From anthropogenic toward natural acidification: Effects of future deposition and climate on recovery in a humic catchment in Norway
Abstract
Five decades of monitoring data (1974–2022) at the acidified forested catchment of Langtjern in southern Norway document strong chemical recovery and browning of surface water, related to changes in sulfur (S) deposition. Further recovery is likely to be impacted by future air quality and climate, through catchment processes sensitive to climate change, where the relative importance of these drivers of recovery is poorly known. Here, we explore the importance of the aforementioned drivers for recovery using the well-established process-oriented Model of Acidification of Groundwater In Catchments (MAGIC) with historical and projected deposition and climate from 1860 to 2100. New in MAGIC are (i) a solubility control of dissolved organic carbon (DOC) from S deposition, which allows inclusion of the role of organic acids in chemical recovery and (ii) climate-dependency of weathering rates. MAGIC successfully described observed chemical recovery and browning, and the change toward organic acid dominated acidification status. Hindcasts of pH predicted lower preindustrial pH than previously modeled with MAGIC (simulated without S-dependency of DOC solubility). Future deposition resulted in limited further recovery. Climate scenarios indicated a substantially wetter future, leading to increased base cation losses and slight surface water reacidification. A sensitivity analysis revealed that a 25%–50% increase of weathering rates was needed to reach preindustrial acid-neutralizing capacity in 2100, provided S deposition is reduced to a minimum. We predict that the limited chemical recovery from reduced S deposition will be counteracted by climate-driven reacidification from base cation losses, but that enhanced weathering rates could partly compensate these losses.
1 INTRODUCTION
Reductions in anthropogenic emissions of sulfur (S) and nitrogen (N) to the atmosphere has resulted in wide-scale chemical recovery of surface waters in acid-sensitive regions in Europe and North America (Garmo et al., 2014; Skjelkvale et al., 2005). Reductions in sulfate (SO4), rather than in N deposition have been the dominant driver of water chemical recovery since N deposition change has been less distinct than for S, and N-retention in catchments diminish its impact on elemental runoff (Watmough et al., 2005).
Chemical recovery of acid-sensitive surface waters includes declines in SO4 concentrations and, following the principle of electroneutrality, associated declines in base cations in addition to increases in pH, alkalinity, acid-neutralizing capacity (ANC), and reduced inorganic aluminum (Al) species (Reuss & Johnson, 1985). Reduced air pollution has also led to widespread browning of surface waters (de Wit et al., 2021; Monteith et al., 2007), where other factors such as precipitation amount also impact dissolved organic matter (DOM) concentrations (de Wit et al., 2016). Increases in DOM are relevant for assessing chemical recovery of surface waters since increases in organic acidity (OA) partly compensate for declines in mineral acidity, for example, declines in strong acid anions (SAAs) (Evans, Monteith, et al., 2008). The organic acid adjusted acid-neutralizing capacity (ANCoaa) was proposed since it describes recovery from acidification in humic lakes better than ANC based on major ions only (Lydersen et al., 2004). Factors that affect surface water acidification on a more local or regional basis are seasalt deposition, through mobilization of protons and aluminum similar to acidification from air pollutants (Hindar et al., 2004), catchment disturbances such as tree die-back from insect attacks, leading to mobilization of S and N (Oulehle et al., 2021), and hydrology through post-drought mobilization of SO4 in peatlands (Clark et al., 2006) and variation in flow-paths (Evans, Reynolds, et al., 2008). Possibly, chemical recovery may be enhanced by climate-induced increased weathering rates (Augustin et al., 2015) resulting in higher rates of base cation replenishment in soils, which have lost base cations through decades of mobilization and leaching due to acid deposition (Watmough et al., 2005).
Because of the high success of emission reductions of S, the potential for further reductions in Europe will be limited (Grennfelt et al., 2020; Schopp et al., 2003). That implies that other factors such as climate and land use will become relatively more important for future chemical recovery of surface waters (Kopacek et al., 2016; Vuorenmaa et al., 2017) and for the time required for reaching preindustrial water quality, if at all possible (Helliwell & Simpson, 2010).
For assessment of future surface water acidification, the process-oriented Model of Acidification of Groundwater In Catchments (MAGIC) (Cosby et al., 1985; Norling et al., 2024) is commonly used (Larssen, 2005; Posch et al., 2019). MAGIC has at its core descriptions of acid–base soil chemistry and elemental mass balances and has been further developed to include forest N cycling and C storage which enable simulation of catchment disturbances and climate change effects on element cycling and surface water chemistry (Oulehle et al., 2015, 2019; Valinia et al., 2021).
A new version of MAGIC called MAGIC-Forest is implemented in the open source model development framework Mobius (Norling et al., 2021) enabling flexible sensitivity analysis and adding of model features (Norling et al., 2024). Simulation of “confounding factors” on expected recovery of surface water is necessary for credible predictions in the current era where factors other than deposition will influence surface water acidification status (de Wit et al., 2023). Until recently, browning of surface waters was not implemented in MAGIC while weathering rates were calibrated to a constant value. In the current era of low S deposition, changes in organic acids and weathering are likely to become increasingly important drivers of surface water acidity status.
Here, we explore how changes in OA and weathering rates affect acidification and recovery in an acidified, humic catchment in Norway, under further reductions of atmospheric deposition and climate change. For this, we use the MAGIC model as a tool with new features, for example, browning (sulfate-control) and user-adjustable weathering rates. Our research objectives were the following: (i) to provide an overview of changing deposition and hydrochemistry at Langtjern; (ii) to reproduce empirical charge balances, including estimates of organic charge; (iii) to reproduce acidification and recovery with MAGIC given observed increases of OA, and produce hindcasts and forecasts of preindustrial and future ANC under changes in acid deposition and climate; (iv) to test how much weathering rates need to increase to reach preindustrial ANC and base saturation in 2100, given changes in acid deposition and climate? The model was calibrated based on five decades of streamwater monitoring at Langtjern.
2 MATERIALS AND METHODS
2.1 Site description
Langtjern is a forested, boreal lake catchment in southeast Norway (Figure 1) (4.8 km2; 510–750 m a.s.l.; 60.371° N, 9.727° E) where monitoring of water chemistry started in 1972 under the national monitoring programs for air pollution effects on surface waters (Lund et al., 2018; Vogt & Skancke, 2022). The catchment includes a lake (0.23 km2) and its land cover is dominated by forest (75% forested with low productivity Scots pine forest and 5% of productive Norway spruce forest), located on shallow mineral soils, and peatland (20%). Since the 1950s, when some small-scale forest harvest was conducted, the catchment has been free from direct human disturbances. The geology consists of till generated from felsic gneisses and granites. Mean annual temperature, precipitation, and discharge (1974–2022) are 2.6°C, 834 mm, and 634 mm, respectively.

2.2 Data sources
2.2.1 Streamwater discharge and chemistry
Annual values for discharge and flow-weighted streamwater chemistry (1974–2022) were derived from the routine monitoring conducted by the Norwegian Institute for Water Research (NIVA) as part of the Norwegian national environmental monitoring programmes (Vogt & Skancke, 2022). Water level is monitored continuously by a weir at the lake outlet and converted to discharge using standardized stage-discharge relationships. Streamwater samples are collected weekly at the lake outlet and analyzed for major chemical components (pH, labile Al, alkalinity), major anions (SO4, Cl, NO3), major cations (Ca, Mg, Na, K), and TOC at the accredited laboratory at NIVA. Analytical methods have evolved since monitoring began in 1973 and currently entail automated ion-chromatography. In 1986, monitoring of TOC started. ANC is calculated as the sum of base cations (SBC) minus the sum of strong acidic anions. ANC corrected for OA, ANCoaa (Lydersen et al., 2004), is calculated as ANC + 1/3 × site density of TOC, where site density is −10.2 μeq mg−1 C. The charge balance of the streamwater chemistry was estimated as the difference between the equivalent sum of all major cations and major anions, and the anion deficit was attributed to OA. Charge density (CD) of TOC was calculated as the anion deficit divided by TOC. Full time series were not available for NH4+ and F−, but earlier data show that their concentrations are low and that they largely cancel each other in terms of equivalent concentrations. Labile Al was assumed to have charge of 2+ while the contribution of bicarbonate was set at zero, an acceptable assumption at pH below 5.5 (mean annual pH at Langtjern <5.3).
Element fluxes at the catchment outlet were calculated for each day using measured water discharges at the sampling point multiplied by the solute concentration interpolated from the weekly samples. The daily calculated data were then aggregated to annual fluxes of each element.
2.2.2 Soil element pools
Soil sampling was done in 1983, 1991, and 2001 with the purpose to estimate catchment soil pools of exchangeable cations. These samplings followed the same design and were reported in Reuss (1990) (1983 sampling), Stuanes et al. (1995) (1991 sampling), and in SFT (2001) (2001 sampling), and compared in Larssen (2005). Standard soil depth for Langtjern was estimated to be 40 cm, including the O horizon. Catchment exchangeable base cation soil pools were calculated by multiplying bulk density with layer depth and soil base cation concentration, where the B horizon was assumed to extend to 40 cm soil depth. Only mineral forest soils were sampled, and the soil base cation stores reported in Larssen (2005) only represent forest soils. Base saturation was 16% in 1983 and 19% in 1991 and 2001 (Table S1).
2.3 Atmospheric deposition
Total annual deposition inputs (the sum of wet and dry deposition) of NH4 and NO3 to MAGIC for the period 1850–2100 were derived from air quality monitoring at the station Brekkebygda (390 m a.s.l., 60.29° N, 9.76° E, 7 km south of Langtjern) for 1974–2022, a hindcast for 1860–1973, projections for 2023–2100, and observed surface water chemistry. The air quality and deposition monitoring for 1974–2022, resulting in annual wet deposition, follows analytical methods, quality control and flux calculations as described in Aas et al. (2022). Total annual deposition inputs of SO4, Cl, and major cations need to be estimated separately to be compatible with the mass balance approach incorporated in the MAGIC model. Also, the measured deposition is not entirely representative for the catchment: dry deposition is not measured, and precipitation amounts may differ from the catchment given the lower elevation of the deposition station. The estimation is done in two steps. First, total deposition of Cl is assumed to be equal to the catchment output of Cl for every year in the monitoring period. Thus, Cl is used as a conservative tracer, assuming that all Cl originates from sea salt aerosols. The total deposition of Cl is assumed to be associated with total deposition of cations and marine sulfate (mSO4) following their relative proportions in seawater (as given by the following molar elemental ratios: Ca/Cl: 0.037, Mg/Cl: 0.196, Na/Cl: 0.856, K/Cl: 0.018, SO4/Cl: 0.103). Second, we assume that catchment outputs of S (in streamwater) equal anthropogenic (SO4*) and marine inputs (mSO4) to the catchment, thus implying there is no SO4 weathering or adsorption. Historical (1860–1973) and future (2023–2100) deposition of sea salts and base cations was assumed to equal their averaged total deposition over 1974–2022. For SO4*, NH4 and NO3, historical deposition was estimated from the modeled deposition for the Langtjern grid square by the European Monitoring and Evaluation Programme (EMEP) based on historical emissions of air pollutants in Europe (Schopp et al., 2003). A scaling factor for the historical SO4* deposition was computed so that the scaled historical total (mSO4 + SO4*) deposition on average matched the observed SO4 flux on the interval 1989–1991, in this case resulting in a scaling factor of 0.98. This 1989–1991 reference interval was selected because it did not show extreme peaks in deposition. Future deposition of SO4*, NH4, and NO3 was assumed to follow the European emissions of the Convention on Long-range Transboundary Air Pollution (CLRTAP) current legislation (CLE) scenario for the Langtjern grid square forward to the year 2050 and then held constant to the year 2100. The scenarios were supplied by the Coordination Centre for Effects of the CLRTAP for the years 2030 and 2050, with 2015 as the reference year. The deposition scenarios were scaled to the Langtjern site based on the ratio between the CLE-scenario data and locally measured data for the year 2015. After 2050, future deposition was assumed constant until year 2100.
2.3.1 Weather data and scenarios
Weather data, including daily average temperature (T, °C), daily precipitation (P, mm), and evaporation (mm), were downloaded using NVE's Grid Time Series (GTS) API. For the period were no monitoring data were available, discharge (Q) was calculated as the difference between precipitation and evapotranspiration. The catchment area of Lake Langtjern was first delineated based on a 10 × 10 m digital elevation model for Norway (https://www.geonorge.no/, DTM 10 Terrengmodell [UTM33]) as described by de Wit et al. (2023). Weather data were then downloaded for each 1 × 1 km grid (partially or entirely) overlapping with the catchment area and then area-weighted averaged to one value for the catchment. The catchment became warmer and wetter during 1974–2022, as indicated by trends (estimated using Sen slopes, Mann-Kendall test; Sen, 1968) in annual T (+0.38°C decade−1 (p < 0.0001)), P (+39 mm decade−1 (p < 0.1) (+5%)), and Q (+40 mm decade−1 (p < 0.01) (+6%)).
We chose scenario Representative Concentration Pathway (RCP) 8.5 over RCP6.0 for projected climate change, since RCP6.0 resulted in changes in climate that were smaller than the empirically-based trends in T, P, and Q. Projected 90% confidence interval of changes in T (delta T), P (delta P), and discharge (delta Q) for scenario RCP8.5 for 2071–2100 relative to reference period 1971–2000 were taken from Hanssen-Bauer et al. (2015) to generate climate data for 2023–2100. The resulting intervals were for delta T: +3.0°C to +5.6°C (median: +4.2°C); for delta P: +8% to +29% (median: +15%); for delta Q: −2% to +16% (median: +8%). The empirically-based averages of T, P, and Q (2002–2022) (3.0°C, 958 mm, 721 mm, respectively) at Langtjern differed from averaged T, P, and Q for the reference period (1971–2000) (2.1°C, 818 mm, 635 mm, respectively). We bias corrected the RCP8.5 scenario for T, P, and Q by adding the differences to year 2023, resulting in the scenario RCP8.5bias_corr. Note that the influence of changes in T is limited to minor changes in chemical constants in the current MAGIC setup. Forest biomass was kept constant, and weathering rates were adjusted independently.
To account for variation in future S and N deposition, related to climate policy that also impacts emissions of S and N (Rafaj et al., 2021), we derived two additional deposition scenarios, for example, CLE_max and CLE_min, for deposition after 2022. CLE_max and CLE_min are linearly diverging from CLE in 2023 to, respectively, reach +20% and −30% in S and N deposition in 2050, after which all deposition scenarios remain constant until 2100.
2.4 Model description and calibration
MAGIC is a process-oriented semi-distributed mass balance model for biochemical processes involving ions and nutrients at catchment scale (Cosby et al., 1985; Cosby et al., 2001), originally used to predict air pollution effects on surface waters and further developed to also include effects of land use and climate change. This setup of MAGIC includes two connected compartments, for example, soil and surface water (stream or lake), where the soil runoff is sent to the surface water compartment. The solutes (SO4, NO3, NH4, Cl, Ca, Mg, Na, and K) concentrations are computed as the mass balance between atmospheric deposition, bedrock weathering, retention (for N species) and export through runoff. A new version of MAGIC, MAGIC-Forest (Norling et al., 2024) with modules for hydrology, forest growth, soil carbon accumulation and SO4-dependent organic matter solubility, has been implemented in the open-source Mobius modeling framework (Norling et al., 2021). The Mobius framework has a modern graphical user interface and scripting for interaction with models, allowing for user-friendly advanced auto-calibration and sensitivity analysis.
In the current MAGIC application, we started from the calibration conducted by Larssen (2005) which described streamwater chemistry responses to reductions in S deposition during the period 1974–2003, using an annual time step. We included monitoring data for the subsequent period 2004–2022. New features in MAGIC used in the current application are (i) the SO4-solubility control of TOC background concentrations (COA of organic acids reduced with factor fSO4 multiplied with SO4), where represents the preindustrial TOC concentration (c for concentration) (Norling et al., 2024) and (ii) a factor w to change weathering rates (only used in a sensitivity analysis; , i.e., the rate stays constant at until year , and increases linearly with slope after that). The values of fSO4 and are derived from the linear relationship (slope and intercept) between yearly means of observed TOC and SO4 concentrations in the lake at Langtjern. Table S2 describes which parameters are fixed and which are calibrated. The parameters describing soil characteristics are based on soil element pools as described above. The model was calibrated manually following three steps, and optimization of the model parameters was done by comparing observed and modeled annual flow-weighted mean concentrations of ANC, H+, labile Al, SO4, NO3, Cl, Ca, Mg, Na, K, and organic acids. During calibration, we also ensured that other modeled variables were showing an acceptable fit, that is, within 10%–20% of empirically-based data such as total and cation-specific soil base saturation. The order of calibration of the various outputs, following previous calibrations of MAGIC (Larssen, 2005), was: (1) SAAs; (2) TOC, H+, labile Al and the SO4-solubility control of TOC; (3) base cations. To ensure proper optimization during calibration, the three-step calibration is repeated several times.
The best set of parameters (all parameters calibrated, see Table S2) was selected using expert judgment anchored in an evaluation of combined performance metrics: the root mean square error (RMSE), the coefficient of determination (r2) and the bias (summation of difference between model and empirical estimate, divided by nr of estimates), in addition to comparison with soil base saturation. The calibration was repeated until all variables were simulated with a bias lower than 2 μeq L−1 and a r2 of at least 0.7 for all solutes, except (i) Na and K which have a minor contribution to the total charge balance and temporal changes and (ii) TOC for which r2 was at least 0.4. The automated optimization implemented in Mobius (Norling et al., 2021) used for the base cation step (Step 3) of the calibration resulted in similar model performance as the manual calibration. However, manual optimization was chosen over automated since the automated optimization did not result in converging parameter values for all calibration steps.
2.5 Predictions and sensitivity analysis
We used the CLE scenario for future deposition in combination with three climate scenarios (ConstClim, constant climate, RCP8.5, and RCP8.5bias_corr) for predictions of future water quality and base saturation until 2100. We also used CLE_max and CLE_min in combination with each climate scenario to assess the effect of variation in deposition on ANC and base saturation.
Regarding weathering, we tested in a sensitivity analysis how much the base cation weathering rates would need to increase to reach preindustrial ANC or soil base saturation by 2090–2100, under CLE deposition and the three climate scenarios. To do this, the weathering factor (see model description) was increased until ANC or total soil base saturation over 2090–2100 reached preindustrial levels (1850–1860). The change in weathering rate was equal for all four base cations and effective from 2000.
3 RESULTS
3.1 Acidification and chemical recovery
Deposition of S and N at Langtjern peaked during the late 1960s at 61 meq m−2 year−1 (9.8 kg S ha−1 year−1) and 63 meq m−2 year−1 (8.8 kg N ha−1 year−1) (averaged over 1965–1969) (Figure 2). After the 1960s, S deposition gradually declined toward the current (2018–2022) level of 6.6 meq m−2 year−1 and is expected to remain more or less constant toward 2100 under the emission scenario of CLE. Deposition of N was more variable, showing two other peaks after the 1960s, and is currently (2018–2022) at 26.5 meq m−2 year−1.

Streamwater chemistry followed patterns that are typical for acidified, but recovering, streams (Figure 3): strong declines in SO4, following the temporal change in S deposition; declines in base cations (and SBC), in particular, in Ca and Mg, and increases in pH and TOC. Labile Al showed the strongest decline. The other SAAs Cl and NO3 also declined, where Cl is currently at an approximately constant level of circa 10 μeq L−1, similar to SO4. Nitrate was present in very low concentrations relative to SO4 because of high retention in the catchment. pH, ANC and ANCoaa are currently (2018–2022) below MAGIC-simulated preindustrial levels (Table 1).

Scenario | Preindustrial | Peak | Current | Future | |||
---|---|---|---|---|---|---|---|
CLE + constant climate | CLE + RCP8.5 | CLE + RCP8.5bias_corr | CLE + RCP8.5bias_corr + increased weathering | ||||
pH | 5.38 | 4.84 | 5.18 | 5.25 | 5.20 | 5.12 | 5.36 |
ANC (μeq L−1) | 73.6 | 8.1 | 63.3 | 68.9 | 65.6 | 61.2 | 73.4 |
ANCoaa (μeq L−1) | 34.7 | −21.8 | 24.0 | 26.8 | 23.3 | 18.8 | 34.5 |
TOC (mg L−1) | 11.2 | 7.4 | 11.0 | 11.3 | 11.2 | 11.2 | 11.2 |
SOA (μeq L−1) | 72.5 | 41.9 | 68.6 | 71.4 | 70.6 | 69.2 | 73.1 |
CD-TOC (μeq mg−1 C) | 6.48 | 5.66 | 6.21 | 6.32 | 6.33 | 6.21 | 6.56 |
SAA (μeq L−1) | 15.8 | 101.6 | 20.2 | 14.5 | 14.5 | 14.7 | 14.7 |
SBC (μeq L−1) | 89.4 | 109.8 | 83.4 | 83.5 | 80.1 | 75.9 | 88.1 |
BS (%CEC) | 23.2 | 20.3 | 19.4 | 20.2 | 19.9 | 17.9 | 22.6 |
- Note: MAGIC predictions are given for the future (2090–2100), for various combinations of air pollution scenario (current legislation policy, CLE) and climate scenarios (constant climate; RCP8.5, RCP8.5bias_corr), all with constant weathering rates; and for the sensitivity analysis with increased weathering rates (RCP8.5bias_corr only).
- Abbreviations: ANC, acid-neutralizing capacity; ANCoaa, organic-acid adjusted acid-neutralizing capacity; BS, base saturation of soil; CD, charge density; SAA, sum of acid anions; SBC, sum of base cations; SOA, sum of organic acids; TOC, total organic carbon.
OA, calculated from the charge balance between SAAs and major cations, increased markedly over time, balancing 33% of the base cations during the end of the 1980s and 73% for 2018–2022 (Figure 4). On an equivalent basis, organic acids have dominated SAA at Langtjern since 2000–2004. CD of TOC increased from 4.3 to 5.7 μeq mg−1 C since the end of the 1980s (Figure 5).


3.2 MAGIC hindcasts of historical acidification and recovery
Model performance for individual and compound variables were evaluated for a set of performance statistics (Table 2). MAGIC described levels and variations for major cations, major anions, pH and ANC well (Figure 3). The variation was described best (r2 > 0.7) for the SAA and its constituents (SO4, Cl, NO3), dominating base cations Ca and Mg, ANC, ANCoaa, organic acids, and pH. These variables also had low normalized RMSE's (<0.7), except for variables associated with, or affected by, OA such as ANCoaa, labile Al, and organic matter CD—these variables also had relatively low values for NSE, reflecting systematic overestimation (organic acids, ANCoaa, CD). The overestimation of OA was primarily related to an overestimation of the TOC CD with on average ca 1 μeq mg−1 TOC (Table 1; Figure 5). The variables that were described with least bias (absolute bias <1) were those that showed least temporal variation (Figure 3), while those described with a large bias (absolute bias >|5|) were ANC, ANCoaa and organic acids.
RMSE | r2 | RMSE/std | Bias | |
---|---|---|---|---|
[Ca2+] | 5.66 | 0.72 | 0.56 | 0.90 |
[Mg2+] | 2.40 | 0.81 | 0.62 | 0.46 |
[Na+] | 2.58 | 0.17 | 0.92 | 0.34 |
[K+] | 0.82 | 0.33 | 0.90 | −0.02 |
[SO42−] | 6.35 | 0.94 | 0.27 | −2.02 |
[Cl−] | 2.16 | 0.72 | 0.52 | −0.18 |
[NO3−] | 0.32 | 0.71 | 0.58 | −0.11 |
[H+] | 2.38 | 0.71 | 0.73 | −1.65 |
pH | 0.07 | 0.71 | 0.60 | 0.03 |
ANC | 8.45 | 0.84 | 0.62 | 6.45 |
ANCooa | 7.10 | 0.85 | 0.92 | 8.76 |
[Aln+] | 1.75 | 0.74 | 0.89 | −1.44 |
TOC | 1.28 | 0.42 | 1.08 | −0.85 |
SOA | 7.46 | 0.78 | 0.72 | −5.38 |
CD-TOC | 0.97 | 0.69 | 1.94 | −0.90 |
SAA | 7.78 | 0.93 | 0.28 | −2.31 |
SBC | 10.30 | 0.67 | 0.62 | 1.68 |
- Abbreviations: ANC, acid-neutralizing capacity; ANCoaa, organic-acid adjusted acid-neutralizing capacity; CD, charge density; SAA, sum of acid anions; SBC, sum of base cations; SOA, sum of organic acids; TOC, total organic carbon.
Since OA is such an important part of the total charge balance, model simulation of other ions is very sensitive to its level. Decreasing of OA in MAGIC leads to lowering of the base cation concentrations, so optimizing OA in MAGIC can come at the cost of good optimization of SBC and therefore pH and ANC. However, the upward trend in OA was described well. This is related to the SO4_dependency of TOC, which is a novel aspect of MAGIC modeling.
The hindcast of all the elements, from 1850 to 1973, was driven by the change in deposition (Figure 2). Preindustrial pH, ANC, ANCoaa, and TOC were higher than current-day values and much higher than their values during the period when acidification peaked (Table 1). Base saturation decreased from 23.2% (preindustrial) to 19.4% (present-day).
3.3 MAGIC projections of future acidification and recovery
EMEP projections of S and N deposition in 2050 under the CLE deposition were 56% and 67% of total S and N deposition, respectively, where 2015 was used as the reference year.
The catchment became warmer and wetter during 1974–2022 (+0.38°C decade−1 (p < 0.0001) and +40 mm discharge decade−1 (p < 0.01)). The climate scenario RCP8.5 predicted lower precipitation than the historical record, which was the main reason for bias-correcting RCP8.5 to RCP8.5bias_corr. That resulted in a higher mean discharge (+80 mm) compared with average discharge for 1974 to 2022 for RCP8.5bias_corr (Figure 6).

The Constant Climate scenario did not result in more chemical recovery in 2050 and 2100 than observed for 2018–2022 (Figure 3), while locally adapted RCP8.5bias_corr and RCP8.5 resulted in a slight reacidification (Table 1). RCP8.5 bias_corr had a slightly stronger effect on acidification than RCP8.5, as a consequence of higher discharge and thus higher base cation export. Chemical acidification status in 2100 was markedly below the preindustrial water quality indicators pH, ANC and ANCoaa. The three future scenarios gave rather similar results, with OA dominating over mineral acidity (Figure 4) and the strongest climate change scenario resulting in a slight reacidification (reduction of ANC by 5.2 μeq L−1; Table 1). The effects of the high (CLE_max) and low (CLE_min) deposition scenarios, for all climate scenarios, were maximum ±1 μeq L−1 for ANC, and ±0.1% base saturation in 2100 (data not shown). By comparison, the differences in ANC and base saturation under the climate scenarios, given the deposition under CLE, were 3–8 μeq L−1 and 1%–2%, respectively (Table 1). The uncertainty in discharge for each scenario (Figure 6) added additional variation in the projections (Table 3). The MAGIC projections indicate that climate change has a stronger effect on future recovery than further reductions in deposition.
Increase (% year−1) in weathering rates (averaged over 2000–2100) | Additional cumulative base cation weathering inputs (in % of inputs under constant weathering, over 2000–2100) | |||||
---|---|---|---|---|---|---|
ConstClim | RCP8.5 | RCP8.5bias_corr | ConstClim | RCP8.5 | RCP8.5bias_corr | |
Soil base saturation | 0.54 | 0.59 (0.51–0.68) | 0.96 (0.87–1.04) | 33 | 36 (30–43) | 68 (59–76) |
ANC | 0.33 | 0.58 (0.44–0.75) | 0.87 (0.73–1.03) | 19 | 35 (26–49) | 59 (47–75) |
- Note: Period considered is 2000–2100. The intervals show interquartile range based on the interquartile range of predicted precipitation under the RCPs.
3.4 Sensitivity analysis
We tested in a sensitivity analysis which increase (from year 2000) in weathering rates was needed to obtain preindustrial base saturation and ANC in 2100 (Table 3) for three climate scenarios and constant deposition. The scenario with the highest precipitation, RCP8.5bias_corr, required the largest increase in weathering (+1% annually, averaged over 2000–2100 a 50% increase) whereas the constant climate scenario required only an +0.5% year−1 increase (25% averaged over 2000–2100). In MAGIC, higher precipitation leads to higher catchment export of base cations which must be compensated by base cation inputs from weathering, which is illustrated by the higher contribution of base cations under RCP8.5bias_corr than for the other two scenarios. The required increased weathering rate for RCP8.5 was comparable to the “constant climate” scenario because of the similar levels of precipitation in both scenarios.
It took a lower increase in weathering to reach preindustrial ANC than preindustrial base saturation because of the far higher mass of exchangeable cations in the soil than in the lake, making water chemistry more sensitive to changes in weathering rates than soil chemistry.
4 DISCUSSION
The monitoring data demonstrate strong chemical recovery from acidification at Langtjern, similar to other strongly acidified surface waters in Norway (de Wit et al., 2023), elsewhere in Europe and North America (Bukaveckas, 2021; Garmo et al., 2014; Houle et al., 2022; Kopacek et al., 2021; Lawrence et al., 2021; Sterling et al., 2022) and in Japan (Sase et al., 2021). Simultaneously, dissolved organic carbon (DOC) at Langtjern has increased and now contributes more to the total anion charge in the streamwater than the SAAs. The site is thus moving toward a state dominated by natural-, rather than anthropogenic acidification. Browning is a common feature for surface waters in boreal regions, primarily related to reduced acid deposition (de Wit et al., 2021; Monteith et al., 2007). Previously, increased DOC was found to be significantly related to long-term declines in SO4 at Langtjern while climate variables controlled seasonal variation (de Wit et al., 2007; Futter & de Wit, 2008). To what extent and at what time scale OA will dominate anthropogenic acidification is not well-known, however. Assumptions about natural levels of OA have considerable implications for assessments of the anthropogenic contribution to surface water acidity (Erlandsson et al., 2011) and the necessity of liming to accelerate recovery (Laudon et al., 2021). Humic CD for DOM increased, as found earlier (de Wit et al., 2007), suggesting that both increased solubility and decreased deprotonation of humic and fulvic acids control the return to naturally acidified systems and preindustrial water quality.
The last MAGIC application for Langtjern was calibrated on the period 1974–2003 (Larssen, 2005). We used the same soil and historical deposition data. In our study, we optimized for 1974–2022 and captured the trend in chemical recovery with hardly any bias, which is a considerable improvement of the Larssen (2005) application where MAGIC underestimated the positive trend in ANC, overestimating peak acidification ANC (by 10 μeq L−1) and underestimating ANC (by 20 μeq L−1) in the early 2000s. Larssen (2005) reported relatively poor model performance for K, Na, Mg, and Cl compared with MAGIC applications for two other acidified catchments in southernmost Norway, which could indicate that the relatively poor simulation of ANC levels at Langtjern were related to other major cations and anions than SO4 and Ca. Also, an internal source of S in the Langtjern catchment (6 meq m−2 year−1) was assumed by Larssen, contrary to our study, which could lead to overemphasizing the importance of background SO4 at the cost of anthropogenic SO4 and anthropogenic acidification. This is illustrated by the substantially higher hindcast for preindustrial ANC in our study (e.g., ANC of 80 μeq L−1; Figure 3) than in Larssen (2005) (e.g., ANC of 40 μeq L−1). Furthermore, the longer time series in our study presents a stronger dataset for model calibration. The inclusion of changing OA in our study, where Larssen assumed constant OA, cannot explain the poorer description of recovery by Larssen since OA counteracts changes in mineral acidity, leading to a lower response in ANC to changes in SAA.
The increases in DOC and OA were not described as well by MAGIC as changes in mineral acidity. The levels of OA at Langtjern were somewhat overestimated possibly indicating that organic acid properties at Langtjern are outside the range for values for organic acid deprotonation reported by Hruska et al. (2003). The organic CD estimated in our study was between 4.4 and 5.7 μeq g−1 C, which agrees with the values previously calculated for Langtjern (de Wit et al., 2007).
The MAGIC-forecasted pH, ANC and ANCoaa under reduced S deposition and current climate were not yet back at preindustrial levels in 2100, which was largely related to depleted base cation stores. Climate change acted through increasing discharge and thereby further depleting base cation stores and thereby leading to a slight reacidification. The RCP8.5 projection has been criticized because the greenhouse gas trajectories require unrealistically high emissions (Hausfather & Peters, 2020). However, the observed changes in precipitation at Langtjern already exceeded the projected change under RCP8.5 and we needed a bias correction to avoid a step-change between the historical and future precipitation record, which illustrates that uncertainties in future climate depend on more factors than emission pathways. Rafaj et al. (2021) estimated a global change in S and N emissions under various climate policies of between +5% and −30% by 2050 compared with 2015, an interval that is mostly likely to be smaller in Europe since air pollution policy has been implemented in Europe since the 1980s. Projected effects of estimated minimum and maximum atmospheric deposition scenarios on surface water ANC and base saturation in 2100 were small compared with projected effects of different climate scenarios, given CLE deposition, suggesting that climate through increased export of base cations is larger driver of future acidification and recovery than deposition.
The projected slight reacidification has already been observed in humus-rich acidified lakes in Eastern Norway (de Wit et al., 2023), which was attributed to OA rather than climate change. However, TOC and OA are also climate-sensitive, in particular to precipitation (de Wit et al., 2016), a feature that is currently not included in MAGIC. Increased catchment base cation export related to discharge is textbook knowledge, but changes in soil base cation stores for periods shorter than three to five decades are typically difficult to measure (Ahrends et al., 2022).
We showed in a sensitivity analysis, using new features of MAGIC, that a 30% increase in weathering rates integrated over the 21st century, resulting in an additional base cation input of +35%, would be required to achieve preindustrial base saturation and associated water quality in 2100. An even higher weathering rate (+44% to +48% over the 21st century) would be necessary to compensate for enhanced soil base cation loss under higher runoff from climate change. While it is clear that mineral weathering rates are higher under forested versus non-forested sites (Berner, 1997), generally higher in moist versus dry climates (West et al., 2005), and are promoted by higher pCO2 and organic acids (Bargrizan et al., 2020), it is difficult to constrain weathering rates based on empirical measurements (Koseva et al., 2010; Kronnas et al., 2019). However, unexpected widespread increases in dissolved silicates (+1% to 2% annually, depending on region) and calcium (−0.5% to +1.5% annually, depending on region) have been observed in Norwegian lakes for the period 1995–2019 (de Wit et al., 2023), possibly providing indirect evidence of recent, enhanced weathering rates. The projected increased weathering rates in MAGIC—+0.5% to +1.0% annually—are thus within the rates of change found for silicates and calcium. However, it should be noted that calcium declined at Langtjern similar to other acid-sensitive surface waters in southern Norway. Also, acidic soils developed in glacial till derived from gneiss, such as those at Langtjern, are usually depleted in easily weatherable minerals (Ali et al., 1995). We conclude that chemical recovery at Langtjern may be impacted—positively and negatively—by climate change through enhanced weathering and increased runoff. Continued monitoring is needed to evaluate which drivers will prove to be strongest.
With its new features for solubility control of DOC from S deposition and adjustable weathering rates, the new, open-source version of MAGIC is an increasingly flexible tool for modeling surface water acidity status, allowing for separate and joint assessments of deposition and climate-induced drivers of water chemical change. Being open source, the model is now more accessible for further refinement and extension, and has better support tools, like the user interface. In the ongoing transition from anthropogenic to natural acidification in surface waters, model predictions indicate that future deposition only leads to limited recovery and that climate change might lead to slight reacidification of surface waters, but with enhanced weathering rates as a possible remediating factor.
ACKNOWLEDGMENTS
This work was conducted as part of the CatchCaN project (the fate and future of carbon in forests), funded by the Technology Agency of the Czech Republic (TA CR) project number TO 01000220 and by NIVAs core funding (Norwegian Research Council, contract nr 342628/L10). We acknowledge EMEP/MSC-W (http://emep.int/mscw) for the provision of historical deposition data and projections of future deposition. Arne Stuanes, Dick Wright, Kari Austnes, and two anonymous reviewers, are gratefully acknowledged for helpful comments on the manuscript.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.