Predicting seed yield in perennial ryegrass using repeated canopy reflectance measurements and PLSR

Abstract

Repeated canopy reflectance measurements together with partial least-squares regression (PLSR) were used to predict seed yield in perennial ryegrass (Lolium perenne L.). The measurements were performed during the spring and summer growing seasons of 2001 to 2003 in three field experiments with first year seed crops using three sowing rates and three spring nitrogen (N) application rates. PLSR models were developed for each year and showed correlation coefficients of 0.71, 0.76, and 0.92, respectively. Regression coefficients showed in these experiments that the optimum time for canopy reflectance measurements was from approximately 600 cumulative growing degree-days (CGDD) to approximately 900 CGDD. This is the period just before and at heading of the seed crop. Furthermore, regression coefficients showed that information about N and water is important. The results support the development of an additional N- and water-application model which will calculate the application rate of N and water according to expected seed yield.