Time trends & mechanism of soil acidification

The effects of acid atmospheric deposition on forest ecosystems have been studied intensively in the past two decades. Measurements of element budgets in forested ecosystems throughout the world have shown that acid deposition may deplete stores of exchangeable base cations in the soil, decrease the soil pH, increase rates of mineral weathering, and release potentially toxic Al into the soil solution. In summary, acid atmospheric deposition can strongly alter the chemical environment for living organisms.In this thesis the mechanisms that may control those phenomena are studied, with emphasis on soil chemical processes. Central in this study are long and comprehensive data sets from continuous measurements (monitoring) of element fluxes in forest ecosystems, specifically from the Solling experimental forest in Germany.<strong>Base Cations</strong>Using over 20 years of data on deposition and soil chemistry at Solling, chapter 2 discusses how long-term changes in atmospheric deposition of acid anions and base cations affect dissolved and exchangeable pools of base cations in the soil, through the process of cation exchange. Positive effects of declining SO ₄ deposition in the 1980s, which potentially reduces leaching of base cations from the soil, were partially offset by (i) continued high concentrations of dissolved SO ₄ in the soil and (ii) declining deposition of the base cations Ca and Mg. The latter appears to result from stricter controls on particle emissions from fuel combustion and industrial processes. Recent studies indicate that declining deposition of base cations is observed in other regions of Europe and North America as well. This is a matter of concern as forests may be sensitive to changes in base cation deposition which is, together with mineral weathering, the most important source of nutrients to forests on acidified soils.In chapter 3 we quantified rates of mineral weathering at Solling, as a source of nutrients and acid neutralization. Estimates of field weathering rates were obtained from, (i) long-term input- output mass balances, correcting for changes in stores of exchangeable base cations in the soil, and (ii) elemental analysis of the soil profile. At Solling, rates of supply of Ca, K and Na by weathering are much lower than by atmospheric deposition; for Mg the contributions are of similar magnitude. This emphasizes (i) the relevance of atmospheric deposition of base cations as a nutrient source to the Solling forests, and (ii) the concern with respect to its present decline. To investigate mechanisms of mineral weathering of Mg and K, a series of laboratory weathering experiments with soils from Solling were conducted.These focused on weathering of illite, the dominant mineral source of K and Mg. A major problem in quantifying rates of mineral weathering is the large discrepancy in rates obtained in laboratory studies and field estimates, with laboratory rates being generally one to three orders higher. In the laboratory, effects of pH, temperature, and soil solution composition on illite weathering were studied. Corrected for the effects of these factors, Mg and K release rates from illite measured in the laboratory were comparable (K) or a factor 2 to 4 higher (Mg) than rates inferred from field methods. It is concluded that field estimates of weathering rates, from elemental analysis of the soil profile or input-output flux balances, remain necessary to obtain reliable estimates of base cation weathering rates.<strong>Anions</strong>In a simplified way, anions in the soil solution can be regarded as carriers of base and acid cations, through charge balance constraints. Understanding the processes that determine dissolved concentrations of these components in the soil is therefore of prior importance. In chapter 4 we discussed the processes that govern the mobility of dissolved Cl, SO ₄ and NO ₃ in the Solling spruce soil over a period of 18 years. For this purpose a simulation model describing soil hydrology and a number of biogeochemical processes was introduced. Using throughfall water and concentrations as input to the model, concentrations of Cl in soil solutions were successfully simulated, a requirement to further study components that interact chemically or biologically with the soil.Sorption of SO ₄ is of significance in the soils at Solling due to comparatively high contents of Al and Fe oxides. By contrast, in the mineralogically poor sandy soils in the Netherlands SO ₄ often behaves chemically nearly conservative. The steep increase in dissolved SO ₄ in the Solling spruce soil during the mid-1970s was modelled adequately using adsorption parameters reported in literature, and was explained by a steep adsorption isotherm and a state of near saturation of SO ₄ sorption sites in the early 1970s, at the start of the monitoring program. Rates of desorption of SO ₄ in the 1980s in response to declining SO ₄ deposition, described by the same model, were less than those measured in the field. In the absence of evidence for alternative controls on dissolved SO ₄ , like jurbanite (AlOHSO ₄ dissolution, the mechanism of SO ₄ desorption remains unclear.The model increased our understanding of the processes that control repeated 3-year cycles of high and low concentrations of N0 ₃ in deeper layers (90 cm) of the Solling spruce soil. Recently, there is increasing interest in N-dynamics in forest ecosystems, as N-saturation has been hypothesized as an additional cause of forest decline. In the Solling spruce soil, concentrations of N0 ₃ water leaving the soil profile increased when the N-flux into the <em>mineral soil,</em> from deposition and net mineralization, exceeded the N uptake capacity of the forest at a threshold value of around 4 kmol ha <sup>-1</SUP>yr <sup>-1</SUP>. At present, total deposition of N on the spruce forest amounts to 3.3 kmol ha <sup>-1</SUP>yr <sup>-1</SUP>and is close to this threshold value. This points to the critical role of N mineralization processes in the litter layer, where large stores of N have accumulated over the past 25 years. The annual release of N from the spruce litter layer was positively correlated with rainfall. This was apparent from large N releases from litter in relatively wet years which subsequently increased NO ₃ concentrations at the 90 cm depth.<strong>Aluminum</strong>The main acid neutralizing process in acid forest soils under high acid loads is the release of AI to the soil solution. Soil acidification models generally describe the solubility of reactive soil Al by equilibrium or kinetic dissolution of Al-trihydroxide (Al(OH) ₃ ), In chapters 5 and 6, we scrutinized the evidence for this reaction mechanism from extensive sets of field data, but found little support for the gibbsite model. By contrast, evidence points to AI binding to soil organic matter as a control of dissolved Al, even when organic matter contents are low. In chapter 6 a new, mechanistic, and yet simple model was presented that explained much of the spread in Al-solubility, as observed in 29 different soil layers from 18 forest soils. The model describes Al-solubility as a function of the degree to which humus binding sites are saturated with Al. It involves only two soil specific parameters, the organic C content of the soil and an estimate of the pool of organically bound soil Al, and may therefore be easily incorporated in current soil acidification models.The soils studied in this thesis generally receive high loads of acid deposition, and acidity is largely neutralized by release of Al in the upper 20 cm of the mineral soil. It has been suggested that in surface layers of these soils depletion of reactive soil Al may result within decades. The depletion hypothesis was put to test by measuring pools of organically bound soil Al in old (around 1980) and new (around 1993) samples from research sites at Hackfort and Gerritsfles in the Netherlands, and Solling in Germany. Indeed, significant decreases with time were found. Independently measured depletion of organically bound soil Al and declining Al solubility, as observed previously from 6 years of field monitoring at Hackfort in the 1980s, proved to be consistent with the humus-Al model presented in this thesis. Rates of depletion of organically bound soil Al at Hackfort, the Netherlands, during a period of about 13 years were lower than expected from long-term monitoring of Al solute fluxes. This suggests that pools other than organically bound Al, i.e. silicates and inorganic Al hydroxides, may contribute to Al mobilization as well. Literature indicates that release of Al from these pools may be kinetically constrained. The current discussion on whether Al activities in solutions of acid forest soils are controlled by kinetically constrained (time dependent) dissolution of inorganic Al hydroxides or by equilibrium binding to soil organic matter is, in my view, clarified when the kinetic and the equilibrium process are regarded as sequential; kinetic release supplying Al to the organic exchanger, which equilibrates with the soil solution.GENERAL CONCLUSIONS1. In the recent past, deposition of SO ₄ , Ca and Mg on the German Solling forests has declined significantly as a result of reduced industrial emissions.2. Results from Solling indicate that soil de-acidification, i.e. an increase in the pool of exchangeable base cations, as a result of declining SO ₄ deposition may be offset by i) continuous release of sorbed SO ₄ from the soil and ii) declining deposition of basecations.3. Weathering rates of Ca, K and Na at Solling, Germany, are much lower than their supply by atmospheric deposition; for Mg the contributions of deposition and weathering are of similar magnitude.4. Illite is the main mineral source of Mg and K at Solling. Although coming from the same mineral source, the two cations are released by different mechanisms. Magnesium release is controlled by chemical kinetic processes and its release is highlypH dependent (R≈(H <sup>+</SUP>) <sup>0.8</SUP>), whereas K is released from interlayers by diffusion, a process that showed independent of pH.6. A six-fold increase of dissolved SO ₄ in the Solling spruce soil (90 cm) in the mid 1970s, and continuously high concentrations thereafter, is explained by 'peaking' SO ₄ deposition loads in the mid-1970s that entered a mineral soil where the SO ₄ sorption capacity was nearly saturated.6. Repeated 3-year cycles of high and low concentrations of NO ₃ in deeper layers (90 cm) of the Solling spruce soil are related to excess release of N from the litter layer in comparatively wet years.7. Strong changes in pH and less strong changes in pAl (-log{Al <sup>3+</SUP>}) in response to varying acid anion concentrations, generally observed in acid forest soils, and the measured increased of CEC with time in the surface soil of the Solling spruce sites, support the hypothesis of Al solubility control by complexation to soil organic matter.8. Data from 29 soil layers of 18 forest soils, covering a range in soil pH, Al activities, and reactive soil Al pools indicate that Al solubility control by complexation to soil organic matter is a general phenomenon, even when organic C contents are low.9. Much of the variation in Al solubility observed in 29 different soil layers of 18 forest soils is explained by the organically bound Al: soil C ratio.10. In surface soils where stores of organically bound Al are low and release rates high depletion of reactive, organically bound, soil Al is significant on a time scale of 12 to 14 years only.11. Independently measured declines in (i) Al solubility in surface soils at Hackfort the Netherlands, from 6 years of monitoring of soil water chemistry and (ii) pools of organically bound soil Al, are consistent with the Al-humus equilibrium modelproposed in this thesis.12. In acid forest soils where dissolved Al activities are controlled by equilibrium complexation to soil organic matter, kinetically constrained release of Al from inorganic Al pools may well contribute to acid neutralization.