geosciences
Article Monitoring and Assessment of Salinity and Chemicals in Agricultural Lands by a Remote Sensing Technique and Soil Moisture with Chemical Index Models
Hashim Ali Hasab 1, Hayder Dibs 2 , Abdulameer Sulaiman Dawood 3, Wurood Hasan Hadi 4, Hussain M. Hussain 4 and Nadhir Al-Ansari 5,* 1 Al-Najaf Technical Institute, Department of Architectural Design and Decoration, Al-Furat Al-Awsat Technical University, Najaf 54001, Iraq; [email protected] or [email protected] 2 Water Resources Faculty, Department of Hydraulic Structures Engineering, Al-Qasim Green University, Babel 51001, Iraq; [email protected] 3 Al-Kufa Technical Institute, Al-Furat Al-Awsat Technical University, Kufa 54001, Iraq; [email protected] or [email protected] 4 Remote Sensing Center, University of Kufa, Kufa 54001, Iraq; [email protected] (W.H.H.); [email protected] (H.M.H.) 5 Department of Civil, Environmental and Natural Resources Engineering, Lulea University of Technology, 971 87 Lulea, Sweden * Correspondence: [email protected]; Tel.: +46-920491858
Received: 18 April 2020; Accepted: 19 May 2020; Published: 29 May 2020
Abstract: Agricultural land in the south of Iraq provides habitat for several types of living creatures. This land has a significant impact on the ecosystem. The agricultural land of Al-Hawizeh marsh covers an area of more than 3500 km2 and is considered an enriched resource to produce several harvests. A total of 74% of this area suffers from a high degree of salinity and chemical pollution, which needs to be remedied. Several human-made activities and post-war-related events have caused radical deterioration in soil quality in the agricultural land. The goal of this research was to integrate mathematical models, remote sensing data, and GIS to provide a powerful tool to predict, assess, monitor, manage, and map the salinity and chemical parameters of iron (Fe), lead (Pb), copper (Cu), chromium (Cr), and zinc (Zn) in the soils of agricultural land in Al-Hawizeh marsh in southern Iraq during the four seasons of 2017. The mathematical model consists of four parts. The first depends on the B6 and B11 bands of Landsat-8, to calculate the soil moisture index (SMI). The second is the salinity equation (SE), which depends on the SMI result to retrieve the salinity values from Landsat-8 images. The third part depends on the B6 and B7 bands of Landsat-8, which calculates the clay chemical index (CCIs). The fourth part is the chemical equation (CE), which depends on the CCI to retrieve the chemical values (Fe, Pb, Cu, Cr, and Zn) from Landsat-8 images. The average salinity concentrations during autumn, summer, spring, and winter were 1175, 1010, 1105, and 1789 mg/dm3, respectively. The average Fe concentrations during autumn, summer, spring and winter were 813, 784, 842, and 1106 mg/dm3, respectively. The average Pb concentrations during autumn, summer, spring, and winter were 4.85, 3.79, 4.74, and 7.2 mg/dm3, respectively. The average Cu concentrations during autumn, summer, spring, and winter were 3.9, 3.1, 4.45, and 7.5 mg/dm3, respectively. The average Cr concentrations during autumn, summer, spring, and winter seasons were 1.28, 0.73, 1.03, and 2.91 mg/dm3, respectively. Finally, the average Zn concentrations during autumn, summer, spring, and winter were 8.25, 6, 7.05, and 12 mg/dm3, respectively. The results show that the concentrations of salinity and chemicals decreased in the summer and increased in the winter. The decision tree (DT) classification depended on the output results for salinity and chemicals for both SE and CE equations. This classification refers to all the parameters simultaneously in one stage. The output of DT classification results can display all the soil quality parameters (salinity, Fe, Pb, Cu,
Geosciences 2020, 10, 207; doi:10.3390/geosciences10060207 www.mdpi.com/journal/geosciences Geosciences 2020, 10, 207 2 of 20
Cr, and Zn) in one image. This approach was repeated for each season in this study. In conclusion, the developed systematic and generic approach may constitute a basis for determining soil quality parameters in agricultural land worldwide.
Keywords: salinity; chemicals; mathematical models; remote sensing data; GIS
1. Introduction The factors that pose a serious threat to ecosystems, water resources, and agricultural land include environmentally destructive development, human non-ethical activities, soil pollution, and inadequate discharge of sewage and industrial wastewaters [1,2]. The surface soil quality of agricultural land and watersheds—such as marshes, lakes, and reservoirs—often varies depending on natural hydrological, biological, chemical, morphological, and sedimentation processes. Salinity, chemicals, heavy metals, and pathogens such as parasites, bacteria, and viruses present in waste materials constitute the most dangerous environmental pollutants [2–9]. Globally, approximately 25,000 deaths occur daily due to diseases related to pollution [10,11]. The huge agricultural area of southern Iraq have a significant impact on the ecosystem. This area is considered an enriched resource for the production of several crops such as rice, wheat, barley, and fruits. However, 74% of the land suffers from problems associated with salinity, chemicals, industrial activities, and toxic metals from the wars of previous decades, which combined have had a significant effect on the soil quality [12–17]. In the past few decades, the agricultural land of southern Iraq has suffered from sewage and industrial wastewater discharge, a lack of integrated agricultural land management, and poor public awareness for agricultural protection [18–20]. An integrated model using remote sensing data and a geographic information system (GIS) can provide a powerful tool for the development of a management processes to assess soil quality and soil surface pollution associated with the agricultural land [21–24]. A general total suspended of normalized difference vegetation index (Ts-NDVI) using vegetation indices from remote sensing data—such as Moderate Resolution Imaging Spectroradiometer (MODIS) and Advanced Very-High Resolution Radiometer (AVHRR)—was developed to monitor soil moisture in the Mongolian Plateau during the period 1981–2012 [25–29]. Soil moisture plays an important role in the hydrological process and is closely related to surface soil temperature at the same soil of the diurnal change. There was a stronger correlation between surface soil moisture (SSM) and surface soil temperature (SST) in two catchments, one in Australia and the other in the United Kingdom [30]. Landsat satellite images and GIS were used to map, monitor, assess, and detect change of wetland dynamics on the Chennai coast in India during three periods (1988–1996, 1996–2006, and 2006–2016) based on the supervised classification method [31–33]. Research has been carried out regarding the monitoring and management strategies of agricultural soils to reduce metal pollution by investigating heavy mineral contamination of Cu, Ni, Co, Cr, Mn, and Pb in agricultural soil of the copper mining regions in India. In addition, researchers have used the enrichment factor, contamination factors, the Nemerow index, pollution loads index, geo-accumulation index, ecological risk index, and inductively coupled plasma-mass spectrometry (ICP-MS, Perkin Elmer DRC-e, Waitham, MA, USA) to determine the concentration of minerals and assess pollution levels in Singhbhum, India [25,34]. An analytical hierarchy process and a geographic information system was used to improve the sustainability and management of agricultural activity by evaluating six soil quality indicators—including pH, cation exchange capacity (CEC), organic carbon (OC) and salinity—in the agricultural area of the Tadla Plain, Morocco. An analytical hierarchy process was also used to identify the weight of each soil quality indicator via a pair-wise comparison matrix and GIS was used to generate a soil quality environmental map [35]. Geosciences 2020, 10, 207 3 of 20
GIS and multi-temporal remote sensing data, such as Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper (ETM) satellite images, have been used to analyze vegetation stress to develop soil salinity maps and determine the content of total dissolved solids (TDS), total suspended solids (TSS), Ca, Mg, Na, K, CO3, H CO3, Cl, NO3, and SO4 in irrigated areas of Syrdarya province, Uzbekistan,Geosciences with2020, 10 a, higherx FOR PEER degree REVIEW of spatial accuracy and in a cost effective way [36]. Another3 of 20 study estimated soil salinity and conducted salinity mapping with considerable accuracy using multi-spectral 3 3 3 4 remotesuspended sensing solids data with(TSS), an Ca, advanced Mg, Na, K, land CO , imager H CO , (ALI)Cl, NO sensor, and SO and in partial irrigated least-square areas of Syrdarya regression province, Uzbekistan, with a higher degree of spatial accuracy and in a cost effective way [36]. model (PLSR), together with soil salinity data collected from the Yellow River Delta in China [37]. Another study estimated soil salinity and conducted salinity mapping with considerable accuracy Another study used different satellite sensors (e.g., TM, MSS, ETM+, SPOT, Terra-ASTER, LISS-III, using multi-spectral remote sensing data with an advanced land imager (ALI) sensor and partial and IKONOS)least-square and regression methods model used (PLSR) to detect,, together monitor with soil and salinity map data of soil collected salinity from in the arid Yellow and River semi-arid regionsDelta [38 in]. China Other [37]. studies Another assessed study used the different metal contamination satellite sensors (e.g., (As, TM, Cr, Cd,MSS, Co, ETM+, Fe, SPOT, Cu, Zn, Terra- Hg, Pb, and Mn)ASTER, of sedimentsLISS-III, and in IKONOS) the Shadegan and methods and Hawr used Al-Azimto detect, monitor wetlands and in map southwestern of soil salinity Iran in basedarid on GIS andand remotesemi-arid sensing regions with [38]. mathematicalOther studies assessed models the [39 –metal41]. contamination (As, Cr, Cd, Co, Fe, Cu, ThisZn, Hg, article Pb, and critically Mn) of sediments evaluates in the the existingShadegan literature and Hawr thatAl-Azim indicate wetlands more in studiessouthwestern are needed Iran to bridgebased the gapson GIS identified and remote in sensing previous with studies. mathematical A comprehensive models [39–41]. survey of the literature revealed that only a fewThis techniques article critically have been evaluates developed the existing for the lite detectionrature that and indicate monitoring more studies of physical are needed and chemical to bridge the gaps identified in previous studies. A comprehensive survey of the literature revealed that soil quality parameters in the marshland area of Iraq. More systematic studies are required to support only a few techniques have been developed for the detection and monitoring of physical and chemical findings from previous studies and improve the accuracy of findings on the soil of agricultural land. soil quality parameters in the marshland area of Iraq. More systematic studies are required to support The aimfindings of this from research previous wasstudies to and examine improve whether the accuracy the integration of findings on between the soil of mathematical agricultural land. models, remoteThe sensing aim of Landsat-8this research OLI was data to examine and GIS whethe couldr provide the integration a powerful between tool mathematical for predicting, models, assessing, monitoring,remote sensing managing, Landsat-8 and OLI mapping data and salinity GIS could and pr chemicalovide a powerful parameters tool for such predicting, as iron (Fe),assessing, lead (Pb), chromiummonitoring, (Cr), andmanaging, copper and (Cu) mapping in the soilssalinity of agriculturaland chemical land parameters in the Al-Hawizehsuch as iron (Fe), marsh lead in (Pb), southern Iraq duringchromium the (Cr), four and seasons copper of (Cu) 2017. in the soils of agricultural land in the Al-Hawizeh marsh in southern Iraq during the four seasons of 2017. 2. Study Area 2. Study Area The study area was the agricultural land that surrounds Al-Hawizeh marsh in southern Iraq, The study area was the agricultural land that surrounds Al-Hawizeh marsh in southern Iraq, covering an area of 3500–4000 km2, as shown in Figure1. It is considered an enriched resource for covering an area of 3500–4000 km2, as shown in Figure 1. It is considered an enriched resource for the the production of several crops that have a significant impact on the ecosystem, as well as a habitat for production of several crops that have a significant impact on the ecosystem, as well as a habitat for biodiversitybiodiversity [13, 14[13,14,18].,18]. The The neglect neglect and and wars wars of of recent recent yearsyears are factors that that have have led led to to an an increase increase of salinityof salinity and heavy and heavy chemical chemical pollution pollution in in this this area. area. Al-Hawizeh Al-Hawizeh marsh marsh is isthe the major major water water source source for for thesethese lands lands [15, 16[15,16,18,20].,18,20].
FigureFigure 1. 1.Geographical Geographical location of of the the study study area. area.
Geosciences 2020, 10, 207 4 of 20
3. Methodology The methodology adopted for this article is presented as a flowchart in Figure2. It includes details regarding satellite images, image pre-processing, soil samples, soil analysis, mathematical models, data validation, image classification, and data visualization. Landsat-8 satellite images are represented as basic remote sensing data collected during 2017. Processed images from Landsat-8 satellite data and ancillary data were synergistically integrated with the mathematical models to predict the salinity and concentrations of chemicals (Fe, Pb, Cu, Cr, and Zn) in the soil for the agricultural land of the study area. The creation of mathematical models was informed by the nature of the satellite image processing and auxiliary data. The mathematical models had two equations: salinity equation (SE) and chemical equation (CE). The SE was based on the soil moisture index (SMI), which depends on the B6 and B11 bands. The CE was based on the clay chemical indices (CCIs), which depends on the B6 and B7 bands. Thedepends outputs of on both the salinityB6 and andB7 bands. chemical The equations outputs of were both the salinity salinity and and chemical chemical equations concentration were datathe salinityobtained and from chemical the spatial concentration distribution data of satellite obtained images. from Thesethe spatial data weredistributi categorizedon of satellite using a images. decision Thesetree (DT) data of were image categorized classification using to classify a decision the salinitytree (DT) and of chemical image classification parameters onto satelliteclassify imagesthe salinity over andfour chemical seasons. parameters Data visualization on satellite was images used to over check four the seasons. salinity, Data Fe, Pb,visualization Cu, Cr, and was Zn used classification to check theresults salinity, on the Fe, maps Pb, with Cu, the Cr, appropriate and Zn classificati specificationson results as output on resultsthe maps of data with validations. the appropriate specifications as output results of data validations.
Satellite Soil Samples
Images Pre-Processing Soil Analysis by (AAS)
Chemicals Data Salinity Data
Mathematical Models Process
Clay Chemicals Indices (CCIs) Soil Moisture Index (SMI)
Chemicals Equation (CE) Salinity Equation (SE)
No Data Validation (DV) Yes
Decision Tree (DT) Classification
Data Visualization (DV)
FigureFigure 2. 2. FlowchartFlowchart of of research research methodology. methodology.
1
Geosciences 2020, 10, 207 5 of 20
3.1. Satellite Images and Pre-Processsing
The remote sensing data used satellite images from Landsat-8 OLI at longitude 47◦3200–47◦4500 E and latitude 31◦3000–31◦4200 N with a resolution of 30 m. The satellite images of the study area were synchronized with soil samples collected from the agricultural land during 2017. The mathematical models of this study were based on the B6, B7 and B11 bands of the satellite image, as shown in Table1. Accurate pre-processing of Landsat-8 satellite images, including atmospheric correction, is the first important step of the research methodology. This process involves radiometric calibration, dark subtraction, and internal average reflectance (IAR) calibration, layer stacking, georeferencing, image enhancement, and region of interest (ROI) determination. The output of this process provides suitable images. These data were combined with mathematical models for assessing and monitoring the spatial distribution of salinity and chemicals in the study area.
Table 1. Endowment of Landsat-8 (OLI and TIRS) band details [42,43]
Band Wavelength Useful for Mapping Discriminates moisture content of soil Band 6—Short-wave Infrared (SWIR) 1 1.57–1.65 and vegetation; penetrates thin clouds Improved moisture content of soil and vegetation Band 7—Short-wave Infrared (SWIR) 2 2.11–2.29 and thin cloud penetration 100 m resolution. Improved thermal mapping Band 11—TIRS 2 11.5–12.51 and estimated soil moisture
3.2. Soil Sample Collection Soil samples were collected at 15 cm depth using a clean shovel. These samples were collected from 60 stations positioned across different regions, covering the whole of the agricultural land by using the Global Positioning System (GPS). These soil sampling stations were numbered S1–S60, as illustrated in Figure3. All soil samples underwent laboratory testing using the atomic absorption spectrophotometer (AAS) technique to measure the concentration values for salinity and chemicals
(Fe, Pb, Cu, Cr, andGeosciences Zn). These 2020, 10, x dataFOR PEER were REVIEW used to determine the status of soil quality. 5 of 20
2.5 0 2.5 5 7.5 10
Figure 3. Satellite image showing the soil sampling stations. Figure 3. Satellite image showing the soil sampling stations. 3.3. Analysis of Soil Samples Soil types present in Iraq are Alfisol (5%), Mollisol (5%), Vertisol (10%), Inceptisol (15%), Entisol (20%), and Aridisol (45%). The soils in the agricultural area of Al-Hawizeh marsh are dominated by relatively fertile alluvial soils within the Entisol soil order based on the classification scheme of the United States for Iraq soils from 1965 [44–47]. All data were tested using the atomic absorption spectrophotometer (AAS) technique. AAS is a technique used to measure trace elements present in soil samples, in which free gaseous atoms absorb electromagnetic radiation at a specific wavelength to produce a measurable signal. The absorption signal is proportional to the concentration of those free absorbing atoms in the optical path. Therefore, for AAS measurements, the sample must be converted into gaseous atoms, usually by the application of heat to a cell using an atomizer. The main type of atomizer used in AAS-based analytical techniques is the flame AAS (FAAS), which provides analytical signals in a continuous fashion [48,49]. The samples were grinded in a motor and sieved. Using 1 mm, the sample (1 g) was digested with concentrated nitric acid in a sand bath until dry. Then, deionized water was added and the sample was filtered then complete the filtrate to 25 mL with deionizer water. AAS was used to determine the chemical concentrations at specific wavelengths along with a hallow cathode lamp (HCL). Chemical speciation was determined by AAS after extraction of the metal with different chemical reagents. The soil analysis results for the salinity and chemicals represent the real data, demonstrating the status of the soil quality in the agricultural land. The concentration values for the salinity and the chemicals during four seasons were lowest in the summer and highest in the winter, as shown in Figures 4 and 5. The minimum, maximum, and average of salinity and chemicals during the four
Geosciences 2020, 10, 207 6 of 20 Geosciences 2020, 10, x FOR PEER REVIEW 6 of 20
3.3.seasons Analysis are ofshown Soil Samples in Table 2. The mathematical models used these data to combine with satellite images and assess the salinity and chemical values for the four seasons of 2017. Soil types present in Iraq are Alfisol (5%), Mollisol (5%), Vertisol (10%), Inceptisol (15%), Entisol (20%), and AridisolTable 2. Laboratory (45%). The measurements soils in the of agricultural salinity and areachemicals of Al-Hawizeh over four seasons marsh in are2017 dominated by relatively fertile alluvial soils within the Entisol soil order based on the classification scheme of Minerals (mg/dm3) Autumn Summer Spring Winter the United States for Iraq soils from 1965 [44–47]. Min 230 1012 913 98 All dataSalinity were tested usingMax the atomic8975 absorption 17,321 spectrophotometer 7438 (AAS) technique. 4109 AAS is a technique used to measureAvg trace elements1209 present in soil1069.4 samples, in which1163.8 free gaseous1747 atoms absorb electromagnetic radiation atMin a specific wavelength282 to produce 62 a measurable 310 signal. 1101 The absorption signal is proportionalIron to theMax concentration 35,011 of those free4123 absorbing atoms10,951 in the optical2978 path. Therefore, for AAS measurements, theAvg sample must867.92 be converted738.9 into gaseous atoms, 819.8 usually by1070.4 the application of heat to a cell using an atomizer.Min The main0.65 type of atomizer0.53 used in AAS-based0.32 analytical0.38 techniques is the flame AASLead (FAAS), whichMax provides12.3 analytical signals 8.14 in a continuous 11.87 fashion [21.4548,49 ]. The samples were grindedAvg in a motor4.9 and sieved. Using4.11 1 mm, the4.53 sample (1 g) was6.41 digested with concentrated nitric acid in aMin sand bath until0.45 dry. Then,0.57 deionized water0.08 was added0.29 and the sample was filtered thenCopper complete Max the filtrate24.65 to 25 mL with 16.94deionizer water. 24.62 AAS was used 38.21 to determine the chemical concentrationsAvg at specific wavelengths4.52 along3.9 with a hallow4.26 cathode lamp (HCL).6.1 Chemical speciation was determined byMin AAS after0.04 extraction of the 0.12 metal with 0.022 different chemical 0.023 reagents. The soilChromium analysis resultsMax for the salinity8.27 and chemicals 3.53 represent 5.2 the real data, 3.13 demonstrating the status of the soil qualityAvg in the agricultural1.23 land.0.93 The concentration1.09 values2.45 for the salinity and the chemicals during fourMin seasons were1.8 lowest in the 5.2 summer and 2.8highest in the winter, 2.87 as shown in Figures4 andZinc5. The minimum,Max maximum,2.13 and average28.65 of salinity and41.5 chemicals during3980 the four seasons are shown in TableAvg2. The mathematical8.012 models7.14 used these data7.52 to combine 10.3 with satellite images and assess the salinity and chemical values for the four seasons of 2017.
Salinity Iron (Fe)
2000 1747 1800 1600 1400 1209 1163.8 1200 1069.4 1070.4 1000 867.92 819.8 738.9 800 600 400 200 0 Autumn Summer Spring Winter Salinity 1209 1069.4 1163.8 1747 Iron (Fe) 867.92 738.9 819.8 1070.4
FigureFigure 4. 4.Laboratory Laboratory measurements measurements for for the the four four seasons seasons of of 2017, 2017, salinity salinity and and iron. iron.
Geosciences 2020, 10, 207 7 of 20 Geosciences 2020, 10, x FOR PEER REVIEW 7 of 20
12 10.3 10
8.012 7.52 8 7.14 6.41 6.1 6 4.9 4.52 4.534.26 4.11 3.9 4 2.45 2 1.23 0.93 1.09
0 Autumn Summer Spring Winter Lead (Pb) 4.9 4.11 4.53 6.41 Copper (Cu) 4.52 3.9 4.26 6.1 Chromium (Cr) 1.23 0.93 1.09 2.45 Zinc (Zn) 8.012 7.14 7.52 10.3 FigureFigure 5. 5.Laboratory Laboratory measurementsmeasurements for for the the four four seasons seasons of 2017, of 2017, Pb, Cu, Pb, Cr, Cu, and Cr, Zn. and Zn.
3.4. MathematicalTable 2. LaboratoryModels measurements of salinity and chemicals over four seasons in 2017.
The developedMinerals mathematical (mg/dm model3) canAutumn retrieve the Summer soil quality Spring parameters Winter (salinity, Fe, Pb, Cu, Cr, and Zn) from Landsat-8 data. Soil quality parameter estimations are based on the integration Min 230 1012 913 98 between ground data, satellite images after processing and mathematical models. The developed Salinity Max 8975 17,321 7438 4109 model is based on the soil moisture index (SMI) and clay chemical indices (CCIs). The input of this Avg 1209 1069.4 1163.8 1747 model is based on the ground dataMin and post-processed 282 satellite 62 image. 310 These models 1101 were introduced by [38,50]. Iron Max 35,011 4123 10,951 2978 The model consists of four parts:Avg The first 867.92 one depends 738.9 on the 819.8 B6 and B11 1070.4 bands of Landsat-8 images, which calculates the SMI;Min the second part 0.65 is the salinity 0.53 equation 0.32 (SE), which 0.38 depends on the output of SMI to retrieveLead the salinityMax values from 12.3 Landsat-8 8.14 images. 11.87 The third 21.45 part depends on the B6 and B7 bands of Landsat-8 images,Avg which 4.9 calculates 4.11 the CCI. 4.53The fourth 6.41part is the chemical equation (CE), which depends onMin the output 0.45of CCI to retrieve 0.57 the 0.08chemical values 0.29 (Fe, Pb, Cu, Cr, and Zn) from Landsat-8Copper images. TheseMax four components 24.65 16.94are described 24.62 below. 38.21 Avg 4.52 3.9 4.26 6.1 3.4.1. Soil Moisture Index (SMI) Min 0.04 0.12 0.022 0.023 Chromium Max 8.27 3.53 5.2 3.13 The SMI equation was used toAvg calculate soil 1.23 moisture, 0.93 depending 1.09 on the relation 2.45 between the B6 and B11 bands of the Landsat-8Min images, as 1.8shown in Equation 5.2 (1). 2.8 The B6 (short 2.87 wave infrared (SWIR)) band discriminatesZinc vegetatiMaxon and moisture 2.13 content 28.65 of the 41.5 soil by penetrating 3980 thin clouds with a wavelength range of 1.57–1.65Avg μm. 8.012Meanwhile, 7.14 the B11 (TIRS- 7.522) band 10.3 estimates the soil moisture and the thermal mapping in a wavelength range of 11.5–12.51 μm [42,43,51]. (B11 − B6) (TIRS)2−(SWIR)1 (1) 3.4. Mathematical Models SMI = = (B11 + B6) (TIRS)2+(SWIR)1 The developed mathematical model can retrieve the soil quality parameters (salinity, Fe, Pb, Cu, Cr, and3.4.2. Zn) Salinity from Equation Landsat-8 (SE) data. Soil quality parameter estimations are based on the integration betweenThe ground SE calculates data, satellite the salinity images value after from processing the Landsat-8 and mathematicalimages, depending models. on the The relation developed modelbetween is based the output on the values soil moisture of the SMI index from the (SMI) Landsa andt-8 clay images chemical and the indices salinity (CCIs).values of The the ground input of this modeldata, is basedas illustrated on the in ground Equations data (2) and and post-processed (3), and shown satellitein Figure image. 6, where These the y models-axis represents were introduced the by [38salinity,50]. values from the ground measurements while the x-axis represents the values of the SMI from theThe satellite model images. consists of four parts: The first one depends on the B6 and B11 bands of Landsat-8 images, which calculates the SMI; the second part is the salinity equation (SE), which depends on the output of SMI to retrieve the salinity values from Landsat-8 images. The third part depends on the B6 and B7 bands of Landsat-8 images, which calculates the CCI. The fourth part is the chemical Geosciences 2020, 10, 207 8 of 20 equation (CE), which depends on the output of CCI to retrieve the chemical values (Fe, Pb, Cu, Cr, and Zn) from Landsat-8 images. These four components are described below.
3.4.1. Soil Moisture Index (SMI) The SMI equation was used to calculate soil moisture, depending on the relation between the B6 and B11 bands of the Landsat-8 images, as shown in Equation (1). The B6 (short wave infrared (SWIR)) band discriminates vegetation and moisture content of the soil by penetrating thin clouds with a wavelength range of 1.57–1.65 µm. Meanwhile, the B11 (TIRS-2) band estimates the soil moisture and the thermal mapping in a wavelength range of 11.5–12.51 µm [42,43,51].
(B11 B6) (TIRS)2 (SWIR)1 SMI = − = − (1) (B11 + B6) (TIRS)2 + (SWIR)1
3.4.2. Salinity Equation (SE) The SE calculates the salinity value from the Landsat-8 images, depending on the relation betweenGeosciences the output2020, 10, x values FOR PEER of REVIEW the SMI from the Landsat-8 images and the salinity values of the8 of ground20 data, as illustrated in Equations (2) and (3), and shown in Figure6, where the y-axis represents the salinity values from the groundSE measurements = f ×Exp g × whileSMI + the i x-axis represents the values of(2) the SMI (TIRS)2−(SWIR)1 (3) from the satellite images. SE = f ×Exp g× + i (TIRS)2+(SWIR)1 SE = f Exp[ g SMI ] + i (2) Here, SE represents the required values× of salinity× from the satellite images; SMI represents the " # values of the SMI from the satellite images, base(dTIRS on the)2 output(SWIR results)1 of Equation (1); and f, g, and i SE = f Exp g − + i (3) are constants that depend on both× the values× (ofTIRS the )SMI2 + from(SWIR the)1 satellite images and the salinity values of the ground data. The base e of the exponential function is equal to 2.718.
4500
∗ 4000 y = f * e (g x) +i
3500 SE = f * Exp (g* SMI)+i 3000
2500
2000
1500
1000
500 Salinity value from ground measurements (gm/dm3) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Soil moisture indexes (SMI) value from image
FigureFigure 6. Salinity 6. Salinity equation equation (SE)(SE) based on on the the relationsh relationshipip between between soil moisture soil moisture index values index of values the of the satellitesatellite imagesimages and and salinity salinity values values from from the the ground ground measurements. measurements.
Here,3.4.3. Clay SE representsChemical Indices the required (CCIs) values of salinity from the satellite images; SMI represents the valuesClay of thechemical SMI fromindices the compose satellite all images,of the compon basedent on chemicals—such the output results as Pb, of Fe, Equation Cu, Cr, and (1); Zn— and f, g, and ithatare are constants found in that thedepend clay soil onof agricultural both the values land. ofCCIs the calculate SMI from the the content satellite of clay images chemicals and thein the salinity valuessoil of based the ground on the data.relationship The base between e of the bands exponential B6 and B7 function of the isLandsat-8 equal to images, 2.718. as shown in Equation (4). The B6 (SWIR-1) band discriminates the moisture content of soil and vegetation by penetrating thin clouds with a wavelength ranging from 1.57 to 1.65 μm. The B7 (SWIR-2) band improves the moisture content detection of soil and vegetation through thin cloud penetration with a wavelength ranging from 2.11 to 2.29 μm [42,43,51]. B6 (SWIR)1 (4) CCIs = = B7 (SWIR)2
3.4.4. Chemical Equation (CE) The CE retrieves values of chemicals (Pb, Fe, Cu, Cr, and Zn) based on the relationship between the output results of the CCIs from the Landsat-8 images and the ground data, as illustrated in Equations (5) and (6) and shown in Figure 7, where the y-axis represents the chemical values of the ground measurements and the x-axis represents the CCIs values from the satellite images. CE = p × (CCIs) +h (5) (SWIR)1 (6) CE = p × +h (SWIR)2
Geosciences 2020, 10, 207 9 of 20
3.4.3. Clay Chemical Indices (CCIs) Clay chemical indices compose all of the component chemicals—such as Pb, Fe, Cu, Cr, and Zn—that are found in the clay soil of agricultural land. CCIs calculate the content of clay chemicals in the soil based on the relationship between bands B6 and B7 of the Landsat-8 images, as shown in Equation (4). The B6 (SWIR-1) band discriminates the moisture content of soil and vegetation by penetrating thin clouds with a wavelength ranging from 1.57 to 1.65 µm. The B7 (SWIR-2) band improves the moisture content detection of soil and vegetation through thin cloud penetration with a wavelength ranging from 2.11 to 2.29 µm [42,43,51].
B6 (SWIR)1 CCIs = = (4) B7 (SWIR)2
3.4.4. Chemical Equation (CE) The CE retrieves values of chemicals (Pb, Fe, Cu, Cr, and Zn) based on the relationship between the output results of the CCIs from the Landsat-8 images and the ground data, as illustrated in Equations (5) and (6) and shown in Figure7, where the y-axis represents the chemical values of the ground measurements and the x-axis represents the CCIs values from the satellite images.
CE = p (CCIs)n + h (5) × !n (SWIR)1 Geosciences 2020, 10, x FOR PEER REVIEW CE = p + h9 of 20 (6) × (SWIR)2 wherewhere CE represents CE represents the the concentration concentration values values ofof thethe chemicals (Pb, (Pb, Fe, Fe, Cu, Cu, Cr, Cr, and and Zn) Zn)acquired acquired from from the satellitethe satellite images; images; CCIs CCIs represents represents the claythe clay chemical chemic indicesal indices from from the the satellite satellite images images based based on on the the output resultsoutput of Equation results of (4); Equationp and h(4);are p constants and h are thatconstants depend that on depend the chemical on the chemical values of values the ground of the data ground data and the indices of the clay chemicals from the satellite images, respectively; and the and the indices of the clay chemicals from the satellite images, respectively; and the power n is an integer. power n is an integer. All of the previous steps of the model must be repeated for each season of All of the previous steps of the model must be repeated for each season of study. study.
0.80
0.75 y = p * (x)n + h
0.70 CE = p * (CCIs)n + h
0.65
0.60
0.55
0.50
0.45
Chemicals value from ground measurements (gm/dm3) measurements ground from value Chemicals 1.6 1.7 1.8 1.9 2.0 2.1 2.2 Clay Chemical Indices (CCIs) from image
FigureFigure 7. Chemical 7. Chemical equation equation (CE) (CE) based based on on thethe relationshiprelationship between between the the clay clay chemicals chemicals indices indices from from the satellitethe satellite imagery, imagery, and and the the chemical chemical values values of of thethe groundground measurements. measurements.
3.5. Image Classification Several image classification techniques can be used to classify the satellite images, such as maximum likelihood (ML), minimum distance (MD), artificial neural network (ANN), support vector machine (SVM), and decision tree (DT) [4,24]. The present study classified soil quality parameters using images of agricultural land in Al-Hawizeh marsh during four seasons of 2017 depending on the output results of SE and CE equations. DT classification was used in this study. It is an advanced approach for image classification, which depends on the layered or stratified approach to solve the problems of distinguishing between spectral classes [4,24]. DT uses data from many sources to make a single decision tree classifier. This performs multistage classification using a series of binary decisions to place pixels into classes. Input data can be obtained from various sources and data types. The results of the decisions are classes that can be saved as trees to apply them for other datasets [52]. This classification refers to all parameters simultaneously in one stage. The output of DT classification results can display all soil quality parameters in one image. Thus, DT in this case is considered as the optimal approach. This approach was repeated for each season in this study.
3.6. Data Visualization The output results are presented as DT classifications for both SE and CE equations after performing validation. The output of the DT results are presented as one image for salinity, Fe, Pb, Cu, Cr, and Zn. This procedure was performed for each season.
Geosciences 2020, 10, 207 10 of 20
3.5. Image Classification Several image classification techniques can be used to classify the satellite images, such as maximum likelihood (ML), minimum distance (MD), artificial neural network (ANN), support vector machine (SVM), and decision tree (DT) [4,24]. The present study classified soil quality parameters using images of agricultural land in Al-Hawizeh marsh during four seasons of 2017 depending on the output results of SE and CE equations. DT classification was used in this study. It is an advanced approach for image classification, which depends on the layered or stratified approach to solve the problems of distinguishing between spectral classes [4,24]. DT uses data from many sources to make a single decision tree classifier. This performs multistage classification using a series of binary decisions to place pixels into classes. Input data can be obtained from various sources and data types. The results of the decisions are classes that can be saved as trees to apply them for other datasets [52]. This classification refers to all parameters simultaneously in one stage. The output of DT classification results can display all soil quality parameters in one image. Thus, DT in this case is considered as the optimal approach. This approach was repeated for each season in this study.
3.6. Data Visualization The output results are presented as DT classifications for both SE and CE equations after performing validation. The output of the DT results are presented as one image for salinity, Fe, Pb, Cu, Cr, and Zn. This procedure was performed for each season.
4. Results and Discussion The mathematical models were combined with the Landsat-8 satellite images and synchronized with the ground data to retrieve soil quality parameters for agricultural land at Al-Hawizeh marsh. The values of soil quality parameters (salinity, Fe, Pb, Cu, Cr, and Zn) are evaluated from Landsat-8 data based on the mathematical models. These values indicate the status of soil quality for agricultural lands of Al-Hawizeh marsh from the satellite images during four seasons of 2017. Table3 summarizes the evaluated soil quality parameters.
Table 3. Evaluated soil quality parameters values from images during four seasons of 2017.
Minerals (mg/dm3) Autumn Summer Spring Winter Min 196 992 813 83 Salinity Max 9235 18,214 7944 4628 Avg 1175 1010 1105 1789 Min 274 56 290 1013 Iron Max 35,531 4604 11,234 3818 Avg 813 784 842 1106 Min 0.47 0.47 0.37 0.29 Lead Max 14 9 13 25 Avg 4.85 3.79 4.74 7.2 Min 0.42 0.5 0.05 0.26 Copper Max 26 18.4 26 40 Avg 3.9 3.1 4.45 7.5 Min 0.03 0.08 0.015 0.017 Chromium Max 10.7 4.3 5.9 3.73 Avg 1.28 0.73 1.03 2.91 Min 1.6 4.9 2.7 2.74 Zinc Max 2.95 32 46 4028 Avg 8.25 6 7.05 12 Geosciences 2020, 10, 207 11 of 20
4.1. Salinity Values Figure8 shows the seasonal variation of soil salinity obtained using the SMI model, which is observed to be lowest during summer and highest during winter. The average salinity concentrations during autumn, summer, spring, and winter are 1175, 1010, 1105, and 1789 mg/dm3, respectively, determined by the soil moisture index. This indicated a reduction of SMI in the summer and rise in the winter. The SMI values are decided by the spectral signatures of Landsat-8 Short Wave Infrared (SWIR-1) band B6 and Thermal Infrared Sensor (TIRS-2) band B11. These images (bands) are considered Geosciences 2020, 10, x FOR PEER REVIEW 11 of 20 as suitable and sensitive as far as the soil moisture indexes are concerned. Salinity concentrations displayed a positive relationship with the SMI values [18,42,43,51].
1900 Salinity 1800
1700
1600
1500
1400
1300
1200
Salinity Value (mg/dm3) 1100
1000
900 Autumn Summer Spring Winter
Seasons
Figure 8. Concentration values for salinity from images during four seasons of 2017.
4.2. IronFigure Values 8. Concentration values for salinity from images during four seasons of 2017.
4.2. IronFigure Values9 shows the seasonal variation of iron content obtained using the CCIs model, which is observed to be lowest during summer and highest during winter. The average Fe concentrations Figure 9 shows the seasonal variation of iron content obtained using the CCIs model, which is during autumn, summer, spring, and winter are 813, 784, 842, and 1106 mg/dm3, respectively, which observed to be lowest during summer and highest during winter. The average Fe concentrations was determined by the clay chemical indices. This indicated a reduction of the clay chemicals in during autumn, summer, spring, and winter are 813, 784, 842, and 1106 mg/dm3, respectively, which the summer and a rise during winter. The CCIs values are decided by the spectral signatures of was determined by the clay chemical indices. This indicated a reduction of the clay chemicals in the Landsat-8 Short Wave Infrared (SWIR-1) band B6 having and Short Wave Infrared (SWIR-2) band B7 summer and a rise during winter. The CCIs values are decided by the spectral signatures of Landsat- with wavelength range 2.11–2.29 µm. These images (bands) are considered as suitable and sensitive 8 Short Wave Infrared (SWIR-1) band B6 having and Short Wave Infrared (SWIR-2) band B7 with as far as the CCIs are concerned. Fe concentrations displayed a positive relationship among the clay wavelength range 2.11–2.29 μm. These images (bands) are considered as suitable and sensitive as far chemicals indices values [18,42,43,51]. as the CCIs are concerned. Fe concentrations displayed a positive relationship among the clay chemicals indices values [18,42,43,51].
1150 Iron (Fe) 1100
1050
1000
950
900
850 Iron (Fe) Value(mg/dm3)
800
750 Autumn Summer Spring Winter
Seasons
Figure 9. Concentration values for (Fe) from images during four seasons of 2017.
Geosciences 2020, 10, x FOR PEER REVIEW 11 of 20
1900 Salinity 1800
1700
1600
1500
1400
1300
1200
Salinity Value (mg/dm3) 1100
1000
900 Autumn Summer Spring Winter
Seasons
Figure 8. Concentration values for salinity from images during four seasons of 2017.
4.2. Iron Values Figure 9 shows the seasonal variation of iron content obtained using the CCIs model, which is observed to be lowest during summer and highest during winter. The average Fe concentrations during autumn, summer, spring, and winter are 813, 784, 842, and 1106 mg/dm3, respectively, which was determined by the clay chemical indices. This indicated a reduction of the clay chemicals in the summer and a rise during winter. The CCIs values are decided by the spectral signatures of Landsat- 8 Short Wave Infrared (SWIR-1) band B6 having and Short Wave Infrared (SWIR-2) band B7 with wavelength range 2.11–2.29 μm. These images (bands) are considered as suitable and sensitive as far as the CCIs are concerned. Fe concentrations displayed a positive relationship among the clay chemicals indices values [18,42,43,51]. Geosciences 2020, 10, 207 12 of 20
1150 Iron (Fe) 1100
1050
1000
950
900
850 Iron (Fe) Value(mg/dm3)
800
750 Autumn Summer Spring Winter Geosciences 2020, 10, x FOR PEER REVIEW 12 of 20 Seasons
4.3. Lead Values Figure 9. Concentration values for (Fe) from images during four seasons of 2017. Figure 9. Concentration values for (Fe) from images during four seasons of 2017. 4.3.The Lead Pb Values concentrations that were evaluated using CCIs model during the four seasons were lowest in summer and highest in winter (Figure 10). The average Pb concentrations during autumn, The Pb concentrations that were evaluated using CCIs model during the four seasons were summer, spring, and winter were 4.85, 3.79, 4.74, and 7.2 mg/dm3, respectively, determined by CCIs. lowest in summer and highest in winter (Figure 10). The average Pb concentrations during autumn, This indicated a reduction the indices of the clay chemicals in the summer and a rise during winter. summer, spring, and winter were 4.85, 3.79, 4.74, and 7.2 mg/dm3, respectively, determined by CCIs. Pb concentrations displayed a positive relationship with the CCIs values [18,42,43,51]. This indicated a reduction the indices of the clay chemicals in the summer and a rise during winter. Pb concentrations displayed a positive relationship with the CCIs values [18,42,43,51].
7.50 Lead (pb) 7.25 7.00 6.75 6.50 6.25 6.00 5.75 5.50 5.25 5.00 4.75 4.50
Lead (Pb) Value (mg/dm3) 4.25 4.00 3.75 3.50 Autumn Summer Spring Winter
Seasons
Figure 10. Concentration values for (Pb) from image during four seasons in 2017. Figure 10. Concentration values for (Pb) from image during four seasons in 2017. 4.4. Copper Values 4.4. CopperFigure Values 11 shows the seasonal variation of Cu content obtained using the CCIs model, which was observedFigure to11 beshows lowest the duringseasonal summer variation and of highestCu content during obtained winter. using The the average CCIs model, Cu concentrations which was observedduring autumn,to be lowest summer, during spring summer and winterand highes weret3.9, during 3.1, 4.45,winter. and The 7.5 mgaverage/dm3 ,Cu respectively, concentrations which duringwas determined autumn, summer, by the CCIs.spring Thisand winter indicated were a reduction3.9, 3.1, 4.45, in theandCCIs 7.5 mg/dm in the3, summer respectively, and awhich rise in wasthe determined winter. Cu concentrationsby the CCIs. This displayed indicated a positive a reduction relationship in the CCIs with in CCIs the summer values [18 and,42 ,a43 rise,51 ].in the winter. Cu concentrations displayed a positive relationship with CCIs values [18,42,43,51].
8.0 Copper (Cu) 7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
Copper (Cu) Value (mg/dm3) Value (Cu) Copper 3.5
3.0
2.5 Autumn Summer Spring Winter
Seasons
Figure 11. Concentration values for (Cu) from image during four seasons.
Geosciences 2020, 10, x FOR PEER REVIEW 12 of 20
4.3. Lead Values The Pb concentrations that were evaluated using CCIs model during the four seasons were lowest in summer and highest in winter (Figure 10). The average Pb concentrations during autumn, summer, spring, and winter were 4.85, 3.79, 4.74, and 7.2 mg/dm3, respectively, determined by CCIs. This indicated a reduction the indices of the clay chemicals in the summer and a rise during winter. Pb concentrations displayed a positive relationship with the CCIs values [18,42,43,51].
7.50 Lead (pb) 7.25 7.00 6.75 6.50 6.25 6.00 5.75 5.50 5.25 5.00 4.75 4.50
Lead (Pb) Value (mg/dm3) 4.25 4.00 3.75 3.50 Autumn Summer Spring Winter
Seasons
Figure 10. Concentration values for (Pb) from image during four seasons in 2017.
4.4. Copper Values Figure 11 shows the seasonal variation of Cu content obtained using the CCIs model, which was observed to be lowest during summer and highest during winter. The average Cu concentrations during autumn, summer, spring and winter were 3.9, 3.1, 4.45, and 7.5 mg/dm3, respectively, which was determined by the CCIs. This indicated a reduction in the CCIs in the summer and a rise in the winter.Geosciences Cu concentrations2020, 10, 207 displayed a positive relationship with CCIs values [18,42,43,51]. 13 of 20
8.0 Copper (Cu) 7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
Copper (Cu) Value (mg/dm3) Value (Cu) Copper 3.5
3.0
2.5 Autumn Summer Spring Winter Geosciences 2020, 10, x FOR PEER REVIEW 13 of 20 Seasons
4.5. Chromium ValuesFigure 11. Concentration values for (Cu) from image during four seasons.
4.5.The Chromium Cr concentrationsFigure Values 11. Concentration were evaluated values using for (Cu) CCI frsom and image were during lowest four during seasons. summer and highest during winter (Figure 12). The average Cr concentrations during autumn, summer, spring, and winterThe were Cr 1.28, concentrations 0.73, 1.03, wereand 2.91 evaluated mg/dm using3, respectively, CCIs and were determined lowestduring by CCIs. summer This indicated and highest a reductionduring winter of CCIs (Figure in the 12 ).summer The average and a Cr rise concentrations in the winter. during Cr concentrations autumn, summer, displayed spring, a and positive winter 3 relationshipwere 1.28, 0.73, with 1.03, the andclay 2.91chemical mg/dm index, respectively, values [18,42,43,51]. determined by CCIs. This indicated a reduction of CCIs in the summer and a rise in the winter. Cr concentrations displayed a positive relationship with the clay chemical index values [18,42,43,51].
3.2 3.0 Chromium(Cr) 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 Chromium (Cr) Value (mg/dm3) Value Chromium(Cr) 0.8 0.6
Autumn Summer Spring Winter
Seasons
Figure 12. Concentration values for (Cr) from image during four seasons. Figure 12. Concentration values for (Cr) from image during four seasons. 4.6. Zinc Values 4.6. ZincFigure Values 13 demonstrates the seasonal variation of Zn contents obtained using CCIs model, which was observed to be lowest during summer and highest during winter. The average Zn concentrations Figure 13 demonstrates the seasonal variation of Zn contents obtained using CCIs model, which during autumn, summer, spring, and winter were 8.25, 6, 7.05, and 12 mg/dm3, respectively, which is was observed to be lowest during summer and highest during winter. The average Zn concentrations decided by the clay chemical indices. This indicated a reduction of CCIs in the summer and a rise in during autumn, summer, spring, and winter were 8.25, 6, 7.05, and 12 mg/dm3, respectively, which the winter. Zn concentrations displayed a positive relationship with CCIs values [18,42,43,51]. is decided by the clay chemical indices. This indicated a reduction of CCIs in the summer and a rise in the winter. Zn concentrations displayed a positive relationship with CCIs values [18,42,43,51].
12.5 Zinc (Zn) 12.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5
Zinc (Zn) Value (mg/dm3) Value (Zn) Zinc 7.0 6.5 6.0 5.5 Autumn Summer Spring Winter
Seasons
Figure 13. Concentration values for (Zn) from image during four seasons.
Geosciences 2020, 10, x FOR PEER REVIEW 13 of 20
4.5. Chromium Values The Cr concentrations were evaluated using CCIs and were lowest during summer and highest during winter (Figure 12). The average Cr concentrations during autumn, summer, spring, and winter were 1.28, 0.73, 1.03, and 2.91 mg/dm3, respectively, determined by CCIs. This indicated a reduction of CCIs in the summer and a rise in the winter. Cr concentrations displayed a positive relationship with the clay chemical index values [18,42,43,51].
3.2 3.0 Chromium(Cr) 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 Chromium (Cr) Value (mg/dm3) Value Chromium(Cr) 0.8 0.6
Autumn Summer Spring Winter
Seasons
Figure 12. Concentration values for (Cr) from image during four seasons.
4.6. Zinc Values
Figure 13 demonstrates the seasonal variation of Zn contents obtained using CCIs model, which was observed to be lowest during summer and highest during winter. The average Zn concentrations during autumn, summer, spring, and winter were 8.25, 6, 7.05, and 12 mg/dm3, respectively, which is decided by the clay chemical indices. This indicated a reduction of CCIs in the summer and a rise in the winter. Zn concentrations displayed a positive relationship with CCIs values [18,42,43,51]. Geosciences 2020, 10, 207 14 of 20
12.5 Zinc (Zn) 12.0 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 7.5
Zinc (Zn) Value (mg/dm3) Value (Zn) Zinc 7.0 6.5 6.0 5.5 Autumn Summer Spring Winter
Seasons
Figure 13. Concentration values for (Zn) from image during four seasons. Figure 13. Concentration values for (Zn) from image during four seasons. From the previous results, we found that all values of salinity and chemical parameters decreased in the summer and increased in the winter based on the SMI and CCIs values in the soil of the agricultural land. The values of SMI and CCIs are dependent on the amount of water flowing from Al-Hawizeh marsh. This indicated a reduction in flow of water amounts during the summer and a rise in the winter. The high temperatures and subsequent drying of soil in the summer led to a further decrease in the SMI and CCIs values in the summer. All of these results exceeded the maximum allowable concentration of salinity and chemicals in agricultural land based on Iraq’s standards of soil for agriculture (Law No. 25/1967, environmental protection and public agricultural lands soil from pollution) [18].
5. Image Classifications Figure 14 presents the minimum and the maximum spatial distribution of salinity and chemical (Fe, Pb, Cu, Cr, and Zn) contents for soils in the agricultural lands of Al-Hawizah marsh obtained using the mathematical models during all four seasons. Figure 14A represents the spatial distribution with the minimum and maximum concentration values for salinity, Fe, Pb, Cu, Cr, and Zn in the autumn, which were determined to be (196–9235), (274–35,531), (0.47–14), (0.42–26), (0.03–10.7), and (1.6–2.95) mg/dm3, respectively. Figure 14B represents the spatial distribution with the minimum and maximum concentration values for salinity, Fe, Pb, Cu, Cr, and Zn in the summer, which were (992–18,214), (56–4604), (0.47–9), (0.5–18.4), (0.08–4.3), and (4.9–32) mg/dm3, respectively. Figure 14C represents the spatial distribution of the minimum and maximum concentration of salinity, Fe, Pb, Cu, Cr, and Zn in the spring, which were (813–7944), (290–11234), (0.37–13), (0.05–26), (0.015–5.9), and (2.7–4.6) gm/dm3, respectively. Figure 14D represents the spatial distribution of the minimum and maximum salinity, Fe, Pb, Cu, Cr, and Zn in the winter, which were (83–4628), (1013–3818), (0.29–25), (0.26–40), (0.017–3.73), and (2.74–4028) mg/dm3, respectively. Figure 15 illustrates two-dimensional image layers for salinity and chemicals distribution during all four seasons based on the proposed models. The salinity values were observed to be lowest during the summer and highest during the winter, which were decided by the SMI, indicating a reduction of the soil moisture indexes during summer and a rise in the winter. The soil moisture index values were decided by the spectral signatures of the Landsat-8 Short Wave Infrared (SWIR-1) band B6 and the Thermal Infrared Sensor (TIRS-2) band B11, which are considered suitable and sensitive as far as SMI is concerned. These salinity concentrations displayed a positive relationship among the soil moisture index values. [42,43,51]. All chemical values were determined using the CCIs and indicated a reduction of the clay CCIs in the summer and a rise during winter. The CCIs were determined by the spectral signatures of the Landsat-8 SWIR-1 band B6 and SWIR-2 band B7, which are considered suitable and sensitive for this purpose. The chemical concentrations displayed a positive relationship with CCI values [19]. Geosciences 2020, 10, 207 15 of 20 Geosciences 2020, 10, x FOR PEER REVIEW 15 of 20
Figure 14. Cont.
Geosciences 2020, 10, 207 16 of 20 Geosciences 2020, 10, x FOR PEER REVIEW 16 of 20
FigureFigure 14. 14.Salinity Salinity and and chemical chemical spatial spatial distributionsdistributions of the the agricultural agricultural lands lands based based on on the the proposed proposed modelsmodels during during all all four four seasons: seasons: ( A(A)) autumn, autumn, ((B) summer, ( (C)) spring, spring, and and (D (D) )winter. winter.
Geosciences 2020, 10, 207 17 of 20 Geosciences 2020, 10, x FOR PEER REVIEW 17 of 20
FigureFigure 15. Two-dimensionalTwo-dimensional image image layers layers for for salinity salinity and and chemical chemical distribution distribution during during all four all seasons four seasonsbased on based the proposed on the proposed models. models. 6. Conclusions 6. Conclusions The agricultural land surrounding Al-Hawizeh marsh in southern Iraq is considered a habitat The agricultural land surrounding Al-Hawizeh marsh in southern Iraq is considered a habitat provider for biodiversity and an enriched resource for the production of several crops. The neglect provider for biodiversity and an enriched resource for the production of several crops. The neglect and wars of recent years are factors that may have led to an increase of salinity, iron, lead, copper, and wars of recent years are factors that may have led to an increase of salinity, iron, lead, copper, chromium, and zinc in the soil. Techniques have been developed for the detection and monitoring of chromium, and zinc in the soil. Techniques have been developed for the detection and monitoring of physical and chemical parameters in soils. physical and chemical parameters in soils. AAS is a technique that can be used to measure the salinity and chemicals in soils. The mathematical AAS is a technique that can be used to measure the salinity and chemicals in soils. The models of this study were based on the B6, B7, and B11 bands of Lasndsat-8 satellite imagery. mathematical models of this study were based on the B6, B7, and B11 bands of Lasndsat-8 satellite These bands were suitable to combine with the mathematical models for assessing and monitoring imagery. These bands were suitable to combine with the mathematical models for assessing and the spatial distribution of salinity and chemicals from satellite imagery. It is concluded that monitoring the spatial distribution of salinity and chemicals from satellite imagery. It is concluded the integration of mathematical models with remote sensing data and GIS can provide a powerful tool that the integration of mathematical models with remote sensing data and GIS can provide a for predicting, assessing, monitoring, managing, and mapping salinity and chemicals components. powerful tool for predicting, assessing, monitoring, managing, and mapping salinity and chemicals The seasonal variation of average salinity and chemicals was obtained using the SE and CE components. equations. The salinity values were evaluated by using the SE during all four seasons. The average The seasonal variation of average salinity and chemicals was obtained using the SE and CE salinity concentrations during autumn, summer, spring, and winter were 1175, 1010, 1105, and 1789 equations. The salinity values were evaluated by using the SE during all four seasons. The average mg/dm3, respectively. These values were determined by the SMI. This indicated a reduction of the SMI salinity concentrations during autumn, summer, spring, and winter were 1175, 1010, 1105, and 1789 mg/dm3, respectively. These values were determined by the SMI. This indicated a reduction of the SMI values in the summer and a rise in the winter. These salinity concentrations displayed a positive relationship with SMI values.
Geosciences 2020, 10, 207 18 of 20 values in the summer and a rise in the winter. These salinity concentrations displayed a positive relationship with SMI values. The chemical concentrations were obtained using the CE equation. These values were observed to be lowest during the summer and highest during the winter. The average values for (Fe) during the autumn, summer, spring, and winter were 813, 784, 842, and 1106 mg/dm3, respectively. The average values for (Pb) during the autumn, summer, spring, and winter were 4.85, 3.79, 4.74, and 7.2 mg/dm3, respectively. The average values for (Cu) during the autumn, summer, spring and winter were found to be 3.9, 3.1, 4.45, and 7.5 mg/dm3, respectively. The average values for (Cr) during the autumn, summer, spring, and winter were found to be 1.28, 0.73, 1.03, and 2.91 mg/dm3, respectively. The average values for (Zn) during the autumn, summer, spring, and winter were found to be 8.25, 6, 7.05, and 12 mg/dm3, respectively. These values were determined using the CCIs in the soil. This indicated a reduction of the CCIs in the summer and a rise in the winter season. These chemicals concentrations displayed a positive relationship with values of the CCIs. To conclude on the salinity and chemical values obtained based on the SMI and CCIs results, the SMI and CCIs were dependent on the amount of water coming from Al-Hawizeh marsh during the four seasons. This indicated a reduction in water during summer and an increase in winter. This was exacerbated by the high temperatures and dry soil in the summer. The salinity and chemical concentrations were above the maximum allowable concentration in agricultural land based on Iraq’s standards for agricultural land pollution. The DT classification scheme provided good results for this study. The DT classification depended on the output results of salinity and chemicals for both the SE and CE equations. This classification refers to all the parameters simultaneously in one stage. The output of the DT classification can display all the soil quality parameters (salinity, Fe, Pb, Cu, Cr, and Zn) in one image. This approach was repeated for each season. In conclusion, the developed systematic and generic approach may constitute a basis for determining the soil quality parameters in agricultural land worldwide.
Author Contributions: Conceptualization, H.A.H., H.D., A.S.D., W.H.H., and H.M.H.; Methodology, H.A.H., H.D., W.H.H., A.S.D., and H.M.H.; Validation, H.A.H., H.D., A.S.D., N.A.-A., W.H.H., and H.M.H.; Formal analysis, H.A.H., H.D., W.H.H., A.S.D., and H.M.H.; Investigation, H.A.H., H.D., W.H.H., A.S.D., and H.M.H.; Data corruption, A.H.H., H.D., W.H.H., A.S.D., and H.M.H.; Writing—original draft preparation, all authors; Writing—review and editing, all authors; Visualization, N.A.-A.; Project administration, N.A.-A. All authors read and agreed to the published version of the manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflict of interest.
References
1. Al-Handal, A.; Hu, C. MODIS observations of human-induced changes in the Mesopotamian Marshes in Iraq. Wetlands 2015, 35, 31–40. [CrossRef] 2. Azab, A.M. Integrating GIS, remote sensing; mathematical modelling for surface water quality management in irrigated watersheds. Ihe Delft Inst. Water Educ. 2012, 24, 75. 3. Kerekes, J.P.; Baum, J.E. Full-spectrum spectral imaging system analytical model, Geoscience; Remote Sensing. IEEE Trans. 2005, 43, 571–580. 4. Mather, P.M.; Koch, M. Remote Sensing: Basic Principles: Computer Processing of Remotely-Sensed Images: An Introduction; John Wiley & Sons: Hoboken, NJ, USA, 2011; p. 67. 5. Ongley, E.D. Water quality management: Design, financing; sustainability considerations-II. In Proceedings of the 2000 Invited Presentation at the World Bank’s Water Week Conference: Towards a Strategy for Managing Water Quality Management, Washington, DC, USA, 2–4 April 2000; Volume 4, p. 34. 6. Ustin, S. Manual of Remote Sensing: Remote Sensing for Natural Resource management; Environmental Monitoring; John Wiley & Sons: Hoboken, NJ, USA, 2004; Volume 6, pp. 23–56. 7. Zacharias, I.; Gianni, A. Hydrodynamic; Dispersion Modeling as a Tool for Restoration of Coastal Ecosystems, Application to a Re-Flooded Lagoon. Environ. Model. Softw. 2008, 23, 751–767. [CrossRef] Geosciences 2020, 10, 207 19 of 20
8. Bahrawi, J.A. Evaluation of distinctive normalized difference vegetation indices in soil erosion estimation using remote sensing concepts in WadiYalamlam, Saudi Arabia. Indian J. Geo Mar. Sci. 2018, 47, 2087–2093. 9. Lu, D.; Mausel, P.; Brondizio, E.; Moran, E. Change detection techniques. Int. J. Remote Sens. 2004, 25, 236–240. [CrossRef] 10. Azab, A. Integrating GIS, Remote Sensing; Mathematical moDelling for Surface Water Quality Management in Irrigated Watersheds; IHE Delft Institute for Water Education: Delft, The Netherlands, 2012; p. 21. 11. Mujumdar, N. World development report, 2000/2001: Attacking poverty. Indian J. Agric. Econ. 2011, 56, 146. 12. Green, E.A. Hydropolitics in the Middle East; US Policy; DTIC Document: Fort Belvoir, VA, USA, 1993; Volume 3, p. 87. 13. Lowi, M.R. Rivers of conflict, rivers of peace. J. Int. Aff. 1995, 49, 123–144. 14. Maltby, E. An Environmental & Ecological Study of the Marshlands of Mesopotamia, Draft Consultative Bulletin: AMAR appeal Trust; University of Exeter, Wetland Ecosystems Research Group: London, UK, 1994; Volume 3, pp. 19–23. 15. Nicholson, E.; Clark, P. Iraqi Marshlands: Politico’s Pub; AMAR International Charitable Foundation: London, UK, 2003; Volume 2, p. 45. 16. Partow, H. The Mesopotamian Marshlands: Demise of an Ecosystem. Nairobi (Kenya): Division of Early Warning; Assessment, United Nations Environment Programme; UNEP publication UNEP/DEWA: Nairobi, Kenya, 2001; Volume 4, pp. 1–3. 17. Hayder, D.; Shattri, M.; Noordin, A.; Biswajeet, P. Band-to-band registration model for near-equatorial Earth observation satellite images with the use of automatic control point extraction. Int. J. Remote Sens. 2015, 36, 2184–2200. 18. Hasab, H.A.; Jawad, H.A.; Dibs, H.; Hussain, H.M.; Al-Ansari, N. Evaluation of Water Quality Parameters in Marshes Zone Southern of Iraq Based on Remote Sensing; GIS Techniques. Water, AirSoil Pollut. 2020, 231, 1–11. [CrossRef] 19. Schwarte, C. Environmental Protection in islamic Law: An Overview on Potential Influences for Legal Developments in Iraq. Local Environ. 2003, 8, 567–576. [CrossRef] 20. Hashim, A.; Marghany, M. Developing a hydrological model for water quality in Iraq marshes zone using Landsat-TM. J. IopscienceEarth; Environ. Sci. 2016, 37, 012073. [CrossRef] 21. Ammenberg, P.; Flink, P.; Lindell, T.; Pierson, D.; Strombeck, N. Bio-optical modelling combined with remote sensing to assess water quality. Int. J. Remote Sens. 2002, 23, 1621–1638. [CrossRef] 22. Goodchild, M.F.; Steyaert, L.T.; Parks, B.O. GIS; Environmental Modeling: Progress; Research Issues; John Wiley & Sons: Hoboken, NJ, USA, 1996; Volume 25, p. 183. 23. Dellapenna, J.W.; Nicholason, E.; Clark, P. The Iraqi Marshlands; A Human Environmental Study; AMAR International Charitable Foundation: London, UK, 2001; Volume 7, p. 123. 24. Mather, P.M.; Koch, M. Computer Processing of Remotely-Sensed Images: An Introduction; John Wiley & Sons: Hoboken, NJ, USA, 2011; p. 95. 25. El-Din, M.S.; Gaber, A.; Koch, M.; Ahmed, R.S.; Bahgat, I. Remote Sensing Application for Water Quality Assessment in Lake T imsah, Suez Canal, Egypt. J. Remote Sens. Technol. 2013, 1, 61. [CrossRef] 26. Guerrero, D.V.C.; NO, U.G.A. The Anacostia River: Ecological Studies of Water Pollution Biology; DC Water Resources Research Center: Washington, DC, USA, 1991. 27. Liu, H.; Zheng, L.; Yin, S. Multi-perspective analysis of vegetation cover changes; driving factors of long time series based on climate; terrain data in Hanjiang River Basin, China. Arab. J. Geosci. 2018, 11, 509. [CrossRef] 28. Cao, X.; Feng, Y.; Wang, J. Remote sensing monitoring the spatio-temporal changes of aridification in the Mongolian Plateau based on the general Ts-NDVI space, 1981–2012. J. Earth Syst. Sci. 2017, 126, 58. [CrossRef] 29. Jensen, J.R. Introduction to Digital Image Processing: A Remote Sensing Perspective, 3rd ed.; Prentice Hall: Upper Saddle River, NJ, USA, 2005; pp. 239–247. 30. Zhuo, L.; Han, D.; Dai, Q. Exploration of empirical relationship between surface soil temperature; surface soil moisture over two catchments of contrasting climates; land covers. Arab. J. Geosci. 2017, 10, 410. [CrossRef] 31. Roy, A.; Keesari, T.; Sinha, U.K.; Sabarathinam, C. Delineating groundwater prospect zones in a region with extreme climatic conditions using GIS; remote sensing techniques: A case study from central India. J. Earth Syst. Sci. 2019, 128, 201. [CrossRef] Geosciences 2020, 10, 207 20 of 20
32. Dehmordi, L.M.; Savari, A.; Dostshenas, A.; Asgari, H.M.; Abasi, A. Remote chlorophyll-a, SST; kd490 retrieval in Northwest Persian gulf using landsat 8 satellite data. Indian J. Geo Mar. Sci. 2018, 47, 148–169. 33. Jacintha, T.; Rajasree, S.R.; Kumar, J.D.; Sriganesh, J. Assessment of wetland change dynamics of Chennai coast, Tamil Nadu, India, using satellite remote sensing. Indian J. Geo Mar. Sci. 2019, 48, 1258–1266. 34. Giri, S.; Singh, A.K.; Mahato, M.K. Metal contamination of agricultural soils in the copper mining areas of Singhbhum shear zone in India. J. Earth Syst. Sci. 2017, 126, 49. [CrossRef] 35. Ennaji, W.; Barakat, A.; El Baghdadi, M.; Oumenskou, H.; Aadraoui, M.; Karroum, L.A.; Hilali, A. GIS-based multi-criteria land suitability analysis for sustainable agriculture in the northeast area of Tadla plain (Morocco). J. Earth Syst. Sci. 2018, 127, 79. [CrossRef] 36. Platonov, A.; Noble, A.; Kuziev, R. Soil salinity mapping using multi-temporal satellite images in agricultural fields of syrdarya province of Uzbekistan Developments in Soil Salinity Assessment; Reclamation. Springer 2013, 112, 87–98. 37. Fan, X.; Liu, Y.; Tao, J.; Weng, Y. Soil salinity retrieval from advanced multi-spectral sensor with partial least square regression. Remote Sens. 2015, 7, 488–511. [CrossRef] 38. Allbed, A.; Kumar, L. Soil salinity mapping; monitoring in arid; semi-arid regions using remote sensing technology: A review. Adv. Remote Sens. 2013, 5, 43–52. [CrossRef] 39. Haith, D.A.; Tubbs, L.J. Watershed loading functions for nonpoint sources. J. Environ. Eng. Div. 1981, 107, 121–137. 40. Abuelaish, B.; Olmedo, M.T.C. Scenario of land use; land cover change in the Gaza Strip using remote sensing; GIS models. Arab. J. Geosci. 2016, 9, 274. [CrossRef] 41. Nasirian, H.; Irvine, K.; Sadeghi, S.M.T.; Mahvi, A.H.; Nazmara, S. Assessment of bed sediment metal contamination in the Shadegan; Hawr Al Azim wetlands, Iran. Environ. Monit. Assess. 2016, 188, 1–15. [CrossRef] 42. USGS. Landsat 8 Operational Land Imager (OLI); Thermal Infrared Sensor (TIRS); Frequently Asked Questions about the Landsat Missions; USGS. Science for a Changing World: Reston, VA, USA, 2016; pp. 1–6. 43. USGS. Using the USGS Landsat-8 Product. USGS. Science for a Changing World; Geological Survey in Vandenberg Air Force Base, California on an Atlas-V rocket in USA; U.S. Department of the Interior: Washington, DC, USA, 2013. 44. Muhaimeed, A.S. Status of Soil Information in Iraq, University of Baghdad College of Agriculture. Int. J. Digit. Earth 2014, 4, 434–443. 45. Omer, Tara Mohamed Anwar. Country Pasture/Forage Resource Profile. University of Sulaimaniyah. Available online: http://www.fao.org/ag/agp/AGPC/doc/Counprof/Iraq/Iraq.html (accessed on 13 December 2019). 46. Panagos, P.; Jones, A.; Bosco, C.; Senthil Kumar, P.S. European digital archive on soil maps (EuDASM): Preserving important soil data for public free access. Int. J. Digit. Earth 2011, 4, 434–443. [CrossRef] 47. Selvaradjou, S.-K.L.; Montanarella, O.; Spaargaren, D. Dent.European Digital Archive of Soil Maps (EuDASM)—Soil Maps of Asia DVD-ROM version. EUR 21823 EN. Off.Off. Publ. Eur. CommunitiesLuxemb. 2005, 7, 134–143. 48. Maurya, A.; Kesharwani, L.; Mishra, M.K. Analysis of Heavy Metal in Soil through Atomic Absorption Spectroscopy for Forensic Consideration. Int. J. Res. Appl. Sci. Eng. Technol. 2018, 6, 231–242. [CrossRef] 49. Addis, W.; Abebaw, A. Determination of heavy metal concentration in soils used for cultivation of Allium sativum L.(garlic) in East Gojjam Zone, Amhara Region, Ethiopia. Cogent Chem. 2017, 1, 1419422. [CrossRef] 50. Khattab, M.F.; Merkel, B.J. Application of Landsat 5; Landsat 7 images data for water quality mapping in Mosul Dam Lake, Northern Iraq. Arab. J. Geosci. 2014, 7, 3557–3573. [CrossRef] 51. Pour, A.B.; Hashim, M. Hydrothermal alteration mapping from Landsat-8 data, SarCheshmeh copper mining district, south-eastern Islamic Republic of Iran. J. Taibah Univ. Sci. 2015, 9, 155–166. [CrossRef] 52. Guide, E.U.S. ENVI on-line software user’s manual. ITT Visual Information Solutions 2008. Arab. J. Geosci. 2015, 8, 154–167.
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).