I used the above mentioned equation to calculate biomass of a species. I have dbh data from the forest and biomass tables from previous work. The table provides biomass for foliage, branches and stems which were calculated using above mentioned equation. Tree dbh 2 to 25cm were used to calculate biomass for tree components but they did not provide log correction and constant value "a" and "b". I used all values to calculate constant values "a" and "b" but could not manage to find that. Maybe I need to make a log correction, could anybody help how to find log correction?
I do not get exactly your point. Are you asking about bias correction when backtransforming the log estimates for different biomass components?
Then, you have several options, that is, a number of correction factors that may be used. One bias correction factor widely utilized is the one proposed by Baskerville (1972) (Use of logarithmic regression in the estimation of plant biomass. Canadian Journal of Forest Research 2: 49-53.). It is calculated as the standard error (SE) of the regression to the power of 2 divided by 2, that is, cf = SE^2/2. This correction factor is applied when backtransforming from the log scale to the original units, that is, when you apply the exponential function: EXP(a+b*ln(dbh)+SE^2/2). I have used it in several papers. Hope this helps.
I have used it for example in this paper: Shater, Z.; de Miguel, S.; Kraid, B.; Pukkala, T.; Palahí, M. 2011. A growth and yield model for even-aged Pinus brutia Ten. stands in Syria. Annals of Forest Science, 68(1): 149-157. In this particular case, Baskerville's correction factor was applied to a self-thinning model to predict survival/mortality of a pine species.
Nevertheless, in practice, I usually try to avoid the use of correction factors by directly fitting nonlinear models in the original scale (that is, without any log transformation). If you are interested, I have used this approach in several papers as well.
You can check following literature.
Patrick Addo-Fordjour and Zakaria B. Rahmad. Development of Allometric Equations for Estimating Above-Ground Liana Biomass in Tropical Primary and Secondary Forests, Malaysia. International Journal of Ecology. doi: 10.1155/2013/658140
Joseph Mascaro, Creighton M. Litton, R. Flint Hughes, Amanda Uowolo, and Stefan A. Schnitzer. Minimizing Bias in Biomass Allometry: Model Selection and Log-Transformation of Data. BIOTROPICA. DOI:10.1111/j.1744-7429.2011.00798.x
Thank you Sergio
A question?
Do you mean: correction factor should be multiplied with model parameters (constant, b1)?
If RMSE is used to compare models, is it necessary to multiply correction factor in RMSE?
- Badges
- Science topic
- Similar topics
- Correction