INTRODUCTION
Salinization is an edaphological process that frequently affects the soils of arid and semi-arid regions according to Wang et al. (2020), which limits the number of species that can be cultivated in these soils. The salinity of a soil is determined by the set of all soluble salts contained in the soil and can be estimated by measuring the electrical conductivity (EC) of a solution extracted from the soil.
They are commonly evaluated through laboratory tests, which leads to a greater expenditure of time and resources (El Harti et al., 2016). Given this situation, the use of remote sensing or remote sensing becomes more feasible. This novel technique, with great potential for collecting soil data, has been used for the detection of salinity both spatially and temporally (Allbed et al., 2014).
Their study is based on the spectral characteristics of the soil. Soil salinity monitoring is based on measuring the increase in soil reflectance as the salinization refracted in the short and long range bands of visible infrared increases (Sidike et al., 2014). In research carried out in Cuba, (Lau et al. 2003, 2005) confirmed the use of remote sensing as a supporting tool for the management of soil salinity in crops through salinity maps, given that it is a stress factor in sugarcane cultivation.
Due to the above, the objective of the research was to determine the relationship between soil salinity and the vegetation state of sugarcane cultivation in areas of the Sugar Mill UEB “Urbano Noris”, in Holguín Province by using remote sensing.
MATERIALS AND METHODS
The investigation was carried out in November 2020. To download the satellite image, the United States Geological Survey (USGS) was accessed, in grid 011/046 and the one referring to September 19, 2020 was taken (LC08_L1TP_011046_20200919_20201006_01_T1. tar) of the Landsat 8 OLI / TIRS satellite which has the following characteristics (Table 1).
Bands | Color of bands | Wavelength (µm) | Resolution (meters) |
---|---|---|---|
1 | Coastal spray | 0.433-0.453 | 30 |
2 | Blue | 0.450-0.515 | 30 |
3 | Green | 0.525-0.600 | 30 |
4 | Red | 0.630-0.680 | 30 |
5 | Near infrared (NIR) | 0.845-0.885 | 30 |
6 | Infrared Short wave (SWIR1) | 1.560-1.660 | 30 |
7 | Infrared Short wave (SWIR2) | 2.100-2.300 | 30 |
8 | Panchromatic | 0.500-0.680 | 15 |
9 | Cirrus | 1.360-1.390 | 30 |
10 | Thermal infrared (TIRS) 1 | 10.30-11.30 | 100 |
11 | Thermal infrared (TIRS) 2 | 11.50-12.50 | 100 |
Because of their cumulus and time, Landsat images have a spatial resolution of 30 m, temporal resolution of 16 days, and a ground cover of 185 km. For the analysis of the satellite images, the QGis 3.10A Coruña software was projected in the WGS 84 UTM Zone 18 North coordinate system where the radiometric correction was made and the spectral indices of vegetation and salinity were calculated (Table 2).
Spectral Index | Expression | References |
---|---|---|
|
(BNir-BRed)/( BNir+BRed) |
Rouse |
|
(BSWIR1-BSWIR2)/(BSWIR1+BSWIR2) |
Bannari |
For the determination of the relationship between soil salinity and sugarcane crop, it was considered that the cultivated soil is Vertisol type with glyic properties according to Hernández et al. (2015). A completely randomized sampling was carried out in the units of sugarcane blocks from one to ten in Sugar Mill UEB “Urbano Noris”. Fifty sampling points separated at a distance of 100 meters (Figure 1) were taken for the measurings whose values were located in the vegetation and salinity maps.
After extracting the values of the sampling points, a database was created to determine the relationship between the NDVI and the SI, applying linear regression analysis and the characteristics of its statistics in the software Stargraphics Plus 5.0. as statistics tool.
RESULTS AND DISCUSSION
The determination of the vegetative index NDVI is shown in Figure 2. There, it is observed that it oscillates between the values of -1 to 1 and, according to Rawashdeh' classification (2012) for values from 0 to 0.5, there is little vegetation in the study area.
The reflectivity of the vegetation covers is determined by optical characteristics and spatial distribution of all its constituents, which include the soil on which the vegetation is, as well as their proportions (Gilabert et al., 1997). Meera et al. (2015) refer that the decrease of water content in the soil due to various reasons in the ecosystem, causes a tendency to dissipate the crop greenness and, therefore, the NDVI values. One of the causes of the variation in the state of the vegetation is the absorption of energy (electromagnetic radiation absorbed inside the sensor). This energy is a process that is quantized, these features are located at specific wavelengths, which depend on the presence of certain components in the material according to Baret (1995), thus, for example, features that are a consequence of electronic transitions such as those due to the presence of iron oxides or the presence of chlorophyll are located in the visible region of the spectrum, while those due to rotational transitions such as those of the ion OH- occur in the near infrared area.
Remote sensing of salinity from NDVI has been widely used, since the state of the vegetation is presented according to salinity stress (Allbed et al., 2014). Figure 3 illustrates the determination of the salt index; for which Elhag (2016) reports that the values indicate the state of soil salinity varies from -1 (high presence of salts in the soil) to 1 (low presence of salts in the soil). In the visible region, the pigments of the leaves absorb most of the light they receive; in the near infrared these substances are quite transparent. For this reason, the cultivation of sugarcane in a good vegetative state offers low reflectivity in the red band of the spectrum and high in the near infrared, so that the greater the vigor that the crop presents, the greater the contrast between the reflectance values captured in both bands will be.
Low reflectance values in the near infrared and an increase in the visible one indicate that the sugarcane crop is in poor condition, one of whose causes is the high salinity of the soil (Soca, 2015).
Platonov et al. (2013) describe that in order to identify the state of the vegetation in the face of saline stress, the analysis of large areas is necessary due to the variation in the use and management of the soil by producers (Muller, 2017). Figure 4 refers to the linear regression analysis between the SI and NDVI spectral indices, for which a strong negative correlation of - 84.41% was obtained expressed in the Pearson coefficient and a determination of 71.25% for which there is a strong inverse dependence between both spectral indices of soil and vegetation.
Ding & Yu (2014) and Ivushkin et al. (2017) affirm that the salinity of the soil captured by remote sensing is related to vegetation indices, which have direct effects on the spectral information, given by the appearance of dark areas. In addition, the vegetation indices and the saline spectral index are commonly used as indicators to detect changes in the state of the vegetation cover. Scudiero et al. (2015) state that in arid and semi-arid areas the most frequent errors in determining the relationship between the content of salts and the state of the vegetation by remote sensing are the structure of the plant canopy, the type of soil management, the annual rainfall and temperature and the type of clay in the soil.
CONCLUSIONS
The use of NDVI as an indicator of the state of the vegetation showed the presence of vast areas under stress with values lower than 0.5, just as the saline index showed a high proportion of soils with high salt content, with negative indices from -1 to 0. The use of remote sensing to determine soil salinity showed that between these variables there is a negative correlation of -88, 41% and a determination of 71.25%, which defines an inverse dependence between them.