Evaluation of Dust Effects on Spectral Behavior of Plants Using Remote Sensing Data

Dust is one of the important climatic phenomena that has occurred in arid and semi-arid regions of the world. In recent years, one of the most important environmental issues in the Middle East and Iran is the occurrence of dust phenomena that affects a variety of factors, including human health, plants and other living organisms, economic conditions, and etc. One of the effective factors in soil stabilization and reducing the amount of dust particles in the air is vegetation and especially agricultural products, which play a significant role in the environmental cycle, human life and the alive creatures. However, plants will suffer from tension and disease through the influence of the dust occurrence. Hence, the main objective of this research is study the effect of dust on the plant in Ahwaz city One of the methods, which reduce costs and prevent direct tests on the plants and thus reduce the needed time, is to use remote sensing techniques. Therefore, two types of data including spectroscopy in different measurement conditions and satellite images (Landsat8) were used to obtain required data in this research. The ASD Field spek 3 spectrometer with a range of 350 to 2500 nm and a resolution of 1 nm was used in this study. For this purpose, 20 leaves of Spindle tree (Shamshad) plant were prepared for spectrophotometry. The leaves of this plant were examined in a dark room and based on specific standards and in each of the conditions, five replicates of the spectrophotometry were performed. Then, noise of spectrophotometric data was removed by wavelet transform. In the next step, the bands that were affected by the dust were identified by using the obtained spectra. Afterward, different vegetation indices such as Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Divergence Vegetation Index (DVI) and Ratio Vegetation Index (RVI) were extracted and analyzed in proportion to the image bands. In the next step, nonparametric spectrum analyzes (e.g. periododogram, Welches and multispepers) were used to analyze the effect of dust on signal strength. The analysis was performed individually on measured data in the laboratory environment and outdoors. The CR method illustrated that the best wavelengths for detecting vegetation free of dust and dust cover is from about 450 to 750 nm and 1500 to 2500 nm. Therefore, the vegetation behavior at the time before and after the occurrence of dust in the study area at 450 to 750 nm wavelengths was investigated using Landsat satellite imagery. Based on this, the results of the studied vegetation indices using the red and infrared band obtained after the occurrence of dust showed that the dependence of the indices NDVI, DVI, RVI before the occurrence of dust is 0.302, -0.47, and 0.35 respectively. Also, the dependence of the EVI index (which has three bands) obtained for the time before the occurrence of dust and the actual values ​​of EVI before the occurrence of dust was equal to 0.69, which is a relatively good correlation between the EVI values and measures the values ​​of this index before the occurrence of dust event. Also, spectrophotometry results demonstrated that the signal strength decreases with increasing dust. This result can be deduced for both field and laboratory spectra. According to the accuracy obtained for the vegetation indices, it was determined that using the images can not detect the effect of dust on the plant, properly. While comparison of the results of triodegraph, welch and multispeed methods showed that only periodogram signal processing is not suitable for detecting the effect of dust on the plant signal, and the other two methods (multifrequency and volcano) are superior to identify the effect of dust on the plant spectrum. The accuracy obtained for the periododogram in the laboratory environment and in the external environment were respectively 83% and 69%, the accuracy obtained from the Welch method in the laboratory environment was 60% and in the natural external environment was 97%, and the accuracy obtained from the multi-tipper method was 87% in the laboratory environment and 88% in the natural environment. According to the results, it can be observed that the dust on the plant can be determined better by using Welches and Multitipper methods.