PSVIII-22 Non-Destructive Estimation of Crude Protein in Mombasa grass using morphometry, colorimetry, and spectrophotometry.

Abstract This study explores the potential of non-destructive tools for estimating nitrogen and crude protein content in Mombasa grass (Megathyrsus maximus). The non-invasive instruments include a measuring tape, Pantone® color chart, smartphone camera with the ColorDetector application, chlorophyll meter, and SCiO™ NIR micro-spectrometer. Current standard methods for determining nitrogen and crude protein content in plants, such as the Kjeldahl and Dumas methods, are destructive, costly, and time-consuming. Therefore, producers and professionals would benefit greatly from rapid, simple, and non-destructive methods that facilitate real-time decision-making in the field on plant and animal nutrition, enhancing the precision on investment in livestock production. The study was conducted on a 990 m² open-field rotational grazing paddock, divided into three plots of 330 m² each. Each plot received varying levels of nitrogen fertilization: 0 kg of N/ha/cycle (No N), 16,7 kg of N/ha/cycle (Low N), and 50 kg of N/ha/cycle (High N). Two sampling cycles were carried out on the same plants with a 21-day recovery period between cycles. In each sampling, 20 plants, tillers, and fully expanded leaves were sampled per plot, totaling 120 samples. Morphometric measurements included leaf length, leaf width, and plant height. Colorimetric analysis was performed using the Pantone® color scale and the ColorDetector mobile application, analyzing the RGB chromatic coordinates. Spectrophotometric readings were obtained using a SPAD-502 Plus chlorophyll meter and a SCiO™ NIR micro-spectrometer. Predictive models were developed using these non-destructive measurements as independent variables, with nitrogen content—determined by the Dumas method and converted to crude protein using the factor N × 6.25—as the dependent variable (Table 1). Morphometric models demonstrated high precision, with coefficients of determination exceeding 0.70. Leaf width alone explained 86% of the variation in crude protein content, suggesting it as a reliable indicator of protein value due to its ease of measurement in the field and precision. Models based on the Pantone® color scale and ColorDetector application exhibited high predictive power, achieving R² values of 0.91 and 0.96, respectively. To facilitate field applications, color charts in shades of green were developed for rapid nitrogen and crude protein estimation (Figure 1 and 2). The regression generated from the SPAD index indicated a low capacity of this instrument to predict nitrogen and crude protein in Mombasa grass (R²=0.51). In contrast, the model derived from SCiO™ NIR micro-spectrometer data showed moderate precision, with an R² of 0.69. These results underscore the viability of non-destructive methods for creating centimeter and colorimetric charts, enabling real-time, cost-effective, and accessible nitrogen and crude protein estimation in M. maximus grasses.