1 Introduction

Coastal landforms are highly dynamic in nature and are continuously modified by natural and anthropogenic processes (Li et al. 2001 and Murali et al. 2015). Shoreline is influenced by coastal processes like waves and the resultant near-shore circulation, sediment characteristics and beach forms (Kumar et al. 2010). Global population living in the coastal areas is likely to be increased to 600 million by the year 2100 (Nicholls and Mimura 1998). Urbanization and industrialization, population growth, trade and tourism are exerting continuous stress on the resources of the coastal region, thereby exposing people to coastal hazards. Hydrodynamic, geomorphic, tectonic and climate forces induce shoreline changes, thus making the coastal zone a vulnerable region. Shoreline is one of the most important terrain features of coastal zone and is defined as the line of contact between land and sea. Although it is easy to define shoreline, it is difficult to capture, as it has never been stable in either long-term or short-term positions (Nayak 2002; Boak and Turner 2005). Processes, such as tidal flooding, sea level rise, winds, erosion–deposition, land subsidence, regular wave action and periodic storm surges, are mainly responsible for gradual as well as abrupt changes in the coastal region (Nayak 2002; Thom and Hall 1991; Pethick 2001; Kumar et al. 2010). The shoreline and geomorphological features are indicators of the changes that have taken place in the coastal region. Further, natural hazards like Tsunami and periodic cyclones along the Indian coast have resulted in the geomorphological and shore line changes (Cooper et al. 2004; Scheffers et al. 2005). An ever increasing coastal population and thereby the demand for qualitative information on the coastal hazards necessitate regular shoreline mapping for an effective coastal zone management. Therefore, periodic monitoring of shorelines and geomorphic features is essential to understand the temporal and spatial changes in the shoreline and other coastal processes (Moran 2003).

Several studies were carried out to understand various coastal processes, in particular mapping of geomorphological landforms and shoreline change analysis using satellite data (Kumari et al. 2012; Mujabar and Chandrasekhar 2011; Tonisso et al. 2012; Murali et al. 2013, Sudha Rani et al. 2015). Temporal multispectral satellite data enable us to identify the coastal segments that are experiencing erosion or accretion, which can provide valuable information for coastal zone management. Present study is focused on the coastal region of Odisha, which is on the eastern coast of India. In the previous studies carried out by Kumar et al. (2010), Pritam and Acharya (2010) and Mukhopadhyay et al. (2012) pertaining to Odisha coast, shoreline changes using remote sensing and statistical techniques were addressed. Ramesh et al. (2011) also reported the dynamic nature of the shoreline along the Odisha coast incorporating in situ and satellite data. Although shoreline analysis along Odisha coast was carried out by several researchers, dynamic nature of shoreline necessitates the periodic mapping of geomorphological and shorelines changes using geospatial technologies. One of the major advantages of using remote sensing data for shoreline and geomorphology feature identification is its ability to generate information in spatial and temporal domain, which is crucial for coastal zone management (Saraf and Choudhary 1999).

Shoreline is highly dynamic in the deltaic regions as well as in the land fall areas of tropical cyclones (TC). Tropical cyclone affects the land cover and shoreline in the coastal areas. Storm surge, torrential rain and violent wind associated with tropical cyclones are major reasons for such disturbance in the coastal regions. Although the number of TCs is less in the BOB in comparison with other tropical oceans, their impact on the region is high due to shallow depth of the Bay and flat coast that produces much larger storm surges (Niyas et al. 2009). The extent of damage due to the cyclone and initial regeneration of the forest from the damaged state can also be identified using remote sensing techniques (Andrew and Elmore 2000). Bhomik and Cabral (2013) studied the impacts of cyclone on floristic diversity of Sunderbans, and the regeneration of mangrove forests affected by hurricane Andrew was studied by Baldwin et al. (2001). In particular, normalized difference vegetation index (NDVI) is frequently used to assess the vegetation cover over time (Rouse et al. 1973; Tucker 1979; Barrett and Curtis 1999; Lin et al. 2003; Ravishankar et al. 2004; Li et al. 2010; Julien et al. 2011). Nayak (2002) carried out change detection of vegetation cover along the Odisha coast using NDVI and observed that mangroves were not totally uprooted but had shed their leaves and within 2 months had reverted to their original state. Thus, in the present study, spatio-temporal analysis of shoreline changes and mapping of geomorphological features along Odisha coast using geospatial technologies were carried out along with decadal land cover change analysis with reference to TC landfall locations.

2 Study area description

The study area is geographically situated between 17°49′N and 22°34′N latitude and 81°27′E and 87°29′E longitude (Fig. 1). The length of the coast line is around 480 km running along six districts, viz. Balasore, Bhadrak, Kendrapada, Jagatsinghpur, Puri and Ganjam where important rivers such as Mahanadi, Brahmini, Baitarani and Subarnarekha discharge large volume of sediments into the BOB. Asia’s largest brackish water lagoon, the Chilka, mangrove forest with 672 km2 extension, the Bhitarkanika wildlife sanctuary, the world’s largest known nesting beaches of olive turtles, the Gahirmatha and the Rushikulya which play an important role in the economy of the Odisha state, are some of the significant ecological features falling in this coastal zone. Puri, Balasore, Cuttack and Ganjam districts of Odisha coast were frequently affected by several TCs and also due to Tsunami in 2004. These natural forces have altered the features such as coastal landforms, the width of beaches or position of the coastline. Therefore, identification of vulnerable areas along Odisha coast is essential for sustainable coastal development and planning.

Fig. 1
figure 1

Location map of the study area

Full size image

3 Methodology

3.1 Data used

Landsat (TM and MSS) satellite data for the years 1990, 1999 and 2009 were used in the present study. These datasets are used to map shoreline changes over the study period (1990–2009) and also for monitoring the land cover changes along the Odisha coast. The Survey of India (SOI) topographical maps (1:50,000) are used as reference maps. The details of the satellite data used in the study are given in Table 1.

Table 1 Details of the satellite data considered for the study
Full size table

3.2 Image processing

Absolute radiometric calibration is a prerequisite for multi-temporal and multi-sensor studies in order to reduce the effects due to changes in the sensor characteristics, atmospheric condition, solar angle and sensor view angle (Hall et al. 1995; Du et al. 2002; Coppin et al. 2004; Chen et al. 2005; Chandler et al. 2009). These corrections involve the conversion of the digital numbers (DN) to up-welling radiance and subsequently into the top-of-atmosphere (TOA) reflectance that helps in the temporal analysis. In this study, conversion of DN to at-satellite spectral reflectance was calculated according to the methods proposed by Markham and Baker (1986) and Chandler et al. (2009). Multi-spectral TOA images thus obtained were enhanced using digital image processing methods like contrast stretching for each band. Further, the false colour composite (FCC) is generated with suitable band combinations for analysis.

3.3 Delineation of decadal changes in the shoreline during 1990–2009

Delineation of shoreline and mapping was carried out based on the Survey of India (SOI) toposheet of 1972 and Landsat images for the years 1990, 1999 and 2009 in the GIS environment. Temporal changes in the shorelines for these years were identified using overlay analysis. The root-mean-square error achieved was less than 0.3 pixels. During overlay analysis, delineated shoreline of each image of the study area was compared with the remaining years and also with the shoreline delineated from the reference map of the year 1972. The methodology for computing gain and loss along the coast was adopted from Eludoyin et al. (2011). Quick assessment of the depositional or erosion activity along the coast could be obtained using overlay analysis.

3.4 Delineation of coastal geomorphological units

Coastal geomorphic unit boundaries were delineated by on-screen visual interpretation and digitization in the GIS platform. The Landsat TM image for the year 2009 was enhanced, and principal components were derived to segregate the information in the data. Principal component (PC) images are often used for landform delineation as it segregates correlative and non-correlative information in different bands. The false colour composite (FCC) was prepared using principal components derived from Landsat data and was used for delineation of geomorphic landforms within the 5 km zone from the coast. Major landforms in the study area viz. beach, beach ridge, lagoons, mangroves, tidal flats and pediment were delineated using visual image interpretation keys such as tone, texture, colour, pattern, shape, location and association. The geomorphological units were mapped for the coastal districts of Odisha as per NGLM (2010).

3.5 Decadal changes in the land cover along east coast of Odisha

For mapping of land cover, NDVI is used in the present study, which indicates plant “greenness” or photosynthetic activity, and is one of the most commonly used vegetation indices (Tucker 1979). Photosynthetically active vegetation, in particular, absorbs radiation in the red band (R) while reflecting much of it, in the near-infrared region (NIR). In this study, at-satellite spectral reflectance was used for computing NDVI according to the methods proposed by Markham and Baker (1986). Vegetation that is dead or stressed reflects more in R than in NIR regions. NDVI at pixel basis is computed as follows:

$${\text{NDVI}} = \left( {{\text{NIR}}{-}R} \right)/\left( {{\text{NIR}} + R} \right)$$
(1)

The NDVI varies between −1 to 1, while negative NDVI (values approximately −1) corresponds to deep water; values close to zero (−0.1 to 0.1) generally correspond to barren rock, sand, or snow. Low, positive values represent scrub and grassland (~0.2–0.4), while high values indicate temperate and tropical rainforests (values approaching 1). Thus, in the present study, land cover in terms of NDVI was generated for the years 1990, 1999 and 2009. The overall methodology of the present study is shown as a flowchart (Fig. 2).

Fig. 2
figure 2

Schematic representation of the methodology adopted in the present study

Full size image

4 Results and discussion

4.1 Delineation of decadal changes in the shoreline during 1990–2009

Decadal shorelines changes were delineated from satellite images for the years 1990, 1999 and 2009. These changes were analysed with reference to SOI toposheets at 1:50,000 scale. Significant changes were observed at few locations like mouth of Chilka Lake, (Fig. 3) and Mahanadi delta (Fig. 4) and Paradeep, Kendrapada district (Fig. 5). The notable changes are being the downward shift of mouth of Chilka Lake (Fig. 3) and formation of a new barrier island in the Mahanadi delta (Fig. 4) and at Kendrapada (Fig. 5). Dynamic behaviour of the mouth of the Chilka Lake can be attributed to constant wave and tidal action on a weak shoreline.

Fig. 3
figure 3

Shoreline change analysis between 1972 and 2009 and downward shift in Chilka mouth, Puri

Full size image
Fig. 4
figure 4

Shoreline change analysis between 1972 and 2009 in deltaic region of Mahanadi, Odisha

Full size image
Fig. 5
figure 5

Shoreline change analysis between 1972 and 2009 in deltaic region of Kendrapada, Odisha

Full size image

The change in shoreline during the entire study period, 1972–2009, is computed in terms of the area of land eroded or deposited along all the coastal districts of Odisha. From the study, it was observed that the northern part of Puri district had eroded in the range of 500–800 m, while erosion in the southern part was in the range of 50–200 m. Along the Ganjam coast, moderate erosion was observed at a rate of less than 2 m/year. Accretion in the range of 65–350 m was observed along the northern part of Jagatsinghpur district, while an erosion of 50–300 m was found along the southern part. Erosion of the range of 450–700 m was observed in the central part of Kendrapada district, and 190–300 m erosion was found towards the northern part. Mild erosion was noted almost along the coasts of Bhadrak and Balasore districts with an exception of the northern part of Balasore where minor deposition was observed.

East coast of India was more prone to tropical cyclones and induced storm surges, in general, and Odisha coast in particular. Hence, detailed analysis was carried out for the more probable cyclone landfall locations (Table 2), which are situated in the coastal districts of Odisha state namely Kendrapada, Balasore, Paradeep, Ganjam and Puri. The cyclone-induced storm surges and coastal wave pattern along the Odisha coast are termed to be moderate with wave height ranges between 1.25 and 1.40 m as stated by Kumar et al. (2010). Erosion and accretion areas were delineated for the periods 1990–1999, 1999–2009 and 1990–2009 whose results are presented in the Figs. 6, 7 and 8, respectively. From these figures, it is observed that shoreline along the Odisha coast was changing dynamically during the study period. From the analysis, it was observed that erosion and accretion processes were distinct near Rushikulya River, Ganjam. A new beach was formed due to accretion process near Devi River in Puri district and erosion near Chandbali, Balasore. The erosion and accretion of land along the coast were computed using temporally delineated shorelines (Table 3). Among the TC landfall sites considered in the present study, area under erosion (loss) was observed to be high at Paradeep (~53 %) and Puri (~41 %). The differences between erosion and accretion areas in the study region can be attributed to the frequency of cyclone landfalls that play an important role in changing the coastal geometry, which are interdependent (Johnny 2010). Apart from the impact of cyclones on the Odisha coast, sea level rise at a rate of more than 1 mm/year along the districts of Ganjam, Puri, Jagatsinghpur and Balasore could also influence the shoreline changes (Kumar et al. 2010).

Table 2 Cyclones along Odisha coast during the period 1964–2013
Full size table
Fig. 6
figure 6

Erosion and accretion during the period 1990–1999

Full size image
Fig. 7
figure 7

Erosion and accretion during the period 1999–2009

Full size image
Fig. 8
figure 8

Erosion and accretion during the period 1990–2009

Full size image
Table 3 Area gained/lost during the study period along the Odisha coast
Full size table

4.2 Geomorphology along Odisha coast

Geomorphological features along the 5 km buffer of Odisha coast were visually interpreted from the Landsat TM image of November, 2009 (Fig. 9). Beaches appear bright in the imagery, because the sand has high reflectance compared to the adjoining vegetated area. Geomorphological features were categorized under coastal landform (965 sq. km), deltaic landforms (600 sq. km) and offshore islands (4 sq. km) along the Odisha coast. Coastal geomorphological features were mapped for the coastal districts of Odisha, namely Balasore, Bhadrak, Kendrapada, Jagatsinghpur, Puri and Ganjam (Fig. 10a–f). Mahanadi delta constitutes the stabilized part of the study area. Geomorphologically, the study area was classified into beach, beach ridge, estuarine island, inter-tidal plain, older coastal plain, pediment plain, swale complex, lagoon, off shore bars, pediment, etc.

Fig. 9
figure 9

Coastal geomorphological units along Odisha coast within 5 km from the coast line

Full size image
Fig. 10
figure 10figure 10figure 10

a Coastal geomorphological units along Balasore district, Odisha. b Coastal geomorphological units along Bhadrak district, Odisha. c Coastal geomorphological units along Kendrapada district, Odisha. d Coastal geomorphological units along Jagatsinghpur district, Odisha. e Coastal geomorphological units along Puri district, Odisha. f Coastal geomorphological units along Ganjam district, Odisha

Full size image

Older coastal plain and swale complex features were more dominant in Balasore district covering an area of 193.6 and 121.4 sq. km, respectively (Fig. 10a). Figure 10b shows the geomorphology map of Bhadrak district, in which older coastal plain (146.95 sq. km), younger coastal plain (84.15 sq. km) and tidal flat (40.82 sq. km) features were more dominant. In Kendrapada district, mangrove swamp (162.34 sq. km), younger coastal plain (121.05) and tidal flat (102.48 sq. km) were found to be the dominant classes (Fig. 10c); along the Jagatsinghpur (Fig. 10d) the younger coastal plain (57.31 sq. km) and older coastal plain (45.71 sq. km) were major classes. Along the coast of Puri district, swale complex of an area of 284.71 sq. km is the most dominant class owing to the presence of the Chilka Lake in the district (Fig. 10e). The Ganjam coast is dominated by pediment with an area of 71.74 sq. km (Fig. 10f). From the geomorphological analysis within the 5 km buffer from the coast, it is observed that swale complex, older coastal plain and tidal flat constitute the major classes with a total area of 498, 460 and 350 sq. km, respectively. These categories are based on the origin of the landform and the processes influencing the units.

4.3 Land cover analysis

Land cover changes along the Odisha coast were assessed in terms of NDVI for 3 years 1990, 1999 and 2009. The spatial distribution of land cover classes within the 100-km buffer from the coast is shown as Fig. 11. Further, land cover changes during three time periods in terms of different NDVI categories are presented in Table 4. During the period 1990–1999, NDVI values observed to increase in the range of 0–0.5, while they are observed to decrease in the range of 0.5–0.9. Land cover analysis during the period 1999–2009 indicates that the NDVI values decreased in the range of 0–0.4 and increased in the range of 0.4–0.8. Generally, NDVI values in the range of 0–0.5 refer to agricultural areas; thereby, fluctuations in this range can be attributed to land use pattern. However, NDVI changes in the 0.5–1.0 range refer to highly vegetated areas like plantations, forests and mangroves. Hence, it can be stated that the land cover loss happened in the high vegetated areas during the period 1990–1999 and regeneration during the period 1999–2009.

Fig. 11
figure 11

Spatial distribution of land cover in the 100 km buffer from the coast line a 1990, b 1999, c 2009

Full size image
Table 4 Land cover changes in terms of NDVI along the Odisha coast during the period 1990–2009
Full size table

Land cover changes for coastal districts namely Balasore, Bhadrak, Kendrapada, Jagatsinghpur, Puri and Ganjam were also computed (Fig. 12). District-wise NDVI distribution has shown that the land cover during 1990 was predominantly less than 0.5 in all the coastal districts except Ganjam and Kendrapada, where NDVI is observed to be greater than 0.5 in 25 and 3.5 % area of these districts, respectively. The presence of agricultural plantations (~ 63 % of the area) in Ganjam (Ramesh et al. 2011) and mangrove forests in Kendrapada contributed towards higher NDVI values. The total mangrove forest area in Odisha, as per the Atlas of Mangrove Wetlands of India, published in 2004, was approximately 215 sq. km, out of which the Bhitarkanika mangroves in Kendrapada district consisted of 202 sq. km. (93.95 %). During the year 1999, the NDVI under 0.5 decreased across all the districts, and consequently, the increase in the NDVI classes greater than 0.5 was observed. Significant increase is observed in Ganjam district where 98 % of the district is shown to be greater than 0.5, which could be due to increased agricultural plantations. However, in the year 2009, NDVI in the districts of Balasore and Bhadrak is analogous to 1999 in both the classes. Remaining coastal districts have shown approximately 20 % increase in NDVI greater than 0.5, which may be attributed to the increase in the forest cover during the period 1995–2010 (Reddy et al. 2012). This indicates regeneration process after super-cyclone in the year 1999, which is further substantiated by the afforestation programmes of the forest department of Odisha state.

Fig. 12
figure 12

Land cover changes in the coastal districts of Odisha state during 1990–2009

Full size image

5 Conclusions

Periodical monitoring of the shoreline change is imperative for sustainable management of coastal zone and its resources. Geo-spatial technologies help to identify and map the change in the position of shoreline on various time scales. Different processes that include tides, winds, regular wave action, erosion–deposition, land subsidence, periodic tropical cyclones, storm surges and tsunamis are responsible for the shoreline change and geomorphology of the coastal region. In order to address these issues, a detailed study has been carried out comprising shoreline change, geomorphological and land cover changes, with the high-resolution satellite data spanning over two decades (1990–2009) along the Odisha coast of India.

Spatial and temporal image analysis of shoreline changes revealed notable changes at cyclone land fall locations along the Odisha coast. It is also possible that, due to cyclones, the sediments get eroded, transported and deposited from one place to another, thereby changing the coastal geometry. Present study reveals the southward shift of mouth of Chilka Lake and formation of a new barrier island in the Mahanadi delta. The shoreline change analysis has shown that the coastal districts of Puri, Kendrapada and Jagatsinghpur are showing moderate to high erosion, while accretion is observed along Ganjam, Bhadrak and Balasore. From the geomorphology of the study area, it is observed that geomorphological features along the Odisha coast are categorized under coastal landform (965 sq.km), deltaic landforms (600 sq.km) and offshore islands (4 sq. km). The geomorphic units present in the coastal districts were categorized based on the origin of the landform and the associated process governing the dominance of the specific geomorphic units. Such information is vital for the coastal management and planning for the given region. Further, long-term analysis of the land cover in terms of NDVI has indicated that the land cover loss happened in the high vegetated areas during the period 1990–1999 and regeneration during the period 1999–2009.