Arab J Geosci (2017) 10:103 DOI 10.1007/s12517-017-2904-5 ORIGINAL PAPER Utilizing Landsat-8 data in mapping of sabkha, mangroves, and land covers in Jizan coastal plain, southwestern Saudi Arabia Ali A. Khawfany1 & Mahmoud A. Aref1,2 & Mohammad I. Matsah3 & Rushdi J. Taj1 Received: 18 January 2016 /Accepted: 13 February 2017 # Saudi Society for Geosciences 2017 Abstract Digital image processing methods were applied on useful as a base reference for further studies by geologists, Landsat-8 data to differentiate different natural land covers engineering geologists, hydrogeologists, and governmental and urban areas in Jizan, Red Sea Coast, Saudi Arabia. The decision leaders. computer-aided works were combined with field investigation to confirm the achieved results. The false-color image with Keywords Image processing . Field study . Jizan . Saudi bands 5, 4, and 3 in RGB, respectively, was found useful in Arabia . Geomorphic features delineating between sabkha and non-sabkha areas. The sabkha areas are characterized by hues of light brown color, and the non-sabkha areas have hues of light grey to light Introduction bluish color. A false-color image with bands 5, 6, and 4 clearly differentiates between mangroves in hues of red color and In the industrialized world, more than 50% of the population other vegetation in hues of orange color. The principal com- lives within 1 km of the sea coasts. As the population expands, ponent analysis (PCA) image clearly shows the wadi drainage continued anthropogenic pressures are placed on coastal re- and estuary in several shades of green color; the coastal gions (NOAA 1995). Coastal resource management is becom- sabkha areas have hues of blue, which range from light blue ing increasingly important, and the ability to provide large- (denoting dry sabkha) to dark blue (denoting wet sabkha), and scale synoptic assessments of valued resources is vital (Toole sand dunes are shown in fuchsia color. The results of this et al. 2000). As the population and urbanization expand, con- study identify features such as urban populated areas, land tinued strategic planning is required. This planning depends vegetation, moisture intensity in the sediments, shallow and on the study of many factors, one of them being the study of deep waters, reef islands, sabkha and non-sabkha areas, salt land covers. dome, shores, tidal flats, wadi drainage, mangrove habitat, the Many research works have been carried out concerning the width of shallow seawater, and land-seawater contacts. geomorphic features of the Red Sea coastal areas of Saudi Several maps are produced in this study showing the distribu- Arabia. These studies were based on the combination of re- tion of the abovementioned features. These maps will be mote sensing techniques and sedimentological characteristics of several environments (e.g., Al Saifi and Qari 1996; Bantan 1999; Al-Washmi et al. 2002; Qari and Basyoni 2003; Al- * Ali A. Khawfany Hazmi 2006;Rifaatetal.2010; Alharbi et al. 2011). Some [email protected] works are concerned with the application of remote sensing techniques to shed light on the large-time-scale environmental 1 Department of Petroleum Geology and Sedimentology, Faculty of changes (Awari and Mullah 2010; Youssef et al. 2012), the Earth Sciences, King Abdulaziz University, Jeddah, Saudi Arabia natural engineering geologic problems (Al-Modayan 2012), 2 Department of Geology, Faculty of Science, Cairo University, and the distribution of some bioaccumulation such as reefs Giza, Egypt and mangroves (Kumar et al. 2010; Rowlands et al. 2012; 3 Department of Structural Geology and Remote Sensing, Faculty of Elsebaie et al. 2013). The objective of this study is to apply Earth Sciences, King Abdulaziz University, Jeddah, Saudi Arabia different image processing techniques on the Landsat-8 image 103 Page 2 of 18 Arab J Geosci (2017) 10:103 to identify the natural and urban land covers and land uses and than it would be otherwise. The average annual precipitation to distinguish the morphologic features of the Jizan coastal is 11.6 mm (Elsebaie et al. 2013). The evaporation rate is plain along the Red Sea coast, in the southwestern part of relatively high (156 cm/year) (Abdelrahman and Ahmad Saudi Arabia. 1995). In the Jizan area, Blank et al. (1987) described three phys- iographic regions: the Red Sea shelf, the coastal plain, and the The study area foothills of the Tihama escarpment. The coastal plain of the Jizan area is represented by supratidal sabkha basins, offshore The study area, which is a part of Jazan province, is located bars, islands, mangrove swamps, and shallow lagoons. The between lat. 16° 34′ and 17° 16′ N and long. 42° 20′ and 42° plain is about 10 km wide and covered by Quaternary aeolian 52′ E; it represents the coastal plain of Jizan City (Fig. 1). Jizan sand, alluvial sand and gravel, loess, and flood plain silt de- City, the main town of Jazan province, is situated on the west- posits. The only relief on the plain is provided by a salt dome, ern coastal plain of Saudi Arabia. It has an arid tropical climate which may reach up to 70 m above mean sea level at Jizan with an average annual temperature of 30.4 °C and an average City. annual humidity of about 70% (PME 2015). High humidity Sabkha is an Arabic word meaning salt-enriched flat from coastal lagoons makes the climate even less bearable area(s). The sabkha salt may be clearly visible in the form of Fig. 1 The study area is shown on a part of Landsat-8 satellite image Arab J Geosci (2017) 10:103 Page 3 of 18 103 hygroscopic crystals or as a crust on the surface. The presence interpretation of the phenomenon under study. Six enhance- of salts is generally manifested by appearance of moisture or ment techniques were used in this study including band com- darker tones that distinguishes a sabkha from other areas. bination, convolution filtering, histogram equalization, princi- Even when a sabkha becomes dry, the moist appearance is pal component analysis (PCA), decorrelation stretch, and retained (Bahafzullah et al. 1993). band ratioing. Each technique aimed at four-band combina- tions were used including a true-color composite (TCC) where band 4 was assigned to red, band 3 was assigned to green, and Methodology band 2 was assigned to blue (R4, G3, B2). A false-color com- posite (FCC) has band combinations (R5, G4, B3), (R5, G6, Landsat-8 data B4), and (R7, G6, B4). The convolution filtering technique is the process of averaging small sets of pixels across an image This work depends on the Landsat-8 scene of Path 167, Row by moving a window of coefficients. It is used to change the 048, which is an image that consists of 11 spectral bands spatial frequency characteristics of an image. Histogram (Table 1) and covers an area of about 170 × 189 km2.The equalization is a nonlinear stretch that redistributes pixel study area (Fig. 1) is a subset of the full scene that is approx- values so that there is approximately the same number of imately 6100 km2. pixels with each value within a range; this includes contrast stretching and brightness enhancement. PCA is used to com- Digital image processing press bands of redundant data that are similar, and produces new bands of data of high variance, whereas the decorrelation Digital image processing procedures were carried out using stretch applies a contrast stretch to the principal components ERDAS IMAGINE 8.4 software (Smith and Brown 1999). of an image. These methods usually produce colorful FCC The objective is to process and enhance the image as to produce images emphasizing the variations between surface materials a better image that can be visually interpreted (Lillesand and which leads to better way of discrimination between earth Kiefer 1994). These techniques have proven valuable for differ- materials (Smith and Brown 1999). ent geological applications such as mapping of lithological units, lineaments, structures, and coastal plain mapping and manage- ments (Mohammad et al. 2001; Yamaguchi and Naito 2003; Field study Ninomiya 2003;El-Badawi2006; Moufaddal and Rifaat 2006; Awari and Mullah 2010;Rifaatetal.2010; Alharbi et al. 2011; Many field visits were carried out to the study area in August Sagheer et al. 2011;PourandHashim2012; Van der Meer et al., 2013 and March 2015 to confirm the obtained digital image 2012; Elsebaie et al. 2013;Sagheer2013). processing results. The location of each feature was carefully Generally, the processing of digital image data can be recorded, visited, studied, and photographed. The combina- achieved using different techniques, one of which is image tion between Landsat image processing and field data was enhancement which involves techniques for increasing the done to investigate and identify accurately several geomor- visual distinction between features in a scene and aids in better phic features. Table 1 The Landsat-8 bands used in the present study (Web Bands Wavelength Resolution reference 1) (μm) (m) Operational Land Imager (OLI) Band 1—coastal aerosol 0.43–0.45 30 Band 2—blue 0.45–0.51 30 Band 3—green 0.53–0.59 30 Band 4—red 0.64–0.67 30 Band 5—near infrared (NIR) 0.85–0.88 30 Band 6—SWIR 1 1.57–1.65 30 Band 7—SWIR 2 2.11–2.29 30 Band 8—panchromatic 0.50–0.68 15 Band 9—cirrus 1.36–1.38 30 Thermal Infrared Sensor (TIRS) Band 10—thermal infrared (TIRS) 1 10.60–11.19 100(30)a Band 11—thermal infrared (TIRS) 2 11.50–12.51 100(30)a a TIR bands are acquired in 100 m resolution and they are re-sampled to 30 m 103 Page 4 of 18 Arab J Geosci (2017) 10:103 Results and discussion and vegetation, and the distinguishing of different forest types can be derived from band 2 (Web reference 2).