Figure 2. Hydrograph decomposition tracking flow variation.
Grab samples of stream water were collected at each visit (approximately weekly) several meters upstream from the respective stilling ponds in high-density polyethylene 250 mL bottles. More frequent samples were obtained during storms from 1983 to 1990 at TM9, collected with an automatic sampler (Avila et al. 1992).
3.2 Atmospheric deposition data
Atmospheric deposition was monitored from 1978 through 2019 with some interruptions. The location and the mode of sampling collection varied in this period as described in Figure 1. Parallel bulk and wet deposition sampling allowed comparison between the two methods: high correlation between them was found for the different ions, BD exceeding WD by around 37% for base cations (Izquierdo and Avila 2012).
3.3 Analytical methods
Stream water and atmospheric deposition samples were taken to the laboratory the same day of collection. Sample processing and analysis are described in Avila (1996), Rodà et al. (1993) and Avila et al. (2020).
Annual mean concentrations were volume-weighted for all ions except for H+. At Montseny, with weekly rainwater pH values varying between 4.0 and 8.0, alkalinity/acidity is the conservative property to average. Volume-weighted mean alkalinity was then transformed to mean pH with the equation: pH = log (alk) + 5.2 (units in µeq L-1) which is derived from the equilibrium equations of the CO2 -carbonate system (Stumm and Morgan, 1981).
Deposition samples with African dust were filtered and chemical (ICP-MS and ICP-AOE), mineralogical (X-Ray diffraction; Avila et al. 1997) and genomic analysis of the particulates in the filters was later performed. DNA was extracted using the Mobio PowerSoil DNA Isolation Kit (Mobio Laboratories). PCR and high-speed multiplexed SSU rRNA gene Illumina MiSeq sequencing were carried out for 16S and 18S rRNA genes (https:// rtsf.natsci.msu.edu/).
Element input or export from the catchment was calculated as the sum of the element inputs or exports for each sample in the period of interest (e.g. month, year). Inputs are calculated as the product of bulk or wet deposition concentrations by precipitation. Export for a sample Si is calculated as the product of the element concentration in the sample (Si concentrations in µeq/L) by the water drained in the period starting at the point midway from Si to the previous sample Si-1 and ending at the point midway from Si to the next sample Si+1 (water yield in mm = L/m2). This mode of calculation has been found adequate for weekly or biweekly stream sampling frequencies, as is here the case (Rekolainen et al. 1991, Swistock et al. 1997). Drainage is also expressed per unit catchment area (drainage in mm) in the hydrology files.
Principal results
4.1 Streamwater chemistry
The Torrent de la Mina streamwaters are well buffered, with Na+ and Ca2+ co-dominating as main cations and alkalinity as the main counterpart, both in TM0 and TM9. Mean pH was 7,47 and 7,54 for TM0 and TM9 respectively. Solute variation at TM9 was explained as the mixing of groundwater and deep-soilwater components, contributing 70 and 30% of annual runoff, respectively (Piñol et al. 1992, Avila et al. 1996). Both catchments responded rapidly to changes in atmospheric deposition: an alkalinity increase was found associated to declining sulfate deposition (Avila and Rodà 2012). This responsiveness may be explained by a low residence time of water in these catchments compared to other geographical areas (Bernal et al. 2013). Nitrogen was strongly retained in the catchments, with only 2% of total N deposition being exported as inorganic N in streamwater. However, higher nitrate export occurred during the wettest months (November and March). Moreover, this pattern showed an increasing trend in time, suggesting the onset of N saturation in the catchments (Avila et al. 2020).
Torrent de la Mina data have been used for modelling the stream chemistry response to forestry practices (Neal et al. 1995 ) and to climate change (Avila et al. 1996). More recently, the stream hydrological response to climate change scenarios from IPCC and to land use changes was modeled in the context of LIFE-Tritó project. The aim was to predict the vulnerability of the critically endangered Montseny newt (Calotriton arnoldi ) to expected habitat changes in future scenarios (Ledesma et al. 2019). In this work, the combination of hydrological modelling, climatology and species ecology knowledge provides an example of the use of data gathering to help in the management of an endangered species.
Atmospheric deposition
Rainwater in Montseny is rarely acidic, since sulfate and nitrate acid anions are largely neutralized by base cations and ammonium (Rodà et al. 1993). Annual mean pH values ranged from 4,92 to 7,18 with a median of 6,79. All annual mean pH values below 5,50 occurred before 1990. Frequent episodes of African dust transport (Avila and Peñuelas 1999) are important contributors of alkalinity and of dissolved and particulate elements to this site (Avila et al. 1997, 1998; Castillo et al. 2017). Back trajectory analysis has revealed that the main atmospheric transport pathway to Montseny is from the Atlantic Ocean, delivering dilute and circumneutral rainwaters while acid episodes predominated in local and European trajectories (Izquierdo et al. 2012a). Nitrogen deposition was higher with air masses coming from Mediterranean/ south-east European provenances (Izquierdo et al. 2014). Significant declining trends of S deposition (since the 1980s; Avila and Rodà 2002) and N deposition (starting in the mid-2000s) have been found (Aguillaume et al. 2016; Avila et al. 2020).
The role of African episodes as contributors of phosphorus deposition in Montseny and in the near Mediterranean coast has been analysed by Izquierdo et al. (2012b) and Longo et al. (2014). Because of concern on the negative eutrophying effects of N deposition, particular attention has been devoted to characterise the N inputs to Montseny (Rodà et al. 2002, Avila et al. 2002, García-Gómez et al. 2018). Throughfall sampling and leaf washing experiments were undertaken at different times along the recording period to distinguish dry deposition from canopy exchange at the canopy level (Rodrigo et al. 2003, Avila et al. 2017). Lately, the role of the leaf microbiome was assessed using a multiple isotope approach (involving δ15N, δ18O and Δ17O in NO3- in wet deposition and throughfall) and quantification of amoA genes: microbial nitrification contributed but the NO3 dominant flux in throughfall was atmospheric (Guerrieri et al. 2020).
The Montseny catchments have shown distinctive characteristics compared to more temperate catchments, such as: (1) the hydrology is strongly controlled by evapotranspiration, which accounts for 2/3 of the incoming precipitation (Piñol et al. 1999), (2) nevertheless, there is a quick hydrologic response, as residence time of water in the catchment is estimated around 4-5 months (Bernal et al. 2013),(3) inorganic N is strongly retained in the system, with nitrate being the only inorganic nitrogen form detected in stream waters (detection limit = 0,5 µeq/L) mostly during stormflows; as a consequence, 98% of DIN deposition is retained in the catchments (Avila et al. 2020). The lack of relationship between DIN inputs and outputs in the Montseny catchments contrasts with the finding that N deposition is a strong driver of stream N export in temperate catchments (Templer et al. 2022). On the other hand, similarly to other catchments worldwide and corresponding to a general declining sulfate deposition trend, sulfate concentrations in Montseny stream waters have decreased, while alkalinity has increased (Avila and Rodà 2012) providing well buffered waters for these catchments lying on silicate bedrock.
Data sets and availability
We provide two datasets of open data for further exploration and hypothesis testing. Files are provided for: (1) stream water flow and stream water chemistry, doi:10.5281/zenodo.7228249; and (2) atmospheric deposition, doi: 10.5281/zenodo.7228266.