254
or soil water collectors (e.g., Binkley et al. 1992;
Joslin et al. 1992; Joslin and Wolfe 1992; Johnson
and Lindberg 1992; Mitchell et al. 1992; Likens et
al. 1994; Johnson 1995). Lysimetry has been used
in experimental settings where the experiments,
such as fertilization with N, are performed on
stands smaller than a watershed (e.g., Currie et al.
1996; Christ et al. 1995; Rustad et al. 1996; Magill
et al. 1997; Gundersen 1998 and other nitrogen saturation experiment (NITREX) sites). In these studies, streamwater outputs would integrate over too
large an area, making the fertilization signal hard
to detect, and riparian or in-stream denitrification
and processing could also confound results. Lysimetry is also used when streams are not present
(e.g., Lajtha et al. 1995) due to high infiltration
rates. Lysimetry is perhaps most appropriate when
the soil solution per se needs to be measured in
studies of the terrestrial part of a watershed. For
example, questions concerning the role of plant uptake versus soil sorption in watershed-level studies
of nutrient retention might best be answered using
lysimeters so that confounding effects of stream or
riparian zone denitrification are omitted. However,
due to variable and significant in-stream processing
of some nutrients such as N, lysimeter results
should be used cautiously for predictions of stream
or drinking water quality.
The term lysimeter has had several meanings in
the soil science literature; for the purpose of this
chapter, a lysimeter is a sampler, inserted into the
soil, that collects in situ soil solution water. Lysimeters collect water either with (tension) or without
(zero-tension) applied tension to extract water. Tension lysimeters are generally smaller and relatively
easy to install. They are most commonly made of
ceramic or Teflon and glass mixes with a fine pore
size to filter the soil solution before it is extracted.
In theory, tension can be applied to match that of
the soil water moving via gravitational or matric
flow, and thus these lysimeters should collect water
from the soil matrix as well as saturated, or macropore flow. Zero-tension lysimeters (ZTLs) can only
collect saturated flow, although this is, presumably,
the majority of water flow through the soil. These
are constructed to have significantly larger collection areas than tension lysimeters, whose collection
area is not easily measured, and thus ZTLs have the
advantage of sampling a known, potentially quite
wide area. Because ZTLs should be relatively large,
Kate Lajtha
they are inserted into the soil via a soil pit adjacent
to the soil column that is under investigation.
There are advantages and disadvantages to each
type oflysimeter (summarized in Lajtha et al. 1999,
along with a general description of purchasing or
constructing different types of lysimeters). Zerotension lysimeters are more difficult to install and
cause more disturbance to a site, especially if they
are to be installed at depth. They tend to collect
significantly less soil water at depth, and thus investigators have generally restricted the use of
ZTLs to surface horizons. A number of studies have
compared concentrations of elements collected by
these different types of lysimeters, and summaries
can be found in Litaor (1988), Marques et al.
(1996), and Lajtha et al. (1999). Although several
studies have found that soil solutions collected with
tension lysimeters had higher concentrations of
many ions, as one would expect since they should
collect a more tightly bound fraction of soil water,
this varied a great deal depending on the ion examined and the site. In general, there have not been
clear patterns of differences between lysimeter
types across the many studies, although nitrate appears to often be elevated in tension versus zerotension lysimeters.
Lysimeters collect soil solution, and thus are not,
per se, flux measurements of ecosystem losses.
Even ZTLs only collect a variable fraction of total
solution flux. Measurements of solute concentration must be coupled with a water-balance model
in order to calculate fluxes below the rooting zone.
It is likely that this calculation introduces one of
the larger errors in lysimetry; it is certainly easier
to directly measure stream hydrology and water
chemistry than to model water flux. Coupled ecosystem and hydrologic models must, of course, be
customized to each site, and require significant
knowledge about the climatology of the site (e.g.,
PnET-II [Aber et al. 1995], WaVES [Dawes and
Hatton 1993]). Chloride and bromide have been
used as inert tracers through ecosystems as a way
of bypassing modeling water flux, although the CImethod has recently been called into question
(Chorover et al. 1994).
It is critical to realize that there are few reasons
to believe that soil solution leachates would or
should match patterns of stream water chemistry.
In theory, lysimeters placed below the rooting zone
of an ecosystem collect soil water that would oth-
or soil water collectors (e.g., Binkley et al. 1992;
Joslin et al. 1992; Joslin and Wolfe 1992; Johnson
and Lindberg 1992; Mitchell et al. 1992; Likens et
al. 1994; Johnson 1995). Lysimetry has been used
in experimental settings where the experiments,
such as fertilization with N, are performed on
stands smaller than a watershed (e.g., Currie et al.
1996; Christ et al. 1995; Rustad et al. 1996; Magill
et al. 1997; Gundersen 1998 and other nitrogen saturation experiment (NITREX) sites). In these studies, streamwater outputs would integrate over too
large an area, making the fertilization signal hard
to detect, and riparian or in-stream denitrification
and processing could also confound results. Lysimetry is also used when streams are not present
(e.g., Lajtha et al. 1995) due to high infiltration
rates. Lysimetry is perhaps most appropriate when
the soil solution per se needs to be measured in
studies of the terrestrial part of a watershed. For
example, questions concerning the role of plant uptake versus soil sorption in watershed-level studies
of nutrient retention might best be answered using
lysimeters so that confounding effects of stream or
riparian zone denitrification are omitted. However,
due to variable and significant in-stream processing
of some nutrients such as N, lysimeter results
should be used cautiously for predictions of stream
or drinking water quality.
The term lysimeter has had several meanings in
the soil science literature; for the purpose of this
chapter, a lysimeter is a sampler, inserted into the
soil, that collects in situ soil solution water. Lysimeters collect water either with (tension) or without
(zero-tension) applied tension to extract water. Tension lysimeters are generally smaller and relatively
easy to install. They are most commonly made of
ceramic or Teflon and glass mixes with a fine pore
size to filter the soil solution before it is extracted.
In theory, tension can be applied to match that of
the soil water moving via gravitational or matric
flow, and thus these lysimeters should collect water
from the soil matrix as well as saturated, or macropore flow. Zero-tension lysimeters (ZTLs) can only
collect saturated flow, although this is, presumably,
the majority of water flow through the soil. These
are constructed to have significantly larger collection areas than tension lysimeters, whose collection
area is not easily measured, and thus ZTLs have the
advantage of sampling a known, potentially quite
wide area. Because ZTLs should be relatively large,
Kate Lajtha
they are inserted into the soil via a soil pit adjacent
to the soil column that is under investigation.
There are advantages and disadvantages to each
type oflysimeter (summarized in Lajtha et al. 1999,
along with a general description of purchasing or
constructing different types of lysimeters). Zerotension lysimeters are more difficult to install and
cause more disturbance to a site, especially if they
are to be installed at depth. They tend to collect
significantly less soil water at depth, and thus investigators have generally restricted the use of
ZTLs to surface horizons. A number of studies have
compared concentrations of elements collected by
these different types of lysimeters, and summaries
can be found in Litaor (1988), Marques et al.
(1996), and Lajtha et al. (1999). Although several
studies have found that soil solutions collected with
tension lysimeters had higher concentrations of
many ions, as one would expect since they should
collect a more tightly bound fraction of soil water,
this varied a great deal depending on the ion examined and the site. In general, there have not been
clear patterns of differences between lysimeter
types across the many studies, although nitrate appears to often be elevated in tension versus zerotension lysimeters.
Lysimeters collect soil solution, and thus are not,
per se, flux measurements of ecosystem losses.
Even ZTLs only collect a variable fraction of total
solution flux. Measurements of solute concentration must be coupled with a water-balance model
in order to calculate fluxes below the rooting zone.
It is likely that this calculation introduces one of
the larger errors in lysimetry; it is certainly easier
to directly measure stream hydrology and water
chemistry than to model water flux. Coupled ecosystem and hydrologic models must, of course, be
customized to each site, and require significant
knowledge about the climatology of the site (e.g.,
PnET-II [Aber et al. 1995], WaVES [Dawes and
Hatton 1993]). Chloride and bromide have been
used as inert tracers through ecosystems as a way
of bypassing modeling water flux, although the CImethod has recently been called into question
(Chorover et al. 1994).
It is critical to realize that there are few reasons
to believe that soil solution leachates would or
should match patterns of stream water chemistry.
In theory, lysimeters placed below the rooting zone
of an ecosystem collect soil water that would oth-