Decreasing DOC trends in soil solution along the hillslopes at two IM sites in southern Sweden — Geochemical modeling of organic matter solubility during acidification recovery

https://doi.org/10.1016/j.scitotenv.2010.09.023Get rights and content

Abstract

Numerous studies report increased concentrations of dissolved organic carbon (DOC) during the last two decades in boreal lakes and streams in Europe and North America. Recently, a hypothesis was presented on how various spatial and temporal factors affect the DOC dynamics. It was concluded that declining sulphur deposition and thereby increased DOC solubility, is the most important driver for the long-term DOC concentration trends in surface waters. If this recovery hypothesis is correct, the DOC levels should increase both in the soil solution as well as in the surrounding surface waters as soil pH rises and the ionic strength declines due to the reduced input of SO42− ions. In this project a geochemical model was set up to calculate the net humic charge and DOC solubility trends in soils during the period 1996–2007 at two integrated monitoring sites in southern Sweden, showing clear signs of acidification recovery. The Stockholm Humic Model was used to investigate whether the observed DOC solubility is related to the humic charge and to examine how pH and ionic strength influence it. Soil water data from recharge and discharge areas, covering both podzols and riparian soils, were used. The model exercise showed that the increased net charge following the pH increase was in many cases counteracted by a decreased ionic strength, which acted to decrease the net charge and hence the DOC solubility. Thus, the recovery from acidification does not necessarily have to generate increasing DOC trends in soil solution. Depending on changes in pH, ionic strength and soil Al pools, the trends might be positive, negative or indifferent. Due to the high hydraulic connectivity with the streams, the explanations to the DOC trends in surface waters should be searched for in discharge areas and peat lands.

Introduction

Numerous studies report increased concentrations of dissolved organic carbon (DOC) during the last two decades in boreal lakes and streams in Europe, Canada and the US (Erlandsson et al., 2008, Evans et al., 2005, Monteith et al., 2007, Skjelkvåle et al., 2005). Many different processes have been proposed to explain these trends e.g. hydrometeorological conditions (Erlandsson et al., 2008, Sarkkola et al., 2009), recovery from acidification due to reduced sulphur deposition (Dawson et al., 2009, Monteith et al., 2007), land cover (Laudon et al., 2009, Sarkkola et al., 2009), forest management and land use (Laudon et al., 2009, Löfgren et al., 2009b, Yallop and Clutterbuck, 2009) etc. Recently, Clark et al. (2010) presented a unifying hypothesis on how various spatial and temporal factors affect the DOC dynamics. They concluded that declining sulphur deposition and thereby increased DOC solubility, is the most important driver for the long-term DOC concentration trends in surface waters, but that the variability between sites is influenced by a multitude of spatial and temporal factors (Clark et al., 2010).

Soils and surface waters in the historically most polluted southern Sweden show clear signs of recovery from acidification since the early 1990s (Karltun et al., 2003, Löfgren et al., 2009a, Skjelkvåle et al., 2005). However, if the recovery hypothesis is correct, the DOC levels should increase both in the soil solution as well as in the surrounding surface waters as soil pH rises and the ionic strength declines due to the reduced input of SO42− ions. In contrast to surface waters, however, the soil waters in forested recharge areas (number of monitoring sites (nsites) = 68, 50 cm soil depth, podzols) exhibit no change (nsites = 32) or decreasing (nsites = 31) DOC concentrations during the period 1986–2008 in southern Sweden, indicating increased coagulation of DOC in the upper soil horizon (Zetterberg and Löfgren, 2009; Löfgren and Zetterberg in prep.). In Norway, similar results were obtained, with no change or decreasing DOC trends during the period 1996–2006 in soil water (nsites = 18) at 15 and 40 cm soil depth in podzols (Wu et al., 2010). In contrast, increased DOC concentrations were found in soil water at two sites during the period 1994–2007 in the Czech Republic. The latter studies represent soil water under the forest floor at Lysina and in the mineral topsoil at Pluhuv (Hruska et al., 2009). Positive DOC trends were also found at 5–20 cm soil depth (nsites = 9, moorlands and forests) during the period 2000–2005 in the UK (Buckingham et al., 2008).

Hruska et al. (2009) concluded that the DOC trends in both soil and surface waters were explained by changes in ionic strength, rather than acidity, while Buckingham et al. (2008) considered the UK time series too short, for making a coupling to the surface water DOC trends. Wu et al. (2010) proposed that small changes in the atmospheric deposition during the investigation period could explain the diverging DOC trends in soil and surface waters. However, they also put forward the possibility of competition between mineral anions and DOC for adsorption sites on oxide surfaces, causing a simultaneous decrease of the DOC and SO42− concentrations. Zetterberg and Löfgren (2009) hypothesized that processes in discharge areas and peat lands rather than dry soils uphill govern the surface water DOC trends.

The solubility of DOC is likely to be determined by a number of different biological, chemical and hydrological processes (see Clark et al., 2010 and references therein), but the acidification recovery theory is primarily coupled to the chemical and physical properties of organic matter in soils and water. According to classical DLVO theory for colloidal stability, the surface potential of a charged colloid may be the single most important factor determining its dispersion into the water phase (e.g. Weng et al., 2002). A high surface potential results in more interactions with water molecules and thus a high water solubility. Therefore, different models have been forwarded that relates the DOC solubility either to the surface potential or to the net charge, which is closely related to the surface potential.

Tipping and Woof (1990) suggested a model for DOC dissolution from soils that assume a nonlinear relationship between the DOC concentration and the net humic charge. According to this model, an increased net charge leads to an increasing DOC concentration. The net charge is calculated using an advanced geochemical model that accounts for the acid–base and metal complexation properties of the organic matter, such as WHAM (Tipping and Woof, 1990), NICA-Donnan (Weng et al., 2002) or SHM (Stockholm Humic Model; Gustafsson, 2001). The model of Tipping and Woof (1990) was slightly modified by Lofts et al. (2001) for the WHAM model and by Lumsdon (2004) for the NICA-Donnan model, and after optimization for individual soils it was found to work well in most cases for predicting the DOC concentration, although difficulties were observed in particular for some mineral soils (c.f. the Discussion section).

A slightly different approach was taken by Weng et al. (2002) who instead related the DOC solubility to the value of a Donnan potential calculated by the NICA-Donnan model, assuming that the Donnan potential was closely related to the surface potential of the humic colloids. These authors found that the magnitude of DOC solubility was related to the Donnan potential in five of six soils, but that acid sandy soils seemed to deviate from the general rule.

If the DOC concentration is related to the net humic charge, it may provide a tool to understand why the DOC concentrations show no trend or decrease in Swedish soil waters simultaneously with acidification recovery.

The aim of this project was to set up a geochemical model to calculate the development of the net humic charge with time at the Swedish integrated monitoring (IM) sites Aneboda and Kindla, to investigate whether the observed DOC solubility in soils is closely related to the humic charge, and if so, use the model to examine the factors influencing the humic charge and thus the DOC solubility. The SHM model was tested on soil water data from one transect along the hillslope in each catchment, covering the time period 1996–2007. The transects extend from recharge to discharge areas, making it possible to estimate the net humic charge in both podzols and riparian soils.

Section snippets

Site descriptions

Locations and maps of the IM sites Aneboda (19.6 ha, N57°05′, E14°32′) and Kindla (19.1 ha, N59°45′, E14°54′) are shown in Fig. 1, Fig. 2, respectively. Both sites are protected Norway spruce (Picea abies) forests, not affected by forestry during the last century (Lundin et al., 2001). The bedrock consists of granite and glacial till is the dominant parent material with quartz and feldspar (albite, plagioclase, and microcline) as the most abundant minerals. At Aneboda, the annual mean temperature

Methods

The soil and surface water samplings are part of the ordinary IM-program (Lundin et al., 2001, http://www.environment.fi/default.asp?node=6329&lan=en). Since 1996, stream water is sampled biweekly at each catchment outlet (Fig. 2). In both catchments, lysimeters (ceramic cups P80, 1 μm cut-off) were installed along a hillslope in 1994 (Fig. 2). In the Aneboda transect, the lysimeters were installed at distances of approximately 1–6 m and 20 m from the stream. Additionally, a group of lysimeters

Model assumptions

The software Visual MINTEQ (Gustafsson, 2009) employing the Stockholm Humic Model (Gustafsson, 2001) was used to study the acid–base and complexation behavior of organic matter in soils. The overall net charge (Z) of the soil organic matter was assumed to influence the DOC mobilization (see Introduction). At each sampling occasion, the pH value as well as the measured dissolved concentrations of ions (Ca2+, Mg2+, SO42−, total Al etc.) was entered as fixed in the model, meaning that the model

Results

The soil solution was highly acidic with pH < 5 and negative ANC at all sites except for in the discharge area at Aneboda and in one of the riparian lysimeters (6203) at Kindla (Table 1a, Table 1b). In both catchments, Na+ was the dominating cation, while Ca2+ and Mg2+ were of the same levels and generally less than half the concentrations of Na+. The Ca2+ and Mg2+ concentrations increased downslope and exhibited soil solution concentrations close to the stream levels in the riparian soils.

Discussion

An objective with this study was to understand the soil water trends in DOC solubility using the Visual MINTEQ geochemical model. Since DOC solubility cannot be simulated directly, the modeled net charge Z was used as a proxy, using the assumption that Z would be directly related to DOC. For individual lysimeters there was indeed a rather strong relationship between these two variables (Fig. 4). Despite this, the model-generated Z values indicated largely unchanged DOC concentrations for the

Conclusions

In summary, this model exercise has highlighted the following factors of importance for the observed decrease of DOC in many lysimeters of the Aneboda and Kindla sites:

  • -

    The increased net charge following the pH increase (because of increased acid dissociation) was in many cases counteracted by a decreased ionic strength, which acted to decrease the net charge and hence the DOC solubility.

  • -

    Particularly at the Kindla site, the pH increase induced reduced solubility of DOC despite no or little

Acknowledgement

This research has been financed by the Swedish Environmental Protection Agency.

References (36)

  • J.J.C. Dawson et al.

    Is the composition of dissolved organic carbon changing in upland acidic streams?

    Environ Sci Technol

    (2009)
  • M. Erlandsson et al.

    Thirty-five years of synchrony in the organic matter concentrations of Swedish rivers explained by variation in flow and sulphate

    Global Change Biol

    (2008)
  • Gustafsson JP. Visual MINTEQ ver. 3.0. Web: http://www.lwr.kth.se/English/OurSoftware/vminteq/index.htm. Last accessed...
  • J.P. Gustafsson et al.

    Modeling salt-dependent proton binding by organic soils with the MICA-Donnan and Stockholm Humic models

    Environ Sci Technol

    (2005)
  • J.P. Gustafsson et al.

    Cation binding in a mor layer: batch experiments and modelling

    Eur J Soil Sci

    (2003)
  • J.P. Gustafsson et al.

    Aluminium solubility mechanisms in moderately acid Bs horizons of podzolized soils

    Eur J Soil Sci

    (2001)
  • J.P. Gustafsson et al.

    Modeling metal binding to soils: the role of natural organic matter

    Environ Sci Technol

    (2003)
  • D.R. Helsel et al.

    Statistical measures in water research

    (1992)
  • Cited by (34)

    View all citing articles on Scopus
    View full text