Appropriate search techniques to estimate Weibull function parameters in a Pinus spp. plantation

Abstract

The Weibull function, a continuous probability distribution, is widely used for diameter distribution modelling, in which parameter estimation performance is affected by stand attributes and fitting methods. The Weibull cumulative distribution function is nonlinear, and classical fitting methods may provide a not optimal solution. Invoking the use of artificial intelligence by metaheuristics is reasonable for this optimisation task. Therefore, aimed and compared (1) the metaheuristics genetic algorithm and simulated annealing performance over the moment and percentile methods; (2) the hybrid strategy merging the metaheuristics tested and the percentile method and, (3) the metaheuristics fitness functions under basal area and density errors. A long-term experiment in a Pinus taeda stand subjected to crown thinning was used. According to our findings, all methods have a similar performance, independent of the thinning regimes and age. The pattern of the estimated parameters tends to be acceptable, as b increases over time and c increases after thinning. Overall, our findings suggest that methods based on metaheuristics have a higher precision than classical methods for estimating Weibull parameters. According to the classification test, the methods that involved simulated annealing stood out. The hybrid method involving this metaheuristic also stood out, with accurate estimates. Classical methods showed the poorest performance, and we therefore suggest the use of simulated annealing due to its faster processing time and high-quality solution.

Keywords

Forest management / Diameters distribution / Weibull function / Genetic algorithm / Simulated annealing

References

[1] Aakala T, Fraver S, D’Amato AW, Palik BJ. Influence of competition and age on tree growth in structurally complex old-growth forests in northern Minnesota, USA. For Ecol Manag, 2013, 308: 128-135.

[2] Abbasi B, Eshragh Jahromi AH, Arkat J, Hosseinkouchack M. Estimating the parameters of Weibull distribution using simulated annealing algorithm. Appl Math Comput, 2006, 183: 85-93.

[3] Adeyemi AA, Adesoye PO. Evaluation of some Weibull parameter estimation methods for characterizing stem diameter distribution in a Tropical Mixed Forest of Southern Nigeria. J Stat Sci Appl, 2016, 4: 257-275.

[4] Akgül FG, Şenoğlu B, Arslan T. An alternative distribution to Weibull for modeling the wind speed data: inverse Weibull distribution. Energy Convers Manag, 2016, 114: 234-240.

[5] Alvares CA, Stape JL, Sentelhas PC, Moraes Gonçalves JL, Sparovek G. Köppen’s climate classification map for Brazil. Meteorol Z, 2013, 22: 711-728.

[6] Arias-Rodil M, Diéguez-Aranda U, Álvarez-González JG, Pérez-Cruzado C, Castedo-Dorado F, González-Ferreiro E. Modeling diameter distributions in radiata pine plantations in Spain with existing countrywide LiDAR data. Ann For Sci, 2018, 75: 1-12.

[7] Assad A, Deep K. A hybrid harmony search and simulated annealing algorithm for continuous optimization. Inf Sci (Ny), 2018, 450: 246-266.

[8] Bailey R, Dell T. Quantifying diameter distributions with the Weibull function. For Sci, 1973, 19: 97-104.

[9] Berger RW, Lawrence K. Estimating Weibull parameters by linear and nonlinear regression. Technometrics, 1974, 16: 617-619.

[10] Bettinger P, Graetz D, Boston K, Sessions J, Chung W. Eight heuristic planning techniques applied to three increasingly difficult wildlife planning problems. Silva Fenn, 2002, 36: 561-584.

[11] Bullock BP, Burkhart HE. Juvenile diameter distributions of loblolly pine characterized by the two-parameter Weibull function. New For, 2005, 29: 233-244.

[12] Burkhart HE, Tomé M. Modeling forest trees and stands, 2012, Dordrecht: Springer 261 297

[13] Calzado-Carretero A, Torres-Alvarez E. Modelling diameter distributions of Quercus suber L. stands in “Los Alcornocales” Natural Park (Cádiz-Málaga, Spain) by using the two-parameter Weibull function. For Syst, 2013, 22: 15-24.

[14] Cao QV. Predicting parameters of a weibull function for modeling diameter distribution. For Sci, 2004, 50: 682-685.

[15] Cheong KH, Koh JM. A hybrid genetic-Levenberg Marquardt algorithm for automated spectrometer design optimization. Ultramicroscopy, 2019, 202: 100-106.

[16] Courchelle V, Soler M, González-Arribas D, Delahaye D. A simulated annealing approach to 3D strategic aircraft deconfliction based on en-route speed changes under wind and temperature uncertainties. Transp Res Part C Emerg Technol, 2019, 103: 194-210.

[17] Craine JM, Dybzinski R. Mechanisms of plant competition for nutrients, water and light. Funct Ecol, 2013, 27: 833-840.

[18] Diamantopoulou MJ, Özçelik R, Crecente-Campo F, Eler Ü. Estimation of Weibull function parameters for modelling tree diameter distribution using least squares and artificial neural networks methods. Biosyst Eng, 2015, 133: 33-45.

[19] Fonton NH, Atindogbé G, Akossou AY, Missanon BT, Fadohan B, Lejeune P. Modeling the distribution of marketable timber products of private teak (Tectona grandis L.f.) plantations. Open J For, 2013, 3: 115-121.

[20] Franzin A, Stützle T. Revisiting simulated annealing: a component-based analysis. Comput Oper Res, 2019, 104: 191-206.

[21] Guevara M, Olmedo GF. Model evaluation in digital soil mapping. In: soil organic carbon mapping cookbook, 2018, Rome: Food and Agriculture Organization of the United Nations (FAO) 220

[22] Holland JH. Adaptation in natural and artificial systems, 1975, Ann Arbor: University Michigan Press 183

[23] Huang JR. Studies on the suitability of Weibull distribution in Masson pine plantations. Guizhou For Sci Technol, 2000, 28: 7-13.

[24] Jin YF, Yin ZY, Shen SL, Hicher PY. Selection of sand models and identification of parameters using an enhanced genetic algorithm. Int J Numer Anal Methods Geomech, 2016, 40: 1219-1240.

[25] Křivý I, Tvrdík J. The controlled random search algorithm in optimizing regression models. Comput Stat Data Anal, 1995, 20: 229-234.

[26] Leduc DJ, Matney TG, Belli KL, Baldwin VCJ. Predicting diameter distributions of longleaf pine plantations: a comparison between artificial neural networks and other accepted methodologies. USDA For Serv Res Pap, 2001, SRS-25: 1-18.

[27] Lei Y. Evaluation of three methods for estimating the Weibull distribution parameters of Chinese pine (Pinus tabulaeformis ). J For Sci, 2008, 54: 566-571.

[28] Li X, Liang E, Gricar J, Prislan P, Rossi S, Cufar K. Age dependence of xylogenesis and its climatic sensitivity in Smith fir on the south-eastern Tibetan Plateau. Tree Physiol, 2013, 33: 48-56.

[29] Liao TW, Li G. Metaheuristic-based inverse design of materials – a survey. J Mater, 2020, 6: 414-430.

[30] Liu C, Beaulieu J, Prégent G, Zhang SY. Applications and comparison of six methods for predicting parameters of the Weibull function in unthinned Picea glauca plantations. Scand J For Res, 2009, 24: 67-75.

[31] Loetsch F, Zöhrer F, Haller KE. Forest inventory, 1973 2 BLV Verlagsgesellschaft: Munich 469

[32] Mafarja MM, Mirjalili S. Hybrid whale optimization algorithm with simulated annealing for feature selection. Neurocomputing, 2017, 260: 302-312.

[33] Maltamo M, Puumalainen J, Päivinen R. Comparison of beta and weibull functions for modelling basal area diameter distribution in stands of Pinus sylvestris and Picea abies. Scand J For Res, 1995, 10: 284-295.

[34] Maltamo M, Mehtätalo L, Valbuena R, Vauhkonen J, Packalen P. Airborne laser scanning for tree diameter distribution modelling: a comparison of different modelling alternatives in a tropical single-species plantation. Forestry, 2018, 91: 121-131.

[35] Millard SP (2018) EnvStats: package for environmental statistics. https://CRAN.R-project.org/package=EnvStats

[36] Nokoe S, Okojie JA. Relationship of stand attributes of some plantation mahoganies with estimated weibull parameters. Ecol Model, 1984, 24: 231-240.

[37] Nord-Larsen T, Cao QV. A diameter distribution model for even-aged beech in Denmark. For Ecol Manag, 2006, 231: 218-225.

[38] Ooi H (2019) doParallel: Foreach Parallel Adaptor for the ‘parallel’ package. https://CRAN.R-project.org/package=doParallel

[39] Ou Q, Lei X, Shen C. Individual tree diameter growth models of larch–spruce–fir mixed forests based on Machine Learning Algorithms. Forests, 2019, 10: 1-20.

[40] Pogoda P, Ochał W, Orzeł S. Modeling diameter distribution of black alder (Alnus glutinosa (L.) Gaertn.) stands in Poland. Forests, 2019, 10: 1-16.

[41] Poudel KP, Cao QV. Evaluation of methods to predict Weibull parameters for characterizing diameter distributions. For Sci, 2013, 59: 243-252.

[42] Pretzsch H. Forest dynamics, growth and yield, 2009, Heidelberg: Springer 423 491

[43] R CORE TEAM. R: a language and environment for statistical computing, 2019, Vienna: R Foundation for Statistical Computing . http://www.R-project.org

[44] Razali AM, Salih AA, Mahdi AA. Estimation accuracy of Weibull distribution parameters. J Appl Sci Res, 2009, 5: 790-795.

[45] Reynolds MR, Burk TE, Huang W-C. Goodness-of-fit tests and model selection procedures for diameter distribution models. For Sci, 1988, 34: 373-399.

[46] Rinne H. The Weibull distribution: a handbook, 2008, New York: Chapman and Hall/CRC 808

[47] Rouvinen S, Kuuluvainen T. Tree diameter distributions in natural and managed old Pinus sylvestris-dominated forests. For Ecol Manag, 2005, 208: 45-61.

[48] Sánchez-Salguero R, Linares JC, Camarero JJ, Madrigal-González J, Hevia A, Sánchez-Miranda Á, Ballesteros-Cánovas JA, Alfaro-Sánchez R, García-Cervigón AI, Bigler C, Rigling A. Disentangling the effects of competition and climate on individual tree growth: a retrospective and dynamic approach in Scots pine. For Ecol Manag, 2015, 358: 12-25.

[49] Shifley S, Lentz E. Quick estimation of the three-parameter Weibull to describe tree size distributions. For Ecol Manag, 1985, 13: 195-203.

[50] Shiver BD. Sample sizes and estimation methods for the Weibull distribution for unthinned slash pine plantation diameter distributions. For Sci, 1988, 34: 809-814.

[51] Stankova TV, Zlatanov TM. Modeling diameter distribution of Austrian black pine (Pinus nigra Arn.) plantations: a comparison of the Weibull frequency distribution function and percentile-based projection methods. Eur J For Res, 2010, 129: 1169-1179.

[52] Tashkova K, Šilc J, Atanasova N, Džeroski S. Parameter estimation in a nonlinear dynamic model of an aquatic ecosystem with meta-heuristic optimization. Ecol Model, 2012, 226: 36-61.

[53] Taylor KE. Summarizing multiple aspects of model performance in a single diagram. J Geophys Res Atmos, 2001, 106: 7183-7192.

[54] Vanclay JK. Modelling forest growth and yield: applications to mixed tropical forests, 1994, Wallingford: CAB International 1 13

[55] Weber MD, Leemis LM, Kincaid RK. Minimum Kolmogorov–Smirnov test statistic parameter estimates. J Stat Comput Simul, 2006, 76: 195-206.

[56] Weiskittel AR, Hann DW, Kershaw JA, Vanclay JK. Forest growth and yield modeling, 2011, Chichester: Wiley 16 36

[57] Wen CJW, Liu X, Cheng X. Parameter evaluation of 3-parameter Weibull distribution based on adaptive genetic algorithm. Proceedings of the 2017 2nd international conference on machinery, electronics and control simulation (MECS 2017), 2017, Paris: Atlantis Press 426 431

[58] Yalçınkaya A, Şenoğlu B, Yolcu U. Maximum likelihood estimation for the parameters of skew normal distribution using genetic algorithm. Swarm Evol Comput, 2018, 38: 127-138.