4. Methods of Estimating Belowground Net Primary Production
69
have become the accepted norm for all ecological
papers published in the refereed literature. A similar standard for estimates of BNPP would provide
a useful way for readers to assess the importance
of particular estimates in the range of numbers they
encounter.
Currently, there is no standard by which small
and large estimates can be judged. We have been
told that more elaborate methods produce "better"
estimates (e.g., Long et al. 1989, 1992). However,
because these "better" estimates often entail making measurements of more variables, they often
have higher uncertainty associated with them.
There can be large differences among methods that
are the result of an interaction between the kind of
calculation made and the amount of variability in
the input data. The calculations made for the Publicover and Vogt (1993) method involved addition,
subtraction, and division by a constant. The relationship between the uncertainties of the input and
output variables was linear. In the cases of the nitrogen balance and minirhizotron methods, both involved division by a variable. In the case of the
nitrogen balance method, the denominator (nitrogen concentration) will always be very small compared with the numerator (amount of nitrogen allocated to fine roots), which means that the estimate
of NFRP will be very sensitive to the estimate of
the nitrogen concentration in fine roots. For the
minirhizotron method, the data from Hendrick and
Pregitzer (1993) suggest that the numerator and denominator will be of similar magnitudes, which
should increase the stability of the calculation.
Decisions about which method to choose to estimate BNPP should be based upon the objective
of the study, which will dictate the kind of estimate
that is required. The final choice should include a
consideration of the potential of the method to produce an estimate with an acceptable level of
uncertainty.
Summary
There is no best way to estimate BNPP. All of the
methods have conceptual strengths and weaknesses, numerous practical considerations, and differing relationships between variability of the input
variables and uncertainty in the estimate of BNPP.
Each of these factors should be evaluated relative
to the objectives of the particular project in arriving
at a decision about which method to use. In some
cases, the decision will be quite simple. The practical considerations in using the carbon isotope
turnover method in a mature forest will lead to its
quick elimination as a possibility. Other decisions
will be more difficult, such as deciding between the
biomass or minirhizotron methods in a grassland.
While there are substantial practical considerations
associated with each method, there are few that
completely overwhelm the decision-making process. The initial costs are different, the amounts and
types of field work required are different, the
amounts of analysis after the field sampling differ,
and the relationships between the variability of the
input variables and the uncertainty in the estimates
of BNPP differ between the methods. Each of these
factors will have to be considered in relation to the
project objectives before a decision can be reached.
Acknowledgments I thank Daniel Milchunas and
Osvaldo Sala for many stimulating discussions
about below ground net primary production. I also
thank Martyn Caldwell, Dennis Knight, F.w. Smith
and Daniel Milchunas for helpful comments on an
earlier draft. Funding for this work was provided
by the Shortgrass Steppe Long Term Ecological
Research project which is supported by the National Science Foundation (BSR 90-13888) and the
Colorado State University Experiment Station
(1-50661).
References
Aber, J.D.; Melillo, J.M.; Nadelhoffer, J.K.; McClaugherty, C.A.; Pastor, J. Fine root turnover in forest
ecosystems in relation to quantity and form of nitrogen
availability: A comparison of two methods. Oecologia
66:317-321; 1985.
Aerts, R.; Berendse, F.; Klerk, N.M.; Bakker, C. Root
production and root turnover in two dominant species
of wet heathlands. Oecologia 81:374-378; 1989.
Binkley, D.; Ryan, M.G. Net primary production and nutrient cycling in replicated stands of Eucalyptus saligna and Albizia facaltaria. For. Ecol. Manage.
112:79-85; 1998.
Bohm, W. Methods of Studying Root Systems. Berlin:
Springer-Verlag; 1979.
Box, G.E.P.; Hunter, W.G.; Hunter, 1.S. Statistics for experimenters: An introduction to design, data analysis,
and model building. New York: Wiley Sons; 1978:8789.
69
have become the accepted norm for all ecological
papers published in the refereed literature. A similar standard for estimates of BNPP would provide
a useful way for readers to assess the importance
of particular estimates in the range of numbers they
encounter.
Currently, there is no standard by which small
and large estimates can be judged. We have been
told that more elaborate methods produce "better"
estimates (e.g., Long et al. 1989, 1992). However,
because these "better" estimates often entail making measurements of more variables, they often
have higher uncertainty associated with them.
There can be large differences among methods that
are the result of an interaction between the kind of
calculation made and the amount of variability in
the input data. The calculations made for the Publicover and Vogt (1993) method involved addition,
subtraction, and division by a constant. The relationship between the uncertainties of the input and
output variables was linear. In the cases of the nitrogen balance and minirhizotron methods, both involved division by a variable. In the case of the
nitrogen balance method, the denominator (nitrogen concentration) will always be very small compared with the numerator (amount of nitrogen allocated to fine roots), which means that the estimate
of NFRP will be very sensitive to the estimate of
the nitrogen concentration in fine roots. For the
minirhizotron method, the data from Hendrick and
Pregitzer (1993) suggest that the numerator and denominator will be of similar magnitudes, which
should increase the stability of the calculation.
Decisions about which method to choose to estimate BNPP should be based upon the objective
of the study, which will dictate the kind of estimate
that is required. The final choice should include a
consideration of the potential of the method to produce an estimate with an acceptable level of
uncertainty.
Summary
There is no best way to estimate BNPP. All of the
methods have conceptual strengths and weaknesses, numerous practical considerations, and differing relationships between variability of the input
variables and uncertainty in the estimate of BNPP.
Each of these factors should be evaluated relative
to the objectives of the particular project in arriving
at a decision about which method to use. In some
cases, the decision will be quite simple. The practical considerations in using the carbon isotope
turnover method in a mature forest will lead to its
quick elimination as a possibility. Other decisions
will be more difficult, such as deciding between the
biomass or minirhizotron methods in a grassland.
While there are substantial practical considerations
associated with each method, there are few that
completely overwhelm the decision-making process. The initial costs are different, the amounts and
types of field work required are different, the
amounts of analysis after the field sampling differ,
and the relationships between the variability of the
input variables and the uncertainty in the estimates
of BNPP differ between the methods. Each of these
factors will have to be considered in relation to the
project objectives before a decision can be reached.
Acknowledgments I thank Daniel Milchunas and
Osvaldo Sala for many stimulating discussions
about below ground net primary production. I also
thank Martyn Caldwell, Dennis Knight, F.w. Smith
and Daniel Milchunas for helpful comments on an
earlier draft. Funding for this work was provided
by the Shortgrass Steppe Long Term Ecological
Research project which is supported by the National Science Foundation (BSR 90-13888) and the
Colorado State University Experiment Station
(1-50661).
References
Aber, J.D.; Melillo, J.M.; Nadelhoffer, J.K.; McClaugherty, C.A.; Pastor, J. Fine root turnover in forest
ecosystems in relation to quantity and form of nitrogen
availability: A comparison of two methods. Oecologia
66:317-321; 1985.
Aerts, R.; Berendse, F.; Klerk, N.M.; Bakker, C. Root
production and root turnover in two dominant species
of wet heathlands. Oecologia 81:374-378; 1989.
Binkley, D.; Ryan, M.G. Net primary production and nutrient cycling in replicated stands of Eucalyptus saligna and Albizia facaltaria. For. Ecol. Manage.
112:79-85; 1998.
Bohm, W. Methods of Studying Root Systems. Berlin:
Springer-Verlag; 1979.
Box, G.E.P.; Hunter, W.G.; Hunter, 1.S. Statistics for experimenters: An introduction to design, data analysis,
and model building. New York: Wiley Sons; 1978:8789.