19. Nutrient Manipulations in Terrestrial Ecosystems
Isotopically labeled fertilizers are most commonly enriched. The minimum level of enrichment
needed depends on how subtle the effects of a treatment are, and the rates of cycling and loss of nutrients, especially in relation to how long after addition the effects must be detected. For short-term
experiments looking at more obvious changes, the
less expensive isotopically depleted fertilizers or
low-enrichment fertilizers can be used (see Power
et al. 1986).
Nitrogen and S have stable isotopes that are valuable in fertilizer-labeled studies as well as in tracing
the fate of atmospheric deposition. The radioactive
N isotope has too short of a half-life to be useful in
fertilizer applications. Phosphorus and S also have
radioactive forms. The P isotope is short-lived, and
is generally only used for short-term partitioning
experiments and to assess soil adsorption (He and
Zhu 1998). For a review of P isotope techniques,
see Di et al. (1997). Of all of the nutrients, P fertilizer tends to be retained in the system for the
longest amount of time, but due to the lack of a
stable isotope, and the short half-lives ofthe radioactive isotopes, it is impossible to trace the fate of
P fertilizer for any appreciable time.
Sulfur has both radioactive and stable isotopes,
allowing much more flexibility in the types of questions that can be asked. Either form can be used to
track the fate of fertilizer S (Shock et al 1983) and
to determine the effect of S fertilization on S dynamics (Eriksen 1996; Ghani et al. 1993).
Nitrogen-IS is the most common isotopically labeled fertilizer used. It has been used in many studies to elucidate fate (Nannipieri et a11985; Bristow
et al. 1987), adsorption (Foster et al 1985), and
long-term versus short-term fertilizer availability
and dynamics (Preston and Mead 1994; Chabrol et
al. 1988; see Schimel 1993 for a good review). It
has also been used to trace the fate of fertilizer nutrients in litter over time (Power et al. 1986; Muller
1988). (For a review of 15N fertilizer methods, see
Hauck et al. 1994.)
The use of 15N fertilizers has led to interesting
findings. For example, it is often assumed that microbial immobilization of fertilizer is a short-term
mechanism for ecosystem retention, and microbial
N will eventually tum over and become available
to plants. Several studies tracing the fate of 15N
fertilizer did not see this microbial N being made
available to plants, even over an extended period
301
of time (Proctor and Mead 1994; Rutherford and
Juma 1992). Another surprising result from 15N
studies was that the ecosystem retention of 15N in
a jack pine forest was dominated by chemical fixation of 15NH3, which was substantially higher than
microbial immobilization (Foster et al. 1985).
While isotopes can be invaluable tools, interpretations must be made carefully. Studies of 15N fertilizers have often shown that fertilizing increases
plant uptake of unlabeled soil N, which has been
attributed to increases in soil organic N mineralization (Pilbeam et al 1997) and increases in plant
uptake rates (Leon et al. 1995). These increases in
native organic N uptake may not necessarily be due
to increased availability or uptake, but may be an
artifact of fertilizer application. Microbes may have
immobilized the flush of 15N fertilizer, allowing native soil N that would normally be immobilized to
be taken up by plants. This pool substitution could
lead to an overestimate of the effect of fertilizer on
N cycling, and an underestimate of nutrient uptake
by plants in response to fertilization. (Hart et al
1986; Jenkinson et al. 1985). Interpretation of isotope data must also consider isotope fractionation
in uptake and soil processes, especially when using
depleted, or low-enrichment fertilizer. This is especially true in systems with even small amounts
of NH3 volatilization, which discriminates very
strongly against 15N, and will enrich the soil 15N
pool relative to the unlabeled pool. In such systems,
the use of 15N03 rather than urea or ammonium is
recommended (Nommik et alI994).
These problems in interpretation of isotopic fertilization experiments can be minimized by using
trace levels of highly enriched label to trace the
natural movement of nutrients in ecosystems. For
example, to resolve the question of whether the increase in unlabeled plant N described above is an
artifact of fertilizer additions, or a result of N addition altering N cycling, unlabeled fertilizer additions could be followed by trace amounts of highly
enriched 15N to document nutrient partitioning and
gross rates of nutrient transformations. Tracer additions should be high enough for the isotope to be
detectable for the duration of the experiment, but
not so high as to perturb natural process rates by
"fertilization" (Jackson et al. 1989; Davidson et al.
1991). Low-level 15 N addition can determine rates
of microbial immobilization in the absence of roots
(Davidson et al. 1991; Hart et al. 1994); the parti-
Isotopically labeled fertilizers are most commonly enriched. The minimum level of enrichment
needed depends on how subtle the effects of a treatment are, and the rates of cycling and loss of nutrients, especially in relation to how long after addition the effects must be detected. For short-term
experiments looking at more obvious changes, the
less expensive isotopically depleted fertilizers or
low-enrichment fertilizers can be used (see Power
et al. 1986).
Nitrogen and S have stable isotopes that are valuable in fertilizer-labeled studies as well as in tracing
the fate of atmospheric deposition. The radioactive
N isotope has too short of a half-life to be useful in
fertilizer applications. Phosphorus and S also have
radioactive forms. The P isotope is short-lived, and
is generally only used for short-term partitioning
experiments and to assess soil adsorption (He and
Zhu 1998). For a review of P isotope techniques,
see Di et al. (1997). Of all of the nutrients, P fertilizer tends to be retained in the system for the
longest amount of time, but due to the lack of a
stable isotope, and the short half-lives ofthe radioactive isotopes, it is impossible to trace the fate of
P fertilizer for any appreciable time.
Sulfur has both radioactive and stable isotopes,
allowing much more flexibility in the types of questions that can be asked. Either form can be used to
track the fate of fertilizer S (Shock et al 1983) and
to determine the effect of S fertilization on S dynamics (Eriksen 1996; Ghani et al. 1993).
Nitrogen-IS is the most common isotopically labeled fertilizer used. It has been used in many studies to elucidate fate (Nannipieri et a11985; Bristow
et al. 1987), adsorption (Foster et al 1985), and
long-term versus short-term fertilizer availability
and dynamics (Preston and Mead 1994; Chabrol et
al. 1988; see Schimel 1993 for a good review). It
has also been used to trace the fate of fertilizer nutrients in litter over time (Power et al. 1986; Muller
1988). (For a review of 15N fertilizer methods, see
Hauck et al. 1994.)
The use of 15N fertilizers has led to interesting
findings. For example, it is often assumed that microbial immobilization of fertilizer is a short-term
mechanism for ecosystem retention, and microbial
N will eventually tum over and become available
to plants. Several studies tracing the fate of 15N
fertilizer did not see this microbial N being made
available to plants, even over an extended period
301
of time (Proctor and Mead 1994; Rutherford and
Juma 1992). Another surprising result from 15N
studies was that the ecosystem retention of 15N in
a jack pine forest was dominated by chemical fixation of 15NH3, which was substantially higher than
microbial immobilization (Foster et al. 1985).
While isotopes can be invaluable tools, interpretations must be made carefully. Studies of 15N fertilizers have often shown that fertilizing increases
plant uptake of unlabeled soil N, which has been
attributed to increases in soil organic N mineralization (Pilbeam et al 1997) and increases in plant
uptake rates (Leon et al. 1995). These increases in
native organic N uptake may not necessarily be due
to increased availability or uptake, but may be an
artifact of fertilizer application. Microbes may have
immobilized the flush of 15N fertilizer, allowing native soil N that would normally be immobilized to
be taken up by plants. This pool substitution could
lead to an overestimate of the effect of fertilizer on
N cycling, and an underestimate of nutrient uptake
by plants in response to fertilization. (Hart et al
1986; Jenkinson et al. 1985). Interpretation of isotope data must also consider isotope fractionation
in uptake and soil processes, especially when using
depleted, or low-enrichment fertilizer. This is especially true in systems with even small amounts
of NH3 volatilization, which discriminates very
strongly against 15N, and will enrich the soil 15N
pool relative to the unlabeled pool. In such systems,
the use of 15N03 rather than urea or ammonium is
recommended (Nommik et alI994).
These problems in interpretation of isotopic fertilization experiments can be minimized by using
trace levels of highly enriched label to trace the
natural movement of nutrients in ecosystems. For
example, to resolve the question of whether the increase in unlabeled plant N described above is an
artifact of fertilizer additions, or a result of N addition altering N cycling, unlabeled fertilizer additions could be followed by trace amounts of highly
enriched 15N to document nutrient partitioning and
gross rates of nutrient transformations. Tracer additions should be high enough for the isotope to be
detectable for the duration of the experiment, but
not so high as to perturb natural process rates by
"fertilization" (Jackson et al. 1989; Davidson et al.
1991). Low-level 15 N addition can determine rates
of microbial immobilization in the absence of roots
(Davidson et al. 1991; Hart et al. 1994); the parti-