to an average of 17 d. Among the years with the least precipitation, 2021/2022 was the warmest year, which in combination makes 2021/2022 the most drought-affected year.

Meteorological drought is generally defined as a prolonged period during which precipitation is significantly below normal or expected levels for a given location and season, resulting in a deficit of moisture relative to climatic norms (Wilhite and Glantz, 1985). The longest period of summer meteorological drought (defined here as consecutive days without rainfall) observed between 1987 and 2024 was 19 d in 2018, followed by 17 d in 2022, 2013 and 1990. Nevertheless, the summer of 2022 was exceptional because this first drought period from 2 July to 19 July was very quickly followed by a second period of 24 d during which daily rainfall did not exceed 0.7 mm (between 21 July and 13 August 2022; see the red band in Fig. 2A). Notably, a rainfall event of 8.9 mm on 20 July 2022 separated these two meteorological drought events (Fig. 2A). The summer months of July and August, with only 61 mm of rainfall, were the driest ever recorded at the site (Fig. S1), representing a decrease of nearly 70 % relative to normal conditions.The particularly dry month of July (10.2 mm cumulated; Fig. S1; Fig. 2A) was amplified by the previous dry March and May months with respectively 33 and 36 mm cumulated, compared with the 102 mm/month usually observed in the Strengbach watershed.

To assess the impact of rainfall deficit on soil water content, BILHYDAY was used with meteorological data. The water content of the three intermediate layers (denoted by their centers at 7.5, 20 and 45 cm) and the relative extractable soil water (REW) during the 2021–2022 hydrological year are depicted in Fig. 2B–C. The simulations showed that the water deficit began in the middle of May in the 7.5 cm depth layer, and the few rainy events in the first half and last week of June did not prevent it from falling to a minimum value of 0.14 in mid-July. The 20 cm depth layer experienced a gradual drying out from May to 22 June; after the rains at the beginning of July, the minimum water content was recorded on 30 July reflecting a significant water deficit (Fig. 2B). The water content remained constant and low in these layers until 16 August 2022. The water content decrease was slower in the 45 cm depth layer with a minimum on 14 August 2022. Following the return of the rainy season, there was an increase in soil water content during the second half of August, particularly in the second layer, while recharge was slower for deeper layers. This was followed by a rapid rise in September, coinciding with the return of heavier rains, resulting in a value of 0.42 and 0.35 at 7.5 and 20 cm depth respectively (Fig. 2B). During the summer of 2022, a total of 25 d of water deficit occurred with the relative extractable soil water below the threshold of 0.4 (Fig. 2C). Trees transpiration increased during spring and summer, reaching its maximum (2.5 mm d⁻¹) in the middle of June and then sharply decreased (Fig. 2C). During the water stress period from mid-July to mid-August, tree transpiration reached a very low value due to stomatal closure or downregulation, and this continued until rain returned.

The field monitoring of soil solution concentrations in the Strengbach catchment allowed for the observation of long-term variations of major cations concentrations through the 2015–2023 period. The time series exhibited annual seasonality in the chemical signal, as well as significant anomalies during the exceptionally severe drought of summer 2022 with numerous outlier concentrations (Figs. 3 and S2, Table 1).

The most extreme chemical anomaly was observed at 30 cm depth on 30 August 2022 with an exceptionally high conductivity (94.1 µS cm⁻¹) and high DOC (66.81 ppm-C). These values were well above the inter-annual variability and were the highest ever registered at this depth, considered as outliers because they exceed Q3+3× IQR (Table 1; Fig. S3). The soil solution sample of 30 August integrated the period from 20 July to 30 August coinciding with the return of rainfall after the long dry period (Fig. 2). This high conductivity corresponded to peaks of concentration of many ions such as K⁺, H⁺, Ca²⁺, Mg²⁺, Al, Mn²⁺, Fe³⁺, NO3-, NH4+, Cl⁻ except for Na⁺, SO42- and Si. These peaks were specially marked at 30 cm (Fig. 3) and were confirmed as outlier concentrations for K⁺, Ca²⁺, Mg²⁺, Al, and NO3- on the 30 August 2022 (Table 1). Focusing on major cations at 30 cm depth, the K⁺, Ca²⁺, Mg²⁺ and Al concentrations increased twice successively between 21 June 2022 and 30 August 2022 (Fig. 3). K⁺, Ca²⁺, NH4+ also reached outlier concentrations on the 21 June 2022 (Table 1; Fig. S3). The highest concentrations ever registered during 2015–2023 period in 30 cm depth soil solution were reached for the second peak on the 30 August 2022 (Fig. 3, Table 1). K⁺ concentration exhibited exceptionally high concentrations also at 5 and 10 cm depth.

The 30 cm depth soil solution anomaly during 2022 drought was also characterized by the lowest pH of 4.1 observed on 30 August 2022 over the 2015–2023 period (Fig. 3G). Notably, this low pH anomaly occurred only at 30 cm. However, high pH anomalies were observed at 30 and 60 cm on 21 June 2022 (4.1 and 6.6) and at 10 cm on 30 August 2022 with a pH equal to 5.8. Those values were the maximum pH ever observed at each depth during the period 2015–2023.

For nitrogen-related ions (NO3-, NH4+), a first peak in NH4+ concentrations was observed on 21 June 2022 at 30 cm and 60 cm, followed by a second peak of lower concentrations on 30 August 2022 at 30 cm (Fig. 3I). The NO3- concentration peak of 0.117 mmol L⁻¹ at 30 cm occurred only on 30 August 2022 (Fig. 3J). For superficial soil solutions, at 5 and 10 cm depth, high concentrations were observed on 30 August 2022 for NH4+ (0.10 and 0.16 mmol L⁻¹ respectively) and for NO3- (0.28 and 0.14 mmol L⁻¹ respectively).

At 30 cm depth, the three points between June and August 2022, which were particularly affected by drought events, were significantly shifted to the right of the plot along the MDS1 axis (Fig. 4). The samples from the summer 2018, which was also a drought episode (cf. 1), also show statistically different coordinates at this depth. Conversely, the values for soil solutions in summer 2022 at the three other depths (5, 10, and 60 cm) did not show any atypical values, except for those in August but only at 5 and 10 cm (Fig. 4). However, these values were less atypical than those observed for soil solutions at a depth of 30 cm (Fig. 4). These observations confirm that the impact of droughts on soil solution chemistry was the most important at 30 cm depth. The NMDS scores for 30 cm soil solutions showed that MDS1 exhibits the highest positive value for NH4+, followed by NO3+, meaning a particular anomaly concerning nitrogen species (Table 2). K⁺ and Ca²⁺ concentrations also had high scores along the MDS1 axis, indicating that the dynamics of cationic mineral nutrients were also atypical during 2022 drought.

Concentrations in throughfall are highly variable under the studied spruce plot, with numerous concentration peaks, often higher during summer than winter (Fig. 5). Throughfall of 16 August 2022 (red cross in Fig. 5) corresponded to the end of the dry period with the return of rainfall on the 14 August 2022 (Fig. 2A). On this date, throughfall exhibited the highest concentration of NH4+ (0.597 mmol L⁻¹; Fig. 4) and high concentrations of K⁺ (0.290 mmol L⁻¹).

The qualitative indicators of dissolved organic matter presented significant anomalies during the 2022 summer drought. At 30 cm depth, the FI was 1.32 on 21 June 2022, considered as an outlier because the value was below (Q1-3× IQR = 1.42; Table 3). FI also showed an anomaly at 5 and 10 cm during this drought period with the lowest and atypical values of about 0.62 on 20 July 2022 and 0.72 on 21 June 2022 respectively (Fig. 6). Low BIX values of 0.60 and 0.58 on 21 June 2022 and 20 July 2022 respectively were measured whereas the average value was 0.73 ± 0.11 at 30 cm depth over the 2020–2023 period. HIX reached a minimum value of 2.83 on 30 August 2022, only at 30 cm depth, whereas no significant anomalies were noted at any other depths.

While most chemical concentrations showed a peak during the drought of 2022, Na⁺ and Si exhibited no significant peak (Fig. 3). Their average concentrations during the period between 21 June and 30 August 2022 were 0.074 and 0.038 mmol L⁻¹ respectively, which were within the range of inter-annual variability with a mean concentration during 2015–2023 period of 0.066 ± 0.025 mmol L⁻¹ (median ± IQR) and 0.061 ± 0.046 mmol L⁻¹ for Na⁺ and Si respectively (Table 1).

In the soils from the Strengbach catchment, sodium is mainly included in albite, a primary mineral from the granite (Fichter et al., 1997). Na is not considered to precipitate or integrate in secondary phases (Chou and Wollast, 1985). It represents only 0.6 % of the cationic exchange capacity and its concentration is negligible in the litter compared to other elements (Oursin et al., 2023). Thus, the main sources of sodium in soil solution are the throughfall and the dissolution of albite. The concentrations of Na⁺ in throughfall did not seem to be affected by drought, the concentration on the 16 August 2022, date of rain return after drought, was 0.033 mmol L⁻¹, which was comparable with the mean concentration of 0.042 ± 0.021 (Supplement Table S2). As no significant consequences of drought on Na⁺ concentration in soil solution were observed, we can propose that drought does not affect albite dissolution in soil. Moreover, the study of Oursin et al. (2023) showed that the contribution of primary mineral weathering in the chemical signature (Ca, Mg and K concentrations) of soil solution from the same site was low and even negligible when compared with other fluxes (alteration of secondary minerals, cation exchange, litter degradation), except for Na. This lack of impact on albite dissolution may be extended to other primary minerals composing granite. Consequently, it can be proposed that the impact of drought on the weathering of all primary minerals is negligible.

In theory, thermodynamic and kinetic of mineral dissolution depend highly on soil temperature and water content (Lasaga et al., 1994; Kump et al., 2000). Studies modelling the effect of climate change on future forest soil weathering rate proposed that a global increase in temperature can increase the chemical weathering. However, because of soil moisture limitation, summer droughts limit the expected gain in weathering associated with higher temperatures (Belyazid et al., 2022; Kronnäs et al., 2023). Thus, severe water deficits in soil tend to reduce weathering rates.

The concentrations of Si in rain and throughfall are below the detection limit (Pierret et al., 2019) and the concentration of Si in litter is negligible compared with other elements (Oursin et al., 2023). Therefore, Si mainly originates from mineral weathering. Hydrological modelling of soil water content (Fig. 2B) exhibited an important diminution during June 2022 in the third layer (between 10 and 30 cm) and during August in the fourth layer (30 to 60 cm), consistent with the Si concentrations with significant reduction on 21 June 2022 at 30 cm and on 30 August 2022 at 60 cm (Fig. 3F). In addition, two minimal Si concentrations, at 30 and 60 cm depth, were observed in the summers 2018 and 2022, the two main drought events of the period 2015–2023, during which impacts on forests were observed (Gharun et al., 2024; Knutzen et al., 2025). So that we can deduce a reduction in silicate mineral dissolution due to low water content during drought. Silica is mainly controlled by clay weathering in the soil of the Strengbach catchment, as shown by geochemical modelling (Goddéris et al., 2006; Beaulieu et al., 2020). As the reduction was visible for Si and not for Na, it is suggested that clay mineral dissolution is reduced during drought events. As the dissolution kinetic is higher for secondary minerals compared to primary, a reduction of secondary minerals dissolution should have more pronounced consequences on the chemical composition of the solution (Drever and Clow, 1995). Thus, water deficit in soils leads to a decrease in weathering rate, as proposed by Belyazid et al. (2022) and Kronnäs et al. (2023), but only visible on secondary minerals such as clays. Silica was the only element showing a reduction in concentration during drought. However, chemical soil solutions signals exhibited positive peaks in many elements such as K⁺, Ca²⁺, Mg²⁺ and Al (Fig. 3). These peaks cannot be explained by an expected reduction in weathering due to water deficit, which would produce the opposite effect. Therefore, other processes will be discussed in the following sections.

Rainfall passing through the canopy, known as throughfall, differs from open field precipitation because of its interactions with leaf surfaces. Throughfall integrates wet and dry atmospheric deposits as well as biological interactions with the canopy during leaf or needle leaching (Ulrich, 1983; Berger et al., 2008; Eisalou et al., 2013; Pierret et al., 2019; Ponette-González et al., 2020). Compared to rain, throughfalls are enriched in each chemical element. On the basis of a monitoring over the 1986–2012 period, Pierret et al. (2019) showed that throughfalls under spruces are enriched by a factor of 15 for K⁺, 5 for Ca²⁺ and Mg²⁺ and 2 for NH4+ compared to open field precipitations from the same studied site, the Strengbach catchment. K⁺ in throughfalls comes mainly from biological excretion by leaves, whereas Ca²⁺ and Mg²⁺ are a mixture between wet and dry atmospheric deposits with a weak influence of biological exchanges (Lovett and Lindberg, 1984; Kopáček et al., 2009; Adriaenssens et al., 2012; Pierret et al., 2019). NH4+ mainly comes from atmospheric deposition, which might be from NH₃ atmospheric emission due to agricultural activities (Avila et al., 2017; Pierret et al., 2019). But NH4+ is involved in numerous biological processes, can also be excreted, absorbed by leaves or organisms living on the leaves (Woods et al., 2012; Ponette-González et al., 2020).

K⁺ dynamics are highly affected by the biological functioning of trees. Reduction of rainfall or drought might reduce the mobility of K⁺ within the canopy by decreasing foliar leaching (Sardans and Peñuelas, 2015; Schlesinger et al., 2016; Touche et al., 2024). Consequently, biologically excreted potassium accumulated on spruce needles during the 2022 dry period. Upon the return of rain, these accumulated elements were leached, leading to high concentrations in the throughfall (Fig. 5). Thus, the exceptionally high peak in K⁺ observed in soil solutions at 5, 10, and 30 cm on 30 August 2022 can be linked to the high K⁺ content in throughfall on 16 August 2022. Indeed, K⁺ concentration on the 30 August 2022 soil solution at 5 cm depth (integrating 20 July to 30 August period and including the rain episode of 16 August; Fig. 2) was 0.331 mmol L⁻¹, similar to the K⁺ concentration in throughfall on the 16 August 2022 (Fig. 3B; 0.290 mmol L⁻¹), thus supporting the important contribution of throughfall in superficial soil solution K⁺ concentration. Similar observations of K⁺ concentrations peak in soil solutions at different depths have been reported on drought experiments on European spruce plots (Lamersdorf et al., 1998) but also in Canadian boreal forest with very high K⁺ concentrations in the soil solution at 30 cm and in throughfall, as observed by Houle et al. (2016). This vertical propagation of K⁺ in the soil profile was likely driven by the increased concentration observed in the throughfall. This spread of K⁺ diminished between 30 and 60 cm, probably because this cation was sorbed, exchanged, precipitated or uptaken by trees. K⁺ was the only cation that exhibited a maximum peak in both throughfall and surface soil solutions (5 and 10 cm deep). This specificity of K⁺ is linked to the important contribution of throughfall in K⁺ soil solution concentration compared to other elements.

NH4+ showed a strong positive anomaly in the throughfall for 16 August 2022, with concentration of 0.597 mmol L⁻¹, more than 10 times the average for the period 2015–2022 of 0.055 mmol L⁻¹ (Fig. 5). This ammonium peak was also found in soil solutions at depths of 5 and 10 cm on 30 August 2022 (Fig. 3). Similarly to K⁺, the ammonium flux in throughfalls resulted from the leaching of spruce needles during the first rain after drought. However, due to the high reactivity of ammonium and nitrogen species in soils (as nitrification converts ammonium into nitrite and/or nitrate), the NH4+ concentrations of soil solutions at 5 and 10 cm depth were significantly lower (0.075 and 0.150 mmol L⁻¹) than in throughfall (0.597 mmol L⁻¹). More frequent sampling of soil solutions would have provided a better match between the chemical composition of throughfall (sampled every two weeks) and that of soil solutions (sampled every six weeks), particularly for better observation of the behaviour of highly reactive species such as nitrates and ammonium. It can still be argued that, after a drought period, throughfall represents an important source of ammonium for soil solutions.

Since 2015, other peaks of high concentrations in the throughfall series have been observed, but not in the soil solutions (Figs. 3, 5). This highlights that additional processes must be involved to explain why, in 2022, K, Ca or Mg accumulated in soil solutions, whereas in other periods, part of the throughfall appeared to be mobilized in the soil.

Soil solutions contain elements that are essential plant nutrients such as K⁺, Ca²⁺, Mg²⁺, NO3-, NH4+ (Ingestad, 1979). K is the most abundant cation in plant cells and a limiting element for plant growth (Leigh and Jones, 1984; Sardans and Peñuelas, 2015). Ca and Mg are also two essential nutrients for trees, found in very small quantities in Strengbach soils and maintained in sufficient concentration in the soil solution through recycling by vegetation (Ovington, 1959; Fichter et al., 1997; Beaulieu et al., 2020; Oursin et al., 2023). Beaulieu et al. (2020) simulated biogeochemical processes under spruce plot in the Strengbach catchment, and they estimated that tree uptake is the most important flux of Ca and Mg in the system, two times higher than the flux coming from the throughfall. In contrast to K, the uptake flux is two times lower than the flux coming from throughfall. So, trees uptake large quantities of Ca²⁺, Mg²⁺, K⁺, NH4+, NO3-, and influence nutrient concentrations in solution. The field data showed an increase of Ca²⁺, Mg²⁺, K⁺, NH4+ and NO3- concentrations in the soil solutions, especially at 30 cm depth, during or after the drought (Fig. 3). We hypothesize that these increases in concentrations were linked to a decrease in biological uptake by vegetation.

The reactions of trees to drought, such as stomate closure or reducing leaf area (needle loss), for instance, decrease water loss through transpiration and so decrease water uptake by roots (Bréda et al., 2006; Ditmarova et al., 2010; Hesse et al., 2024). The decrease in tree transpiration during the drought period, simulated by the model BILHYDAY used in this study (Fig. 2), implies a reduction in water uptake by tree roots. This was also observed in the Strengbach watershed, with measurements of reduced sap flow during an experimental water deficit on spruce trees (Lu et al., 1995). In addition, an experimental water deficit study carried out on a spruce stand at the same site showed that the composition of the xylem sap was lower in nutrients content (Ca, Mg and K) during the water deficit period (Dambrine et al., 1993). Thus, trees take up less Ca, Mg and K during drought. Reduced nutrient uptake can explain the lower biomass concentrations found in numerous experimental water deficit studies on forest plots. For example, lower concentrations of K, B and Mg were observed after droughts in spruce needles (Nilsen, 1995), but also in the biomass of other tree species (Sardans et al., 2008; Touche et al., 2022).

However, soil solution collected by zero-tension lysimeters is the fraction of soil water participating in gravitational flow, circulating mainly in the macropores (Singh et al., 2018). It cannot be considered similar to water collected by vegetation, since roots also collect water from microporosity with a certain suction pressure (Rüdinger et al., 1994). Furthermore, gravitational solutions have a very different chemical composition from other fractions of soil water collected under a certain pressure (Haines et al., 1982; Watmough et al., 2013), also under the spruce forests of the Strengbach watershed (Ranger et al., 1993). But the soil on this plot is very sandy, with a sand content ranging from 50 % on the surface to 70 % at depth, and are relatively homogeneous (Oursin et al., 2023). Thus, the studied soil is very porous (more macro- than micro-porosity) and well-drained (more gravity flow than capillarity), therefore we assume that gravitational soil solution is representative of the whole soil water. In addition, gravity soil solutions enable input-output balances to be calculated, given that the soil solution at a depth of 30 cm can be considered as water leaving the root zone of spruce trees (Ranger et al., 1993; Marques et al., 1996). The chemical composition of the solution thus collected depends on all the biogeochemical processes taking place in the soil and can indirectly indicate a change in root uptake.

Thus, the decrease in biological uptake by trees during drought may explain the maximum concentrations observed for Ca, Mg, K, nitrate and ammonium in soil solutions, particularly at a depth of 30 cm, which corresponds to the depth after the maximal root density zone, in the first 20 cm of the mineral soil for spruces (Schmid and Kazda, 2002; Borja et al., 2008). It explains why high concentrations were found at 30 cm depth particularly for Ca and Mg, the two nutrients most impacted by trees uptake.

A DOC increase was observed in soil solution at 30 cm depth during and after drought period, with a maximum concentration of 66.81 ppm-C after the rain return (Table 1; Fig. 3). On other sites, similar DOC peaks have been reported under spruce plots, after a water deficit experiment at 20 cm depth (Lamersdorf et al., 1998) or after 2006 drought at 90 cm depth (Schulze et al., 2011). The DOC increase during the drought period could have been induced by different organic inputs. Indeed, Gaul et al. (2008) reported that water deficit can lead to an increase in fine root mortality up to 61 %, not compensated by an increase in fine root production on a Norway spruce forest in Germany. This root mortality, also observed in Norway spruce located in Sweden subjected to drought (Persson et al., 1995), can represent a carbon source to soil organic matter pool in a temperate ecosystem.

Moreover, another source of carbon in upper soil horizons in response to drought could stem from stress and mortality in bacteria and fungi. As reported in many studies, water deficit leads to a decrease in organic matter degradation (Schimel et al., 2007; Boczón et al., 2021; Gehring et al., 2017; Bogati and Walczak, 2022). Indeed, even if heterotrophic respiration is controlled by temperature in wet soils, it is limited by water deficit during dry period (Curiel Yuste et al., 2007). The degradation decrease could lead to accumulation of organic carbon in superficial soil layer. On the other hand, the high concentration of DOC observed after the rain return in soil solutions could be the result of fast-moving water and transport of DOC accumulated in superficial soil layer during drought period through soil profile. A high dissolution of organic matter would also lead to an input of other elements contained in this organic matter to the soil solution. Therefore, this mechanism likely contributed to the observed increase in Ca, Mg or K concentrations in soil solutions at 30 cm after rewetting, in addition to other mechanisms disturbed during drought.

To better understand this modification of DOC dynamics, optical characterization by fluorescence is employed. Three fluorescence indices are frequently used to characterize natural organic matter in water: the fluorescence index (FI, McKnight et al., 2001), the biological index (BIX, Huguet et al., 2009) and the humification index (HIX, Zsolnay et al., 1999). HIX was defined by Zsolnay et al. (1999) and is an indicator sensitive to the quantity of aromatic molecules in solution, but also to the type of molecules, i.e., more or less aromatic, heavy, complex (Serène et al., 2025). HIX was relatively high before and during the drought in 30 cm depth soil solution (median of 8.00; Table 3). It then fell sharply when the rain returned after the long dry period (HIX = 2.83 on 30 August 2022; Fig. 6) during the DOC peak. A similar dynamic between DOC and aromaticity during and after drought has been observed in several streams in boreal ecosystems (Tiwari et al., 2022). As these large aromatic molecules precipitate at acid pH (Swift, 1996), the decrease in pH may explain the decrease in HIX on 30 August 2022. HIX is often correlated with DOC (Pearson coefficient = 0.54 on all four depths of the plot), but the solution showing the DOC peak after the drought was relatively low in aromatic molecules (Figs. 3, 6), reflecting the atypical composition of this solution, reinforcing the hypothesis of a new source of organic matter originating from biological stress on soil organisms, rather than usual litter decomposition.

The FI in superficial waters and groundwaters is used to distinguish sources of terrestrially or microbially derived fulvic acids. McKnight et al. (2001) propose index values of 1.4 for terrestrial origin and of 1.9 for microbial origin. In our study, FI averaged 1.74 in soil solution at 30 cm depth over the period 2020–2023 (Table 3), which could highlight predominantly microbial sources of organic material. However, a decrease of FI to 1.3 was observed in June 2022 at 30 cm depth (Fig. 6), an index value consistent with predominantly terrestrial sources. The low contribution of microbial source to organic material could be explained by a reduction of fungal and bacterial activity due to low water content during the May–June period (Fig. 2B), as mentioned in Sect. 4.5. The index value increased to 1.8 in soil solution with the rain return, indicating a recovery of regular microbial activity (Fig. 6).

The BIX is commonly used as a proxy of biological activity in aquatic environments (Parlanti et al., 2000; Huguet et al., 2009). High values of BIX suggest a high biological activity and so an organic matter recently produced by microbial activity, while the low values are supposed to stem from relatively low biological activity. The BIX values decreased in soil solution at 30 cm depth during the May–June period, then increased upon rewetting (Fig. 6), consistent with the FI value evolution.

The drought characterized by a decrease in soil water content from May to July (Fig. 2), led to a DOM with a lower microbial signature. This supports the hypothesis of a decrease in soil microbial and fungal activity during drought. On 30 August 2022, after the soil rewetting, the DOM regained its usual microbial source signature, reflecting a resumption of biological activity once favorable water conditions were restored.

While a NO3- peak was only observed after the return of rainfall, NH4+ exhibited two distinct dynamics: high concentrations at 30 and 60 cm in June and high concentrations at 5 and 10 cm after rewetting in August. These temporal dynamics are explained below in chronological order.

During the water deficit period in soil (Fig. 2), a peak of NH4+ concentration on 21 June 2022 (Fig. 3) occurred only at 30 and 60 cm depth. The disturbance of the nitrogen cycle during drought has been the subject of numerous historical studies, notably with experimental droughts on various spruce plots on acid soils in Europe, where canopy roofs were installed (Lamersdorf et al., 1998). They all observed high concentrations of ammonium and dissolved organic carbon in soil solution. Gangloff et al. (2014) observed that ammonium peaks in soil solutions are also linked to specific periods of bacterial activity and to hydrological conditions, in a spruce plot from the Strengbach catchment. This supports the idea of an accumulation of organic matter during water deficit due to stress on the soil biosphere (cf. 4.4). The mortality of bacteria and fungi may release cell content that is rich in nitrogen-containing organic matter. This organic nitrogen is degraded to ammoniacal nitrogen by ammonification (Kuypers et al., 2018). The accumulation of NH4+ is attributed to the interruption of nitrification during water deficits, whereas ammonification is less sensitive, as demonstrated by experiments on spruce soils carried out by Hentschel et al. (2007) or Chen et al. (2011). Thus, the high concentrations of NH4+ at deeper horizons, i.e. 30 and 60 cm depth, can be related to the punctual stress of the soil microfauna. This is consistent with the FI and BIX index, which showed a less microbially derived signature at the corresponding period (cf. 4.5).

After the return of rainfall (30 August 2022), the NH4+ concentrations peaked in soil solutions at 5 and 10 cm depths. This can be attributed to the throughfall inputs, as previously discussed (cf. 4.2). At the same time, high nitrate concentrations were observed in soil solutions at all depths (Fig. 3). After soil rewetting, conditions became favorable for nitrifying bacteria: high ammonium concentration, high pH, high temperature and sufficient water content. This led to significant development of nitrifying populations at 30 cm depth, as observed by Krüger et al. (2021) after the 2018 drought under a beech plot. This resulted in high nitrification after rain returned, leading to the observed high NO3- concentrations in soil solutions. In addition, nitrate accumulation in soil solutions was also exacerbated by the lower tree uptake (cf. 4.3).

Thus, periods of water deficit and the first episodes of rewetting, by modifying bacterial activity, severely disrupt the cycle of nitrogen and nitrogen nutrients (NH4+ and NO3-) concentrations in soil solutions.

The most acidic pH (4.1) recorded in soil solution at 30 cm depth during the 2015–2023 monitoring period was observed on 30 August 2022. In contrast, a significantly higher pH (5.2) was measured on June 21, coinciding with low soil moisture content (Fig. 2). The ammonification consumes protons and consequently increases pH (Kuypers et al., 2018). This process likely explains the highest pH of the 2015–2023 period at 30 cm with a value of 5.21 observed on 21 June 2022 (Fig. 3). Conversely, the low pH coincided with exceptionally elevated nitrate concentrations observed in August (Fig. 3) because nitrification is acidifying (Kuypers et al., 2018). The sharp pH variation illustrates the strong influence of the nitrogen cycle on the pH of those soil solutions. In addition, this low value of pH can also be related to the high DOC concentration, which was the highest ever measured at 30 cm depth (Fig. 3). Indeed, DOC is composed mainly of organic acids, contributing to acidification (Strobel et al., 2001).

Furthermore, this acidification has consequences for the dynamics of some elements in the soil, as protons can desorb cations from soil exchangeable surface, in particular Al³⁺, Ca²⁺, Mg²⁺ and K⁺ (Ulrich et al., 1980; Fest et al., 2005; Meng et al., 2019). This process resulted in the release of those cations into soil solution and likely contributed to the elevated concentrations in solution at 30 cm depth (Fig. 3). This is especially noticeable with Al concentrations, with the highest concentration ever observed during the 2015–2023 period at 30 cm depths on 30 August 2022 (Fig. 3). It raises issues for tree health, as Al³⁺ is toxic in high concentrations (Ulrich et al., 1980; De Wit et al., 2001). The desorption of cationic nutrients such as Ca²⁺, Mg²⁺ and K⁺ is of particular concern in acid soils that are already low in cationic nutrients, such as those in the Strengbach catchment (Probst et al., 1992; Dambrine et al., 1998; Oursin et al., 2023).