Nitrogen content of rice leaves as predicted by SPAD, NIB, and spectral reflectance measurements

Witt, C.; Pasuquin, J.M.C.A.C.

Nitrogen content of rice leaves as predicted by SPAD, NIB, and spectral reflectance measurements

2003

Witt, C. | Pasuquin, J.M.C.A.C. (Intenational Rice Research Inst. DAPO Box 7777, Metro Manila (Philippines))

The accurate estimation of leaf nitrogen (N) content in rice is of equal importance in research and extension. Farmers need real-time information on the plant's need for N to achieve high yields and efficient fertilizer use. The objective of this study was to compare the accuracy of a chlorophyll meter (Minolta SPAD-502) and a Near Infrared Reflectance (NIR, Satake Corporation) meter with spectral reflectance (SR) measurements (Minolta CM-3700-d) of rice leaves in assessing leaf N. A two-season field experiment was conducted at IRRI's Experimental Farm in 2001 using 10 modern, high yielding varieties grown in three different N levels. Measurements were taken at three growth stages. Across all seasons, varieties and growth stages, SPAD explained only 40% and NIR only 34% of the variation in leaf N based on dry weight (N sub dw). Both SPAD and NIR explained 79% of the variation in N sub dw only after adjusting readings for leaf thickness (specific leaf weight). The potential of SR measurements was evaluated using a model for predicting leaf N content based on Partial Least Squares (PLS) regression using all visible spectral data (400-700 nm) without variable selection (SAS V.8.02, SAS Institute, NC, USA). A model having 15 latent variables was able to account for over 99% of the predictor variation (reflectance) and over 96% of the response variation (leaf N content). In conclusion, SR measurements provide the most accurate estimate of leaf N content without the need for destructive sampling. SR measurements are currently further used to develop quality control measures for the production of the leaf color chart, an on-farm tool for efficient N management jointly developed by IRRI [Los Banos, Laguna, Philippines] and PhilRice [Philippine Rice Research Inst., Munoz, Nueva Ecija, Philippines]

Show more [+]