GAP FREQUENCY AND CANOPY ARCHITECTURE OF SUGAR-BEET AND WHEAT CROPS

Gap frequency of sugar beet and wheat canopies was measured using downward looking hemispherical photography. This technique accounts for gap frequency (P(o)) variation with both zenith theta and azimuth phi angles. A range of leaf area indices was obtained by varying the sowing date, the sowing density or the measurement date. Results show that the exponential law (P(o)(theta, phi) = e(-K(theta,phi).L)) accurately describes variations with leaf area index (L) for each direction (theta, phi). Row architecture had very little effect on gap frequency azimuthal variations. A dispersion coefficient (lambda(o)) was directly computed with a Markov chain model applied on sugar beet crops using observed gap frequencies, leaf area indices and leaf inclination. The lambda(o) coefficient value was close to 1.2 for zenith angles below 60-degrees, and then decreased for higher zenith angles. Assuming the dispersion coefficient lambda(o) to be independent of the zenith angle theta allowed us to fit both the lambda(o) value and the average leaf inclination when it was not measured. Fitted values of leaf inclination and lambda(o) indicated that the wheat crop was rather erectophile with a regular leaf arrangement (lambda(o) = 1.56). The sugar beet canopy was more planophile and very close to the random leaf arrangement (lambda(o) = 1.04). These results contrasted with those found when lambda(o) was assumed to vary with theta and with the observed leaf inclination value. The discrepancy was partly explained by the fact that the fitted values of lambda(o) absorbed most of the experimental errors and violations of the implicit hypothesis. However, this study provides a simple semi-empirical way to describe gap frequencies of crops throughout their growing cycle or for various densities or sowing dates, using Markov chain processes.