256
soil depths in zero-tension lysimeters and stream
concentrations after fertilization in the Femow Experimental Forest, West Virginia. C horizon soil solution concentrations of most elements appeared
comparable to stream concentrations. Because soil
solution concentrations declined with depth, soil
solution concentrations from below the B horizon
were higher than stream concentrations. This has
implications for lysimetry studies, as many experiments have installed lysimeters below the B horizon, with the justification that the majority of roots
rarely extend below the B horizon. Both the soil
solution and stream water responded similarly to
fertilization, although trends appeared more clearly
in the stream water data. As part of the NITREX
study, Kjonaas et al. (1998) compared plot-scale
lysimetry and whole-catchment N export data both
before and after N fertilization of a small catchment
in Sweden. Although N losses measured using tension lysimeters were higher before N fertilization,
losses at the plot and watershed scales were similar
after fertilization. Curiously, although plot-scale
lysimeters collected more dissolved organic nitrogen (DON), catchment-level runoff data showed
greater dissolved inorganic nitrogen (DIN) levels,
perhaps simply due to differences between the plot
with installed lysimeters and the rest of the catchment. Finally, Seely et al. (1998) compared concentrations of different N species in lysimeters
placed below the rooting zone in sandy soils and in
groundwater outputs directly below the sites (Fig.
16.3). Concentrations of all N species declined with
passage to groundwater, but the attenuation of
DON was greater than the attenuation of either
NO; or NHt, and the decline in DON in these
soils below the rooting zone, and thus below a zone
of any significant organic matter accumulation, was
greater than the attenuation seen in upper mineral
and organic zones. Although DON is generally the
predominant form of N leaching from most forested
watersheds (Hedin et al. 1995; Sollins and McCorison 1981; Johnson and Lindberg 1992), clearly
budgets of DON and thus N will differ markedly
depending on the output measurement used.
Such differences in the estimate of ecosystem
outputs need to be noted in literature compilations,
although many data syntheses, such as the European ECOFEE and ENSF compilations, have combined studies that use both lysimeter and stream
water measures of output N (Gundersen 1995, Dise
Kate Lajtha
and Wright 1995). Studies of N saturation often address the question of eutrophication of surface or
drinking water; clearly in these cases stream water
outputs are the critical variable (e.g., Murdoch and
Stoddard 1992; Stoddard 1994, Baron et al. 1994).
Other studies of N saturation potential at the watershed level also analyze stream outputs (Durka et
al. 1994; Hedin et al. 1995; Lepisto et al. 1995;
Nodvin et al. 1995), and many conclusions about
the relationship between N inputs and outputs, and
current N saturation models (i.e. Aber 1992) have
been made. Many studies, such as the European
NITREX study or the U.S. Integrated Forest Study
(IFS) network, measured losses of N compounds
from either fertilized or unfertilized plots via lysimeters (Johnson and Lindberg 1992; Aber et al.
1993; Tietema et al. 1995; Gundersen and Rasmussen 1995; Kjonaas et al. 1998) and also draw
conclusions about the relationship between N
leaching and anthropogenic N loading. Because
systematic errors can be introduced when merging
results from these different types of ecosystem budgets, it will be important to keep leaching estimates
separated by collection method.
Monolith Lysimetry
and Sandbox Experiments
A completely different approach to the element
mass-balance budget is to create an artificial ecosystem under field conditions. Isolated pits filled
either with native soil or sand and variously instrumented to measure leaching losses have been used
for decades in soil and agricultural research (e.g.,
Young et al. 1996) and are variously referred to as
monolith lysimeters, weighing lysimeters, unconfined lysimeters, or sandboxes. These microecosysterns can allow for a nearly complete budgeting of
inputs and outputs. In the San Dimas lysimeter experiment (San Dimas Experimental Forest, Glendora, CA) (Ulery et al. 1995; Quideau et al. 1996),
large earthen-walled pits were filled in 1937 with
homogenized soil material that was sampled and
archived at each 7.5 cm depth, and planted into
different monocultures. These have been used for
studies of weathering under different species, soil
aggregate stability, and changes in soil C, N, and
exchangeable cations over time, but these were not
equipped to measure complete input and output
budgets.
soil depths in zero-tension lysimeters and stream
concentrations after fertilization in the Femow Experimental Forest, West Virginia. C horizon soil solution concentrations of most elements appeared
comparable to stream concentrations. Because soil
solution concentrations declined with depth, soil
solution concentrations from below the B horizon
were higher than stream concentrations. This has
implications for lysimetry studies, as many experiments have installed lysimeters below the B horizon, with the justification that the majority of roots
rarely extend below the B horizon. Both the soil
solution and stream water responded similarly to
fertilization, although trends appeared more clearly
in the stream water data. As part of the NITREX
study, Kjonaas et al. (1998) compared plot-scale
lysimetry and whole-catchment N export data both
before and after N fertilization of a small catchment
in Sweden. Although N losses measured using tension lysimeters were higher before N fertilization,
losses at the plot and watershed scales were similar
after fertilization. Curiously, although plot-scale
lysimeters collected more dissolved organic nitrogen (DON), catchment-level runoff data showed
greater dissolved inorganic nitrogen (DIN) levels,
perhaps simply due to differences between the plot
with installed lysimeters and the rest of the catchment. Finally, Seely et al. (1998) compared concentrations of different N species in lysimeters
placed below the rooting zone in sandy soils and in
groundwater outputs directly below the sites (Fig.
16.3). Concentrations of all N species declined with
passage to groundwater, but the attenuation of
DON was greater than the attenuation of either
NO; or NHt, and the decline in DON in these
soils below the rooting zone, and thus below a zone
of any significant organic matter accumulation, was
greater than the attenuation seen in upper mineral
and organic zones. Although DON is generally the
predominant form of N leaching from most forested
watersheds (Hedin et al. 1995; Sollins and McCorison 1981; Johnson and Lindberg 1992), clearly
budgets of DON and thus N will differ markedly
depending on the output measurement used.
Such differences in the estimate of ecosystem
outputs need to be noted in literature compilations,
although many data syntheses, such as the European ECOFEE and ENSF compilations, have combined studies that use both lysimeter and stream
water measures of output N (Gundersen 1995, Dise
Kate Lajtha
and Wright 1995). Studies of N saturation often address the question of eutrophication of surface or
drinking water; clearly in these cases stream water
outputs are the critical variable (e.g., Murdoch and
Stoddard 1992; Stoddard 1994, Baron et al. 1994).
Other studies of N saturation potential at the watershed level also analyze stream outputs (Durka et
al. 1994; Hedin et al. 1995; Lepisto et al. 1995;
Nodvin et al. 1995), and many conclusions about
the relationship between N inputs and outputs, and
current N saturation models (i.e. Aber 1992) have
been made. Many studies, such as the European
NITREX study or the U.S. Integrated Forest Study
(IFS) network, measured losses of N compounds
from either fertilized or unfertilized plots via lysimeters (Johnson and Lindberg 1992; Aber et al.
1993; Tietema et al. 1995; Gundersen and Rasmussen 1995; Kjonaas et al. 1998) and also draw
conclusions about the relationship between N
leaching and anthropogenic N loading. Because
systematic errors can be introduced when merging
results from these different types of ecosystem budgets, it will be important to keep leaching estimates
separated by collection method.
Monolith Lysimetry
and Sandbox Experiments
A completely different approach to the element
mass-balance budget is to create an artificial ecosystem under field conditions. Isolated pits filled
either with native soil or sand and variously instrumented to measure leaching losses have been used
for decades in soil and agricultural research (e.g.,
Young et al. 1996) and are variously referred to as
monolith lysimeters, weighing lysimeters, unconfined lysimeters, or sandboxes. These microecosysterns can allow for a nearly complete budgeting of
inputs and outputs. In the San Dimas lysimeter experiment (San Dimas Experimental Forest, Glendora, CA) (Ulery et al. 1995; Quideau et al. 1996),
large earthen-walled pits were filled in 1937 with
homogenized soil material that was sampled and
archived at each 7.5 cm depth, and planted into
different monocultures. These have been used for
studies of weathering under different species, soil
aggregate stability, and changes in soil C, N, and
exchangeable cations over time, but these were not
equipped to measure complete input and output
budgets.