Published January 11, 2016

Journal of Environmental Quality

TECHNICAL REPORTS Ecosystem Restoration

Magnesium Isotope Variations to Trace Liming Input to Terrestrial Ecosystems: A Case Study in the Vosges Mountains B. Emile Bolou-Bi, Etienne Dambrine, Nicolas Angeli, Benoît Pollier, Claude Nys, François Guerold, and Arnaud Legout*

I

n northern and central European forests, large-scale

Abstract

terrestrial liming with calcium (Ca) and magnesium (Mg) carbonates was initiated in the 1980s to fight soil and stream acidity and to improve forest nutrition and stream biodiversity. While acid deposition has decreased, Mg and Ca deposition has also decreased (e.g., Hedin and Likens, 1996; Vuorenmaa, 2004). On the other hand, the demand for bio-energy and the use of “environmental friendly” materials has strongly increased supporting short rotation sylviculture and whole-tree harvesting (Ericsson, 2004; Puech, 2009). These new pressures may significantly affect soil Ca and Mg stores especially in forest ecosystem where the fertility level is originally low. Liming is presently used in that case to maintain or improve forest soil fertility, at least in eastern France and Germany. Although the positive effects of liming on ecosystem functioning are recognized, the processes involved are complex and partly misunderstood. Liming generally improves the forest nutrition in Ca and Mg (Bonneau et al., 1992; Huettl and Zoettl, 1993) by (i) delivering plant-available Ca and Mg (Wilmot et al., 1996; Moore et al., 2000; Bridgham and Richardson, 2003), and/or (ii) because liming improves humus mineralization (Bridgham and Richardson, 2003; Piatek et al., 2009), which increases the nutrient concentration of soil solution, and/or (iii) because liming raises pH and base saturation and decreases Al toxicity to roots (Cronan et al., 1989; De Wit et al., 2010). The rate and depth of Ca and Mg penetration in soil depends on lime solubility (Hindar, 2005; Westling and Zetterberg, 2007), on climate, and on physical, chemical, and biological soil properties, which are in turn influenced by liming. At the catchment level, the effects of liming on stream water chemistry vary with the proportion of catchment limed (Dalziel et al., 1994; Hindar et al., 2003), soil acidity and buffering capacity (Hindar, 2005; Lofgren et al., 2009), and catchment hydrology (Brahmer, 1994).

Liming with Ca and Mg carbonates is commonly used to reduce soil and stream acidity and to improve vegetation growth and nutrition in forests. Ten years ago, dolomite lime was experimentally applied to a forest catchment on granite in the Vosges Mountains (northeast France), which is characterized by acid soils and drained by an acid stream. The average Mg isotope composition of the dolomite lime (−1.75‰) was low compared with that of tree foliage (−0.70‰), granite and deep soil layers (−0.40‰), and stream water (−0.80‰) in the control catchment. After liming, the exchangeable Mg concentrations in surface soil layers, which were initially very low, increased, and the Mg isotope composition decreased (up to −0.60‰). The decrease was smaller in deeper layers but not in proportion to the increase in exchangeable Mg content, suggesting contributions from mineralization of organic matter and/or displacement of exchangeable Mg from surface layers. Before application, Mg concentration in beech and fir leaves was low, and that of 1-yrold fir needles was lower than that in current needles. Internal Mg translocation within fir needles also resulted in a lower d26Mg of older needles. Three years after dolomite application, the Mg isotope composition of plant leaves was lower than that in the control catchment; this decrease (up to −1.00‰) was attributed to direct uptake of Mg from dissolving dolomite. Liming doubled the concentration of Mg in the stream, whereas the Mg isotope composition decreased correspondingly from −0.80 to −1.20‰, indicating a fast transfer of dolomite Mg to the stream. Our findings indicate that monitoring of d26Mg may be a promising tool to study the fate of dolomitic inputs in terrestrial and aquatic ecosystems.

Core Ideas • Ten years ago, dolomite lime was applied to a forested catchment on granite. • Liming induced an increase of Mg content in tree leaves, soil pools, and stream water. • Liming induced a decrease of δ26 Mg in the ecosystem compartments (soil, tree, stream). • Monitoring of Mg isotope variations appears to be an efficient tracer of dolomite inputs.

B.E. Bolou-Bi, B. Pollier, C. Nys, and A. Legout, INRA, UR 1138 Biogéochimie des Ecosystèmes Forestiers (BEF), Champenoux, 54280 Nancy, France; E. Dambrine, UMR 042 Carrtel INRA-Université de Savoie Centre Interdisciplinaire Scientifique de la Montagne, Belledonne 226, 73376 Le Bourget-du-Lac Cedex, France; N. Angeli, INRA, UMR 1137 INRA-UdL Ecologie et Ecophysiologie Forestières; Plateforme Technique d’Ecologie Fonctionnelle (OC 081), Champenoux, 54280 Nancy, France; F. Guerold, Université de Lorraine, CNRS UMR 7360, Laboratoire Interdisciplinaire des Environnements Continentaux (LIEC), Campus Bridoux, Rue du Général Delestraint, 57070 Metz, France; B.E. Bolou-Bi, current address: Université Paris-Est Créteil Val de Marne, IEES- Institut d’Ecologie et des Sciences de l’Environnement de Paris, Département Sol et Eau, Equipe DIIM- Diversité des ingénieurs et interactions microbiennes. Assigned to Associate Editor Matthew Polizzotto.

Copyright © 2015 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. 5585 Guilford Rd., Madison, WI 53711 USA. All rights reserved. J. Environ. Qual. 45:276–284 (2016) doi:10.2134/jeq2015.02.0096 Received 13 Feb. 2015. Accepted 4 Aug. 2015. *Corresponding author ([email protected]).

Abbreviations: DOC, dissolved organic carbon.

276

Hence, predicting the fate and the effects of liming products on forest ecosystems remains rather complicated, and new tracers are welcome. In fact, knowledge about biogeochemical cycling changes induced by liming remains incomplete mainly because pools and fluxes of elements are often studied without considering the partition between lime dissolution products and native elements. Natural Mg isotope variations are currently developed to constrain Mg sources and cycling in terrestrial ecosystems (Bolou-Bi et al., 2012; Opfergelt et al., 2012, 2014; Mavromatis et al., 2014) and summarized by Schmitt et al. (2012). Because Mg isotope compositions of carbonate rocks (from −5.00 to −1.00‰), including dolomite (−2.50 to −1.00‰) (Galy et al., 2002; Brenot et al., 2008; Jacobson et al., 2010) are lower than that of silicate rocks (−0.75 to +0.40‰) (Huang et al., 2009; Bolou-Bi et al., 2009; Li et al., 2010; Liu et al., 2010; Ryu et al., 2011), Mg isotope variations induced by liming may be used to investigate the fate of the Mg released in soil. Thus, this study aims (i) to assess the effects of dolomite lime application on soil exchangeable Mg, tree Mg nutrition, and stream water Mg and (ii) to study Mg isotope variations induced by liming in the same ecosystem compartments. A practical objective of this pilot study was to develop a relevant approach to trace the Mg added with dolomite in acidic catchments.

Materials and Methods Site Description

in relation to the compacted till layer. Before liming operations, soils were mapped at the 1/10,000 scale, and no major difference was found between the two catchments, except that the peat area was larger at the Longfoigneux catchment.

Lime Composition and Liming Operations The Longfoigneux catchment was limed with 2.80 t ha-1 of a fine (5% (Galy et al., 2003), the corresponding sample was passed again through the resin using the last step of separation chemistry to reach a ratio below 5%. Soil, rock, and vegetation samples are systematically passed five times on resin at the last step; for water samples, the first step using the anionic resin was deleted and passed two times on resin at the last step. The purified Mg solutions obtained were evaporated to dryness. Residues were diluted with 0.05 mol L-1 HNO3 at 80 to 100 ppb before introduction into the Nu Instruments MC ICP– MS mass spectrometer located at the Ecole Normale Supérieure (Lyon, France). Magnesium isotope ratios were measured using the standard-sample bracketing technique with DSM3 standard solution (Galy et al., 2003). Data are expressed in d notation as the parts per thousand (‰) deviation from the standard (DSM3): dxMg = {[(xMg/24Mg)sample/(xMg/24Mg)DSM3] - 1} × 103 where x is either mass 26 or 25. The procedure was checked through the reference material analyses. During the chemical Mg isotope purification, seawater (BCR 403) and doped Cambridge-1 standard with various cations were treated as all Journal of Environmental Quality

samples. Seawater yielded d26Mg values of −0.81 ± 0.04‰ (2 SD; n = 4), similar to overall reported values seawater in the literature (−0.83 ± 0.11‰; n > 200) (Ling et al., 2011). Doped Cambridge-Mg yielded d26Mg values of −2.59 ± 0.06‰ (2 SD; n = 4). In addition, during the Mg isotope ratio measurements, pure Cambridge-1 was repeatedly analyzed and yielded −2.61 ± 0.09‰ (2 SD; n = 10). Both results of Cambridge-1 (doped or pure) display a value similar to overall analyzed Cambridge-1 in the literature (−2.63 ± 0.09‰; n = 1425; data compiled from the literature).

Results Lime Products The two liming products, referred to as dolomite 10 and dolomite 35, contained 6 and 21% Mg, respectively (Table 1). Overall, 135 kg ha−1 of Mg was applied as dolomite, with about 105 kg as dolomite 35 (78% of the applied tracer). Because the Mg supplied with the gypsum (CaSO4) and KCl was less than 0.10% of the total, this contribution was neglected. The Mg isotope ratios of dolomites were −1.62 and −2.25‰ for dolomite 35 and dolomite 10, respectively. These values fall within the range of d26Mg values of dolomite (−2.50 to −1.00‰) reported in the literature (i.e., Brenot et al., 2008; Jacobson et al., 2010). The average weighted isotope composition of the Mg added with dolomite was −1.75‰.

Soils Granite and Bulk Soil The chemical properties of granite and soils in the control and limed catchments are presented in Table 1. The average granite contained 0.79% MgO, close to that (0.90%) reported by Nedeltcheva et al. (2006) for this granite. Soil MgO contents and variations in relation to depth were very close for both catchments, with a continuous decrease from 0.57 ± 0.09% in C horizons to 0.25 ± 0.01% in Ah horizons. This trend was similar for other base cations. The Mg isotope ratio of granite was −0.42‰, which falls in the range of granitic rocks analyzed (Shen et al., 2009). Bulk soils developed on both catchments showed similar variations of

d26Mg. In addition, the d26Mg of deep horizons (Bs2 and C) were close to d26Mg of the granite. The d26Mg of Bs1 horizons was the highest (d26Mg values of −0.15 and −0.12‰ for the control and limed catchment, respectively) for both soil profiles. The eluvial E horizon and Ah horizons displayed average bulk soil d26Mg values of −0.24 and −0.19‰, respectively, for both soil profiles.

Soil Exchangeable Pools At the catchment scale (six plots average), soil exchangeable Mg contents before liming (from about 0.02 cmol+ kg−1 in the C horizon to 0.14 cmol+ kg−1 in the Ah horizon) were very low but were close in both catchment soils (Fig. 2). After liming, soil exchangeable Mg contents increased at all depths and reached 0.72 cmol+ kg−1 in the Ah and 0.16 cmol+ kg−1 in the C horizons in 2006. At the plot scale (A4 and T5), soil exchangeable Mg also increased from 2006 to 2010 down to the Bs1 horizon (Fig. 3A). In the control soil (T5), the d26Mg exch ranged from −0.78‰ in the C horizon to −0.60‰ in the Ah horizon with much lower values around −1.00‰ in the lower Bs2 horizon in 2006 and 2010. These d26Mg exch values are lower that the bulk soil and granite d26Mg values (Fig. 3B), as already observed by Bolou-Bi et al. (2012) and Opfergelt et al. (2014) for soils developed on sandstone and basalt, respectively. After liming, the d26Mg exch in the Ah horizon decreased by −0.60‰ compared with the control soil (Fig. 3B). In the E horizon, the d26Mg exch decreased progressively by −0.40 and −0.60‰ in 2006 and 2010, respectively, compared with the control soil. For the Bs1 horizon, d26Mg exch was not significantly different from the control soil in 2006 but decreased by −0.30‰ in 2010. In the deepest soil horizons (Bs2 and C), d26Mg exch values (−0.83‰) remained close to the control soil d26Mg exch values (−0.87‰).

Tree Leaves Before liming (2001–2002), the average values of Mg content of fir needles in the control catchment were higher than that in the limed catchment, whereas the Mg content in beech leaves was similar for both catchments (Fig. 4). After liming, in 2006 and 2010 for fir, but only in 2010 for beech, Mg contents

Table 1. Selected chemical properties of bulk soil horizons at control and limed plots, granite and dolomites, and respective Mg isotope compositions. Code Control T5 Control T5 Control T5 Control T5 Control T5 Limed A4 Limed A4 Limed A4 Limed A4 Limed A4 Ganite “Ventron” Dolomite D10† Dolomite D35†

Horizon

Depth

Ah E Bs1 Bs2 C Ah E Bs1 Bs2 C – – –

cm −15 −30 −45 −60 −100 −15 −30 −45 −60 −100 – – –

MgO

CaO

Na2O

K2O

—————————————— % —————————————— 0.24 0.21 1.64 4.33 0.26 0.23 1.87 4.38 0.31 0.26 2.04 4.44 0.52 0.29 2.37 4.36 0.63 0.32 2.61 5.19 0.26 0.26 1.75 3.38 0.26 0.19 1.92 3.82 0.32 0.19 2.09 3.91 0.46 0.21 2.42 3.95 0.50 0.30 3.20 4.36 0.79 1.12 3.76 4.53 10 – – – 35 – – –

d26Mg (±0.10‰)

d25Mg (±0.05‰)

−0.18 −0.24 −0.12 −0.43 −0.35 −0.20 −0.24 −0.15 −0.38 −0.46 −0.42 −2.25 −1.62

−0.09 −0.11 −0.07 −0.22 −0.19 −0.11 −0.12 −0.09 −0.17 −0.24 −0.21 −1.17 −0.84

† From Black Forest, Germany. Journal of Environmental Quality 279

Stream Water Before liming, the stream water of both catchments had low and similar Mg concentrations (Fig. 6; Table 2). Compared with the control, Mg concentration peaked just after dolomite application and decreased strongly in the following year to reach a level approximately twice as high as in the control since 2005. Streams were not gauged during this period of the study, but we used a gauging station (Rupt sur Moselle) 20 km downstream the same river to assess the effects of hydrology on d26Mg. The set of sampling dates includes two very low-flow (0.5 m3 s−1) periods (27 July 2010 and 30 May 2011), a summer high flow (5.5 m3 s−1) event (24 Aug. 2010), and two medium-flow periods (Table 2). No winter high flow (which rises above 20 m3 s−1 at Rupt sur Moselle) peak was sampled. The average d26Mg of the stream in the control catchment (−0.81‰) Fig. 2. Range of soil exchangeable Mg contents before liming (2002) in six plots in was close to the exchangeable d26Mg values of deep the control and limed catchments (light gray); after liming (2006) in six plots of the limed catchment (dark grey). soil horizons (−0.87%) (Fig. 7). In 2011–2012, the Mg isotope ratio of the stream in the limed catchbecame higher in the limed catchment (Fig. 4, 5A). As usually ment was about −0.40‰ lower compared with the control seen (Le Goaster et al., 1991), second-year needles as well as stream (Fig. 7). At both control and limed catchments, stream litterfall were depleted in Mg compared with the newly grown d26Mg was higher in summer than in spring and peaked on organs. 24 August, during a summer flood event, in relation to an Since 2006, d26Mg values of fir needles and beech leaves increase in DOC (Table 2). In addition, variations of d26Mg at the limed plot were lower than vegetation in the control were strongly and positively related to DOC. plot, and the difference in Mg content was larger in 2006 than 26 in 2010 (Fig. 5B). Interestingly, a large difference in d Mg between beech leaves in the limed and control plots occurred in 2006, whereas their Mg content was not different. In fir, Recovery of the Dolomite Mg Applied second year-needles as well as litterfall were depleted in heavy at the Catchment Scale Mg compared with the newly grown organs. Hence, translocaCompared with the exchangeable Mg stored in soil (40–60 tion of Mg from old needles or leaves supplied preferentially −1 kg ha ) before liming (Angéli, 2006), the amount of Mg added heavy Mg to newly grown organs or reserve tissues, as observed with dolomite was 135 kg ha−1. In 2004–2005, Mg drainage was by Bolou-Bi et al. (2012). −1 1.70 kg ha yr−1 from the control catchment (Angéli, 2006). During the two first years after liming, 8.90 kg ha−1 yr−1 Mg was

Discussion

Fig. 3. Soil exchangeable Mg content (A) and d26Mg variation (B) with soil depth in control (C) and limed (L) plots in 2006 and 2010. 280

Journal of Environmental Quality

Fig. 4. Range of Mg contents in leaves and needles in six plots of the control and limed catchments. Fir needles and beech leaves are indicated in light gray and black, respectively. Samples were collected in 2002, 2006, and 2010; values for 2001, 2005, and 2009 are related to second-year needles.

drained from the limed catchment. The additional output due to liming was thus 14.40 kg ha−1 Mg for the 2004–2005 period. Water fluxes were not monitored after 2005. Taking into account that, since 2005, Mg concentrations were almost twice higher in the limed stream than in the control, an additional output of about 2.00 kg ha−1 yr−1 Mg due to liming can be estimated. This drainage represents 10 kg ha−1 for the 2006–2010 period. According to our calculations, this drainage was thus about 25 kg ha−1 Mg from 2003 to 2010, which represents less than 20% of the liming input. The remaining Mg should thus be located mainly on the soil cation exchange capacity and immobilized in trees. In 2010, exchangeable Mg pools for the 0- to 60-cm layers were 48 kg ha−1 in the control catchment and 137 kg ha−1 in the limed catchment; this difference represents 65% of the liming input. Trees should have taken up about 3 kg ha−1 yr−1 of Mg (Van der Heijden et al., 2013), which represents 20 kg ha−1 of Mg from 2004 to 2010 (about 5% of the liming input). According to this budget, the Mg taken up by trees should mainly originate from dolomite.

Mg Isotope Variations to Assess the Dolomite-Liming Effect on Forest Ecosystem Compared with the traditional approach where the effects of liming are measured by assessing the change in Mg and

Fig. 5. Magnesium content (A) and Mg isotope composition (d26Mg) (B) of fir needles (first and second year), beech leaves, and of mixed litterfall at control (white symbols) and limed (black symbols) plots (A4 and T5). Samples were collected in 2002, 2006, and 2010; values for 2001, 2005, and 2009 are related to second-year needles. The d26Mg of leaves collected in 2002 was not analyzed. Magnesium concentration uncertainties were ±5% in 0.05 mol L−1 HNO3.

Ca concentrations and stores in the ecosystem, the isotopic approach provides additional information about the fate of the Mg released by the dolomite. The average mass weighted Mg isotopic composition was −1.75‰, and that of the most soluble dolomite 10 was −2.25‰. During the first year after dolomite application, the isotope ratio of the Mg released by the dolomite probably falls into that range but closer to the most soluble dolomite. After 7 yr, we assume that all the dolomite powder was dissolved because of the high acidity of soils and large rainfall amount in these catchments. Because the leaching of dolomite does not induce Mg isotope fractionation (Brenot et al., 2008), the mean isotopic composition of the dolomite may be used for estimation of Mg fluxes derived from the dolomite in different compartments of the ecosystem.

Dolomite Lime Contribution to Exchangeable Soil Pools As already described in other studies, liming progressively increased soil exchangeable Mg from surface to depth

Journal of Environmental Quality 281

soil organic matter mineralization. This possible contribution would be in agreement with an observed decrease in the thickness of the O horizon after 10 yr, as reported by Kreutzer (1995). Conversely, a contribution of litter Mg (d26Mg = −1.21 and −1.91‰ for control and limed catchment, respectively) (Fig. 5B) is also unlikely, unless a strong fractionation occurs on mineralization. At depths deeper than 45 cm, a direct contribution of dolomite Mg is quite unlikely because exchangeable Mg strongly increased (Fig. 3), whereas no change in d26Mg occurred. This suggests that the change in exchangeable Mg contents in these soil horizons resulted from the displacement of exchangeable Mg from upper horizons.

Dolomite Lime Contribution to Tree Nutrition For trees, the difference in Mg content of leaves and needles, before and after liming, indicate that dolomite application improved the Mg status of trees, as observed elsewhere (Wilmot et al., 1996; Moore et al., 2000; Bridgham and Richardson, 2003). The decrease in d26Mg leaf demonstrates that leaf Mg was taken up preferentially from the dissolving dolomite rather than from soil exchangeable Mg Fig. 6. Time variation (2002–2010) of magnesium concentrations (mmol L−1) in the stream (Opfergelt et al., 2014). This result is also quite water of control and limed catchments. Magnesium concentration uncertainties were ±1% interesting in light of the results of Van der in H2O. LF, Longfoigneux catchment; WS, Wassongoutte catchment. Heijden et al. (2013), who showed no change (Belkacem and Nys, 1997; Ranger et al., 1994). From the in the d26Mg leaf in the 2 yr after labeling with change in exchangeable Mg stored and isotopic composition 26 Mg of the exchangeable Mg store of the humus layer of a very in each soil horizon between 2002 and 2010, and assuming acid soil. It suggests that the translocation of Mg from soil to that the change in d26Mg was only related to dolomite Mg, leaf may be reduced when the soil-available Mg level is too low. we can compute theoretical values of d26Mg exch of −1.48 and Translocation of Mg from old needles of fir seems to be lower −0.95‰ for the Ah and E horizons in 2006 (against meain 2010 in the limed catchment than in the control catchment, sured values of −1.28 and −0.91‰, respectively) and −1.53, according to leaf Mg content and d26Mg. −1.30, and −1.38‰ for the Ah, E, and Bs horizons in 2010, Dolomite Lime Contribution to the Stream respectively (against measured values of −1.28, −1.21, and Liming doubled the Mg concentration in the stream, as −1.00‰, respectively). Taking into account the spatial variobserved by others (Hindar et al., 2003; Clair and Hindar, ability of soil exchangeable Mg and the measurement uncer2005), and decreased the Mg isotope ratio by about −0.40‰. tainties, the theoretical d26Mg fit well the respective measured The control stream had d26Mg values ranging from −0.93 to values. The disagreement in the Ah horizon may derive from Table 2. Major element concentrations and Mg isotope compositions of stream water samples from the control and limed catchments in 2010–2011. Stream

Date

Flow†

27 July 2010 24 Aug. 2010 21 Sept. 2010 05 Apr. 2011 30 May 2011 27 July 2010 24 Aug. 2010 21 Sept. 2010 5 Apr. 2011 30 May 2011

m s 0.5 5.5 2.0 3.3 0.5 0.5 5.5 2.0 3.3 0.5 3

Control Control Control Control Control Limed Limed Limed Limed Limed

pH

−1

Mg

Ca

Si

Al

S-SO4

DOC‡

——————————— mmol L ——————————— 8.64 11.98 123.55 4.82 24.23 7.82 12.97 85.81 8.89 19.12 7.41 10.98 97.20 3.71 24.91 5.76 9.48 79.04 4.08 23.76 6.99 6.74 118.56 3.34 25.65 13.99 17.96 120.34 2.22 29.48 13.17 15.22 78.69 5.19 24.06 13.99 16.72 100.05 1.85 28.74 14.81 14.72 77.62 1.85 30.69 17.69 20.21 115.00 2.22 40.54

ppm 3.65 6.42 3.27 3.05 2.6 2.52 4.24 2.54 2.37 2.27

−1

5.5 5.34 5.82 5.58 5.46 5.61 5.7 5.97 6.09 6.62

d25Mg d26Mg ——— ‰ ——— −0.75 −0.39 −0.68 −0.34 −0.78 −0.41 −0.93 −0.49 −0.93 −0.48 −1.19 −0.6 −0.99 −0.52 −1.22 −0.63 −1.31 −0.68 −1.32 −0.63

† The water flow data were measured at the gauging station of Rupt sur Moselle, 20 km downstream from the catchments. Measurement uncertainties (2 SD) of d26Mg and d25Mg were 0.10 and 0.05‰, respectively. ‡ Dissolved organic carbon. 282

Journal of Environmental Quality

−0.68‰, falling in the range of soil exchangeable d26Mg values. The average ratio (−0.81‰) was close to that found in rivers draining silicate catchments (Brenot et al., 2008; Tipper et al., 2012; Bolou-Bi et al., 2012). If the Mg in excess in the limed stream was derived from the dolomite (average d26Mg, −1.75‰), the limed stream d26Mg should vary between −1.13 and −1.42‰, which is in good agreement with the measured values. Bolou-Bi et al. (2012) observed, on a sandstone catchment in the same region, a decrease in d26Mgwhen flow increased, which is not the case here. stream The relation between d26Mg stream and DOC suggests that the main source of Mg at high flow in summer stands in surface horizons, which release DOC, whereas deep horizons contribute more at low flow (Bolou-Bi et al., 2012; Mavromatis et al., 2014). However, the fact that d26Mg increases during the summer high flow is striking because the surface horizons of the limed soil studied are more depleted in d26Mg (−1.20‰) compared with deep soil horizons. There are two possible explanations. (i) Stream water partly derives from wet peaty soils along the stream. Organic matter from the peaty zones may control the stream water base cation chemistry Fig. 7. Variation (2002–2010) of d26Mg (‰) in the stream water of control and limed (Ledesma et al., 2013). These wet peaty soils could catchments. LF, Longfoigneux catchment; WS, Wassongoutte catchment. act as a cation exchanger that preferentially retains des Forêts (ONF) for their contribution to this study; and three the heavy Mg isotope. At high flow, the provision of dilute anonymous reviewers and the associate editor who helped improve this manuscript. The BEF unit is supported by the French National Research water could desorb the heavy Mg isotopes bound to organic Agency through the Laboratory of Excellence ARBRE (ANR-12matter and could drain it with DOC, as suggested by the relaLABXARBRE-01). tion with Mg isotopes. (ii) The spatial variability at the catchment scale may explain our results. Only one soil profile was References studied with isotopic tools, which may not be representative Angéli, N. 2006. Evolution de la composition chimique des ruisseaux vosgiens: for the catchment at different hydrological stages, especially Analyse rétrospective et effet d’un amendement calco-magnésien. Thèse de when lime is applied. Doctorat, Université Henri Poincaré, Nancy, France.

Conclusions Dolomite lime application induced a fast and long-lasting increase of Mg content and decreased d26Mg in tree leaves on soil exchangeable pools and in stream water. The change in d26Mg was mostly in agreement with a major contribution of Mg from the dolomite, except in deep horizons where displacement of exchangeable Mg from upper layers was probably the main source of the additional Mg. Measurement of the Mg isotope ratio improves the understanding of Mg routes throughout the ecosystem, but care should be taken in sampling soil spatial variability in relation to hydrology and all major ecosystem compartments. In addition, Mg isotope fractionation by organic matter binding and mineralization should be further studied to use this tool quantitatively. The coupling of the Ca and Mg isotopes could also be an interesting way, in the future, to better understand the impact of liming on the environment.

Acknowledgments The authors thank Région Lorraine, The Conseil Général des Vosges, the Agence de l’Eau Rhin-Meuse (AERM), the Zone Atelier Moselle (ZAM), and the French ANR program (ANR-07-BDIV-007-01 Recover project) for funding the present study; l’Office National

Belkacem, S., and C. Nys. 1997. Effects of liming and gypsum regimes on chemical characteristics of an acid forest soil and its leachates. Ann. Sci. For. 54:169–180. doi:10.1051/forest:19970204 Bolou-Bi, E.B., N. Vigier, A. Poszwa, J.-P. Boudot, and E. Dambrine. 2012. Effects of biogeochemical processes on magnesium isotope variations in a forested catchment in the Vosges Mountains France. Geochim. Cosmochim. Acta 87:341–355. doi:10.1016/j.gca.2012.04.005 Bolou-Bi, E.B., N. Vigier, A. Brenot, and A. Poszwa. 2009. Magnesium isotopic composition of plants and rocks reference materials. Geostand Geoanal Res. 33:95–109. doi:10.1111/j.1751-908X.2009.00884.x Bonneau, M., G. Landmann, and M. Adrian. 1992. La fertilisation comme remède au dépérissement des forêts en sol acide: Essais dans les Vosges. Rev. For. Fr. 44:207–223. doi:10.4267/2042/26318 Brahmer, G. 1994. Effects of whole catchment liming and Mg addition on soil water and runoff at two forested watersheds in the Black Forest Germany. For. Ecol. Manage. 68:47–60. doi:10.1016/0378-1127(94)90137-6 Brenot, A., C. Cloquet, N. Vigier, J. Carignan, and C. Francelanord. 2008. Magnesium isotope systematics of the lithologically varied Moselle river basin, France. Geochim. Cosmochim. Acta 72:5070–5089. doi:10.1016/j. gca.2008.07.027 Bridgham, S.D., and C.J. Richardson. 2003. Endogenous versus exogenous nutrient control over decomposition and mineralization in North Carolina peatlands. Biogeochemistry 65:151–178. doi:10.1023/A:1026026212581 Clair, T.A., and A. Hindar. 2005. Liming for the mitigation of acid rain effects in freshwaters: A review of recent results. Environ. Rev. 13:91–128. doi:10.1139/a05-009 Cronan, S., R. April, J. Bartlett, and R. Bloom. 1989. Aluminum toxicity in forests exposed to acidic deposition: The ALBIOS results. Water Air Soil Pollut. 48:181–192. doi:10.1007/BF00282377

Journal of Environmental Quality 283

Dalziel, T.R.K., E.J. Wilson, and M.V. Proctor. 1994. The effectiveness of catchment liming in restoring acid waters at Loch Fleet, Galloway, Scotland. For. Ecol. Manage. 68:107–117. doi:10.1016/0378-1127(94)90142-2 De Wit, A., D. Eldhuset, and J. Mulder. 2010. Dissolved Al reduces Mg uptake in Norway spruce forest: Results from a long-term field manipulation experiment in Norway. For. Ecol. Manage. 259:2072–2082. doi:10.1016/j. foreco.2010.02.018 Ericsson, K. 2004. Bioenergy policy and market development in Finland and Sweden. Energy Policy 32:1707–1721. doi:10.1016/ S0301-4215(03)00161-7 Galy, A., O. Yoffe, and P.E. Janney. 2003. Magnesium isotope heterogeneity of the isotopic standard SRM 980 and new reference materials for magnesium isotope ratio measurements. J. Anal. At. Spectrom. 18:1352– 1356. doi:10.1039/b309273a Galy, A., M. Bar-Matthews, L. Halicz, and R.K. O’Nions. 2002. Mg isotopic composition of carbonate: Insight from speleothem formation. Earth Planet. Sci. Lett. 201:105–115. doi:10.1016/S0012-821X(02)00675-1 Hedin, L.O., and G.E. Likens. 1996. Atmospheric dust and acid rain. Sci. Am. 275:88–92. doi:10.1038/scientificamerican1296-88 Hindar, A., R.F. Wright, P. Nilsen, T. Larssen, and R. Hogberget. 2003. Effects on stream water chemistry and forest vitality after whole-catchment application of dolomite to a forest ecosystem in southern Norway. For. Ecol. Manage. 180:509–525. doi:10.1016/S0378-1127(02)00647-3 Hindar, A. 2005. Whole-catchment application of dolomite to mitigate episodic acidification of streams induced by sea-salt deposition. Sci. Total Environ. 343:35–49. doi:10.1016/j.scitotenv.2004.09.040 Huang, F., J. Glessner, A. Ianno, C. Lundstrom, and Z. Zhang. 2009. Magnesium isotopic composition of igneous rock standards measured by MC-ICPMS. Chem. Geol. 268:15–23. doi:10.1016/j.chemgeo.2009.07.003 Huettl, R.F., and H.W. Zoettl. 1993. Liming as a mitigation tool in Germany’s declining forests-reviewing results from former and recent trials. For. Ecol. Manage. 61:325–338. doi:10.1016/0378-1127(93)90209-6 Jacobson, A.D., Z. Zhang, C. Lundstrom, and F. Huang. 2010. Behavior of Mg isotopes during dedolomitization in the Madison Aquifer, South Dakota. Earth Planet. Sci. Lett. 297:446–452. doi:10.1016/j.epsl.2010.06.038 Kreutzer, K. 1995. Effects of forest liming on soil processes. Plant Soil 168169:447–470. doi:10.1007/BF00029358 Le Goaster, S., E. Dambrine, and J. Ranger. 1991. Croissance et nutrition minérale d’un peuplement d’épicéa sur sol pauvre: I. Evolution de la biomasse et dynamique d’incorporation des éléments minéraux. Acta Ecol. 12:771–789. Ledesma, J.L., T. Grabs, M.N. Futter, K.H. Bishop, H. Laudon, and S.J. Köhler. 2013. Riparian zone control on base cation concentration in boreal streams. Biogeosciences 10:3849–3868. doi:10.5194/bg-10-3849-2013 Li, W.Y., F.Z. Teng, S. Ke, R.L. Rudnick, S. Gao, F.Y. Wu, and B.W. Chappell. 2010. Heterogeneous magnesium isotopic composition of the upper continental crust. Geochim. Cosmochim. Acta 74:6867–6884. doi:10.1016/j.gca.2010.08.030 Ling, M.X., F. Sedaghatpour, F.Z. Teng, P.D. Hays, J. Strauss, and W. Sun. 2011. Homogeneous magnesium isotopic composition of seawater: An excellent geostandard for Mg isotope analysis. Rapid Commun. Mass Spectrom. 25:2828–2836. doi:10.1002/rcm.5172 Liu, S.A., F.Z. Teng, Y. He, S. Ke, and S. Li. 2010. Investigation of magnesium isotope fractionation during granite differentiation: Implication for Mg isotopic composition of the continental crust. Earth Planet. Sci. Lett. 297:646–654. doi:10.1016/j.epsl.2010.07.019 Lofgren, S., N. Cory, T. Zetterberg, P.E. Larsson, and V. Kronna. 2009. The long-term effects of catchment liming and reduced sulphur deposition on forest soils and runoff chemistry in southwest Sweden. For. Ecol. Manage. 258:567–578. doi:10.1016/j.foreco.2009.04.030

284

Mavromatis, V., A.S. Prokushkin, O. Pokrovsky, J. Viers, and O.A. Korets. 2014. Magnesium isotopes in permafrost-dominated Central Siberian larch forest watersheds. Geochim. Cosmochim. Acta 147:76–89. doi:10.1016/j. gca.2014.10.009 Moore, J.D., C. Camireì, and R. Ouimet. 2000. Effects of liming on the nutrition, vigor and growth of sugar maple at the Lake Clair Watershed, Quebec, Canada. Can. J. For. Res. 30:725–732. doi:10.1139/x00-009 Nedeltcheva, T., C. Piedallu, J.C. Gégout, J.M. Stussi, J.P. Boudot, N. Angeli, and E. Dambrine. 2006. Influence of granite mineralogy, rainfall, vegetation and relief on stream water chemistry Vosges mountains, north eastern France. Chem. Geol. 231:1–15. doi:10.1016/j.chemgeo.2005.12.012 Opfergelt, S., K.W. Burton, and R.B. Georg. 2014. Magnesium retention on the soil exchange complex controlling Mg isotope variations in soils, soil solutions and vegetation in volcanic soils, Iceland. Geochim. Cosmochim. Acta 125:110–130. doi:10.1016/j.gca.2013.09.036 Opfergelt, S., R.B. Georg, B. Delvaux, Y.M. Cabidoche, K.W. Burton, and A.N. Halliday. 2012. Mechanisms of magnesium isotope fractionation in volcanic soil weathering sequences, Guadeloupe. Earth Planet. Sci. Lett. 341-344:176–185. doi:10.1016/j.epsl.2012.06.010 Piatek, K.B., P. Munasinghe, W.T. Peterjohn, M.B. Adams, and J.R. Cumming. 2009. Oak contribution to litter nutrient dynamics in an Appalachian forest receiving elevated nitrogen and dolomite. Can. J. For. Res. 39:936– 944. doi:10.1139/X09-028 Puech, J. 2009. Mise en valeur de la forêt française et développement de la filière bois. Rapport remis à Monsieur Nicolas Sarkozy, Président de la République. Ranger, J., A. Mohamed, and D. Gelhaye. 1994. Effet d’un amendement calcomagneìsien associeì ou non aÌ une fertilisation sur le cycle biogeìochimique des eìleìments nutritifs dans une plantation d’Eìpiceìa commun Picea abies Karst: Deìpeìrissante dans les Vosges. Ann. Sci. For. 51:455–475. doi:10.1051/forest:19940503 Ryu, J.S., A.D. Jacobson, C. Holmden, C.C. Lundstrom, and Z. Zhang. 2011. The major ion, d44/40Ca, d44/42Ca, and d26/24Mg geochemistry of granite weathering at pH = 1 and T = 25°C: Power-law processes and the relative reactivity of minerals. Geochim. Cosmochim. Acta 75:6004–6026. doi:10.1016/j.gca.2011.07.025 Schmitt, A.D., N. Vigier, D. Lemarchand, R. Millot, P. Stille, and F. Chabaux. 2012. Processes controlling the stable isotope compositions of Li, B, Mg and Ca in plants, soils and waters: A review. C. R. Geosci. 344:704–722. doi:10.1016/j.crte.2012.10.002 Shen, B., B. Jacobsen, C.T.A. Lee, Q.Z. Yin, and D.M. Morton. 2009. The Mg isotopic systematics of granitoids in continental arcs and implications for the role of chemical weathering in crust formation. Proc. Natl. Acad. Sci. USA 106:20652–20657. doi:10.1073/pnas.0910663106 Tipper, E.T., E. Lemarchand, R.S. Hindshaw, B.C. Reynolds, and B. Bourdon. 2012. Seasonal sensitivity of weathering processes: Hints from magnesium isotopes in a glacial stream. Chem. Geol. 312-313:80–92. doi:10.1016/j. chemgeo.2012.04.002 Vuorenmaa, J. 2004. Long-term changes of acidifying deposition in Finland 19732000. Environ. Pollut. 128:351–362. doi:10.1016/j.envpol.2003.09.014 Van der Heijden, G., A. Legout, A.J. Midwood, C.A. Craig, B. Pollier, J. Ranger, and E. Dambrine. 2013. Mg and Ca root uptake and vertical transfer in soils assessed by an in situ ecosystem-scale multi-isotopic 26Mg & 44Ca tracing experiment in a beech stand Breuil-Chenue, France. Plant Soil 369:33–45. doi:10.1007/s11104-012-1542-7 Westling, O., and T. Zetterberg. 2007. Recovery of acidified streams in forests treated by total catchment liming. Water Air Soil Pollut. Focus 7:347–356. doi:10.1007/s11267-006-9107-5 Wilmot, T.R., D.S. Ellsworth, and M.T. Tyree. 1996. Base cation fertilization and liming effects on nutrition and growth of Vermont sugar maple stands. For. Ecol. Manage. 84:123–134. doi:10.1016/0378-1127(96)03743-7

Journal of Environmental Quality

Magnesium Isotope Variations to Trace Liming Input to Terrestrial Ecosystems: A Case Study in the Vosges Mountains.

Liming with Ca and Mg carbonates is commonly used to reduce soil and stream acidity and to improve vegetation growth and nutrition in forests. Ten yea...
978KB Sizes 0 Downloads 5 Views