Search

Hyperspectral remote sensing data have certain disadvantages as well as being a widely used tool for investigating biophysical and biochemical characteristics in grasslands due to its many advantages. Most importantly, some external influences have negative effects on the signals obtained from the canopy. Studies conducted in recent years have revealed that one of these negative externalities is the flowering on the canopy. The purpose of this study is to show how spectral reflectance readings are affected in samples with different flowering rates. The following procedure in the given order was carried out, and this procedure was repeated for a total of 46 measurements from within 10 quadrats: (1) placing quadrats of 50 × 50 cm on selected sampling areas, (2) performing spectral measurements in the quadrats, (3) measuring the chlorophyll content, (4) taking photographs of the quadrats and (5) subtilization of some of the flowers in the quadrats. Vegetation indices were also generated from collected spectral data during the data processing stage, and the flowering rate in each canopy was determined by the supervised classification method. Relations between flowering rate and spectral data were analyzed by Pearson's correlation coefficient and linear regression models. The results show that there was a linear relationship between the flowering rate and the spectral reflectance in the red and green regions, whereas there was no statistically significant relationship with the reflectance in the NIR region. Moreover, all vegetation indices, especially REP, were affected from flowering variations. This effect was found to be lower in heterogeneous samples and higher in homogeneous samples. Evidence was found that the basic factor governing this effect was the fact that the flowers formed an obstacle to the detection of chlorophyll content by covering a certain part of the canopy.

Aboveground biomass (AGB) is one of the key indicators of aboveground net primary productivity (ANPP). The aim of this study is to demonstrate the potential of hyperspectral remote sensing techniques to predict AGB in grasslands. In order to reach this goal, biomass properties with different ecological features and altitudes of 550 m, 1200 m, and 1400 m above sea level were investigated. Twenty-one biomass samples and hyperspectral measurements were collected from each region and a total of 63 samples were analyzed. Linear and nonlinear regression models were generated to analyze the relationships between biomass and hyperspectral vegetation indices (VIs). The results showed strong relationships between VIs and biomass variations. However, dense biomass samples indicated weaker relationships with VIs due to saturation phenomena. Findings based on the measured data showed that AGB (except dense biomass) can be estimated with high accuracy using hyperspectral vegetation indices.