Modelling joint distribution of tree diameter and height using Frank and Plackett copulas

Abstract

Bivariate distribution models are veritable tools for improving forest stand volume estimations. Their accuracy depends on the method of construction. To-date, most bivariate distributions in forestry have been constructed either with normal or Plackett copulas. In this study, the accuracy of the Frank copula for constructing bivariate distributions was assessed. The effectiveness of Frank and Plackett copulas were evaluated on seven distribution models using data from temperate and tropical forests. The bivariate distributions include: Burr III, Burr XII, Logit-Logistic, Log-Logistic, generalized Weibull, Weibull and Kumaraswamy. Maximum likelihood was used to fit the models to the joint distribution of diameter and height data of Pinus pinaster (184 plots), Pinus radiata (96 plots), Eucalyptus camaldulensis (85 plots) and Gmelina arborea (60 plots). Models were evaluated based on negative log-likelihood (−ΛΛ). The result show that Frank-based models were more suitable in describing the joint distribution of diameter and height than most of their Plackett-based counterparts. The bivariate Burr III distributions had the overall best performance. The Frank copula is therefore recommended for the construction of more useful bivariate distributions in forestry.

Keywords

Bivariate distributions / Frank copula / Plackett copula / Diameter / Height

References

[1] Burr IW. Cumulative frequency functions. Ann Math Stat, 1942, 13: 215-232.

[2] Fisher NI (1997) Copulas. In: Encyclopedia of statistical science, update vol 1. Wiley, New York, pp 159–163

[3] Frank MJ. On the simultaneous associativity of F(x, y) and x +  y −  F(x, y). Aequ Math, 1979, 19: 194-226.

[4] Genest C, Mackay J. The joy of copulas: bivariate distributions with uniform marginals. Am Stat, 1986, 40: 280-283.

[5] Genest C, Rivest LP. Statistical inference procedures for bivariate Archimedean copulas. J Am Stat Assoc, 1993, 88(423): 1034-1043.

[6] Gorgoso-Varela JJ, Garcia-Villabrille JD, Rojo-Alboreca A, Kv Gadow, Alvarez-Gonzalez JG. Comparing Johnson’s S_BB, Weibull and Logit-Logistic bivariate distributions for modeling tree diameters and heights using copulas. For Syst, 2016, 25(1): 1-5.

[7] Gove JH, Ducey MJ, Leak WB, Zhang L. Rotated sigmoid structures in managed uneven-aged northern hardwood stands: a look at the Burr type III distribution. Forestry, 2008, 81(2): 161-176.

[8] Gumbel EJ. Bivariate logistic distributions. J Am Stat Assoc, 1961, 56: 335-349.

[9] Hafley WL, Schreuder HT. A useful bivariate distribution for describing stand structure of tree heights and diameters. Biometrics, 1977, 33: 471-478.

[10] Jin XJ, Li FR, Jia WW, Zhang LJ. Modelling and predicting bivariate distributions of tree diameter and height. Scientia Silvae Sinicae (Lin Ye Ke Xue), 2013, 49(6): 74-82.

[11] Johnson NL. Bivariate distributions based on simple translation systems. Biometrika, 1949, 36: 297-304.

[12] Knoebel BR, Burkhart HE. A bivariate distribution approach to modeling forest diameter distributions at two points in time. Biometrics, 1991, 47: 241-253.

[13] Kumaraswamy P. A generalized probability density function for double-bounded random processes. J Hydrol, 1980, 46(1–2): 79-88.

[14] Li FS, Zhang LJ, Davis CJ. Modelling the joint distribution of tree diameters and heights by bivariate generalized beta distribution. For Sci, 2002, 48(1): 47-58.

[15] Lindsay SR, Wood GR, Woollons RC. Modelling the diameter distribution of forest stands using the Burr distribution. J Appl Stat, 1996, 23(6): 609-619.

[16] Mardia KV. Families of bivariate distributions, 1970, London: Griffin.

[17] Mitková VB, Halmová D. Joint modelling of flood peak discharges, volume, and duration: a case study of the Danube River in Bratislava. J. Hydrol Hydromech, 2014, 62(3): 186-196.

[18] Mønness E. The bivariate power-normal and the bivariate Johnson’s system bounded distribution in forestry, including height curves. Can J For Res, 2015, 45: 307-313.

[19] Nelson RB. An introduction to copulas, 2006 2 New York: Springer.

[20] Ogana FN, Osho JSA, Gorgoso-Varela JJ. An approach to modeling the joint distribution of tree diameter and height data. J Sustain For, 2018, 37(5): 475-488.

[21] Plackett RL. A class of bivariate distributions. J Am Stat Assoc, 1965, 60: 516-522.

[22] R Core Team (2016) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/. Accessed 30 June 2017

[23] Wang ML (2005) Distributional modeling in forestry and remote sensing. PhD thesis, University of Greenwich, UK

[24] Wang ML, Rennolls K. Tree diameter distribution modelling: introducing the Logit-Logistic distribution. Can J For Res, 2005, 35: 1305-1313.

[25] Wang ML, Rennolls K. Bivariate distribution modelling with tree diameter and height data. For Sci, 2007, 53(1): 16-24.

[26] Wang ML, Rennolls K, Tang S. Bivariate distribution modelling with tree diameter and height: dependency modelling using copulas. For Sci, 2008, 54(3): 284-293.

[27] Wang ML, Upadhay A, Zhang LJ. Trivariate distribution modeling of tree diameter, height, and volume. For Sci, 2010, 56(3): 290-300.

[28] Weibull W. A statistical distribution function of wide applicability. J Appl Mech, 1951, 18: 293-297.

[29] Zucchini W, Schmidt M, Kv Gadow. A model for the diameter-height distribution in an uneven-aged beech forest and a method to assess the fit of such models. Silva Fenn, 2001, 35(2): 169-183.