Abstract
Background and Aims : We determined the effectiveness of attenuated total reflectance Fourier transform infrared spectroscopy as a rapid and low‐cost method for measuring the concentration of nutrients in grapevine petioles.
Methods and Results: Attenuated total reflectance Fourier transform infrared spectra were recorded for 392 dried and ground petiole samples collected from Chardonnay, Semillon and Shiraz vineyards in the Riverina between 2003 and 2010. Predictive models of nutrient concentration obtained with standard reference methods were developed using partial least squares regression. Good predictive models were produced for all macronutrients, with R2 values for nitrogen (N), phosphorus (P), potassium (K), magnesium (Mg), calcium (Ca) and sulfur (S) of 0.945, 0.915, 0.951, 0.961, 0.940 and 0.849, respectively. For sodium (Na) and the micronutrients iron (Fe), manganese (Mn), boron (B), copper (Cu) and zinc (Zn), R2 values were 0.773, 0.750, 0.743, 0.630, 0.612 and 0.835, respectively. When ranked according to the residual predictive deviation, expressed as the ratio of the standard error of performance to the calibration set standard deviation, values of 5.4, 3.8, 3.1, 3.8 and 3.5 were obtained for N, P, K, Mg and Ca, respectively; intermediate values of 2.4, 2.0 and 1.8 were obtained for S, Zn and Na, while Fe, Mn, B, Cu had a value of 1.3 or below.
Conclusion: All macronutrients in grapevine petiole tissue can be determined by attenuated total reflectance Fourier transform infrared spectroscopy with sufficient accuracy to assess grapevine nutritional status against standard interpretive ranges. Certain micronutrients can also be determined, although the predictive models for these elements were not as robust as models for the macronutrients.
Methods and Results: Attenuated total reflectance Fourier transform infrared spectra were recorded for 392 dried and ground petiole samples collected from Chardonnay, Semillon and Shiraz vineyards in the Riverina between 2003 and 2010. Predictive models of nutrient concentration obtained with standard reference methods were developed using partial least squares regression. Good predictive models were produced for all macronutrients, with R2 values for nitrogen (N), phosphorus (P), potassium (K), magnesium (Mg), calcium (Ca) and sulfur (S) of 0.945, 0.915, 0.951, 0.961, 0.940 and 0.849, respectively. For sodium (Na) and the micronutrients iron (Fe), manganese (Mn), boron (B), copper (Cu) and zinc (Zn), R2 values were 0.773, 0.750, 0.743, 0.630, 0.612 and 0.835, respectively. When ranked according to the residual predictive deviation, expressed as the ratio of the standard error of performance to the calibration set standard deviation, values of 5.4, 3.8, 3.1, 3.8 and 3.5 were obtained for N, P, K, Mg and Ca, respectively; intermediate values of 2.4, 2.0 and 1.8 were obtained for S, Zn and Na, while Fe, Mn, B, Cu had a value of 1.3 or below.
Conclusion: All macronutrients in grapevine petiole tissue can be determined by attenuated total reflectance Fourier transform infrared spectroscopy with sufficient accuracy to assess grapevine nutritional status against standard interpretive ranges. Certain micronutrients can also be determined, although the predictive models for these elements were not as robust as models for the macronutrients.
| Original language | English |
|---|---|
| Pages (from-to) | 299-309 |
| Number of pages | 11 |
| Journal | Australian Journal of Grape and Wine Research |
| Volume | 20 |
| Issue number | 2 |
| DOIs | |
| Publication status | Published - Jun 2014 |