Simulation and validation of leaf area prediction model for Picrorhiza kurroa - An endangered medicinal plant of Western Himalaya

Abstract

The study was aimed at modeling of individual leaf area of Picrorhiza kurroa using linear measurements of leaf length (L) and maximum width (W). Leaves were collected from the greenhouse at different time intervals during 2009 and 2011. The actual leaf area (LA) and leaf dimensions were measured with a laser area meter. Different combinations of prediction equations were obtained from length (L), width (W), product of LW to build linear (y = a + bx), quadratic (y = a+bx+cx2), exponential (y = aebx), logrithmatic (y = a+bLnx) and power models (y = axb) for different samples and pooled data compared with earlier models by graphical procedures and statistical criteria root mean square error (RMSE). A linear model having LW as the independent variables (y = 0.333 + 0.603LW) provided the most accurate estimate (R2 = 0.955, RMSE = 0.573, coefficient of variation (CV) = 7.46%) of P. kurroa leaf area. Validation of the regression model having LW of leaves measured in two different experiments during September, 2011 showed that the correlation between measured and predicted values by the use of this equation was very high (R2 = 0.9053), with low RMSE (0.39) and CV (5.44%). Key words: Picrorhiza kurroa, leaf area model, non destructive, validation.