Comparison of soil organic carbon estimation using remote sensing and Near Infrared spectrum in forest and agricultural land uses in Gyan area, Hamadan province
Subject Areas : Natural resources and environmental management
Soheilasadat Hashemi
1
(Assistant prof. Soil Science Department. Malayer University)
Parinaz Abdoli
2
(MSc student, Malayer university)
Keywords: remote sensing, Landuse change, Band7, Near-Infrared,
Abstract :
In this study, the relationship between spectral reflections using Landsat 8 satellite sensors and near infrared spectrum with 48 soil samples were investigated in agricultural and forestry uses in Gyan Nahavand, Hamadan province . Soil samples were collected from 0-30 cm depth, randomly. The analysis of the correlation between main bands, artificial bands and soil surface organic carbon, as well as vegetation indices, composition of indicators and soil surface organic carbon were performed. Spectral analysis of soils using field spectrometer with wavelength range of 350-2500 nm was conducted. After recording the spectra, a variety of pre-processing methods were evaluated. The results showed that in the remote sensing method, only the 11 band shows a significant correlation at the level of 5% with organic carbon in agricultural. Also, band composition (band7/ band8) had a significant correlation at the level of 1% with organic carbon content. Three vegetation indices, NDVI, DVI and RAI with organic carbon showed a significant correlation at the level of 5%. The correlation between the calculated organic carbon in the laboratory and the image in agricultural land use was achieved R2 = 0.36. While the correlation of calculated organic carbon in the laboratory and the image was calculated (R2=0.32) at all points. In the spectroscopy method, the highest correlations were observed at wavelengths of 1404, 1907, and 2216, respectively. Among the fitted models given by the multiple regression, stepwise model is proposed for the estimation of organic carbon, a suitable model. Consequently if the number of samples is very low, the laboratory method may be appropriate, but if the number of samples is too high, the spectroscopy method is appropriate to save time, and in order to save costs. Due to the high cost of spectroscopy in Iran, the method of remote sensing is propose as appropriate method.
_||_