BORROW STRENGTH APPROACH APPLIED TO A GEOSTATISTICAL MODEL TO ESTIMATE VOLUME
DOI:
https://doi.org/10.5902/1980509827739Keywords:
maximum likelihood, precision silviculture, Akaike criterion, forest inventory.Abstract
This study aimed to use the share parameters of the geo-statistical models applied to maximum likelihood estimators to predict the volumes per hectare in three fragments of a Deciduous Forest located in Santa Teresa, RS state, employing the ‘Borrow Strength’ approach. Data were collected in 56 sampling units (S.U) of variable sizes with approximately 250 m2 for a total of nine ha, distributed in a systematic grid of 40 x 40 m. Dendrometric variables from individuals with DBH ≥ 10 cm near the center of the S.U. were measured. Two approaches to the data set were prepared, the first of which considering both areas entirely independent themselves, subdivided into two types: a fit to non-spatial model (NSM) and a fit to the maximum likelihood (ML) not shared (individual adjustment) model. The second approach described the adjustment of the shared as a function of random error or nugget, comprising models: a shared model without fixed nugget (variability between S.U) and a shared model with fixed nugget (variability within S.U) models, using a logarithmic function of M.L applied to the Matèrn family of exponential correlation model. Then, the models were compared using Akaike information criterion (AIC) and by degree of spatial dependence for subsequent preparation of both kriging and prediction surfaces of the selected models. It was observed that the combined volume models to estimate values were higher for the AIC values and spatial dependence with respect to the adjustments for the individual areas. Among the shared models, it was observed that there was a gain in the parameter estimates using the fixed nugget, which resulted in a higher correlation of samples and spatial dependence (AP = 88 m), than the shared models without the fixed nugget (AP = 75 and 66 m). The AIC was efficient because it compared the different levels of proposed adjustments to the methodology of the study, selecting a model with parsimony and compatible with the spatial distribution patterns found in the areas. The use of combined models for data sampling in different areas with the introduction of the error estimate intra-plot (fixed nugget) in the equations of MV can be suggested to increase the correlation between the S.U and combined evaluation of the AIC plus the degree of spatial dependence in estimating dendrometric variables.
Downloads
References
AKAIKE, H. A new look at the statistical model identification. IEEE Transactions on Automatic Control, Boston, v. 19, n. 6, p. 716–723, dec. 1974.
BAS, D.; VAN PUIJENBROEK, R. A. G. Conditional Regression Analysis. Quality & Quantity, Dordrecht, v. 23, n. 1, p. 83-108, feb. 1989.
BOX, G. E. P.; COX, D. R. An Analysis of Transformations. Journal of the Royal Statistical Society, London, v. 26, n. 2, p. 211-252, 1964.
BUONGIORNO, J.; GILLESS, J. K. Decision methods for forest resource management. San Diego: Academic Press, 2003. 429 p.
BRILLINGER, D. Selected works in probability and statístics. Oslo: Springer, 1997. 609 p.
BURNHAM, P. K.; ANDERSON, D. R. Model selection and multimodel inference – a pratical information - theroretic approach. 2. ed. Colorado: Springer, 2002. 347 p.
DIGGLE, P. J.; RIBEIRO JUNIOR, P. J. Model-Based geostatistics. New York: Springer, 2007. 228 p.
EMBRAPA. Sistema brasileiro de classificação de solos. 3. ed. Brasília: EMBRAPA, 2013. 353 p.
FINLEY, A. O.; BANERJEE, S.; MACFARLANE, D. W. A hierarchical model for quantifying forest variables over large heterogeneous landscapes with uncertain forest areas. Michigan: Department of Forestry, 2010. 88 p.
GREGOIRE, T. G.; SCHABENBERGER, O. Fitting bole-volume equations to spatially correlated within-tree data. In: ANNUAL CONFERENCE OF APPLIED STATISTICS IN AGRICULTURE, 6., 1994, Manhatan. Procedings… Kansas, 1994. p. 120-133.
KANGAS, A.; MALTANO, M. Forest inventory – methodology and applications. New York: Springer, 2006. 350 p.
MACHADO, S. A.; FIGUEIREDO FILHO, A. Dendrometria. Curitiba: Os autores, 2003. 309 p.
MALCZEWSKI, J. GIS and multicriteria decision analysis. New York: Wiley, 1999. 298 p.
MATÈRN, B. Spatial Variation. Berlin: Springer, 1986. 594 p.
MATTHIOPOUOS, J. How to be a quantitative ecologist – the ‘A to R’ of green mathematics and statistics. Scotland: Wiley, 2011. 1091 p.
MELLO, J. M. Geoestatística aplicada ao inventário florestal. 2004. 110 f. Tese (Doutorado em Recursos Florestais) - Escola Superior de Agricultura Luiz de Queiroz, Piracicaba, 2004.
MORENO, J. A. Clima do Rio Grande do Sul. Porto Alegre: Secretaria da Agricultura, 1961. 73 p.
PEBESMA, E. J.; BIVAND, R. S. Classes and methods for spatial data in R. R News, Ontario, v. 5, n. 2, 2005.
PINHEIRO, J.; BATES, D. Mixed-Effects Models in S and S-PLUS. New York: Springer, 2000. 530 p.
PRODAN, M. Holzmesslehre. Frankfurt: Sauerlände’s Verlag, 1965. 644 p.
R CORE TEAM. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing, 2013. Disponível em: <http://www.R-project.org/>. Acesso em: 12 abr. 2014.
RIBEIRO JÚNIOR, P. J. Métodos geoestatísticos no estudo da variabilidade espacial dos parâmetros do solo. 1995. 99 f. Dissertação (Mestrado) - Escola Superior de Agricultura "Luiz de Queiroz", Universidade de São Paulo, Piracicaba, 1995.
RIBEIRO JUNIOR, P. J.; DIGGLE, P. J. The geoR package functions for geostatistical data analysis. R News, Ontario, v. 1, n. 2, p. 15-18, 2001.
RONGXIA. L.; PETE, B.; KEVIN, B. Informed development of meta heuristics for spatial forest planning problems. The Open Operational Research Journal, Sharjah, n. 4, p. 1-11, 2010.
RIO GRANDE DO SUL. Secretaria Estadual do Meio Ambiente. Relatório Final do Inventário Florestal Contínuo do Rio Grande do Sul. Porto Alegre: SEMA/RS, 2001. 2 v. 706 p.
VENABLES, W. N.; RIPLEY, B. D. Modern applied statistics with S. 4th ed. New York: Springer, 2002.
