Final Report Summary - ISOCRIT (Application of a Novel Magnesium-Lithium Dual Isotopic Tracer to Biogeochemical Cycles in the Soil Critical Zone)
Project Context and Objectives
Magnesium is an essential macronutrient for all organisms, yet soils worldwide exhibit Mg deficiencies (e.g. Hüttl and Schaaf, 1997). In many temperate and tropical forests, Mg is largely stored in the aboveground biomass, rather than in the soil itself (Armbruster et al., 2002); thus Mg cycles in such forests are extremely sensitive to land use and climate change. Losses of soil nutrients, such as Mg, threaten soil productivity in both forested and agricultural ecosystems.
Mangesium also plays a key role in global climate cycles. The weathering of Ca- and Mg- containing silicate rocks is a significant long-term sink for atmospheric CO2 (Berner et al., 1983), and chemical weathering of Ca-Mg silicates is a primary long-term control on this important greenhouse gas. Thus to estimate and predict CO2 consumption rates, it is important to identify and quantify the sources and fluxes of Ca and Mg that are discharged to the world’s oceans from rivers and the weathering processes and mechanisms that produce those fluxes. Chemical weathering also exerts a key control on ocean chemistry, and Mg is a particularly important aspect of seawater chemistry, given that it controls the nature of carbonate precipitation and the fact that Mg/Ca ratios of foraminifera are an important tool in the reconstruction of past ocean temperatures. Despite its importance, there are significant uncertainties associated with the seawater Mg budget through time, including the magnitude and temporal dynamism of the riverine flux. Stable isotopic tracers such as Mg and Li are increasingly being used to identify sources of these constituents in rivers and oceans (e.g. Huh et al., 1998; Tipper et al., 2006). Similarly, stable isotope systems are also used in marine sediments and sedimentary rocks to deduce paleoclimates, ancient ocean chemistry and circulation, and other aspects of Earth’s history (e.g. Vance et al., 2009). However these interpretations and historical reconstructions are entirely dependent on the processes that may fractionate the isotopes during mineral weathering and (bio)geochemical cycling. Therefore a solid understanding of the isotopic signatures of these elements in terrestrial materials and the mechanisms that fractionate the isotopes during the cycling of these elements, is crucial for both interpreting the historical rock record and predicting future CO2 sequestration.
Stable isotope ratios of elements such as Mg in ocean records have been touted as proxies for continental weathering fluxes; however, Mg isotope ratios are known to fractionate during clay mineral precipitation (e.g. Pogge von Strandmann et al., 2008), and uptake by vegetation (Black et al., 2006; Tipper et al., 2010), suggesting that weathering and/or biogeochemical Mg isotope fractionations in soils may be essential for understanding the relationship between riverine Mg isotope ratios and weathering fluxes. Mg takes a circuitous route from bedrock to ocean, passing through dissolved and solid phases as it undergoes geochemical reactions (e.g. clay formation and dissolution, adsorption onto clay minerals, complexation with organic acids), bio(geo)chemical reactions (e.g. uptake by microorganisms and vegetation, decomposition of organic material, intracellular enzymatic processes), and is transported through soil, vegetation, and groundwater. Considering the spatial and temporal overlap of these processes, it is no simple task to identify and quantify them in natural samples in order to trace the Mg cycle through the environment. Isotopic ratios are an increasingly utilized tool for tracing such processes, but as yet, very little is known about the Mg isotopic fractionations these processes may or may not produce. Furthermore, overlapping or sequential processes that produce isotopic fractionations can overprint earlier or weaker isotopic fractionations, resulting in ephemeral or complicated isotopic signatures (e.g. Buss et al., 2010).
In order to interrogate the Mg cycle in the terrestrial environment, I am analysing Mg isotopes alongside isotopes of Sr, H, O, and Li and a suite of elemental ratios. Riverine Li is predominantly derived from silicate mineral weathering and often substitutes for Mg in primary minerals. Like Mg, Li isotopes are fractionated during precipitation of secondary minerals such as clays (Pogge von Strandmann et al., 2008a; Vigier et al., 2008), but in contrast to Mg (an essential macronutrient), Li is not a nutrient element and is not expected to be fractionated by biological processes. Thus, Li isotope data should allow the separation of the biological component from the geological component of the Mg cycle. I am further developing techniques for Mg and Li isotopic analysis of soils, vegetation, and other environmental samples; developing a novel Mg-Li multi-isotopic tracer of biogeochemical processes, testing for Mg and Li isotope fractionation in a suite of natural samples significantly adding to the dataset in the literature, and producing data for biogeochemical cycling models.
As an intensive case study, we are applying the isotopic dual-tracer technique to a large set of natural samples that I have collected previously from the Luquillo Mountains in Puerto Rico, a US National Science Foundation (NSF) Critical Zone Observatory of which I am a founding PI. This dataset will be used to demonstrate the potential of the dual-tracer technique, and to move forward the models of biogeochemical mineral nutrient cycling that I am developing for the Luquillo Critical Zone Observatory (LCZO). Overall, this work will represent a significant advancement in our understanding of the terrestrial Mg biogeochemical cycle, in particular in tropical soils. Specific objectives in this context are:
1. The development of new sample preparation and analytical procedures for Mg and Li isotopic analysis of soil and vegetation.
2. Creation of a soil and vegetation Li- and Mg-isotope dataset.
3. Establishment and testing of a novel dual-isotopic tracer technique for the mechanistic understanding and quantification of the terrestrial Mg cycle.
4. Application of that approach to the Luquillo watershed of Puerto Rico, thereby elucidating pathways of primary and secondary weathering and constraining mechanisms for nutrient export from tropical catchments to the oceans.
Beyond these scientific objectives, the work involves building stronger international collaborations between the US and the EU, specifically between the University of Bristol and my existing collaborators at the US Geological Survey, Penn State University, the University of California at Berkeley, and the University of Pennsylvania, as well as strengthening links between the US NSF Critical Zone Observatory network and the EU SoilTrEC critical zone observatory network. It also fosters increased intra-European linkages via my ongoing collaborations with European colleagues. These knowledge transfer objectives are :
1. To provide soil and chemical weathering expertise to the Bristol Isotope Group by collaborating with them on their research into the use of isotopic proxies for chemical weathering fluxes, which are strongly influenced by weathering processes in soils, of which they have little knowledge.
2. To develop a Mg-Li dual isotopic tracer and protocols for Mg and Li isotope analysis of soils, soil extracts, pore waters, and vegetation, resulting in publications that contribute to the scientific literature as well as directly expanding the capabilities of the Bristol Isotope Group.
3. To speak about my research at European universities and conferences, including at the host institution.
4. To collaborate with researchers within the UK and train postgraduate students in the application of cross-disciplinary approaches to critical zone science.
5. To work more closely with my collaborators in France and the UK on critical zone research.
6. To become more directly involved with the SoilTrEC project’s European soil critical zone observatories (CZOs), strengthening EU links to the US CZO network, by participating in meetings and workshops hosted by both groups and participating in reciprocal CZO site visits.
Results
Soil pore water concentrations of Mg and K that have been corrected for rain input and evapotranspiration indicate that mineral weathering contributes to the pore water solute compositions in four depth profiles along a slope transect (from ridgetop to riparian zone) in the Bisley watershed of the LCZO. Pore water concentrations are relatively invariant with time below 0.5 m depth (2.0 m depth in the deepest profile). Above these depths, Mg concentrations vary in space and time, with all profiles showing overall higher concentrations in the surface layers, suggestive of biological influence. In contrast, delta26Mg values in the deepest profile show a clear trend towards heavier delta26Mg with increasing depth (-0.83‰ to -0.18‰) suggesting mixing between atmospheric Mg at the surface and a dissolving, isotopically heavier phase at depth. An excursion towards heavier delta26Mg at the soil-saprolite interface (-0.7‰, ca. 1m depth) indicates a change in controlling processes. A similar heavy excursion is present at the same depth all four profiles. This depth coincides with the transition from “true soil” to saprolite and is independent of trends in elemental concentrations, indicating that a fundamental change in the Mg cycle occurs at the soil-saprolite transition. This result demonstrates that Mg isotope ratios can trace biogeochemical soil processes that may not be evident using traditional methods.
Clay minerals are negatively charged in most soils, providing a natural attraction for positively charged cations like Mg. Thus we performed soil extractions using NH4-acetate to determine the amount of Mg (and other elements) present as adsorbed species. Although the soil profiles contain up to 80% clay, most of this clay is kaolinite, a clay with a low cation exchange capacity. Indeed, only about 1-5 wt% of the total Mg in the soil is present as adsorbed species. Adsorbed Mg increases with depth in the profile analysed (the deepest profile, to 9.3 m). The isotopic ratio, delta26Mg, of the adsorbed Mg also increases from 0.6 to 6.4 m depth (becoming heavier with depth), but is variable from 6.4 to 9.3 m. The isotopic signature of the adsorbed Mg is similar to that of the pore water Mg at depth, but is significantly heavier than the pore water Mg above about 6.4 m, likely reflecting the tendency of the heavier isotope to associate with the solid phase. The heavy isotopic excursion in pore water Mg at the soil-saprolite transition (ca. 1.0 m depth) may indicate additional release of adsorbed Mg from clay surfaces into the pore water. Indeed, there less kaolinite at 1.0 m relative to shallower and deeper depths, indicating that there are fewer adsorption sites at that depth.
Oxygen isotope ratios, delta18O, and hydrogen isotope ratios, delta2H, generally decrease (become isotopically lighter) with depth in the 4 pore water profiles. An isotopically light excursion in pore water O and H is present at the soil-saprolite transition in the deepest profile, similar to the isotopically heavy excursion in pore water Mg at the same depth. This result likely reflects a change in the hydrology at this depth and analysis of this data is ongoing.
Strontium isotope ratios, 87Sr/86Sr, reflect the Sr source rather than the process that has occurred. In the deepest profile, the pore water 87Sr/86Sr decreases with increasing depth, from 0.70985 at the surface to 0.70692 at the bottom of the profile. This result is consistent with a mixing between relatively radiogenic atmospheric sources (dust, rain) and less radiogenic weathering inputs (kaolinite, plagioclase, some other primary minerals), with the weathering sources contributing more Sr at deeper depths, relative to atmospheric sources. This is consistent with the results of a study of Sr isotopes in a neighbouring watershed overlying a different lithology (Pett-Ridge et al., 2009). Interestingly, an isotopically heavy (more radiogenic) excursion in Sr isotopes is also observed at ca. 1.0 m depth in the profile. This is attributed to the lower abundance of kaolinite at this depth. One explanation is that at the soil-saprolite transition at ca. 1.0 m depth, biological activity alters the geochemical conditions such that kaolinite becomes unstable and dissolves, releasing isotopically heavy Sr and Mg.
Isotopic data on stream waters from a storm event show that the Mg isotopic composition is lowest (isotopically light) at high stage and increases (becomes isotopically heavier) as the stream recovers to baseflow height. This is consistent with large input of rain during the storm event (either directly to the stream or via quick runoff from the land surface) pushing the isotopic composition towards a lighter Mg signature. As the stream level recovers, deep-sourced, isotopically heavier Mg becomes more prominent in the stream waters. This result is consistent with a similar effect observed in Ge/Si ratios of stream waters during a storm event in the neighbouring granitic watershed (Kurtz et al., 2011). These results suggest that watershed export of Mg is dominated by deep-weathering processes during baseflow with contributions from rain and overland runoff during storms. If this is the case, then Mg isotope signatures in river waters at this site reflect weathering at bedrock-saprolite interfaces and along bedrock fractures mixed with atmospheric inputs rather than biogeochemical processes that occur in the upper layers of the soil profiles. This result is likely to hold true for similar tropical upland watersheds. However, the isotopic signature of the riverine inputs to the oceans may also be altered by weathering and biological processes occurring within rivers and estuaries as well as by input of Mg to the rivers in solid form, via erosion (e.g. landslides, soil creep) and these phenomena also need to be investigated.
The stream water data from the storm event suggests that weathering processes in the deep subsurface are key to the isotopic signature of the riverine outputs. However, the deep subsurface is not a simple layered system in which weathering fronts progress through soil to saprolite to bedrock. In fact, we have found that the subsurface may be made up of numerous weathering fronts surrounding corestones embedded in saprolite and along fracture zones within corestones and bedrock (Buss et al., 2013). Each of these weathering fronts not only produce weathering products with the isotopic signature of the weathering process, but also a flux of mineral energy and nutrients available to subsurface microorganisms. We are thus able to quantitatively relate growth rates of subsurface microorganisms to mineral and bedrock weathering rates (Buss et al., 2005; Buss et al., 2010; Liermann et al., Subm.).
Magnesium is an essential macronutrient for all organisms, yet soils worldwide exhibit Mg deficiencies (e.g. Hüttl and Schaaf, 1997). In many temperate and tropical forests, Mg is largely stored in the aboveground biomass, rather than in the soil itself (Armbruster et al., 2002); thus Mg cycles in such forests are extremely sensitive to land use and climate change. Losses of soil nutrients, such as Mg, threaten soil productivity in both forested and agricultural ecosystems.
Mangesium also plays a key role in global climate cycles. The weathering of Ca- and Mg- containing silicate rocks is a significant long-term sink for atmospheric CO2 (Berner et al., 1983), and chemical weathering of Ca-Mg silicates is a primary long-term control on this important greenhouse gas. Thus to estimate and predict CO2 consumption rates, it is important to identify and quantify the sources and fluxes of Ca and Mg that are discharged to the world’s oceans from rivers and the weathering processes and mechanisms that produce those fluxes. Chemical weathering also exerts a key control on ocean chemistry, and Mg is a particularly important aspect of seawater chemistry, given that it controls the nature of carbonate precipitation and the fact that Mg/Ca ratios of foraminifera are an important tool in the reconstruction of past ocean temperatures. Despite its importance, there are significant uncertainties associated with the seawater Mg budget through time, including the magnitude and temporal dynamism of the riverine flux. Stable isotopic tracers such as Mg and Li are increasingly being used to identify sources of these constituents in rivers and oceans (e.g. Huh et al., 1998; Tipper et al., 2006). Similarly, stable isotope systems are also used in marine sediments and sedimentary rocks to deduce paleoclimates, ancient ocean chemistry and circulation, and other aspects of Earth’s history (e.g. Vance et al., 2009). However these interpretations and historical reconstructions are entirely dependent on the processes that may fractionate the isotopes during mineral weathering and (bio)geochemical cycling. Therefore a solid understanding of the isotopic signatures of these elements in terrestrial materials and the mechanisms that fractionate the isotopes during the cycling of these elements, is crucial for both interpreting the historical rock record and predicting future CO2 sequestration.
Stable isotope ratios of elements such as Mg in ocean records have been touted as proxies for continental weathering fluxes; however, Mg isotope ratios are known to fractionate during clay mineral precipitation (e.g. Pogge von Strandmann et al., 2008), and uptake by vegetation (Black et al., 2006; Tipper et al., 2010), suggesting that weathering and/or biogeochemical Mg isotope fractionations in soils may be essential for understanding the relationship between riverine Mg isotope ratios and weathering fluxes. Mg takes a circuitous route from bedrock to ocean, passing through dissolved and solid phases as it undergoes geochemical reactions (e.g. clay formation and dissolution, adsorption onto clay minerals, complexation with organic acids), bio(geo)chemical reactions (e.g. uptake by microorganisms and vegetation, decomposition of organic material, intracellular enzymatic processes), and is transported through soil, vegetation, and groundwater. Considering the spatial and temporal overlap of these processes, it is no simple task to identify and quantify them in natural samples in order to trace the Mg cycle through the environment. Isotopic ratios are an increasingly utilized tool for tracing such processes, but as yet, very little is known about the Mg isotopic fractionations these processes may or may not produce. Furthermore, overlapping or sequential processes that produce isotopic fractionations can overprint earlier or weaker isotopic fractionations, resulting in ephemeral or complicated isotopic signatures (e.g. Buss et al., 2010).
In order to interrogate the Mg cycle in the terrestrial environment, I am analysing Mg isotopes alongside isotopes of Sr, H, O, and Li and a suite of elemental ratios. Riverine Li is predominantly derived from silicate mineral weathering and often substitutes for Mg in primary minerals. Like Mg, Li isotopes are fractionated during precipitation of secondary minerals such as clays (Pogge von Strandmann et al., 2008a; Vigier et al., 2008), but in contrast to Mg (an essential macronutrient), Li is not a nutrient element and is not expected to be fractionated by biological processes. Thus, Li isotope data should allow the separation of the biological component from the geological component of the Mg cycle. I am further developing techniques for Mg and Li isotopic analysis of soils, vegetation, and other environmental samples; developing a novel Mg-Li multi-isotopic tracer of biogeochemical processes, testing for Mg and Li isotope fractionation in a suite of natural samples significantly adding to the dataset in the literature, and producing data for biogeochemical cycling models.
As an intensive case study, we are applying the isotopic dual-tracer technique to a large set of natural samples that I have collected previously from the Luquillo Mountains in Puerto Rico, a US National Science Foundation (NSF) Critical Zone Observatory of which I am a founding PI. This dataset will be used to demonstrate the potential of the dual-tracer technique, and to move forward the models of biogeochemical mineral nutrient cycling that I am developing for the Luquillo Critical Zone Observatory (LCZO). Overall, this work will represent a significant advancement in our understanding of the terrestrial Mg biogeochemical cycle, in particular in tropical soils. Specific objectives in this context are:
1. The development of new sample preparation and analytical procedures for Mg and Li isotopic analysis of soil and vegetation.
2. Creation of a soil and vegetation Li- and Mg-isotope dataset.
3. Establishment and testing of a novel dual-isotopic tracer technique for the mechanistic understanding and quantification of the terrestrial Mg cycle.
4. Application of that approach to the Luquillo watershed of Puerto Rico, thereby elucidating pathways of primary and secondary weathering and constraining mechanisms for nutrient export from tropical catchments to the oceans.
Beyond these scientific objectives, the work involves building stronger international collaborations between the US and the EU, specifically between the University of Bristol and my existing collaborators at the US Geological Survey, Penn State University, the University of California at Berkeley, and the University of Pennsylvania, as well as strengthening links between the US NSF Critical Zone Observatory network and the EU SoilTrEC critical zone observatory network. It also fosters increased intra-European linkages via my ongoing collaborations with European colleagues. These knowledge transfer objectives are :
1. To provide soil and chemical weathering expertise to the Bristol Isotope Group by collaborating with them on their research into the use of isotopic proxies for chemical weathering fluxes, which are strongly influenced by weathering processes in soils, of which they have little knowledge.
2. To develop a Mg-Li dual isotopic tracer and protocols for Mg and Li isotope analysis of soils, soil extracts, pore waters, and vegetation, resulting in publications that contribute to the scientific literature as well as directly expanding the capabilities of the Bristol Isotope Group.
3. To speak about my research at European universities and conferences, including at the host institution.
4. To collaborate with researchers within the UK and train postgraduate students in the application of cross-disciplinary approaches to critical zone science.
5. To work more closely with my collaborators in France and the UK on critical zone research.
6. To become more directly involved with the SoilTrEC project’s European soil critical zone observatories (CZOs), strengthening EU links to the US CZO network, by participating in meetings and workshops hosted by both groups and participating in reciprocal CZO site visits.
Results
Soil pore water concentrations of Mg and K that have been corrected for rain input and evapotranspiration indicate that mineral weathering contributes to the pore water solute compositions in four depth profiles along a slope transect (from ridgetop to riparian zone) in the Bisley watershed of the LCZO. Pore water concentrations are relatively invariant with time below 0.5 m depth (2.0 m depth in the deepest profile). Above these depths, Mg concentrations vary in space and time, with all profiles showing overall higher concentrations in the surface layers, suggestive of biological influence. In contrast, delta26Mg values in the deepest profile show a clear trend towards heavier delta26Mg with increasing depth (-0.83‰ to -0.18‰) suggesting mixing between atmospheric Mg at the surface and a dissolving, isotopically heavier phase at depth. An excursion towards heavier delta26Mg at the soil-saprolite interface (-0.7‰, ca. 1m depth) indicates a change in controlling processes. A similar heavy excursion is present at the same depth all four profiles. This depth coincides with the transition from “true soil” to saprolite and is independent of trends in elemental concentrations, indicating that a fundamental change in the Mg cycle occurs at the soil-saprolite transition. This result demonstrates that Mg isotope ratios can trace biogeochemical soil processes that may not be evident using traditional methods.
Clay minerals are negatively charged in most soils, providing a natural attraction for positively charged cations like Mg. Thus we performed soil extractions using NH4-acetate to determine the amount of Mg (and other elements) present as adsorbed species. Although the soil profiles contain up to 80% clay, most of this clay is kaolinite, a clay with a low cation exchange capacity. Indeed, only about 1-5 wt% of the total Mg in the soil is present as adsorbed species. Adsorbed Mg increases with depth in the profile analysed (the deepest profile, to 9.3 m). The isotopic ratio, delta26Mg, of the adsorbed Mg also increases from 0.6 to 6.4 m depth (becoming heavier with depth), but is variable from 6.4 to 9.3 m. The isotopic signature of the adsorbed Mg is similar to that of the pore water Mg at depth, but is significantly heavier than the pore water Mg above about 6.4 m, likely reflecting the tendency of the heavier isotope to associate with the solid phase. The heavy isotopic excursion in pore water Mg at the soil-saprolite transition (ca. 1.0 m depth) may indicate additional release of adsorbed Mg from clay surfaces into the pore water. Indeed, there less kaolinite at 1.0 m relative to shallower and deeper depths, indicating that there are fewer adsorption sites at that depth.
Oxygen isotope ratios, delta18O, and hydrogen isotope ratios, delta2H, generally decrease (become isotopically lighter) with depth in the 4 pore water profiles. An isotopically light excursion in pore water O and H is present at the soil-saprolite transition in the deepest profile, similar to the isotopically heavy excursion in pore water Mg at the same depth. This result likely reflects a change in the hydrology at this depth and analysis of this data is ongoing.
Strontium isotope ratios, 87Sr/86Sr, reflect the Sr source rather than the process that has occurred. In the deepest profile, the pore water 87Sr/86Sr decreases with increasing depth, from 0.70985 at the surface to 0.70692 at the bottom of the profile. This result is consistent with a mixing between relatively radiogenic atmospheric sources (dust, rain) and less radiogenic weathering inputs (kaolinite, plagioclase, some other primary minerals), with the weathering sources contributing more Sr at deeper depths, relative to atmospheric sources. This is consistent with the results of a study of Sr isotopes in a neighbouring watershed overlying a different lithology (Pett-Ridge et al., 2009). Interestingly, an isotopically heavy (more radiogenic) excursion in Sr isotopes is also observed at ca. 1.0 m depth in the profile. This is attributed to the lower abundance of kaolinite at this depth. One explanation is that at the soil-saprolite transition at ca. 1.0 m depth, biological activity alters the geochemical conditions such that kaolinite becomes unstable and dissolves, releasing isotopically heavy Sr and Mg.
Isotopic data on stream waters from a storm event show that the Mg isotopic composition is lowest (isotopically light) at high stage and increases (becomes isotopically heavier) as the stream recovers to baseflow height. This is consistent with large input of rain during the storm event (either directly to the stream or via quick runoff from the land surface) pushing the isotopic composition towards a lighter Mg signature. As the stream level recovers, deep-sourced, isotopically heavier Mg becomes more prominent in the stream waters. This result is consistent with a similar effect observed in Ge/Si ratios of stream waters during a storm event in the neighbouring granitic watershed (Kurtz et al., 2011). These results suggest that watershed export of Mg is dominated by deep-weathering processes during baseflow with contributions from rain and overland runoff during storms. If this is the case, then Mg isotope signatures in river waters at this site reflect weathering at bedrock-saprolite interfaces and along bedrock fractures mixed with atmospheric inputs rather than biogeochemical processes that occur in the upper layers of the soil profiles. This result is likely to hold true for similar tropical upland watersheds. However, the isotopic signature of the riverine inputs to the oceans may also be altered by weathering and biological processes occurring within rivers and estuaries as well as by input of Mg to the rivers in solid form, via erosion (e.g. landslides, soil creep) and these phenomena also need to be investigated.
The stream water data from the storm event suggests that weathering processes in the deep subsurface are key to the isotopic signature of the riverine outputs. However, the deep subsurface is not a simple layered system in which weathering fronts progress through soil to saprolite to bedrock. In fact, we have found that the subsurface may be made up of numerous weathering fronts surrounding corestones embedded in saprolite and along fracture zones within corestones and bedrock (Buss et al., 2013). Each of these weathering fronts not only produce weathering products with the isotopic signature of the weathering process, but also a flux of mineral energy and nutrients available to subsurface microorganisms. We are thus able to quantitatively relate growth rates of subsurface microorganisms to mineral and bedrock weathering rates (Buss et al., 2005; Buss et al., 2010; Liermann et al., Subm.).