Modelling crown diameter for Sal trees (Shorea robusta Gaertn. f.) with an artificial neural network
DOI:
https://doi.org/10.58628/JAE-2622-102Keywords:
Machine Learning, Artificial Neural Network, Crown diameter model, Sal treesAbstract
In recent years, Machine Learning algorithms have been frequently used in forest sciences to build Classification and Regression models. Such algorithms are known for their ability to explore complex relationships between the target variable and multiple predictors with minimum tuning and high accuracy. In this article, an Artificial Neural Network (ANN) was used to develop a Crown Diameter (CD) model for Sal trees based on four predictors, viz., Diameter, Height, Basal Area (BA) and Height-to-Diameter ratio (HDR). Trees were growing naturally in the Ranibagh forest area located at the foothills of the Kumaun Himalaya. Results indicated a good model fit (R2adj. = 0.79), i.e., the model explained 79% of the variation in crown diameter based on four predictors. The Pearson Correlation coefficient (r) value between actual and predicted CD came out to be 0.89 (p< 0.001). The residual plots also indicated homoscedasticity. Mean Absolute Percentage Error (MAPE) for the model came out to be 13.26%. HDR was found to be the most influential predictor for the model, followed by tree height, diameter, and Basal Area. One-way Partial Dependence Plot (PDP) demonstrated that diameter, height and BA had a positive impact on CD. In addition, for HDR, the CD-curve increased up to a certain point with an increase in HDR value and then decreased sharply. Two- way PDP further suggested that HDR in combination with Diameter and BA had a negative influence on crown diameter growth, i.e., HDR had masked the positive effect of diameter and BA on CD. The overall model suggested the competitive nature of Sal trees within a forest stand.
References
Ahmed S, Hilmers T, Uhl E, Simoes FT, Ordonez C, Bravo F, Rio MD, Peters RL and Pretzsch H, 2025. From suppressed to dominant: 3D crown- shapes explains the “to grow or wait” growth behaviour in close-to-nature forests, Forest Ecology and Management, 592, https://doi.org/10.1016/j.foreco.2025.122814
Alves LF and Santos FAM,2002. Tree allometry and crown shapes of four tree species in Atlantic rain forest, South- East Brazil, Journal of Tropical Ecology, 18(02): 245- 260. https://doi.org/10.1017/S026646740200216X
Anarbekova G, Ruiz LGB, Akanova A and Sharipova S, 2024. Fine- tuning Artificial Neural Networks to predict pest numbers in grain crops: A case study in Kazakhstan, Machine Learning and Knowledge Extraction, 6, 1154- 1169, https://doi.org/10.3390/make6020054
Arguedas TB, Roddy AB, Coley PD and Kursar T, 2010. Do differences in understory light contribute to species distribution along a tropical rainfall gradient? Oecologia, 166(2): 443- 456, https://doi.org/10.1007/s00442-010-1832-9
Asigbaase M, Dawoe ELK, Abugre S, Kyereh B, 2023. Allometric relationships between stem diameter, height and crown area of associated trees of cocoa agroforests of Ghana, Scientific Reports, 13(1), https://doi.org/10.1038/s41598-023-42219-6
Bartkowicz L, Paluch J and Wertz B, 2025. Comparison of the stem basal area increment of five co existing tree species with different light demands growing in Central European Deciduous Forests with complex vertical structures, Forests, 16(11), https://doi.org/10.3390/f16111700
Bayat M, Ghorbanpour M, Zare R and Jaafari A, 2019. Application of artificial neural networks for predicting tree survival and mortality in the Hyrcanian forest of Iran, Computers and Electronics in Agriculture, 164, https://doi.org/10.1016/j.compag.2019.104929
Bellin M, Tesi G, Marchesani N and Rossi V, 2022. Species distribution modelling and machine learning in assessing the potential distribution of fresh water zooplanktons in northern Italy, Ecological Informatics, 69, https://doi.org/10.1016/j.ecoinf.2022.101682
Bhebhe ZM, Liu X, Zhang Z, Paudyal DR, 2025. Estimation of tree diameter at breast height (DBH) and biomass from allometric models using LiDAR data: A case of the Lake Broadwater Forest in Southeast Queensland, Australia, Remote Sensing, 17(14). https://doi.org/10.3390/rs17142523
Chaturvedi AN and Khanna LS, 1982. Forest Mensuration, International Book Distribution, Dehradun, India, Page 403.
Chon TS, Park YS, Kim JM and Lee BM, 2000. Use of an Artificial Neural Network to predict Population Dynamics of the forest pest pine needle Gall Midge (Diptera: Cecidomyiida), Environmental Entomology, 29(6): 1208- 1215. https://doi.org/10.1603/0046-225X-29.6.1208
Deng L, Wang J, Yin J, Chen Y, Wu S, 2025. Construction of crown profile prediction model of Pinus yunnanensis based on CNN-LSTM- attention method, Frontiers in Plant Sciences, 16. https://doi.org/10.3389/fpls.2025.1567131
Ezenweny JU and Chukwu O, 2017. Effects of slenderness coefficient in crown area prediction for Tectona grandis Linn. F. in Omo Forest Reserve, Nigeria, Current Life Sciences, 3(4): 65- 71. https://doi.org/10.5281/zenodo.996326
Fu L, Sharma RP, Hao K, Tang S, 2017. A generalized interregional nonlinear mixed- effects crown width model for Prince Rupprecht larch in northern China, Forest Ecology and Management, 389: 364- 373
Goodman RC, Phillips OL and Baker TR, 2014. The importance of crown dimensions to improve tropical tree biomass estimates, Ecological Applications, 24(4): 680- 698, https://doi.org/10.1890/13-0070.1
Gracia WS, 2025. Regression analysis and artificial neural networks for predicting pine species volume in community forests, Ecological Information, 89, https://doi.org/10.1016/j.ecoinf.2025.103203
Harris NL, Gibbs DA, Baccini A, Birdsey RA, de Bruin S, Farina M et al., 2021. Global maps of twenty- first century forest carbon fluxes, Nat. Climate Change, 11, 234- 240. https://doi.org/10.1038/s41558-020-00976-6
Huang, H, Wu, D, Fang, L., Zheng, X. (2022). Comparison of multiple machine learning models for estimating the forest growing stock in large scale forests using multi – source data, Forests, 13, 1471. https://doi.org/10.3390/f13091471
Jeelani MI, Tabassum A, Rather I and Gul M, 2023. Neural Network modelling of height-diameter relationships for Himalayan Pine through backpropagation approach, Journal of the Indian Society for Probability and Statistics, 76(3): 169- 178.
Jucker T, Fischer FJ, Chave J, Coomes DA, Caspersen J, Ali A et al., 2025. The globe spectrum of tree crown architecture, Nat. Commun., 16, 4876. https://doi.org/10.1038/s41467-0.5-60262-x
Kang J, Chiung K, Lee SJ and Yim JS, 2021. Relationship of H/D and crown- ratio and tree growth for Chamaecyparis obtusa and Cryptomeria japonica in Korea, Forest Science and Technology, 17(2): 1- 9. https://doi.org/10.1080/21580103.2021.1904009
Kantarcioglu O, Kocaman S and Schindler K, 2023. Artificial Neural Networks for assessing forest fire susceptibility in Turkiye, Ecological Informatics, 75, https://doi.org/10.1016/j.ecoinf.2023.102034
Kumar A, Kumar P, Rongpi R, Ranjan P, Kumari A and Singh A, 2025. Assessing forest fire vulnerability using artificial neural networks in Almora district, Uttarakhand, India, Journal of the Bulgarian Geographical Society, 53: 67- 86. https://doi.org/10.3897/jbgs.e159980
Laarhoven TV, 2017. L2 Regularization versus batch and weight normalization, Computer Science, https://doi.org/10.48550/arxiv.1706.05350
Lam TY, Ducey MJ, 2024. Analysis of the inflection points of height – diameter models, Forest Ecosystems,11, https://doi.org/10.1016/j.fecs.2024.100202
Lau A, Claders K, Bartholomeus H, Martius C, Raumonen P, Herold M, Vicari M, Sukhdeo H, Singh J, Goodman RC, 2019. Tree biomass equations from terrestrial LiDAR: A case study in Guyana, Forests, 10, 527
Lee KY, Chung N and Hwang S, 2016. Application of an artificial neural network (ANN) model for predicting mosquito abundance in urban areas, Ecological Informatics, 36: 172- 180, https://doi.org/10.1016/j.ecoinf.2015.08.011
Liu M, Feng Z, Zhang Z, Ma C, Wang M, Lian BL, Sun R and Zhang L, 2017. Development and evaluation of height diameter at breast height models for native Chinese metasequoia, PLOS ONE, 12(8): e0182170, https://doi.org/10.1371/journal.pone.0182170
Long S, Zeng S and Wang G, 2021. Developing a new model for predicting the diameter distribution of oak forests using an artificial neural network, Annals of Forest Research, 64(2): 3- 20, https://doi.org/10.15287/afr.2020.2060
MacDicken KG, 1997. A guide to monitoring carbon storage in forestry and agroforestry projects, Forest Carbon Monitoring Program, Winrock International Institute of Agriculture Development, page 403.
Mankou GS, Ligot G, Panzou GJL, Boyemba F, Loumeto JJ, Ngomanada A, Obiang D, Rossi V, Sonke B, Yango OD and Fayolle A, 2021. Tropical tree allometry and crown allocation and their relationship with species traits in central Africa, Forest Ecology and Management, 493. https://doi.org/10.1016/j.foreco.2021.119262
Martinez JLF, Santamariaa JC, Campo FC, Antaa MB, Obesob JR and 2013. Tree height prediction approaches for uneven – aged beech forests in northwestern Spain, Forest Ecology and Management, 307(10): 63- 73. https://doi.org/10.1016/j.foreco.2013.07.014
Matsuo T, Ramos MM, Bongers F, Sande MTVD and Poorter L,2021. Forest structure drives changes in light heterogeneity during tropical secondary forest succession, Journal of Ecology, 109(8): 2871- 2884. https://doi.org/10.1111/1365-2745.13680
Mencuccini M, Vilarta JM, Vanderklein D and Hamid HA, 2005. Size- mediated ageing reduces vigour in trees, Ecology Letters, 8(11): 1183- 1190. https://doi.org/10.1111/j.1461-0248.2005.00819x
Myttenaere AD, Golden B, Grand BL and Rossi F, 2015. Using the Mean Absolute Percentage Error for Regression models, Neuro Computing, 192. https://doi.org/10.1016/j.neucom.2015.12.114
Nguyen TD and Katabuchi M, 2025. Saturating allometric relationships reveal how wood density shapes global tree architecture, Journal of Forestry Research, 36: 107. https://doi.org/10.1007/s11676-025-01898-9
Omijeh JE, 2021. Tree stem diameter (DBH) as predictor variable in assessing other growth parameters of Sclerocarya birrea (Anacardiaceae) in a Nigerian Guinea Savanna, Int. Res. Jr. of Applied Sciences, Engineering and Technology, 7(12): 1-11.
Opio C, Jacob N and Coopersmith D, 2000. Height-to-diameter ratio as a competition index for young conifer plantations in northern British Columbia, Canada, Forest Ecology and Management, 137(1): 245- 252. https://doi.org/10.1016/S0378-1127(99)00312-6
Ou Y and Barraza GQ 2023. Modelling Height-Diameter relationship using Artificial Neural Networks for Durango Pine (Pinus durangenesis Martinez) species in Mexico, Forests, 14(8). https://doi.org/10.3390/f14081544
Ozcelik R, Diamantopoulou M, Brookes J, 2014. The use of tree crown variables in over- bark diameter and volume prediction models, Iforest Biogeosciences For., 7:132- 139.
Ozcelik R, Diamantopoulou MJ, Campo FC and Eler U, 2013. Estimating Crimean Juniper tree height using nonlinear regression and artificial neural network models, Forest Ecology and Management, 306: 52-60. https://doi.org/10.1016/j.foreco.2013.06.009
Paula J, Peper E, Mcpherson G, Mori SM, 2001. Equation for predicting diameter, height, crown width and leaf area of San Joaquin valley street trees, Int. Society of Arboriculture, 27(6), 306. https://doi.org/10.48044/jauf.2001.034
Plaga BNE, Bauhus J, Pretzsch H, Pereira MG and Forrester DI, 2024. Influence of crown and canopy structure on light absorption, light use efficiency and growth in mixed and pure Pseudotusga menziesii and Fagus sylvatica forests, Eur. Jr. For. Res., 143: 479- 491. https://doi.org/10.1007/s10342-023-01638-w
Pretzsch, H, Biber P, Uhl E, Dahlhausen J, Rotzer T, Caldentey J, Koike T, Can TV, Chavanne A, Seifert T, Toit BD, Farnden C and Pauleit S,2015. Crown size and growing space requirement of common tree species in urban centres, parks and forests, Urban Forestry and Urban Greening, 14(3): 466- 479. https://doi.org/10.1016/j.ufug.2015.04.006
R Development Core Team, 2025. A Language and Environment for Statistical computing, Available Online, https://www.R-project.org
Reis L, Souza ALD, Reis PCMD and Mazzei L, 2018. Estimation of mortality and survival of individual trees after harvesting wood using artificial neural networks in the amazon rain forest, Ecological Engineering, 112: 140- 147, https://doi.org/1016/j.ecoleng.2017.12.014
Rudnicki M, Silins U and Lieffers VJ, 2004. Crown cover is correlated with relative density, tree slenderness and tree height in Lodgepole Pine, Forest Science, 50(3): 356- 363, https://doi.org/10.1093/forestscience/50.3.356
Safi Y and Bouroumi A, 2013. Predicting of forest fires using Artificial Neural Networks, Applied Mathematical Sciences, 7(5): 271- 286. https://doi.org/10.12988/ams.2013.13025
Sakici OE and Ozdemir G, 2018. Stem taper estimations with artificial neural networks for mixed Oriental Beech and Kazdagi Fir stands in Karabuk region, Turkey, Cerne, 24(4): 439- 451, https://doi.org/10.1590/01047760201824042572
Salehin I and Kang DK, 2023. A review on Dropout Regularization approaches for deep neural networks within the scholarly domain, Electronics, 12(14): 1- 23. https://doi.org/10.3390/electronics12143106
Sandoval S and Acuna E, 2022. Stem taper estimation using Artificial Neural Networks for Nothofagus trees in natural forests, Forests, 13(12), https://doi.org/10.3390/f13122143
Sanquetta C, Piva LRDO, Wojciechowski J and Corte APD, 2017. Volume estimation of Cryptomeria japonica logs in southern Brazil using artificial intelligence models, Southern Forests: Journal of Forest Science, 80(1): 1-8, https://doi.org/10.2989/20702620.2016.1263013
Savo V, Amato L, Bartoli F, Zappltelli I, Caneva G, 2025. Evaluation of main regulating provisioning and supporting ecosystem services of urban street trees: A literature review, Ecosystem Services, 71, https://doi.org/10.1016/j.ecoser.2024.101690
Scrinzi G, Marzullo L and Galvagni D, 2007. Development of a neural network model to update forest distribution data for managed alpine stands, Ecological Modelling, 206(3-4): 331- 346. https://doi.org/10.1016/j.ecolmodel.2007.04.001
Seki M, 2023. Predicting stem taper using artificial neural network and regression models for Scot Pine (Pinus sylvestris L.) in northwestern Turkiye, Scandinavian Journal of Forest Research, 38(1-1): 97- 104, https://doi.org/10.1080/02827581.2023.2189297
Sharma RP, Bilek L, Vacek Z, Vacek S, 2017. Modelling crown width – diameter relationship for Scot pine in the central Europe, Trees, 31: 1875- 1889.
Silva TASE, Tibario FCS, Dodonov P and Matos DMS, 2015. Differences in allometry and population structure between native and invasive populations of a tropical tree, New Zealand Journal of Botany, 53(2): 90- 102. https://doi.org/10.1018/0028825x.2015.1015575
Skudnik M and Jevsenak J, 2022. Artificial Neural Networks as an alternative method for nonlinear mixed- effects models for tree height predictions, Forest Ecology and Management, 507. https://doi.org/10.1016/j.foreco.2022.120017
Souza DC, Jardina KJ, Rodrigues JVFC, Gimenez BO, Rogers A, McDowell N, Walker AP, Higuchi N, Filho IJS and Chambers J, 2021. Canopy position influences the degree of light suppression of leaf respiration in abundant tree genera in the Amazon Forest, Frontiers in Forests and Global Change, 4, https://doi.org/10.3389/ffgc.2021.723539
Sporek M and Sporek K, 2023. Allometric model of crown length for Pinus sylvestris L. stands in South- Western Poland, Forests, 14(9), https://doi.org/10.3390/f14091779
Stoffberg GH, Rooyen MW, Linda MJV, Groeneveld HT, 2008. Predicting the growth in tree height and crown size of three street tree species in the city of Tshwane, South Africa, Urban forestry and Urban Greening, 7(4): 259- 264. https://doi.org/10.1016/j.ufug.2008.05.002
Sumnall MJ, Gracia IR, Carter DR, Albaugh TJ, Campoe OC, Rubilar RA, Alexander B, Cohrs W and Cook RL, 2025. Assessing methods to measure stem diameter at breast height with high pulse density helicopter lase scanning, Remote Sensing, 17(2), https://doi.org/10.3390/rs17020229
Taherdoost H, 2023. Deep Learning and Neural Networks: Decision making implications, Symmetry, 15(9). https://doi.org/10.3390/Sym15091723
Tang L, Hou C, Huang H, Chen C, Zou J and Lin D, 2015. Light interception efficiency analysis based on three- dimensional peach canopy models, Ecological Informatics, 30: 60- 67, https://doi.org/10.1016/j.ecoinf.2015.09.012
Vieira GC, Mendonca ARD, da Silva GF, Zanetti SS, da Silva MM and Santos ARD, 2018. Prognoses of diameter and height of trees of Eucalyptus using artificial intelligence, Science of the Total Environment, 619-620: 1473- 1481. https://doi.org/10.1016/j.scitoenv.2017.11.138
Vospernik S, Monserud RA and Sterba H, 2010. Do individual -tree growth models correctly represent height: diameter ratios of Norway Spruce and Scot pine? Forest Ecology and Management, 260(10): 1735- 1753, https://doi.org/10.1016/j.foreco.2010.07.055
Wang J, Jiang L and Yan Y, 2022. The impact of Climate, competition and their interactions on crown width for three major species in Chines boreal forests, Forest Ecology and Management, 526, 120597. https://doi.org/10.1016/j.foreco.2022.120597
Wang Y, Liu Z, Li J and Cao X, 2024. Assessing the relationship between tree growth, crown size and neighbouring tree species diversity in mixed coniferous and broad forests using crown size competition indices, Forests, 15(4): 633, https://doi.org/10.3390/f15040633
Wright LG, Onodera T, Stein MM, Wang T, Schachter DT, Hu Z and McMohan PL, 2022. Deep physical neural network trained with backpropagation, Nature, 601, 549- 555. https://doi.org/10.1038/s41586-021-04223-6
Yamakawa M, Onoda Y, Kurokawa H, Oguro M, Nakashizuka T and Hikosaka K, 2023. Competitive asymmetry in a forest composed of a shade- tolerant species depends on gap formation, Forest Ecology and Management, 549. https://doi.org/10.1016/j.foreco.2023.121442
Zhang Z, 2018, Artificial Neural Network in Multivariate Time Series Analysis in Climate and Environmental Research; Springer International Publishing: Berlin/Heidelberg, Germany, 1-35, ISBN: 978-3-319-67339-4, https://doi.org/10.1007/978-3-319-67340-0
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
