Scholarly record

WATER QUALITY PROTECTION BY SPECTRAL DETERMINATION OF PRECISE DOSES OF FERTILIZERS ON ARABLE LAND

Václav Nedbal

Abstract

The outflow of inorganic fertilizers from arable land is one of the main sources of surface water pollution. The pollution can be reduced by precise dosing of inorganic fertilizers according to the momentary nutrition need of the cultivated crops. The exact dose of fertilizer can vary considerably from one area to another as the soil conditions can be very variable even in the scope of a single field. The spectral properties of cultivated crops can be used for the calculation of precise doses of fertilizer. In our work we aimed to define the relation between the necessary nitrogen nutrition and the spectral properties of the winter wheat. As a rate of the actual crop nutrition the laboratory analysis of the plants was used to compute the Nitrogen Nutrition Index (NNI). An ensemble of the spectral measurements was used to compute the Red Edge Inflection Point index (REIP) to establish a relation between NNI and REIP indices. For the comparison of results from various technologies we employed the three different devices for spectral measurements (ground spectroradiometer, aerial hyperspectral sensor and satellite multispectral sensor). The results showed a close relation between laboratory-determined values of the crop NNI index and the REIP spectral index for all three measurement methods. Moreover, all three methods showed comparable and consistent results. Ground spectroradiometer measurement showed the closest relation between NNI and REIP indices (R2 = 0.9). Unlike a ground spectroradiometer, the aerial hyperspectral and the satellite multispectral measurements provide also a spatial distribution of the REIP values, so the spatial distribution of NNI values could be established using the previously computed relation. While hyperspectral measurements showed better spatial resolution (5 m per pixel) maintaining relatively good correlation (R2 = 0.82), multispectral measurements with lower spatial resolution (20 m per pixel) showed even a slightly better correlation (R2 = 0.84). Unlike an aerial hyperspectral data, the satellite data are available for free and the scanning frequency is relatively high (2 to 3 datasets a week), so the technology can be used for precise fertilization purposes at low costs. In general, the work shows that the new technologies of spectral evaluation of crops nutrition can reduce the consumption of fertilizer by its precise dosing, so they can be also an effective tool in surface water quality protection.

Publication Impact Profile

• Citations
• CrossRef - Citation Indexes: 1

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2019v/1.4/s02.039

Publication details

Back to publication list