479

Soil salinity assessment in Toba Tek Singh using remote sensing and GIS SHAHID KARIM & EJAZ HUSSAIN Institute of geographical information systems (IGIS), National university of sciences and technology (NUST). Sector H-12, Islamabad, Pakistan. [email protected] , [email protected] Abstract. Soil Salinity is the major land degradation process in arid and semi arid regions and it has affected about 1billion hectares of lands around the world, which represents about 7% of the earth’s continental area. Globally, soil Salinity is spreading at a rate of up to 2 million hectare per year. Pakistan has a total area of 79.6 mha, with 22.0 mha cultivated and 6.28 mha affected by salinity within the irrigation regions. There is a need to develop such methods that should be faster and cheaper for the detection of this problem. The use of remote sensing and GIS can play an important role for the detection and mapping of saline areas in a very short time. Main objective of this study was to develop a methodology for detection and mapping of saline areas with the help of remote sensing and GIS and to correlate and verify this methodology with laboratory analysis of ground samples. This study was conducted to detect and map the saline areas with the help of satellite remote sensing and GIS in Tehsil Toba Tek Singh in central Punjab. Soil samples were collected from field and analyzed in laboratory. Results were interpolated and salinity map of 2011 was developed. Satellite image of this area was classified and a land cover map was developed containing different classes including saline areas. Results were compared and there was 80% correlation between these two methods that is a good indication for this methodology. Keywords: Soil salinity, Remote sensing, Geographic Information System (GIS), NDVI, NDSI. 1. Introduction. Soil Salinization is one of the most common land degradation processes in arid and semi arid regions, where evaporation exceeds over precipitation and it reduces the productivity of agricultural lands adversely. Soil salinity has affected about 1billion hectares of lands around the world, which represents about 7% of the earth’s continental extent. On average 20% of the world’s irrigated lands are affected by salts, but this figure increases more than 30% in countries such as Egypt, Iran and Argentina (Abdelfattah at al, 2009). Estimates of irrigation salinity for top four irrigators in the world are India11%, Pakistan 21%, US 23% and Mexico10% of the irrigated land. At global scale, soil Salinization is spreading at a rate of up to 2.0 mha per year. Pakistan has a total area of 79.6 mha with 22.0 mha cultivated and 6.28 mha affected by salinity within the irrigation regions. A land area between 2 to 3 mha is categorized as wasteland due to high salinity and sodicity. It is estimated that 25% of irrigated land in Punjab and 40% in Sindh are salt affected (Abbas at al ,2010). This picture shows that soil salinity is affecting agricultural land adversely in Pakistan. The problem of detection, monitoring and mapping salt affected soil is very difficult because dynamic processes are involved. Saline areas can be detected through traditional methods but it takes long time and it is not possible to check and monitor the whole area with the help of these traditional methods. There is a need of such method and techniques that provide the required results in a very short time to monitor and cope with this problem. The combination of remote sensing with GIS can play a vital role for the detection and mapping of saline areas in a short time. Remote sensing can help to detect the salt affected lands while the GIS can be helpful to map these areas defining the different classes of salinity. The use of remote sensing and GIS techniques is the best selection for this purpose as this method is cheaper than traditional techniques. Main objective of this study was to develop a methodology for detection and mapping of saline areas with the help of remote sensing and GIS and to correlate and verify this methodology with laboratory analysis of ground samples. The present study was conducted in Tehsil Toba Tek Singh of Toba Tek Singh District located in the centre of the Punjab. It extends from 30° 44´ 25 ´´ N to 31° 07´ 19´´ N and 72° 13´ 43´´ E to 72° 45´ 10´´ E and has an area of 1279 Sq.Km. This area is a part of Rachna Doab (the area lying between river Ravi and river Chenab) and is irrigated by the canal water system of river Chenab. Being an agricultural area this land produces wheat, cotton, sugar cane and other food and cash crops. This area is categorized under hot semi arid climatic zone of Pakistan. The source of rainfall is monsoon that occurs in July and august. In winter there is very less rainfall that is due to western depressions. Therefore all the agriculture depends upon canal system. Due to canal irrigation system this area is facing the problem of water logging and salinity. 2. Materials and Methods a. Data Acquisition and Analysis: The satellite image belongs to Land Sat 4-5 Thematic Mapper of June 2011 was used. The specifications of this sensor are as:

International Conference of GIS-Users, Taza GIS-Days, May 23-24, 2012 Proceeding Book

480

Sensor Information Spatial Resolution

30 m in Band 1-5 and 7(60 m -thermal,)

Spectral Range

0.45 - 12.5 µm

Number of Bands

7

Temporal Resolution Image Size Swath

16 days 183 km X 170 km 183 km Table 1. Sensor Information, Source: glovis.usgs.gov

The overall methodology adopted for this study is as under:

Figure 1. Methodology Two scenes of this sensor were used. The layers were stacked and mosaic was applied on these scenes. The study area was extracted by applying sub setting on mosaic image. The process of supervised classification was performed to get different classes on the basis of pixel reflectance in the study area. Six classes i-e Built up area, Watered fields and bodies, Forest, Saline areas, Barren land and Vegetation were used to develop a land cover map of study area. Saline and non saline areas were highlighted and a salinity map 2011 of this area was developed. Normalized Difference Vegetation Index (NDVI) and Normalized Difference Salinity Index (NDSI) were also calculated to verify the saline areas in Classified Image. For field based analysis a GPS based field survey was conducted. 196 soil samples were collected and analyzed in laboratory. The presence of Electric Conductivity (EC) and pH were tested in the soil samples. The following standard was used to categorize the soil in different levels of salinity. Salinity Classification Soil

EC(dS/m)

Slightly saline

4-8 dS/m

Moderately saline

8-16 dS/m

Strongly saline (Richards, 1954)

More than 16 dS/m Table 2. Salinity classification

The laboratory results were interpolated and a map was generated showing different categories of salinity. This salinity map was then compared with the map developed by image based analysis.

Colloque International des Utilisateurs de SIG, Taza GIS-Days, 23-24 Mai 2012 Recueil des Résumés

481

3. Results and Discussion: In the image based analysis the land cover map of study area was developed with the help of supervised classification. Six classes were used to develop this map.

Figure 2. Land cover map of tehsil Toba Tek Singh

Figure 3. Land cover areas based on classification

This map shows the different land covers with areas. Details of areas under each land cover are shown in the bar graph. These results showed that 78 Sq.km i-e 6% of total area is saline. It should be noted that image only showed the area that was saline and bare land. The areas that were saline in low intensity but covered by vegetation are not shown as saline in image. The intensity of salinity in those areas can only be calculated by field based sampling. Each land cover showed a different spectral response during classification. Signature mean plot of each land cover was produced. This was done with the help of locations of field survey points. Reflectance of each survey point in all seven bands was calculated then average reflectance of survey points in all seven bands was calculated. Results were as under in the table and graph. 0.45 -0.52

0.52 0.60

0.63 0.69

0.76 0.90

1.55 1.75

10.5 12.5

2.08 2.35

Built up area

127

69

83

90

132

175

79

Water Bodies and Fields

121

64

77

91

124

172

72

Forest

105

50

53

95

87

154

39

Saline Areas

144

80

101

97

154

178

97

Barren Land

133

73

91

89

140

181

87

Vegetation

115

59

67

92

112

165

59

Band Range(µm) Classes

Table 3. Spectral reflectance of various Land covers

Figure 4. Spectral reflectance of various Land covers

The result showed that the lowest spectral reflectance was of forest in all bands except band number 4 (near infra red) where water has the minimum reflectance value and forest cover has higher reflectance value. While vegetation cover, watered fields and bodies, built up areas, barren land and saline areas gained higher reflectance values simultaneously in all bands except band number 4 (near infra red) where vegetation cover and forest cover have higher reflectance values than all land covers except saline areas. Saline areas have maximum reflectance values in all bands. In band number 6 the reflectance value was maximum i-e 178 while minimum reflectance value was seen in band number 2 i-e 80. Accuracy assessment was done and Overall Classification Accuracy was 89.17% NDVI and NDSI were also calculated to ensure the saline areas in classified image. For this purpose following formulae were applied on image. NIR – R NDVI =

(0.76to0.9µm) - (0.63to0.69) =

NIR + R

R – NIR (1) , NDSI =

(0.76to0.9µm) + (0.63to0.69)

(0.63to0.69) - (0.76to0.9µm) =

R + NIR

(2) (0.63to0.69) + (0.76to0.9µm)

International Conference of GIS-Users, Taza GIS-Days, May 23-24, 2012 Proceeding Book

482

(Khan.et al, 2005; Deering.et al, 1975).

Figure 5. NDVI in Tehsil Toba Tek Singh

Figure 6. NDSI in Tehsil Toba Tek Singh

Parallel to this methodology GPS based field survey was also conducted to produce the salinity map of study area. For this purpose 196 samples were collected from different locations in study area. On the basis of laboratory analysis of field samples following map was developed in GIS environment.

Figure 7. Areas covered by salinity classes

Figure 8. Area calculation covered by salinity classes

4. Conclusion. The results of salinity classes showed that 64 sq.km areas i-e 5% of total area were highly saline in the study area. The image based supervised classification showed that saline area was 78 sq.km i-e 6% of total area. These results showed 80% correlation between these two methods that is a good correlation for further studies. On the basis of mentioned results it can be concluded that study of saline areas can be conducted through image based classification. It will provide satisfactory results that can be verified by the sampling techniques rather than time consuming and hurdles based field study. References. [1] Abbas, A., & S.Khan. “Using remote sensing techniques for appraisal of irrigated soil salinity”,(2010). [2] Abdelfattah, M., Shahid, S., & Othman, Y; “Soil salinity mapping model developed using RS and GIS-A case study from Abu-Dhabi, United Arab Emirates”. European Journal of scientific research, (2009). 26 (3), 342-351. [3] Buces, F., Siebea, C., Cramb, S., & Palaciob, J; “Mapping soil salinity using a combined spectral response index for bare soil and vegetation: A case study in the former lack Texcoco, Mexico”. Journal of arid environment, (2005). 65, 644-667. [4] Deering D.W, Rouse J.W, Haas J.R.H & Schell J.S; “Measuring forage production of grazing units from Landsat MSS data”. Proc. Tenth International Symposium on Remote Sensing of the Environment. (1975). Ann Arbor. MI, pp. 1169-1178. [5] Dwivedi, R. “Monitoring and the study of effects of image scale on delineation of salt affected soil in the Indo-Gangetic plains”. International journal of remote sensing, (1992). 13, 1527-1536. [6] Dwivedi, R., & Sreenvas.K; “Delineation of salt affected soils and water logged areas in the Indo-Gangetic plains using IRS-IC LISS-III data”. International journal of remote sensing , (1998), 19, 2739-2751. [7] Ghassemi F, Jakeman A.J & Nix, H.A; “Salinization of Land and Water Resources: Human Causes, Extent, Manegment and Case Studies”. (1995). CAB Int., p. 526 [8] Iqbal,F. “Detection of salt affected soil in Rice –wheat area using Satellite Image”. African Journal of Agricultural Research, (2011). Vol. 6(21), 4973-4982.

Colloque International des Utilisateurs de SIG, Taza GIS-Days, 23-24 Mai 2012 Recueil des Résumés

483

[9] Khan, M., & Sato, Y; “Monitoring hydro-salinity status and its impact in irrigated semi-arid areas using IRS-IB LISS-II data”. Asian journal of Geoinformatics, (2001). 1 (3), 63-73. [10] Khan.M.S, Rastoskuev, V., Sato.Y, & Shiozawa, S; “Assesment of hydrosaline land degradation by using a simple approach of remote sensing indicators”. Agricultural water management , (2005). 77, 96-109. [11] Naseri M.Y; “Characterization of salt-affected soil for modelling sustainable land management in semi arid environment: a case study in the Gorgan region of Northeast Iran”, (1998). M.sc. Thesis. Ghent University, Belgium [12] Matternicht, G., & J.A. Zinck; “Remote sensing of soil salinity: potentials and constraints”. Remote sensing of environment, (2003). 85, 1-20. [13] Sharma, R & G Bhargawa; “Landsat imagery for mapping saline soils and wetlands in north-west India”. International journal of remote sensing, (1998). 9, 69-84. [14] Szabolcs I; “Salinization of soil and water and its relation to desertification”. Desertificat. Control Bull. (1992), 21: 32-37.

International Conference of GIS-Users, Taza GIS-Days, May 23-24, 2012 Proceeding Book