Evaluation of nitrogen nutrition status in oil palm based on spectral response and multi-sensor images

Nitrogen (N) plays crucial roles in sustainability of oil palm (Elaeis guineensis) production, environmentally and economically. However, assessing N status of tall perennial crops such as oil palm was complex and not straightforward in comparison to annual crops due to complex N partitioning, age, and larger amount of respiratory loads. The focus of this study was to develop N prediction model to estimate N nutrition status of oil palm. N fertilization rates varied from 0 to 2 kg N palm-1 according to the age classes (immature, young mature and prime mature) nutrient requirements. Growth and spectral parameters measured for this study included height, diameter, leaf area index (LAI), leaf N content, chlorophyll meter reading and spectral reflectance in visible (VIS) and near infra-red (NIR) regions measured from a ground spectroradiometer and satellite images. The growth and spectral parameters responses to N fertilization were age-dependent, where only immature palm showed significant responses to the N fertilization. In comparing the growth and spectral parameters, correlation analysis indicated that the latter was more sensitive to foliar N, especially ones that were acquired from satellite images, and therefore have potential in predicting N nutrition status of oil palm. Foliar N content was found to be less influenced by the palm maturity classes, while chlorophyll meter readings were confounded by age, and would precluded the use of that measurement for N management. Based on these results, spectral reflectance then were employed for indices such as VIS spectra, NIR spectra and combination of VIS and NIR spectra (VIS+NIR) indices to predict N status in both immature and mature palms.Each age group was sensitive to different N indices. The VIS+NIR indices acquired from the ground level sensor such G+R+NIR (R2 = 0.91) and SPAD (R2 = 0.64) indices were found to be significantly useful to assess N status of immature and young mature palms respectively, while NIR (R2 = 0.26) index was the best index for predicting N for prime mature palm. At the canopy level, foliar N were best assessed using different VIS indices such as B+R and GRI for young and prime mature palms, respectively. Additionally, the discriminant analysis (DA) illustrated that the chosen bands for discriminating N sufficiency levels were age-dependent as the accuracies of classification decreased as the data pooled across all maturity classes. The best discriminant function for all maturity classes was credited to the combination of original with narrow band spectra by using the spectroradiometer with accuracies ranging from 76.20 to 98.65%. The most pronounced spectral bands for all three maturity classes were the combination of original and narrow bandwidth; located at 405, 575, 625, 795, 925 and 955 nm. Based on the preliminary study, remote sensing has the opportunity to assist in plantation agronomic management involving nutrient management. The best method in assessing foliar N in oil palm was portrayed by DA. Hence, further studies should be done to develop a` suitable index or equation for predicting foliar N based on the DA findings.

Preview Text FP 2016 3 IR.pdf Download (392kB) | Preview

Actions (login required)

View Item