Open AccessEstimation of leaf area of Erythroxylum citrifolium from linear leaf dimensions
Author(s) -
João Everthon da Silva Ribeiro,
Ester dos Santos Coêlho,
Francisco Romário Andrade Figueiredo,
Sérgio de Faria Lopes,
Manoel Bandeira de Albuquerque
Publication year - 2019
Publication title -
bioscience journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.199
H-Index - 19
eISSN - 1981-3163
pISSN - 1516-3725
DOI - 10.14393/bj-v35n6a2019-41743
Subject(s) - akaike information criterion , mathematics , mean squared error , coefficient of determination , statistics , leaf blade , botany , biology
Erythroxylum citrifolium is a neotropical plant species recorded in all regions of Brazil. Determining leaf area is of fundamental importance to studies related to plant propagation and growth. The objective was to obtain an equation to estimate the leaf area of E. citrifolium from linear dimensions of the leaf blade (length and width). A total of 200 leaf blades were collected in Parque Estadual Mata do Pau-Ferro in the municipality of Areia, state of Paraíba, Northeast Brazil. The models evaluated were: linear, linear without intercept, quadratic, cubic, power and exponential. The best model was determined by the criteria of: high coefficient of determination (R²), low root mean square error (RMSE), low Akaike information criterion (AIC), high Willmott concordance index (d) and a BIAS index close to zero. All of the models constructed satisfactorily estimated the leaf area of E. citrifolium, with coefficients of determination above 0.9050, but the power model using the product between length and width (L*W) ŷ = 0.5966 * LW1.0181 was the best, with the highest values of R² and d, low values of RMSE and AIC, and a BIAS index closest to zero.