Global Change Biology (2012), doi: 10.1111/j.1365-2486.2012.02731.x
Thaw depth determines reaction and transport of inorganic nitrogen in valley bottom permafrost soils T A M A R A K . H A R M S and J E R E M Y B . J O N E S J R . Institute of Arctic Biology, University of Alaska-Fairbanks, 99775, Alaska USA
Abstract Nitrate (NO3–) export coupled with high inorganic nitrogen (N) concentrations in Alaskan streams suggests that N cycles of permafrost-influenced ecosystems are more open than expected for N-limited ecosystems. We tested the hypothesis that soil thaw depth governs inorganic N retention and removal in soils due to vertical patterns in the dominant N transformation pathways. Using an in situ, push–pull method, we estimated rates of inorganic N uptake and denitrification during snow melt, summer, and autumn, as depth of soil–stream flowpaths increased in the valley bottom of an arctic and a boreal catchment. Net NO3– uptake declined sharply from snow melt to summer and decreased as a nonlinear function of thaw depth. Peak denitrification rate occurred during snow melt at the arctic site, in summer at the boreal site, and declined as a nonlinear function of thaw depth across both sites. Seasonal patterns in ammonium (NH4+) uptake were not significant, but low rates during the peak growing season suggest uptake that is balanced by mineralization. Despite rapid rates of hydrologic transport during snow melt runoff, rates of uptake and removal of inorganic N tended to exceed water residence time during snow melt, indicating potential for retention of N in valley bottom soils when flowpaths are shallow. Decreased reaction rates relative to water residence time in subsequent seasons suggest greater export of inorganic N as the soil–stream flowpath deepens due to thawing soils. Using seasonal thaw as a proxy for longer term deepening of the thaw layer caused by climate warming and permafrost degradation, these results suggest increasing potential for export of inorganic N from permafrostinfluenced soils to streams. Keywords: ammonium (NH4+), arctic tundra, boreal forest, denitrification, high-latitude catchments, nitrate (NO3–), nutrient uptake, retention, soil–stream flowpath Received 29 February 2012; revised version received 23 April 2012 and accepted 26 April 2012
Introduction Recent observations of nitrogen (N) transport in boreal and arctic streams suggest that high-latitude N cycles are more open than expected for N-limited ecosystems. Nitrate (NO3–) dominates annual N flux from boreal streams in Alaska draining watersheds that are underlain by discontinuous permafrost (Jones et al., 2005), and NO3– concentration is elevated relative to streams in permafrost-free boreal forest or in regions with continuous permafrost (Petrone et al., 2006). Long-term observation of the Kuparuk River, an arctic river draining continuous permafrost, indicates increasing NO3– flux that is unrelated to trends in stream discharge (McClelland et al., 2007). Such patterns may result from increasing temperature, and in particular thawing of permafrost, but mechanisms linking hydrologic exports of N to permafrost remain untested. N cycles of boreal forest and arctic tundra, two dominant ecosystems of high-latitude regions, are charCorrespondence: Tamara K. Harms, tel. + 907 474 6117, fax + 907 474 6967, e-mail: [email protected]
© 2012 Blackwell Publishing Ltd
acterized by rapid internal recycling relative to small inputs and outputs. Presence of permafrost constrains biological activity to a shallow active layer that thaws seasonally. N limitation of biota combined with a shallow active layer may foster competition for N among plants and soil microorganisms (Jonasson & Shaver, 1999; Kielland et al., 2007), constraining hydrologic export of inorganic N (Petrone et al., 2007a; Yano et al., 2010; Fig. 1). Uptake and assimilation of organic forms of N by many boreal and arctic plant species underscores the severity of N limitation and scarcity of inorganic N (Kielland, 1994; Persson et al., 2003). Rapid assimilative uptake of N by plants and microorganisms is then expected to restrict N availability to nitrifiers and denitrifiers. Small pools of inorganic N in boreal and tundra soils combined with low rates of nitrification support this conceptual model, especially for NO3– (Giblin et al., 1991; Kielland et al., 2007; Fig. 1). Alternatively, denitrifiers may successfully compete for NO3– in moist, organic-rich surface horizons, effectively removing inorganic N from the cycling pool (Ho¨gberg et al., 2006), but few estimates of denitrification have been made in high-latitude soils. 1
2 T. K. HARMS & J. B. JONES
Fig. 1 Hypothesized patterns of N transformation and retention with respect to thaw depth. Shallow soil–stream flowpaths occur in locations where the active layer encompasses only upper soil horizons or during the early thaw season (top panel). Ecosystems or time periods characterized by shallow hydrologic flowpaths would be expected to export N as DON with limited potential for hydrologic losses to exceed inputs of N. Deeper thaw (bottom panel), including late summer and autumn, or where permafrost has been degraded, allows interactions with mineral soils and may promote production of inorganic N, with potential for export.
Despite strong potential for N retention, N export in streams draining catchments underlain by discontinuous permafrost exceeds inputs from atmospheric deposition (Jones et al., 2005). Such patterns are counter to nutrient retention theory, which predicts that N-limited ecosystems retain inorganic N in biomass due to demand by primary producers (Vitousek & Reiners, 1975), with loss of N occurring in organic forms (Hedin et al., 1995). Changing climate of high-latitude ecosystems may be contributing to this discrepancy between theory and observed patterns in N export. Increasing evidence suggests that rising temperature causes characteristic shifts in inorganic N availability in high-latitude catchments. Experimental warming of both boreal and arctic soils results in increased soil inorganic N pools (Schimel et al., 2004; Allison & Treseder, 2008; Natali et al., 2011), likely due to increased rates of N mineralization or nitrification (Nadelhoffer et al., 1991; Rustad et al., 2001; Keuper et al., 2012). Subsequent leaching from experimentally warmed soils (Mack et al., 2004; Buckeridge & Grogan, 2010) suggests potential for export of inorganic N to streams. In addition to the direct effects of warming on soil processes, changes in depth of dominant hydrologic flowpaths through catchments may cause increased export of inorganic N. As active layer depth increases
due to permafrost thaw, water flows through deeper soil strata (Keller et al., 2010) potentially introducing N from mineral soils or recently thawed soils into the actively cycling pool (Keuper et al., 2012). Because the existing conceptual model of terrestrial N cycling at high latitudes focuses on plant–soil interactions, few observations describe N dynamics at depths >20 cm. Deeper soils may contain more inorganic N compared with organic horizons due to absence of uptake by plants and decreased rates of immobilization. Lower C : N content of mineral compared with organic soil horizons may promote N mineralization and nitrification over assimilation by microorganisms (Kaye & Hart, 1997; Ho¨gberg et al., 2006). Furthermore, dissolved organic matter may be preferentially adsorbed to mineral soil particles (Kawahigashi et al., 2006), resulting in dominance of inorganic N in dissolved exports from deep soil–stream flowpaths (Fig. 1). Here, we examined how seasonal increase in thaw depth influences the dominant N transformations within soil–stream flowpaths connecting valley bottom soils to streams in boreal and arctic catchments underlain by permafrost. Frozen soils restrict the flow of water to organic horizons during snow melt and summer, and gradual seasonal thaw results in flow through mineral soils in autumn. We hypothesized that soil thaw depth governs inorganic N retention and removal in soils due to vertical patterns in the dominant N transformation pathways. Specifically, we hypothesized that upper organic soils promote high rates of denitrification and N uptake and therefore provide a sink for N during snow melt. We expected decreased N uptake and removal from soils in autumn, when flows through deeper mineral soils bypass zones of N retention, which would permit export of inorganic N to streams.
Materials and methods
Site description We conducted experiments in the valley bottom soils of a boreal forest and an arctic tundra catchment. Saturated, valley bottom soils have a persistent hydrologic connection to streams (McNamara et al., 1997; Petrone et al., 2007b), and therefore strong potential to influence hydrologic export of N from catchments. The boreal site was within the Caribou-Poker Creeks Research Watershed (CPCRW), one of two primary study sites of the Bonanza Creek Long-Term Ecological Research (LTER) Program located in interior Alaska. Experiments in the CPCRW were located at the base of a northeast-facing hillslope within a 60-km2 catchment. Permafrost in the catchment is spatially discontinuous,
© 2012 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2012.02731.x
NITROGEN CYCLING IN PERMAFROST SOILS 3 but it was present throughout the valley bottom where experiments were conducted. Mean annual air temperature in the watershed is 4 °C, and mean annual precipitation is 410 mm, with 2/3 falling as summer rain. Vegetation in the valley bottom includes shrub birch (Betula glandulosa), cranberry (Vaccinium vitis-idaea and V. oxycoccus), Labrador tea (Ledum groenlandicum and L. decumbens), cloudberry (Rubus chamaemorus), moss (Sphagnum spp.), and tussock-forming grasses and sedges (Eriophorum spp. and Calamagrostis canadensis). Black spruce (Picea mariana) and tamarack (Larix laricina) also occur at low densities. Soils of the valley bottom are classified as subgelic Typic Histoturbels. A thick layer of mosses (~10 cm) covers organic soil horizons (~15 cm) which contain high densities of roots, whereas roots occur at lower densities or are absent in mineral horizons (Ping et al., 2005). The arctic site was within the Upper Kuparuk River basin of northern Alaska, draining 130 km2 at the location of the experiments. This site is within the study area of the Arctic LTER, where mean annual air temperature is 7.4 °C and mean annual precipitation is 300 mm, with greater than half falling as summer rain. Tussock-forming sedges (Eriophorum spp.) dominate the vegetative communities, with mosses (Sphagnum spp.) and other sedges (Carex bigelowii and C. aquatilis) in inter-tussock locations. Shrub cover (Betula nana and Salix lanata) was sparse. Soils are gelisols classified as Ruptic-histic aquiturbels. Moss layers can be up to 25 cm deep in the valley bottom, but are typically ~15 cm, overlying organic soils ~10 cm deep. Roots are sparse or absent in mineral soils (Ping et al., 1998; Borden et al., 2010).
Experimental design We measured rates of N-transforming reactions including NH4+ and NO3– uptake, and denitrification at six replicate locations within each site, in each of three seasons during 2010. “Snow melt” designates experiments conducted in early (boreal) or mid-May (arctic) when snow covered portions of the catchments and stream discharge was near peak flow, “summer” experiments were conducted mid-June (boreal) or late June/early July (arctic) during the peak growing season, and “autumn” experiments were conducted in mid-August (arctic) and early September (boreal) when leaf senescence had begun.
Push–pull method The push–pull method was developed to examine fate and transport of contaminants in aquifers (Istok et al., 1997) and has since been adapted to study nutrient
uptake in groundwater, hyporheic zones, and saturated sediments (Addy et al., 2002; Whitmire & Hamilton, 2005; Clilverd et al., 2008). The method includes addition of a solution containing a reactive and conservative solute to a well (push phase), followed by withdrawal of samples over time (pull phase). Reaction rate is then calculated from the change in the ratio of reactive to conservative solute concentrations. Herein, we adapt this method to assess N uptake and denitrification in saturated soils. We injected soils with solutions of N (KNO3 or (NH4)2SO4, to measure net NO3– or NH4+ uptake, respectively) or 15N (99 atom% K15NO3) to measure denitrification, and NaBr as a conservative tracer. Injected solution used in net uptake experiments contained 0.5 mg N L1, and 0.05 mg N L1 for denitrification experiments. Bromide concentration was 1– 2 mg L1 for all experiments. Sulfur hexafluoride (SF6) was also included as a conservative tracer to measure gas evasion during denitrification experiments; all other injection solutions received 1 mL He L1 solution to maintain similar oxygen conditions among experiments. Tracer solutions were prepared in 1 L collapsible cubitainers. Previous applications of the push–pull technique have amended water withdrawn from the point of injection (Addy et al., 2002), or a local source water assumed to match site-specific chemistry (Whitmire & Hamilton, 2005). We injected deionized water amended with tracers because we did not expect that a sufficient volume of soil water could be withdrawn from each injection point over a reasonable amount of time. Spatial heterogeneity in soil water chemistry (Table 1) precluded use of a single, local source of water. For the processes we measured, the most significant effect of using deionized water is dilution of dissolved organic matter. On average, dissolved organic carbon concentration was 30% of background concentration at the first sampling, and 90% at the end of experiments. Push–pull experiments were conducted at the base of the thaw layer in each season. A stainless steel tube (ID = 2 mm), perforated along the bottom 5 cm with 1.6 mm diameter holes, was connected to a gas-tight fitting, and inserted to the depth of frozen soils. This depth was determined as depth of refusal of the stainless steel tube. The tube was left in place throughout the experiment and was used for both introduction of tracer solutions and withdrawal of samples. Following collection of a background sample, tracer solutions (500 mL) were introduced to soils at a rate of 30 mL min1 (push phase). Samples were withdrawn at a rate of 20 mL min1 (pull phase) at regular intervals over 2– 3 h. For all experiments, each pull phase consisted of a 2 mL rinse followed by collection of 42 mL of water. Samples from net N uptake experiments were filtered
© 2012 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2012.02731.x
4 T. K. HARMS & J. B. JONES Table 1 Physical and ambient chemical attributes of soils and soil water during the study period. Values are means ± 1 SE Boreal
Arctic
Snow melt NO3--N (mg L1) NH4+-N† (mg L1) DON† (mg L1) DOC (mg L1) Soil C : N (atoms) Thaw depth (cm) Soil horizon
0.013 ± 0.230 ± 1.16 ± 47.2 ± 56 ± 9± organic
0.001a 0.072a 0.13a 3.3a 3‡ 1a
Summer 0.011 ± 0.073 ± 0.50 ± 22.7 ± 30 ± 12 ± organic
Autumn 0.001ac 0.029a 0.04a 0.8b 2 1a
0.007 ± 0.081 ± 0.72 ± 23.6 ± 27 ± 26 ± mineral
Snow melt 0.002bc 0.036a 0.14a 2.2b 1 2b
0.012 ± 0.014 ± 0.50 ± 24.4 ± 37 ± 8± organic
0.001a 0.002a 0.08a 2.7b 1‡ 1a
Summer 0.006 ± 0.244 ± 1.38 ± 28.2 ± 23 ± 20 ± organic
P-value*
Autumn 0.001d 0.108ac 0.27b 4.9b 1 1b
0.009 ± 0.790 ± 1.34 ± 30.3 ± 20 ± 40 ± mineral
0.001acd 0.317bc 0.35b 3.7b 1 4c
0.005 0.006
Thaw depth determines reaction and transport of inorganic nitrogen in valley bottom permafrost soils T A M A R A K . H A R M S and J E R E M Y B . J O N E S J R . Institute of Arctic Biology, University of Alaska-Fairbanks, 99775, Alaska USA
Abstract Nitrate (NO3–) export coupled with high inorganic nitrogen (N) concentrations in Alaskan streams suggests that N cycles of permafrost-influenced ecosystems are more open than expected for N-limited ecosystems. We tested the hypothesis that soil thaw depth governs inorganic N retention and removal in soils due to vertical patterns in the dominant N transformation pathways. Using an in situ, push–pull method, we estimated rates of inorganic N uptake and denitrification during snow melt, summer, and autumn, as depth of soil–stream flowpaths increased in the valley bottom of an arctic and a boreal catchment. Net NO3– uptake declined sharply from snow melt to summer and decreased as a nonlinear function of thaw depth. Peak denitrification rate occurred during snow melt at the arctic site, in summer at the boreal site, and declined as a nonlinear function of thaw depth across both sites. Seasonal patterns in ammonium (NH4+) uptake were not significant, but low rates during the peak growing season suggest uptake that is balanced by mineralization. Despite rapid rates of hydrologic transport during snow melt runoff, rates of uptake and removal of inorganic N tended to exceed water residence time during snow melt, indicating potential for retention of N in valley bottom soils when flowpaths are shallow. Decreased reaction rates relative to water residence time in subsequent seasons suggest greater export of inorganic N as the soil–stream flowpath deepens due to thawing soils. Using seasonal thaw as a proxy for longer term deepening of the thaw layer caused by climate warming and permafrost degradation, these results suggest increasing potential for export of inorganic N from permafrostinfluenced soils to streams. Keywords: ammonium (NH4+), arctic tundra, boreal forest, denitrification, high-latitude catchments, nitrate (NO3–), nutrient uptake, retention, soil–stream flowpath Received 29 February 2012; revised version received 23 April 2012 and accepted 26 April 2012
Introduction Recent observations of nitrogen (N) transport in boreal and arctic streams suggest that high-latitude N cycles are more open than expected for N-limited ecosystems. Nitrate (NO3–) dominates annual N flux from boreal streams in Alaska draining watersheds that are underlain by discontinuous permafrost (Jones et al., 2005), and NO3– concentration is elevated relative to streams in permafrost-free boreal forest or in regions with continuous permafrost (Petrone et al., 2006). Long-term observation of the Kuparuk River, an arctic river draining continuous permafrost, indicates increasing NO3– flux that is unrelated to trends in stream discharge (McClelland et al., 2007). Such patterns may result from increasing temperature, and in particular thawing of permafrost, but mechanisms linking hydrologic exports of N to permafrost remain untested. N cycles of boreal forest and arctic tundra, two dominant ecosystems of high-latitude regions, are charCorrespondence: Tamara K. Harms, tel. + 907 474 6117, fax + 907 474 6967, e-mail: [email protected]
© 2012 Blackwell Publishing Ltd
acterized by rapid internal recycling relative to small inputs and outputs. Presence of permafrost constrains biological activity to a shallow active layer that thaws seasonally. N limitation of biota combined with a shallow active layer may foster competition for N among plants and soil microorganisms (Jonasson & Shaver, 1999; Kielland et al., 2007), constraining hydrologic export of inorganic N (Petrone et al., 2007a; Yano et al., 2010; Fig. 1). Uptake and assimilation of organic forms of N by many boreal and arctic plant species underscores the severity of N limitation and scarcity of inorganic N (Kielland, 1994; Persson et al., 2003). Rapid assimilative uptake of N by plants and microorganisms is then expected to restrict N availability to nitrifiers and denitrifiers. Small pools of inorganic N in boreal and tundra soils combined with low rates of nitrification support this conceptual model, especially for NO3– (Giblin et al., 1991; Kielland et al., 2007; Fig. 1). Alternatively, denitrifiers may successfully compete for NO3– in moist, organic-rich surface horizons, effectively removing inorganic N from the cycling pool (Ho¨gberg et al., 2006), but few estimates of denitrification have been made in high-latitude soils. 1
2 T. K. HARMS & J. B. JONES
Fig. 1 Hypothesized patterns of N transformation and retention with respect to thaw depth. Shallow soil–stream flowpaths occur in locations where the active layer encompasses only upper soil horizons or during the early thaw season (top panel). Ecosystems or time periods characterized by shallow hydrologic flowpaths would be expected to export N as DON with limited potential for hydrologic losses to exceed inputs of N. Deeper thaw (bottom panel), including late summer and autumn, or where permafrost has been degraded, allows interactions with mineral soils and may promote production of inorganic N, with potential for export.
Despite strong potential for N retention, N export in streams draining catchments underlain by discontinuous permafrost exceeds inputs from atmospheric deposition (Jones et al., 2005). Such patterns are counter to nutrient retention theory, which predicts that N-limited ecosystems retain inorganic N in biomass due to demand by primary producers (Vitousek & Reiners, 1975), with loss of N occurring in organic forms (Hedin et al., 1995). Changing climate of high-latitude ecosystems may be contributing to this discrepancy between theory and observed patterns in N export. Increasing evidence suggests that rising temperature causes characteristic shifts in inorganic N availability in high-latitude catchments. Experimental warming of both boreal and arctic soils results in increased soil inorganic N pools (Schimel et al., 2004; Allison & Treseder, 2008; Natali et al., 2011), likely due to increased rates of N mineralization or nitrification (Nadelhoffer et al., 1991; Rustad et al., 2001; Keuper et al., 2012). Subsequent leaching from experimentally warmed soils (Mack et al., 2004; Buckeridge & Grogan, 2010) suggests potential for export of inorganic N to streams. In addition to the direct effects of warming on soil processes, changes in depth of dominant hydrologic flowpaths through catchments may cause increased export of inorganic N. As active layer depth increases
due to permafrost thaw, water flows through deeper soil strata (Keller et al., 2010) potentially introducing N from mineral soils or recently thawed soils into the actively cycling pool (Keuper et al., 2012). Because the existing conceptual model of terrestrial N cycling at high latitudes focuses on plant–soil interactions, few observations describe N dynamics at depths >20 cm. Deeper soils may contain more inorganic N compared with organic horizons due to absence of uptake by plants and decreased rates of immobilization. Lower C : N content of mineral compared with organic soil horizons may promote N mineralization and nitrification over assimilation by microorganisms (Kaye & Hart, 1997; Ho¨gberg et al., 2006). Furthermore, dissolved organic matter may be preferentially adsorbed to mineral soil particles (Kawahigashi et al., 2006), resulting in dominance of inorganic N in dissolved exports from deep soil–stream flowpaths (Fig. 1). Here, we examined how seasonal increase in thaw depth influences the dominant N transformations within soil–stream flowpaths connecting valley bottom soils to streams in boreal and arctic catchments underlain by permafrost. Frozen soils restrict the flow of water to organic horizons during snow melt and summer, and gradual seasonal thaw results in flow through mineral soils in autumn. We hypothesized that soil thaw depth governs inorganic N retention and removal in soils due to vertical patterns in the dominant N transformation pathways. Specifically, we hypothesized that upper organic soils promote high rates of denitrification and N uptake and therefore provide a sink for N during snow melt. We expected decreased N uptake and removal from soils in autumn, when flows through deeper mineral soils bypass zones of N retention, which would permit export of inorganic N to streams.
Materials and methods
Site description We conducted experiments in the valley bottom soils of a boreal forest and an arctic tundra catchment. Saturated, valley bottom soils have a persistent hydrologic connection to streams (McNamara et al., 1997; Petrone et al., 2007b), and therefore strong potential to influence hydrologic export of N from catchments. The boreal site was within the Caribou-Poker Creeks Research Watershed (CPCRW), one of two primary study sites of the Bonanza Creek Long-Term Ecological Research (LTER) Program located in interior Alaska. Experiments in the CPCRW were located at the base of a northeast-facing hillslope within a 60-km2 catchment. Permafrost in the catchment is spatially discontinuous,
© 2012 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2012.02731.x
NITROGEN CYCLING IN PERMAFROST SOILS 3 but it was present throughout the valley bottom where experiments were conducted. Mean annual air temperature in the watershed is 4 °C, and mean annual precipitation is 410 mm, with 2/3 falling as summer rain. Vegetation in the valley bottom includes shrub birch (Betula glandulosa), cranberry (Vaccinium vitis-idaea and V. oxycoccus), Labrador tea (Ledum groenlandicum and L. decumbens), cloudberry (Rubus chamaemorus), moss (Sphagnum spp.), and tussock-forming grasses and sedges (Eriophorum spp. and Calamagrostis canadensis). Black spruce (Picea mariana) and tamarack (Larix laricina) also occur at low densities. Soils of the valley bottom are classified as subgelic Typic Histoturbels. A thick layer of mosses (~10 cm) covers organic soil horizons (~15 cm) which contain high densities of roots, whereas roots occur at lower densities or are absent in mineral horizons (Ping et al., 2005). The arctic site was within the Upper Kuparuk River basin of northern Alaska, draining 130 km2 at the location of the experiments. This site is within the study area of the Arctic LTER, where mean annual air temperature is 7.4 °C and mean annual precipitation is 300 mm, with greater than half falling as summer rain. Tussock-forming sedges (Eriophorum spp.) dominate the vegetative communities, with mosses (Sphagnum spp.) and other sedges (Carex bigelowii and C. aquatilis) in inter-tussock locations. Shrub cover (Betula nana and Salix lanata) was sparse. Soils are gelisols classified as Ruptic-histic aquiturbels. Moss layers can be up to 25 cm deep in the valley bottom, but are typically ~15 cm, overlying organic soils ~10 cm deep. Roots are sparse or absent in mineral soils (Ping et al., 1998; Borden et al., 2010).
Experimental design We measured rates of N-transforming reactions including NH4+ and NO3– uptake, and denitrification at six replicate locations within each site, in each of three seasons during 2010. “Snow melt” designates experiments conducted in early (boreal) or mid-May (arctic) when snow covered portions of the catchments and stream discharge was near peak flow, “summer” experiments were conducted mid-June (boreal) or late June/early July (arctic) during the peak growing season, and “autumn” experiments were conducted in mid-August (arctic) and early September (boreal) when leaf senescence had begun.
Push–pull method The push–pull method was developed to examine fate and transport of contaminants in aquifers (Istok et al., 1997) and has since been adapted to study nutrient
uptake in groundwater, hyporheic zones, and saturated sediments (Addy et al., 2002; Whitmire & Hamilton, 2005; Clilverd et al., 2008). The method includes addition of a solution containing a reactive and conservative solute to a well (push phase), followed by withdrawal of samples over time (pull phase). Reaction rate is then calculated from the change in the ratio of reactive to conservative solute concentrations. Herein, we adapt this method to assess N uptake and denitrification in saturated soils. We injected soils with solutions of N (KNO3 or (NH4)2SO4, to measure net NO3– or NH4+ uptake, respectively) or 15N (99 atom% K15NO3) to measure denitrification, and NaBr as a conservative tracer. Injected solution used in net uptake experiments contained 0.5 mg N L1, and 0.05 mg N L1 for denitrification experiments. Bromide concentration was 1– 2 mg L1 for all experiments. Sulfur hexafluoride (SF6) was also included as a conservative tracer to measure gas evasion during denitrification experiments; all other injection solutions received 1 mL He L1 solution to maintain similar oxygen conditions among experiments. Tracer solutions were prepared in 1 L collapsible cubitainers. Previous applications of the push–pull technique have amended water withdrawn from the point of injection (Addy et al., 2002), or a local source water assumed to match site-specific chemistry (Whitmire & Hamilton, 2005). We injected deionized water amended with tracers because we did not expect that a sufficient volume of soil water could be withdrawn from each injection point over a reasonable amount of time. Spatial heterogeneity in soil water chemistry (Table 1) precluded use of a single, local source of water. For the processes we measured, the most significant effect of using deionized water is dilution of dissolved organic matter. On average, dissolved organic carbon concentration was 30% of background concentration at the first sampling, and 90% at the end of experiments. Push–pull experiments were conducted at the base of the thaw layer in each season. A stainless steel tube (ID = 2 mm), perforated along the bottom 5 cm with 1.6 mm diameter holes, was connected to a gas-tight fitting, and inserted to the depth of frozen soils. This depth was determined as depth of refusal of the stainless steel tube. The tube was left in place throughout the experiment and was used for both introduction of tracer solutions and withdrawal of samples. Following collection of a background sample, tracer solutions (500 mL) were introduced to soils at a rate of 30 mL min1 (push phase). Samples were withdrawn at a rate of 20 mL min1 (pull phase) at regular intervals over 2– 3 h. For all experiments, each pull phase consisted of a 2 mL rinse followed by collection of 42 mL of water. Samples from net N uptake experiments were filtered
© 2012 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2012.02731.x
4 T. K. HARMS & J. B. JONES Table 1 Physical and ambient chemical attributes of soils and soil water during the study period. Values are means ± 1 SE Boreal
Arctic
Snow melt NO3--N (mg L1) NH4+-N† (mg L1) DON† (mg L1) DOC (mg L1) Soil C : N (atoms) Thaw depth (cm) Soil horizon
0.013 ± 0.230 ± 1.16 ± 47.2 ± 56 ± 9± organic
0.001a 0.072a 0.13a 3.3a 3‡ 1a
Summer 0.011 ± 0.073 ± 0.50 ± 22.7 ± 30 ± 12 ± organic
Autumn 0.001ac 0.029a 0.04a 0.8b 2 1a
0.007 ± 0.081 ± 0.72 ± 23.6 ± 27 ± 26 ± mineral
Snow melt 0.002bc 0.036a 0.14a 2.2b 1 2b
0.012 ± 0.014 ± 0.50 ± 24.4 ± 37 ± 8± organic
0.001a 0.002a 0.08a 2.7b 1‡ 1a
Summer 0.006 ± 0.244 ± 1.38 ± 28.2 ± 23 ± 20 ± organic
P-value*
Autumn 0.001d 0.108ac 0.27b 4.9b 1 1b
0.009 ± 0.790 ± 1.34 ± 30.3 ± 20 ± 40 ± mineral
0.001acd 0.317bc 0.35b 3.7b 1 4c
0.005 0.006
