ORIGINAL ARTICLE

Estimation of the electrical conductivity of the soil from spectral information in the cultivation of rice (Oryza sativa L.)

0000-0002-8950-0268Ing. Roberto Alejandro García-Reyes^{I  *}  , 0000-0001-7410-6703Dr.C. Mario Damian González Posada-Dacosta^II  , 0000-0002-2436-0591MSc. Juan Alejandro Villazón-Gómez^III  , 0000-0003-2923-5092Dr.C. Sergio Rodríguez-Rodríguez^IV 

^IUniversidad de Holguín, Facultad de Ciencias Naturales y Agropecuarias, Departamento de Ciencias Agropecuarias, Holguín, Cuba.

^IIUniversidad de Granma, Facultad de Ciencias Técnicas, Bayamo, Granma, Cuba.

^IIIUniversidad de Holguín, Facultad de Ciencias Naturales y Agropecuarias, Centro de Estudios para Agroecosistemas Áridos, Holguín, Cuba.

^IVUniversidad de Granma, Facultad de Ciencias Agrícolas, Centro de Estudios de Biotecnología Vegetal, Bayamo, Granma, Cuba.

ABSTRACT

The salinity of the soil constitutes today one of the main degradation processes that affects the lands under irrigation; and especially the production of rice. The objective of the research was to estimate the electrical conductivity of the soil by means of spectral information in the cultivation of rice in the Mayarí municipality, Holguín. The research was developed on a Chromic Vertisol, which are the most prone in the territory to trigger degradation processes such as salinity, given by management practices in rice cultivation. Two semi-empirical models proposed from different spectral indices were used, the NDSI and the IS, which were calculated in the QGis 3.10 software with multispectral images from the Landsat 8 OLI / TIRS sensor. The NDSI and the SI-ASTER showed the lowest determination and negative correlation in both models. The SI and the SI-ASTER produce an overestimation of the electrical conductivity values ​​of the soil (EC ≥ 100%), the NDSI, SSSI.1 and SSSI.2 indicate a moderate content of salts in the soil (EC 20% ˂ EC ≤ 40%). Although the use of spectral salinity indices yielded a high determination, the SI and the SI-ASTER indicated an overestimation of the electrical conductivity existing in the soil, which could be due to the presence of a saturation of the signal captured by the sensor and reflected in the indices obtained, which exceeded the values ​​in which the saline index oscillates.

Key words: Saline Index; Salinity; Remote Sensing

INTRODUCTION

Soil salinity is considered an environmental problem in the world, especially in arid and semi-arid areas (^{Kumar et al., 2018}). Has been reported in irrigated agricultural lands in China, India, the United States and America ^{Zaman et al. (2018)}, and around 20% of the lands under irrigation in the world express severe damage by salinity, low yields, causing soil degradation and also loss of fertility (^{Ali et al., 2019}).

Rice is considered one of the crops sensitive to salinity, and it has been pointed out that from an electrical conductivity of the saturated soil paste extract (Cep) of 3 dS.m^-1, the potential yield declines rapidly in an environment of a 12% for each unit increase in soil Cep ^{Ayers & Westcot, (1985)} cited by ^{Pujol et al. (2009)}. In a traditional way, laboratory analysis is the analysis technique used to characterize soil salinity, which consumes a large amount of time and resources (^{Harti et al., 2016}). Geostatistical interpolation tools are also used ^{Balakrishnan et al. (2011)} for the analysis of the spatial and temporal variability of the information obtained.

Given this situation, the use of remote sensing or remote sensing becomes more feasible. This novel technique with great potential for collecting soil data, and 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. ^{Sidike et al. (2014)}, refer that soil reflectance increases when salinity increases in the visible spectrum of the red and infrared bands.

Among the main difficulties in obtaining spectral information on salinity is due to the presence of factors such as soil cover, organic matter content, and texture affect the acquisition of information in time and space (^{Ding y Yu, 2014}). Due to the aforementioned, the objective of the research was based on the estimation of the electrical conductivity of the soil from semiempirical models of spectral information in rice cultivation in the Mayarí municipality of, Holguín.

MATERIALS AND METHODS

The research was carried out in the rice production area known as Guaro, located in the Mayarí agroecosystem in the province of Holguín. A completely randomized sampling design was used with the elaboration of a grid of 50 meters spacing between projected points in the WGS84 / UTM zone 18 N system (Figure 1).

FIGURE 1 Agroecosystem Mayarí (a), image of the SAS Planet of the Guaro rice production area (b) and Grid of points (c). 

The image used (LC08_L1TP_011046_20181219_20181227_01_T1) for the extraction of spectral information comes from the commercial Landsat 8 OLI / TIRS satellite, downloaded from the site www.usgs.com of the United States Geological Survey. The satellite used has the following characteristics (Table 1). For the acquisition date of the multispectral images, the study area was planted with rice and the average rainfall up to that month was 79.0 mm and 26.6 ° C in temperature.

TABLE 1 Landsat 8 sensor characteristics OLI / TIRS according to (^{United State of Geological Survey, 2016}) 

Bands	Band Color	Wavelength (µm)	Resolution (m)
1	Coastal Aerosol	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 Shortwave (NIR)	0.845 - 0.885	30
6	Infrared Shortwave (SWIR1)	1.560 - 1.660	30
7	Infrared Shortwave (SWIR2)	2.100 - 2.300	30
8	Panchromatic	0.500 - 0.680	15
9	Cirrus	1.360 - 1.390	30
10	Thermal Infrared (TIRS1)	10.30 - 11.30	100
11	Thermal Infrared (TIRS2)	11.50 - 12.50	100

Radiometric correction was performed on the image to eliminate the effects caused by atmospheric factors in the QGis software version 3.10 A Coruña. The spectral index used for the estimation of electrical conductivity are shown in Table 2.

TABLE 2 Spectral index determined for the estimation of the electrical conductivity of the soil 

Spectral salinity index	References
$\mathit{N}\mathit{D}\mathit{S}\mathit{I}=(\mathit{B}4-\mathit{B}5)/(\mathit{B}4+\mathit{B}5)$	Khan et al. (2001)
$SI=SWIR1/SWIR2$	Khan et al. (2001)
$\mathrm{S}\mathrm{I}.\mathrm{A}\mathrm{S}\mathrm{T}\mathrm{E}\mathrm{R}=\mathrm{}(\mathrm{B}4-\mathrm{B}3)/(\mathrm{B}4+\mathrm{B}5)$	Al-Khaier (2003)
$\mathrm{S}\mathrm{S}\mathrm{S}\mathrm{I}1=\mathrm{}\mathrm{B}9-\mathrm{B}10$	Bannari et al. (2009)
$\mathrm{S}\mathrm{S}\mathrm{S}\mathrm{I}.2=(\mathrm{B}9\mathrm{*}\mathrm{B}10-\mathrm{B}10\mathrm{*}\mathrm{B}10)/\mathrm{B}9$	Bannari et al. (2009)

After the digital processing of the image resulting from the determination of the spectral index of salinity, the semi-empirical saline index of the electrical conductivity of the soil was predicted through mathematical models proposed by ^{Bannari et al. (2009)} which are indicated in the Table 3.

TABLE 3 Mathematical models for the calculation of the semiempirical index of electrical conductivity and spectral index for the determination of electrical conductivity from spectral information 

Spectral salinity index	Mathematical models to estimate the EC
NDSI	CE = 627.45x2 + 147.16x + 9.71
SI	CE = 1082.80x2-2202x + 1121
SI.ASTER	CE = 5643x2-117.91 x + 2.18
SSSI.1	CE = 3212x2 + 132.20x + 0.397
SSSI.2	CE = 4521 x2 + 124.5x + 0.407

For the classification of soil salinity, the one proposed by ^{Taylor (1993)} was used, it presents a scale that adopts salinity classes adapted for predictive models of electrical conductivity with values ​​between 0 % and 100% dS. m^-1 respectively. The statistical analysis was carried out in the Statgraphics Plus version 5.1 software in which the linear regression analysis of the spectral information and the estimated values ​​of electrical conductivity was carried out.

RESULTS AND DISCUSSION

In Figure 2 the linear regression analysis of the spectral index vs. electrical conductivity estimated by means of the semi-empirical mathematical models is presented. It can be seen that both the spectral index used and the mathematical models, SI vs CE and SI.ASTER vs CE overestimate the content of salts in the soil and heterogeneity in the estimated values ​​given by the use of different spectral bands for their determination. ^{Elhag (2016)} refers that the values ​​of the spectral index that indicate the state of soil salinity ranges from -1 (high presence of salts) to 1 (low presence of salts in the soil).

FIGURE 2 Linear regression analysis of spectral index vs electrical conductivity estimated by semi-empirical mathematical models. 

The use of linear models to determine the error in the determination of salinity by remote sensing is widely used to reduce the interferences caused by the spatial and temporal variation of salts in the soil (^{Scudiero et al., 2015}). Studies by ^{Ma et al. (2017)} and ^{Bannari et al. (2009}; ²⁰¹⁶⁾, validated the use of linear models for the estimation of electrical conductivity from spectral information of saline index calculated from images from the Landsat 8 OLI / TIRS satellite. In the literature on the use of remote sensors for the determination of saline index, various methods are proposed for their calculation.

^{Khan et al. (2001)} in their research propose the use of bands 3 and 4 of the LISS-II and IRS-1B sensor by means of the ground brightness adjustment index (BI), the NDSI and the IS. On the other hand, ^{Al-Khaier (2003)} determined by means of the NDSI different saline classes in the soil in a semi-arid zone with bands 4 and 5 of the ASTER sensor. In the study carried out by ^{Mashimbye (2013)} for the determination of electrical conductivity in the laboratory, he stated that the best way to detect soil salinity and sodicity through remote sensors is with the use of short-wave infrared-related bands (SWIR1).

Other scientists have directed their study on different types of soils and with salinity levels in which they use the red and near infrared bands for the analysis of vegetation and the types of salts in the soil (^{Howari et al., 2002}).

TABLE 4 Statistical information from the linear regression analysis between the spectral index and the electrical conductivity estimated from semi-empirical mathematical models 

Spectral index vs EC	Coefficient of determination	Correlation coefficient	Standar error	Absolute mean error	Durbin Watson	Model equation
NDSI vs EC	73,5168	-0,6596	6,5097	3,8087	0,0675	EC = -7,5038 - 86,8489*NDSI
SI vs EC	97,3034	0,9864	8,5195	5,8261	0,0518	EC = -820,9512 + 705,5234*SI
SI.ASTER vs EC	73,6875	-0,8584	1351,7802	981,0384	0,0242	EC = 210,0047 - 5286,91*SI.ASTER
SSSI.1 vs EC	96,9478	0,9846	2,7584	1,9690	0,0206	EC = -28,4225 + 0,0201*SSSI.1
SSSI.2 vs EC	97,2118	0,98596	1,5671	1,0441	0,0418	EC = -20,3179 + 0,0199*SSSI.2

Table 4 presents the statistical information of the linear regression analysis between the spectral index and estimated electrical conductivity. The NDSI vs EC model presented the lowest coefficient of determination with 73.5168% and a negative correlation of -0.6596, as did the SI model. ASTER vs EC with -0.8584 which deduces that as this index reaches Positive values ​​decrease the estimated electrical conductivity values. The remaining models presented a high determination and correlation coefficient close to 100%. In an investigation ion made by ^{Bannari et al. (2016)} obtained with the use of semi-empirical models and the NDSI and IS index, coefficients of determination of 70.0% and 67.0% respectively.

Based on the classification proposed by ^{Taylor (1993)}, there are differences between the estimated mean values ​​of electrical conductivity (Figure 3). According to this author, the NDSI and SSSI.2 index show moderate salinity (20% ˂ EC ≤ 40%) while the use of the SI-ASTER and SSSI.1 produce an overestimation with values ​​greater than 100% and evidence of an area with extreme salinity conditions (EC ≥ 100%). According to ^{El-Battay et al. (2017)} indicates the potentiality of the NDSI, SI and SI-ASTER index for estimating the electrical conductivity of the soil in arid agroecosystems, and refers that the SSSI.1 performed an overestimation with average values ​​of 145.24%; which causes confusion when it comes to an investigation of this type.

FIGURE 3 Graphic representation of the estimated mean values ​​of electrical conductivity through the calculated spectral index. 

CONCLUSIONS

The determination of electrical conductivity through the use of spectral information is an indicator of this property in the soil and rice crop conditions studied. Although the use of spectral salinity index yielded a high determination, the SI and the SI-ASTER indicated an overestimation of the electrical conductivity existing in the soil, which could be due to the presence of a saturation of the signal captured by the sensor and reflected in the index obtained, which exceeded the values ​​in which the saline index oscillates.

REFERENCES

ALI, S.K.; ALI, S.A.; WU, Y.; ABBASI, B.; ALPI, F.G.; ALI, L.I.; ZHOU, G.: ¨An assessment of the spatial and temporal distribution of soil salinity in combination with field and satellite data: a case study in Sujawal District¨, [en línea] Agronomy, DOI-https://doi.org/10.3390/agronomy9120869, 9(12): 869, 2019.

AL-KHAIER, F.: Soil salinity detection using satellite remotes sensing, 61pp., Master’s Thesis, International institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands, 2003.

ALLBED, A.; KUMAR, L.; ALDAKHEEL, Y.Y.: ¨Assessing soil salinity using soil salinity and vegetation indices derived from IKONOS high-spatial resolution imageries: Applications in a date palm dominated region¨, [en línea] Geoderma, ISSN-0016-7061, DOI-https://doi.org/10.1016/j.geoderma.2014.03.025, 230: 1-8, 2014.

AYERS, R.S.; WESTCOT, D.W.: Water quality for agriculture, Ed. Food and Agriculture Organization of the United Nations Rome, vol. 29, ISBN-92-5-102263-1, 1985.

BALAKRISHNAN, P.; SALEEM, A.; MALLIKARJUN, N.: ¨Groundwater quality mapping using geographic information system (GIS): A case study of Gulbarga City, Karnataka, India¨, [en línea] African Journal of Environmental Science and Technology, ISSN-1996-0786, DOI-https://doi.org/10.5897/AJEST11.134, 5(12): 1069-1084, 2011.

BANNARI, A.; GUÉDON, A.; EL-GHMARI, A.: ¨Mapping slight and moderate saline soils in irrigated agricultural land using advanced land imager sensor (EO-1) data and semi-empirical models¨, [en línea] Communications in Soil Science and Plant Analysis, ISSN-0010-3624, DOI-https://doi.org/10.1080/00103624.2016.1206919, 47(16): 1883-1906, 2016.

BANNARI, A.: GUEDON, A.; EL-HARTI, A.; CHERKAOUI, F.; EL-GHMARI, A.; SAQUAQUE, A.: ¨Slight and moderate saline and sodic soils characterization in irrigated agricultural land using multispectral remote sensing¨, [en línea] International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, DOI-https://doi.org/10.1080/00103620802432717, 34(Part XXX), 2009.

DING, J.; YU, D.: ¨Monitoring and evaluating spatial variability of soil salinity in dry and wet seasons in the Werigan-Kuqa Oasis, China, using remote sensing and electromagnetic induction instruments¨, [en línea] Geoderma, ISSN-0016-7061, DOI-https://doi.org/10.1016/j.geoderma.2014.07.028, 235: 316-322, 2014.

EL-BATTAY, A.; BANNARI, A.; HAMEID, N.; ABAHUSSAIN, A.: ¨Comparative study among different semi-empirical models for soil salinity prediction in an arid environment using OLI Landsat-8 data¨, [en línea] Advances in Remote Sensing, DOI-DOI: https://doi.org/10.4236/ars.2017.61002, 6(01): 23, 2017.

ELHAG, M.: ¨Evaluation of different soil salinity mapping using remote sensing techniques in arid ecosystems, Saudi Arabia¨, [en línea] Journal of Sensors, ISSN-1687-725X, DOI-https://doi.org/10.1155/2016/7596175, 2016, 2016.

HARTI, E.A.; LHISSOU, R.; CHOKMANI, K.; OUZEMOU, J.-E.; HASSOUNA, M.; BACHAOUI, E.M.; EL GHMARI, A.: ¨Spatiotemporal monitoring of soil salinization in irrigated Tadla Plain (Morocco) using satellite spectral indices¨, [en línea] International Journal of Applied Earth Observation and Geoinformation, ISSN-0303-2434, DOI-https://doi.org/10.1016/j.jag.2016.03.008, 50: 64-73, 2016.

HOWARI, F.; GOODELL, P.; MIYAMOTO, S.: ¨Spectral properties of salt crusts formed on saline soils¨, [en línea] Journal of Environmental Quality, ISSN-0047-2425, DOI-https://doi.org/10.2134/jeq2002.1453, 31(5): 1453-1461, 2002.

KHAN, N.; RASTOSKUEV, V.; SHALINA, E.; SATO, Y.: ¨Mapping salt-affected soils using remote sensing indicators a simple approach with the use of GIS IDRISI¨, En: 22nd Asian Conference on Remote Sensing, Center for Remote Imaging, sensing and Processing (CRISP), National University of Singapore; Singapore, pp. 5, Singapore, 2001.

KUMAR, N.; SINGH, S.K.; PANDEY, H.: ¨Drainage morphometric analysis using open access earth observation datasets in a drought-affected part of Bundelkhand, India¨, [en línea] Applied Geomatics, ISSN-1866-928X, DOI-https://doi.org/10.1007/s12518-018-0218-2, 10(3): 173-189, 2018.

MA, Z.; SHI, Z.; ZHOU, Y.; XU, J.; YU, W.; YANG, Y.: ¨A spatial data mining algorithm for downscaling TMPA 3B43 V7 data over the Qinghai-Tibet Plateau with the effects of systematic anomalies removed¨, [en línea] Remote Sensing of Environment, ISSN-0034-4257, DOI-https://doi.org/10.1016/j.rse.2017.08.023., 200: 378-395, 2017.

MASHIMBYE, Z.E.: Remote sensing of salt-affected soil, 151pp., Ph.D. Thesis, Stellenbosch University, Faculty of AgriSciences, Stellenbosch, South Africa, 2013.

PUJOL, O.R.; HERRERA, P.J.; CID, L.G.; ALARCÓN, L.R.; LÓPEZ, C.G.: ¨Evaluación del funcionamiento hidráulico del drenaje soterrado en áreas arroceras de la zona norte de la provincia Granma, Cuba¨, Revista Ciencias Técnicas Agropecuarias, ISSN-1010-2760, e-ISSN: 2071-0054, 18(3): 30-34, 2009.

SCUDIERO, E.; SKAGGS, T.H.; CORWIN, D.L.: ¨Regional-scale soil salinity assessment using Landsat ETM+ canopy reflectance¨, [en línea] Remote Sensing of Environment, ISSN-0034-4257, DOI-http://dx.doi.org/10.1016/j.rse.2015.08.0260034, 169: 335-343, 2015.

SIDIKE, A.; ZHAO, S.; WEN, Y.: ¨Estimating soil salinity in Pingluo County of China using QuickBird data and soil reflectance spectra¨, [en línea] International Journal of Applied Earth Observation and Geoinformation, ISSN-0303-2434, DOI-https://doi.org/10.1016/j.jag.2013.06.002, 26: 156-175, 2014.

TAYLOR, S.: Dryland salinity: introductory extension notes, Ed. Department of Land and Water Conservation, pp. 63, ISBN-978-0-7310-1124-7, Sydney, Australia, 1993.

UNITED STATE OF GEOLOGICAL SURVEY: ¨Landsat Earth observation satellites: U.S. Geological Survey Fact Sheet 2015-3081¨, [en línea] DOI-https://doi.org/10.3133/fs20153081, 4, 2016.

ZAMAN, M.; SHAHID, S.A.; HENG, L.: Guideline for Salinity Assessment, Mitigation and Adaptation Using Nuclear and Related Techniques, Ed. Springer: New York, vol. 3, pp. 43-53, ISBN-978-3-319-96190-3, NY, USA, 2018.

The mention of trademarks of specific equipment, instruments or materials is for identification purposes, there being no promotional commitment in relation to them, neither by the authors nor by the publisher.

Received: February 01, 2021; Accepted: September 20, 2021

^*Author for correspondence: Roberto Alejandro García-Reyes, e-mail: [e-mail]

Roberto Alejandro García-Reyes, Profesor de la Universidad de Holguín, Facultad de Ciencias Naturales y Agropecuarias, Departamento de Ciencias Agropecuarias, Holguín, Cuba, e-mail: [e-mail]

Mario Damian González Posada-Dacosta, Profesor de la Universidad de Granma, Facultad de Ciencias Técnicas, Bayamo, Cuba, e-mail: [e-mail]

Juan Alejandro Villazón-Gómez, Profesor de la Universidad de Holguín, Facultad de Ciencias Naturales y Agropecuarias, Centro de Estudios para Agroecosistemas Áridos, Holguín, Cuba, e-mail: [e-mail]

Sergio Rodríguez-Rodríguez, Profesor de la Universidad de Granma, Facultad de Ciencias Agrícolas, Centro de Estudios de Biotecnología Vegetal, Bayamo, Cuba, e-mail: [e-mail]

The authors of this work declare no conflict of interests.