Mapping and characterization of salt-affected and waterlogged soils in the Gangetic plain of central Haryana (India) for reclamation and management

IRS LISS III Resource SAT data (2005–07) were integrated with ground truth and soil studies for delineation and characterization of salt-affected and waterlogged soils in the Indo-Gangetic plain of central Haryana. The quality appraisal for salty ground water was also conducted prior to its use for irrigation. Such studies are useful for planning reclamation and management of salt-affected soils and poor quality ground water. Strongly sodic soils were easily identified based on the white to yellowish white tones, high spectral and low NDVI values. Waterlogged areas (surface ponding) were detected based on higher absorption in infrared range. Sodic soils with poor quality ground water showed higher reflectance from dry salts during June and freshly precipitated moist salts in March and October. Sodic soils irrigated with normal ground water showed higher cropping density and higher NDVI values. Moderately and slightly sodic soils showed mixed spectral signatures for salt crusts, moderate cropping density and surface wetness. Soil profile studies indicated higher moisture content at sub-surface depths. The presence of iron and manganese mottles indicated the incidences of water stagnation. Soils with high pHs, ESP, and SAR values and showing the dominance of carbonate and bicarbonates of sodium in the saturation extract indicated sodic nature. Significant presence of CaCO3 concretions at 1 m depth, low organic carbon contents, clay illuviation at sub-surface depth are typical features in sodic soil profiles. Water samples with high pH and SAR values and at places high RSC (Residual Sodium Carbonate) content indicated their sodic nature. Gypsum application is recommended for the reclamation of sodic soils and sodic water. *Corresponding author: A.K. Mandal, Department of Soil and Crop Management, Central Soil Salinity Research Institute, Karnal 132001, Haryana, India E-mail: arupkmondal@gmail.com


PUBLIC INTEREST STATEMENT
The submitted paper highlighted the potential of remote sensing data for natural resource inventory for salt-affected soils and poor quality ground water in the arid and semiarid regions and also used geo-informatics for decision-making and planning in the land reclamation, soil and water management and sustainable land use planning in the Gangetic plain of India.

Global and national distribution
Salt-affected soils are commonly distributed in arid and semiarid climatic zones and covered 1,307 M ha at global scale (FAO/IIASA/ISRIC/ISS-CAS/JRC, 2008). The largest areas of salt-affected soils are in Australia followed by North and Central Asia, South America and South and West Asia. An estimated area of 6.73 M ha salt-affected soils are in India, of which 2.5 M ha is in the Indo-Gangetic plain (Mandal, Obi Reddy, & Ravisankar, 2011;Mandal & Sharma, 2006;National Remote Sensing Agency, 2008;Saxena, Sharma, Verma, Pal, & Mandal, 2004). In central Haryana, four districts Karnal, Kurukshetra, Panipat and Sonepat, were worst affected, showing 52% of the geographical area (TGA) under salt-affected soils (Mandal & Sharma, 2005). Interpretation of Landsat images showed old levees, relict flood plain and poorly drained low-lying flats are common topographic zones with salt infestation along the Gangetic alluvial plain (Manchanda & Iyer, 1983). The introduction of canal irrigation from Western Yamuna Canal (WYC) in Haryana during the 1950s accentuated upward movement of salt by rising water table (Singh, Bundela, Sethi, Lal, & Kamra, 2010). Due to the over use of irrigation water in poorly drained areas, waterlogging and secondary salinization appeared and caused losses in productivity for rice (42%), wheat (38%) and sugarcane (61%) crops (Samra, Singh, & Ramakrishna, 2006). Due to the use of salty ground water (60-70% of TGA) for irrigation, secondary salt enrichment in soil profiles occurred along the Ghaggar and Markanda river plains (Gupta, 2010;Manchanda, 1976;Phogat, Satyavan, & Sharma, 2011).

Modern tools and techniques for diagnosis and assessment of salt-affected soils
Because of the large spectral coverage and discreet bands, remote sensing data have been used for mapping and monitoring salt-affected and waterlogged soils in a time and cost-effective manner (Dwivedi, 2006;Rao et al., 1998;Saxena, 2003;Shrestha, 2006). Mougenot, Pouget, and Epema (1993) easily identified barren, salt-affected soils by high reflectance in the visible range, while studies conducted in thermal, infrared and microwaves ranges were used to characterize hygroscopic characteristics of salts and vegetation-covered soils, respectively. Howari (2003) and Howari, Goodell, and Miyamoto (2002) used spectro-radiometry as a remote sensing tool in visible and near infrared bands to quantify spectral ranges for salt-affected soils with variable salt composition. Khan, Rastoskuev, Sato, and Shiozawa (2005) used ratio indices, spectral properties and digital image classification for mapping hydro-saline land degradation in the Indus basin of Pakistan.

Traditional methods for ground estimation and characterization of degraded soils
Studies conducted by Metternicht and Zinck (1997) showed different approaches for mapping sodium and salt-affected soils, combining digital analysis with field observations and laboratory analysis. They concluded that the main causes of spectral confusions, masking different soil salinity-alkalinity degrees, were the type and abundance of salt-tolerant vegetation cover, topsoil texture and other field properties. Joshi and Sahai (1993), Sharma, Saxena, and Verma (2000) and Verma, Saxena, Barthwal, and Deshmukh (1994) used a similar approach combining remote sensing, ground truth and soil analysis data for mapping coastal salt-affected soils in Saurashtra (Gujarat State) and inland salt-affected soils of Uttar Pradesh State. Such methods are laborious and need concerted efforts for image analysis, collection of ground truth and laboratory analysis of soil and water to integrate for mapping, but produce results and classified outputs of salt-affected soils with higher accuracies. Classification of soils for salinity/alkalinity classes such as slight, moderate and strong, is useful for deciding precise soil reclamation and management options. http://dx.doi.org/10.1080/23312041.2016.1213689

Justification and objectives
The complexity of soil salinity, alkalinity and waterlogging problems in central Haryana and the Gangetic Plain of India were reported by several authors (Gupta, 2010;Mandal & Sharma, 2005;Raj kumar, Ghabru, Singh, Ahuja, & Sharma, 2010;Singh et al., 2010) and is a primary concern for reclamation and management. The complex surface properties of salt-affected and waterlogged soils varied in seasonal imageries causing low mapping accuracies (Sharma, Saxena, Verma, & Mandal, 2008;Verma, Saxena, & Bhargava, 2007). Field validation is therefore necessary for spatial characterization of salinized areas followed by the chemical characterization of soil samples to assess degrees of limitations required for reclamation and management. Keeping in view the use of poor quality ground water for irrigation and its impact on soil degradation (salt enrichment) and reduced crop production (Gupta, 2010), chemical characterization of ground water is also necessary before its use in irrigation. To address these issues, the present study is aimed at the delineation and characterization of degraded (salt-affected and waterlogged) soils, and appraisal of ground water quality in central Haryana useful for planning reclamation and management.

Study area
The study area (29°52′58.32″N to 30°15′34.42″N latitude and 76°25′31.31″ to 77°21′19.19″E longitude) covered administrative boundary of Kurukshetra district of central Haryana (1,530 km 2 ) and lies 253 m above mean sea level. The average annual rainfall is 608 mm, mean winter temperature is 12.7°C and mean summer temperature is 38.5°C. The landform is alluvial under the Gangetic alluvium. The area is drained by the Yamuna, Ghaggar and its tributaries Markanda, Saraswati, Chautang, Tangri and other seasonal streams Sahibi, Dohan and Krishnawati that originate from the Aravalli Hills. The primary source of irrigation is the WYC and Bhakra canal. In the absence of canal irrigation supply ground water from tube wells is commonly used for irrigation. Prolonged irrigation altered the moisture regime and chemical characteristics of soils leading to salt infestations, waterlogging and low productivity (Singh, 2009).  Table 1.
(6) GPS (Lawrence global) for collecting location-data for soil profiles and soil sampling sites, water samples and tube wells.

Image processing and spatial analysis
The pre-processed IRS images for atmospheric corrections (by NRSC) were geo-referenced using the Survey of India topographical maps at 1:50,000 scale. The data from different bands were integrated to prepare a digital mosaic for the study area, using ERDAS software. Different band combinations, B321 (NIR, R, G) and B432 (SWIR, NIR, R) with histogram equalized (256 intervals) stretches were used to develop False Color Composites (FCC) for visual analysis of degraded soils (National Remote Sensing Agency, 2007). Based on the different manifestations of soil salinity such as tone, texture and patterns, the images (Figures 1 and 2) were visually interpreted to identify degraded soils (Table 2). Spectral reflectance was calculated based on the mean reflectance in bands B2, B3 and B4 (Table 1). A principal component analysis was carried out to prepare homogenous data-sets and filters were used to improve sharpness of the images for visual analysis. The spectral response patterns were analyzed for spatial characterization of image elements such as crop, riverine sand, salt-affected and waterlogged soils ( Figure 3). The NDVI values were calculated using the band ratios [(B3 − B2)/ (B3 + B2)] for differentiation of crop and non-crop areas (Figures 1 and 4). A supervised classification of digital data was carried out using a nearest neighborhood operator. An interactive-database was prepared comparing the map units prepared by digital and visual analysis to generate a confusion (error) matrix for accuracy assessment (Table 3). A flow chart showing methodology for mapping salt-affected soils were presented in Figure 5 for clear understanding.

Ground truth studies for soil profile and water quality
A ground truth survey was conducted during March (pre-monsoon) and October (post-monsoon) 2005-07 seasons to authenticate interpreted units in the field and locate salt-affected and waterlogged areas. The areas showing salinity emergence in different topographic zones and land uses such as crop and non-crop areas were studied and the data on status, condition and types of vegetation tolerant, partially tolerant and non-tolerant crops were also recorded. The salinity status at surface and sub-surface depths were obtained from soil profile studies. The field salinity/alkalinity status of soil samples was measured by portable pH and EC meters. The ground truth observations sites, soil profiles/soil and water sampling sites and topographical data on slope, aspects, contours, and related ground control points were collected during the ground truth study and were marked on the topographical maps. Water table depths data were also collected in waterlogged areas under canal irrigation to 1.5 m depth below the surface. Ground water samples were collected from tube wells for detailed chemical analysis. Ground water table depths were also recorded to relate with geology data.
Representative soil profiles (1.5 m depth) were studied to assess status and distribution of soil salinity and alkalinity at 24 sites covering the study area. Soil morphological properties such as soil moisture content, texture, color, structure and drainage were recorded from soil profile studies. Soil samples were collected at representative depths up to 1.2 m and properly stored in polythene and cloth bags to minimize moisture loss and changes in salt composition. These were further air-dried, processed to pass through <2 mm sieve and stored for physical and chemical properties (Table 4).

Studies for physical and chemical properties of soil and water samples
In the laboratory, soil samples were analyzed for physical and chemical properties such as soil reaction (pHs) and electrical conductance (ECe, dS m −1 ); salt composition for soluble Na + , K + , Ca 2+ , Mg 2+ ,       Fifteen ground water samples were collected from different locations to study water quality for agricultural applications (Table 5). These were analyzed for pH iw and EC iw (dS m −1 ), soluble cations and anions (Na + , K + , Ca 2+ , Mg 2+ , CO 2− 3 , HCO − 3 and Cl − ) using the methodology described by Richards (1954). The sodium adsorption ratio (SAR) [Na + /{(Ca 2+ + Mg 2+ )/2} 1/2 ] and residual sodium carbonate 2+ )] values were also calculated for classification of saline and sodic water (Richards, 1954).

Preparation of the thematic layers for base map and degraded soils
The Survey of India topographical maps at 1:50,000 scales, Universal Transverse Mercator (UTM) projections and ILWIS GIS software were used for geo-referencing and digitizing thematic layers for administrative and political boundaries (state/district), infrastructure (roads/railways), irrigation/ drainage (canal/river) and settlements (state/district capitals). These layers were overlaid to prepare a base map for the study area. GPS was used to collect geo-referenced data for soil profile locations, soil and water sampling sites, tube wells and were stored in an attribute table that was linked with the base map. The spatial coverage of interpreted units was delineated using on-screen digitizing and were overlaid on the base map. Distinguishing colors were used for representative map units. These were annotated for scale, north direction, legends, title, boundary coordinates and other cartographic elements (Figures 1, 2 and 6).

Mapping of degraded soils
An integrated approach of image interpretation, ground truth survey and laboratory analysis data for soil physical and chemical properties was used for mapping degraded soils in the Indo-Gangetic plain of Central Haryana ( Figure 5). Legends were developed for mapping degraded soils based on the methodology developed by National Remote Sensing Agency (2007). Categories of salt-affected and waterlogged soils were identified based on the soil physical and chemical properties. The area statistics of map (soil) polygons were used to assess spatial extent of salt-affected and waterlogged soils (Dwivedi, 2006;, 2012.

Image interpretation and ground truth studies of degraded soils
The spatial characteristics of salt-affected and waterlogged soil; natural vegetation and field crops were presented with soil chemical properties and ground water data ( Table 2). The strongly sodic soils (white to yellowish white tone in B321, irregular shape), normal crops (bright red tone, continuous), waterlogged (surface ponding, dark blue to black tone, irregular shape) soils and riverine sands (yellowish white with definite shape along the river bed) were easily detected based on their strong signatures from the visible and infrared bands in IRS data (Figures 1 and 2). The seasonal data showed higher extents of moist salt-affected soils and waterlogged areas (irrigated) during March and October ( Figure 1) and dry salts (salt crust) during June. This may possibly be due to similar reflectance of salt  and sand and the absence of vegetative cover during the dry season. The normal cropped areas were identified by the distinct (bright) red tones at different growth stages, while stressed vegetation was identified by lighter red tones, patchy occurrence and patchy white tones for salt crust in the saline and moist surface in waterlogged areas (Mandal & Sharma, 2010). Riverine sands were identified by the yellowish white tone and spotted natural vegetation along the river course (Figure 1).
Field studies indicated prominent salt crusts, scanty vegetation, scattered salt-tolerant natural vegetation, scrub and pastures in strongly sodic soils and in places, intercepted with forestry plantations for biological reclamation. Moderate and slightly sodic soils appeared as mixed red and gray tones in irrigated areas, the ground truth studies showing patchy salts, scattered crop cover, moist soil surface, low permeability and absence of natural drainage. In dry areas, strongly sodic soils appeared as white patches of barren salt crust underlain by sodic ground water. Moderately sodic soils appeared as tiny white patches and red to dark red tones for crops irrigated with good quality ground water . Partially reclaimed sodic soils showed moderate crop cover and intermittent salt patches in low lying flats and depressions (Mandal, 2012). Slightly sodic soils showed good to very good crop and vegetative covers (Howari, 2003), though the field study reported low productivity due to crop damages in the maturity stage.
In irrigated areas, permanent waterlogged soils (surface ponding) were in the low-lying flats/depressions and appeared as gray to dark gray tones in all seasons (Figure 2). Mixed red and reddish gray tones were identified in the irrigated areas supporting vegetation. Field studies indicated high water table depth (sub-surface waterlogging, WT < 1.5 m depth), crop cover and secondary soil salinization during the post-monsoon season (Mandal & Sharma, 2001, 2010. However, using moderate spatial and spectral resolution of IRS data, the segregation of mixed signatures of water and crop in sub-surface waterlogged areas was difficult . It was authenticated on the strength of ground truth.

Digital analysis of remote sensing data for spectral properties of degraded soils
Spectral analysis of IRS data identified prominent energy absorption for waterlogged areas (surface ponding, SP_S) during October (B3 > B4 > B2) and March (B3 > B4 > B2) (Mandal & Sharma, 2001). The NDVI values were low (0.1-0.3) as a result of low crop cover . In irrigated areas, spectral values ranging from 60 to 148 in B3 and 58 to 66 in B4 indicated high water table depth or sub-surface waterlogging SSW . The NDVI values (0.24-0.34) indicated the presence of stressed vegetation (Dwivedi & Sreenivas, 2002;Joshi, Toth, & Sari, 2002).
Matured winter crops showed higher reflectance of B3 during March (B3 > B4 > B2) while crops in moist soil surface showed higher values of B4 during October. NDVI values showed similar trends during March and October, respectively. The spectral reflectance of riverine sand was high (60-100) due to bare surface ( Figure 3) and low NDVI values (−0.04 to 0.04) which indicated scanty vegetative cover (Figure 4).
Principal component analysis was performed to homogenize digital data and achieve higher accuracy in classification. The principal component coefficients (PC) showed significant relationship between B1 and PC1 (0.524); B2 and PC4 (0.831); B3 and PC1 (0.707); and PC2 (0.683); B4 and PC3 (0.670). PC1 showed 93.5% variance while PC2, 3 and 4 showed 5.91, 0.30 and 0.09% variance, respectively. An average (AVG 3 × 3) filter was used to enhance sharpness of the images for visual analysis and to reduce noises prior to multi-band image classification. The nearest neighboring nine pixels were calculated to assign the values for central pixel to reduce noises and enhanced interpretation of the images.
Digital classification was performed using a supervised classification based on maximum likelihood classifier. Ground truth, laboratory analysis and land use data (field crop, forestry, urban settlement, road, natural water for pond, river and canal) were included as training sets for digital classification. Legacy data such as digitized maps of salt-affected soils, water table depth and quality and other collateral data including topographical maps of the Survey of India were also used as supporting data (Saxena, 2003;Verma, Singh, Sreenivas, Dwivedi, & Mathur, 2004). The salt-affected soils map at 1:250,000 scale was also consulted as supporting data. Clusters of pixels showing average reflectance for B1-B4 in March data were assigned a class name and the sample statistics (feature space) of the training set was generated to provide a visual overview of the separation of classes for the training pixels using a scatter plot for two bands. The feature spaces for B1 and B2, B1 and B4, and B2 and B4 indicated positive relation while B1 and B3, B2 and B3, and B3 and B4 showed partial or null relationships. An interactive (cross) database was prepared using maps prepared from visual analysis and digital classification. A confusion (error) matrix was prepared to assess the accuracy of digital classification (Table 3). The data showed an overall accuracy of 25.4%, average accuracy of 18.0% and reliability 10.5%, respectively. The highest accuracy was shown for slightly sodic soil (34%, reliability 65%) followed by sub-surface waterlogging (27%, reliability 3%), riverine sand (27%, reliability 21%) and moderately sodic soil (11%, reliability 35%), respectively.

Physical and chemical characteristics of salt-affected soils and waters
The field morphological characteristics of four representative soil profiles ranges from deep to very deep, pale brown to dark yellowish brown, sandy loam to sandy clay loam/clay loam texture, medium to strong, coarse to fine angular/sub-angular blocky structure, sticky, plastic to very sticky, very plastic consistency, presence of few to abundant CaCO 3 nodules and moist to wet sub-surface horizons. A few iron and manganese mottles were also found in sub-surface (50 cm) layers of P3 (Markanda plain) and P4 (Ghaggar plain), due to prolonged saturation with water. CaCO 3 concretions (2-5 cm, 10-30%) were found at 1 m depth in P2 and P4. The textural changes occurred from sandy loam to sandy clay loam and sandy clay loam to clay loam at P1, P2, and P4 apparently due to clay illuviation. The silt and clay contents were higher than sand content in P3 and P4 possibly due to lower topographic position.

Distribution of salt-affected and waterlogged soils
The spatial distribution of salt-affected and waterlogged soils is shown in Figure 6 and the extents were presented in Table 6. Slightly sodic soils have the largest area (10,409 ha) covering 61% of the total degraded soil in Kurukshetra district. It is followed by moderately sodic soils 5,697 ha covering 33.6% area and strongly sodic soils that occur in 0.2% of the area. Surface ponding occupies 363 ha (2.1%) while sub-surface waterlogging (203 ha) covers 1.2% area. Riverine sand covers 210 ha (2.1%) along the flood plain of the Markanda River.

Remote sensing studies
The digital analysis of remote sensing data revealed mixed surface properties for salts, soil particles during dry (June) season and complex spectral signatures of moist soil surface and moderate crop cover in salt-affected soils (Khan et al., 2005). The similarity of spectral signatures for village settlements (muddy roof top) and barren salt-affected soils caused spectral confusion during digital analysis. Visual analysis revealed definite shape and sizes of rural settlements that differs from irregular pattern in salt-affected soils (Khan et al., 2005). Mixed gray to reddish gray and mottled red tones indicated waterlogging in cropped areas , which was authenticated during field studies. The linear shape of canals and typical curvilinear meandering rivers differs from stagnant water bodies (waterlogged surface) though these elements showed similar spectral reflectance. Irrigated areas with poor quality ground water showed mixed spectral signatures for poor crop stand (light to red tone) and moist soil surface (light to gray tones). Ground truth studies showed salt enrichment, unfavorable physical properties and poor drainages in soil profiles Sharma & Mondal, 2006). The low reflectance values of irrigated sodic soils in March data (40-60) appeared to be due to surface moisture. Similar results were reported for carbonate rich salts in visible (0.55-0.77 um) and infrared (0.9-1.3 um) ranges (Csillag, Pasztor, & Biehl, 1993;Khan et al., 2005;Rao et al., 1995). The higher NDVI values of moderately sodic soils (0.29-0.52) may be ascribed to higher vegetative cover and also management interventions at selected locations Raghuwanshi, Tiwari, Jassal, Raghuwanshi, & Umat, 2010). The mixed reddish gray to dark gray tone for sub-surface waterlogged areas indicated scattered crop cover, and higher moisture content at soil surface.

Soil studies
Slight to strong soil alkalinity/sodicity indicated variable and complex chemical properties of sodic soils in the Kurukshetra district. The higher soil pHs (P3) at 40 cm depth indicated unfavorable soil physical properties and development of waterlogging. The high soil pHs of P1 (9.6-10.2) and P2 (9.8-10.7) at surface depth also limited its use for arable cropping. The dominance of CO 2− 3 + HCO − 3 anions and high Na + content in P1, P3, P2 and P4 indicated the sodium carbonate and bicarbonate parent materials that favored sodicity development in soils (Bhargava, Sharma, Pal, & Abrol, 1980;Sharma, Mandal, Singh, & Singh, 2011). The low contents of Ca 2+ + Mg 2+ are due to precipitation of calcium carbonates in an alkaline medium (Bhargava & Bhattacharjee, 1982). The texture analysis of P4, P3, P1 and P2 indicated higher clay contents in sub-surface layers that caused restricted drainage and favored waterlogging. Higher CEC values in P2 and P4 is attributed due to higher clay content. The high ESP values showed significant saturation with exchangeable Na + that favored alkali soil formation. The high CaCO 3 contents caused drainage congestion. The soil physical and chemical properties indicated variable alkalinity dominated by alkaline earth metals and poor drainage caused low permeability (Raghuwanshi et al., 2010).
The high pH, RSC and SAR values of water samples indicated their sodic nature dominated by the presence of CO 2− 3 , HCO − 3 and Na + while the presence of Ca 2+ + Mg 2+ and Cl − is also noted. Higher SAR values indicated dominance of Na + ion, causing soils unsuitable for agriculture (Richards, 1954). The critical limits of RSC in T1, T3, T2 and T4 indicated the need for treatment with amendments for irrigation in field crops. Treatment with gypsum is required for water samples with high RSC (T1-T4). Samples with moderate alkalinity (T5 and T6) may be used for the growing salt-resistant varieties.

Reclamation and use potential of salt-affected and waterlogged soils
Sodic soils of the Gangetic plain in Central Haryana are rich in sodium carbonate and bicarbonate salts and showed high ESP and variable soil texture. Strongly sodic soils (P1 and P2) containing high Na 2 CO 3 and NaHCO 3 salts, coarse soil texture and sodic ground water needs gypsum application @ 8-10 t ha −1 to reduce alkalinity in soil and water followed by leaching of excess soluble salts. Moderately sodic soil (P3) containing soluble Na 2 CO 3 and NaHCO 3 salts and fine soil texture can be reclaimed by addition of 4-6 t ha −1 gypsum. Due to high clay content and presence of CaCO 3 concretions, P4 (slightly sodic soil) showed drainage restrictions and waterlogging. It may be used for growing salt-tolerant rice and wheat crops. The addition of FYM in soils and cultivation of Dhaincha (Sesbania sp.) is suggested to improve physical properties, drainage conditions and reduce waterlogging.

Conclusions
Visual and digital analysis of IRS LISS III multi-temporal data was used for identification and delineation of sodic soils and waterlogged areas in the Gangetic plain of Central Haryana. Field validation and laboratory analysis for physical and chemical properties facilitated development of map legends. High values for spectral reflectance were observed from salty surfaces, and higher energy absorption in visible and infrared bands suggested the identification of strongly sodic soils and surface waterlogging. The mixed spectral signatures for salt, scattered crop covers and waterlogging were authenticated by field investigation. Saturation of Na 2 CO 3 and NaHCO 3 salts in soil and ground water caused alkalization and low soil productivity. Fine soil texture and the presence of concretionary calcium carbonate layer at sub-surface depths tended to produce waterlogging. Sodic soils and sodic water can be reclaimed with suitable amendments such as gypsum or pyrite.