Chloride Retention and Release in a Boreal Forest Soil: Effects of Soil

Mar 22, 2006 - The latter of these processes leads to the loss of soil water Cl- due to the biological transformation of Cl- to other chlorine species...
3 downloads 8 Views 210KB Size
Environ. Sci. Technol. 2006, 40, 2977-2982

Chloride Retention and Release in a Boreal Forest Soil: Effects of Soil Water Residence Time and Nitrogen and Chloride Loads D A V I D B A S T V I K E N , * , † P E R S A N D EÄ N , ‡ TERESIA SVENSSON,‡ C A R I N A S T A° H L B E R G , § MALIN MAGOUNAKIS,‡ AND GUNILLA O ¨ BERG‡ Department of Geology and Geochemistry, Stockholm University, 106 91 Stockholm, Sweden, Swedish Institute of Climate Science and Policy Research, Linko¨ping University, 601 74 Norrko¨ping, Sweden, and Department of Water and Environmental Studies, Linko¨ping University, 581 83 Linko¨ping, Sweden

The common assumption that chloride (Cl-) is conservative in soils and can be used as a groundwater tracer is currently being questioned, and an increasing number of studies indicate that Cl- can be retained in soils. We performed lysimeter experiments with soil from a coniferous forest in southeast Sweden to determine whether pore water residence time and nitrogen and Cl- loads affected Clretention. Over the first 42 days there was a net retention of Cl- with retention rates averaging 3.1 mg Cl- m-2 d-1 (68% of the added Cl- retained over 42 days). Thereafter, a net release of Cl- at similar rates was observed for the remaining experimental period (85 d). Longer soil water residence time and higher Cl- load gave higher initial retention and subsequent release rates than shorter residence time and lower Cl- load did. Nitrogen load did not affect Cl transformation rates. This study indicates that simultaneous retention and release of Cl- can occur in soils, and that rates may be considerable relative to the load. The retention of Cl- observed was probably due to chlorination of soil organic matter or ion exchange. The cause of the shift between net retention and net release is unclear, but we hypothesize that the presence of O2 or the presence of microbially available organic matter regulates Cl- retention and release rates.

Introduction Chloride (Cl-) is generally considered to be freely mobile in soils in the sense that Cl- is not retained by soil particles and that Cl- represents a good tracer of soil water and groundwater movement (e.g., refs 1, 2). This conservative nature of Cl- in soils is now being questioned (e.g., refs 3, 4). Field studies and modeling indicate that significant amounts of Cl- are retained in the catchment area (3, 5-10). Various * Corresponding author phone: +46 8 6477324; fax: +46 8 6477855; e-mail: [email protected]. † Stockholm University. ‡ Swedish Institute of Climate Science and Policy Research, Linko¨ping University. § Department of Water and Environmental Studies, Linko ¨ ping University. 10.1021/es0523237 CCC: $33.50 Published on Web 03/22/2006

 2006 American Chemical Society

explanations have been suggested to explain the unexpected loss of Cl-, such as vertical transport patterns (9), geochemical sorption including ion exchange (7, 11), Cl- uptake by vegetation (3, 5), and microbial chlorination of soil organic matter (SOM) (12). The latter of these processes leads to the loss of soil water Cl- due to the biological transformation of Cl- to other chlorine species, while the former processes remove Cl- from soil pore water without any transformation. In this paper, “Cl- retention” denotes the loss of Cl- from soil pore water regardless of the processes responsible for this. Experimental evidence supports the hypothesis that Clis nonconservative in soil by confirming that at least four different processes may cause Cl- retention in soil; however, the relative contribution of each of these processes is unclear. For obvious reasons, geochemical sorption and the uptake of Cl- by vegetation may take place in terrestrial ecosystems. The initially controversial idea of microbial chlorination of SOM is by now well documented (e.g., refs 4, 13-17). In addition, there is some experimental evidence of abiotic chlorination of organic matter (18). Furthermore, there are indications not only that soil retains Cl- but also that soilbound Cl- may occasionally be released from soil as well (in addition to the weathering of rocks, ref 19). Hence, it seems that soil sometimes acts as a sink of Cl- and sometimes acts as a source. However, it is unclear under what conditions soil acts as a Cl- sink and under what conditions it acts as a source (see ref 19), although there are indications that several environmental factors may have an influence (e.g., ref 13; see below). The current study investigates the influence of water residence time as well as nitrogen and Cl- load on Clretention in intact soil cores. The water residence time constrains the time frames of possible reactions between Cl- and soil particles, and there are previous indications that the residence time and/or Cl- load may influence the retention of Cl- in soil (14, 20). Nitrogen load was chosen because previous studies have indicated that increasing nitrogen loads may decrease Cl- retention in soil (13).

Materials and Methods Soil was collected at the Stubbetorp catchment (58°,44′ N, 16°,21′ E) in southeast Sweden in late June 2001. The catchment is covered with coniferous forest dominated by Norwegian spruce (Picea abies) and pine (Pinus sylvestris). The catchment area is 0.87 km2 and has a broken topography and a bedrock that is poor in Cl- (21). The long-term annual mean precipitation in the region is approximately 600 mm, and the annual mean temperature is approximately 6 °C. Undisturbed soil cores (n ) 24, height 15 cm) were collected within a sampling area of 2.5 m2 by pressing polypropylene cylinders into the soil. The soil was of spodosol type and consisted of a thick organic layer (O and A horizons; together, up to 15 cm thick) and to varying extents parts of the E horizon as well. The sampling area was situated in a discharge area of the catchment, where the average depth to groundwater is less than 0.5 m, to minimize the change in hydrology during the experiment. The cylinders had a cross-sectional area of 80 cm2 and were equipped with a stainless steel tip to facilitate penetration of the root system. After the materials were transported to the laboratory, zero tension lysimeters were constructed according to Rodstedth et al. (19) by mounting the soil cores on polypropylene bases covered by PE microfilters (43 µm average pore size). A stainless steel pipe was placed horizontally through each base to allow drainage of the lysimeter. An approximately VOL. 40, NO. 9, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2977

20-cm-long silicon tube (6 mm inner diameter) attached to each steel pipe and leading to a 250-mL E-flask placed below the base of the lysimeters allowed collection of the leachate throughout the incubation period. The flask opening was covered with plastic film to prevent evaporation of water from the leachate. Throughout the incubation period, the lysimeters were irrigated with treatment-specific artificial rain twice a week. The rain was gently poured onto the lysimeters, and care was taken to distribute the water over the whole soil surface. The rain for all treatments contained SO42-, Ca2+, Mg2+, Na+, K+, and H+ corresponding to concentrations in precipitation in the Stubbetorp area (22). Treatments represented high and low levels of water, nitrogen (added as NH4NO3), and Cl- (added as NaCl) load according to a factorial design (Table S1; see Supporting Information). This resulted in eight different treatments including all possible combinations of high and low levels of each factor. For each treatment there were three replicates. The two levels of water load, represented by total water additions throughout the incubation period of 1344 and 4032 mL/lysimeter, respectively, correspond to annual precipitation amounts of 483 and 1449 mm. These amounts in turn correspond to the different precipitation levels on the east and the west coasts of southern Sweden. Since the water could drain freely from lysimeters, the different water loads primarily affected the water residence time. Therefore, the water load treatment is synonymous with a residence time treatment in this study. The two levels of nitrogen load were 1.6 and 5.7 mg N lysimeter-1 (579 and 1931 mg m-2 yr-1). The lower level represents the average inorganic nitrogen deposition in the Stubbetorp area, while the higher level corresponds to the deposition on the west coast of southern Sweden (23). Cl- was provided in the artificial rain in total amounts of 4.0 and 12.1 mg Cl- lysimeter-1 (1449 and 4346 mg m-2 yr-1). The higher value corresponds to the load along the west coast of southern Sweden, while the lower value corresponds to the moderate load in the Stubbetorp area (22). Prior to incubation, the soil was saturated to field capacity by adding artificial rain and allowing excess water to drain before the experiment started. The drained excess water volume was greater than or similar to the original water volume in the lysimeters, indicating that most of the original water was replaced by the artificial rain during the preexperimental establishment of field capacity. The lysimeters were incubated in a dark climate chamber (10 °C, ∼90% humidity) from June 25 until October 29, 2001 (127 days). The leachate was collected weekly and analyzed for Clconcentration by means of potentiometric titration according to the method described in Rodstedth et al. (19). The pH of each leachate was also measured immediately upon collection. Soil water content and soil dry weight were determined after incubation by weighing each lysimeter before and after drying at 70 °C for 4 d. Organic matter content was measured as loss of ignition (LOI) at 550 °C for 8 h. Prior to combustion, the entire content of each soil core was milled to 0.5 mm particle diameter and thoroughly mixed before subportions were used for determining LOI. Statistics. In comparing treatments to evaluate the influence of water residence time and of Cl- and nitrogen loads, we used two-way ANOVA in line with the factorial design of the experiment (24). For tests of correlations or linear relationships, we used Kendall’s τ and the KendallTheil robust line tests to avoid producing biased results because of single outliers or extreme values in the data set (25). The significance level chosen when interpreting results was 0.05. 2978

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 9, 2006

FIGURE 1. Lysimeter water balance in the low and high water load treatments, respectively, during the experiment (i.e., after field capacity had been established). The numbers for irrigation, evaporation, and leachate represent the total water flux during the 127 day experiment period. The standing stock of water varied between lysimeters but remained constant over time in each specific lysimeter throughout the whole experiment. See text for details.

Results Soil Characteristics. The organic matter content (LOI) ranged between 11% and 80% depending on the proportions of organic layer and E-horizon soil in the lysimeter. The pHH2O in the soil was 3.46-4.03. The final water content of the lysimeters ranged from 40% to 78%, with an average of 57%. Hence, since the volume of each core was approximately equal, the soil dry weight varied between the cores, with minimum, maximum, and average dry weights of 197, 893, and 513 g, respectively. Water Balance. The low-water lysimeters were irrigated with a total of 1344 mL of artificial rain (after the field capacity was reached; Figure 1) whereas the high-water lysimeters were irrigated with 4032 mL of artificial rain. The total amount of leachate leaving the low-water lysimeters ranged from approximately 941 to 1105 mL, whereas the total amount of leachate from the high-water lysimeters ranged from 3685 to 3812 mL. Hence, the difference between the amount of water added and the amount of water leaving the lysimeters in both treatments was approximately 300 mL (low water, 239-403 mL; high water, 220-347 mL). The leachate volume corresponded to 70-82% and 91-95% of the total amount of water added to the low-water and high-water lysimeters, respectively. The standing stock of water in the lysimeters at any given time ranged from 522 to 764 mL, and the soil cores did not dry out between the rain events. Accordingly, the total weight of the lysimeters remained fairly constant throughout the incubations. The difference between final and initial weight of each lysimeter indicates an average weight increase of 4.70 g (