Assessing the vertical footprint of reflectance measurements to characterize nitrogen uptake and biomass distribution in maize canopies

Abstract

Evaluation of phenotypic traits of crop plants on a large scale could provide important information to understand their responses to the environment. In this regard, remote sensing methods have shown much promise. However, the effect of the different contributions to spectral reflectance of the different leaf levels of plant canopies remains poorly investigated, despite their potential to improve the precision of canopy information estimation. In this study, we investigated the efficacy of sensor measurements in determining the vertical leaf nitrogen uptake of maize (Zea mays L.). We examined how nitrogen is distributed in the plant canopy, whether or not differences exist among fertilizer application rates, and how deep does the passive reflectance sensor meaningfully provide insight into the plant canopy. Our results, derived from either a sensor system with an oblique and multi view optic as well as SPAD measurements, indicated a convex (bulging outward) distribution of the relative chlorophyll content in maize plants. Similarly, both leaf nitrogen uptake and leaf biomass presented vertical bell shape distribution, although only the former showed qualitative differences among the fertilization treatments in the intermediate canopy leaf levels. By contrast, vertical nitrogen content presented a vertically decreasing gradient from top to bottom and one that was steeper at reduced nitrogen application. The spectral index R780/R740 was positively and curvilinearly related (R2=1.00) to the nitrogen uptake profile of the maize foliage and was able to detect the nitrogen uptake of each leaf level, even at the lowest levels. Yet, despite more than half of the total nitrogen being stored in the stem, the index values were influenced mainly by the foliage. Altogether, our results should help improve nitrogen fertilization recommendations in crop management as well as being useful in precision phenotyping and in improving in crop growth simulation models for architectural modeling.