Geospatial Assessment of Long-Term Changes (1937–2019) in Mangrove Vegetation and Shoreline Dynamics of Godavari Estuary, East Coast of India

Godavari mangroves are the second largest mangrove ecoregion of India located on its eastern coast. Mangroves perform major ecological functions including shoreline stabilization and coastal protection. They have undergone serious alterations, largely induced by human activities resulting in increased vulnerability to coastal flooding. A reliable database on mangroves distribution and temporal changes is essential for their effective management and conservation. Present study analyzes the long-term changes in mangroves, as well as the geomorphological changes along the coast of Godavari Estuary over a span of 82 years (1937–2019), using topographical maps, remotely sensed multi-temporal satellite imageries for five time periods. Results indicated that the rate of land uses conversions of agricultural fields, degraded areas, mudflats and mangrove vegetation into aquaculture significantly increased during the late 1990s. About 40 percent of it was constructed on mangrove areas between 1980 and 2019. The area under aquaculture was estimated to be 6,388.62 ha in 2000, almost doubling-up to 11,924.67 ha in 2019. Mangrove cover in this region increased by 15.95 percent between 2000 and 2019 (including 11.9 percent in Coringa Wildlife Sanctuary) due to increased protection, consequent regeneration and accretion. Shift in Hope Island spit was observed in north and northwest directions, gradually growing nearly 3.84 km. High-resolution declassified satellite data were found very helpful for long-term mangrove mapping and analyzing geomorphological changes; enabling a better understanding of these regions to help decision makers/planners to find better solutions to existing coastal issues. The paper concludes with some recommendations to restore and conserve the existing mangrove cover in the study area.


Introduction
Delta ecosystems, which are regarded as the cradles of civilization in India, have a great economic significance in terms of agriculture, water resources, wetlands, wildlife habitats, fish production and tourism (Day et al., 1995;Rao et al., 2010). The Godavari delta, which is also called the rice bowl of Andhra Pradesh, is a major prograding delta located along the eastern coast of India, fringing the Bay of Bengal (Fig. 1). It is a densely populated zone of intense economic activity which houses 729 persons/km 2 (Dandekar & Thakkar, 2014), about twice the Indian average (382 persons/km 2 ). The vast delta is a complex of marine and coastal ecosystems including a riverine estuary, a large bay, mangrove forests, large parts of which are protected in Coringa Wildlife Sanctuary (CWLS), and multiple distributaries of the river crisscrossing these mangrove forests (Malla, 2020). The East Godavari River Estuarine Ecosystem (EGREE) enclosing Godavari mangroves is the second largest contiguous mangrove region along the eastern coast of India, after the famous Sunderbans in West Bengal (Nandy & Kushwaha, 2011;Shankar, 2021). Thus, the delta has a rich biodiversity and is home to numerous rare species, making it an important region for conservation in India (Malla, 2020).
These coastal and marine ecosystems play a vital role in supporting human welfare by virtue of its immense biological and mineral resources as well as the life-supporting systems they provide (GMR Group, 2020;UNDP, 2011). Economic installations and production operations along the coast, such as aqua farms, agriculture, marine fishing, gas and oil exploration, harbors and ports, and tourism, have been largely contributing to the regional economy. The economic value of various resources provided by a hectare of mangroves is between 2,00,000 to 9,00,000 USD approximately (Nguyen Dang Vu, 2006). According to a research, this service is worth 2,700 USD per hectare in the Godavari Delta, which extrapolates to almost 90,000 USD annually for the entire region (CORDIS, 2003;UNDP, 2011). Besides contributing to the economy for the East Godavari region, the mangrove landscape also plays a crucial role in livelihood development (GMR Group, 2020). It offers innumerable ecosystem services to a population of more than 79,400 people who reside in the 44 mangrove-abutting settlements. These villagers are directly dependent on the mangroves for their livelihoods as well as for firewood, timber, and fodder (Malla, 2020). About 40 percent of the population is actively involved in fishing (UNDP, 2011). The wetlands support the livelihoods of over 3,85,392 fishermen (3,03,000 marine and 82,392 inland) from almost 99 coastal fishermen villages in the area, out of which 76,777 are active fishermen (Fisheries Department, 2021). The number of fishermen families (44,476) in this district was the highest in Andhra Pradesh, in contrast to other districts (Maheswarudu et al., 2014). The five fish landing centers in and around CWLS alone recorded about 4,480 tonnes of fish catch during -2002(UNDP, 2011. In 2019, the East Godavari District recorded a catch of 84,080 tonnes of marine fish and 32,175 of inland fish with a total Gross Value Added (GVA) of INR 828 Crores (APEDB, 2019). For the purpose of manufacturing lime, about 3,600 tonnes of molluscs are annually removed from Coringa mudflats and Kakinada Bay. Bivalve species (Anadara granosa, Placuna placenta, Meretrix sp, Macoma sp) and gastropods (Cerithidea cingulata, Telescopium telescopium) are regularly handpicked (Department of Ocean Development, 2001). The bark of Ceriops decandra is used for dying the fishing nets and the Porteresia coarctata grass is used as fodder .
But being low-lying and fragile ecosystems, these densely populated zones are, of late, subjected to extensive land-use changes and unprecedented environmental degradation mainly driven by the expansion of the aquaculture industry and industrial development that are borne out of conflicting interests of various stakeholders inhabiting the deltas as well as the river catchments, leading to severe coastal erosion, habitat loss and a high risk of flooding, which would be exacerbated in lower areas if there is a rise in the global sea level (Sarma et al., 2001). Available literature (Rao, 2009a) suggest that the mangrove ecosystem on the eastern coast of India is among the most susceptible regional habitats to be exposed to sealevel rise.
Recently, the region has also lost several acres of mangrove vegetation in the proximity of fishing villages near Kakinada town, where the state government authorities have levelled a part of the land for a specific housing scheme (Reddy, 2020). According to the new CRZ notification 2019 (https://ncscm.res.in/pdf_docs/crz-2019.pdf), exemplary power has been given to the state authorities to make infrastructure interventions, also permitting exploratory drilling of oil and natural gas and all associated activities in CRZ-I zone (Ecologically Sensitive Areas). Nearly 100 acres of mangrove land have been levelled so far for this housing scheme, with mangroves being chopped down and salt creeks covered up in mud largely affecting the livelihoods of the fisherfolk (Puzhakkal, 2020;Reddy, 2020).
A study had estimated that more than 8.7 million people and 28,000 km 2 of deltaic area across 33 deltas studied (including Godavari) could suffer from enhanced inundation and increased coastal erosion by 2050, if no mitigation measures are undertaken for the conservation of these ecosystems (Ericson et al., 2006). In this background, it is essential to understand the magnitude of, and the factors responsible for, the degradation of this important coastal ecosystem in the country and its environmental implications on such areas. The present study is such an attempt in this direction which also analyses the potential of using high-resolution declassified CORONA and KeyHole (KH-9) satellite imagery for long-term changes in mangrove vegetation and shoreline change dynamics in combination.
CORONA had been a satellite reconnaissance program providing satellite imageries, originally used by the United States of America during 1959-1972 for surveillance purposes, mainly of the Soviet Union (USSR) and Chinese nuclear and strategic missile programs (Day et al., 1998;Richelson, 2003). KeyHole (KH-9) was the program that followed CORONA and it ended in the 1980s, which was designed to support mapping requirements and exact positioning of geographical points for the military (USGS, 2008). Since this data ranges back to the year 1960s to 1980s, they are often resembling the sole known existing photographs of human footprints and settlements of that time (especially CORONA). This was before digital technology was invented that is currently used in all the polar orbiting remote sensing satellites (Rao, 2009b). Furthermore, since many formerly natural areas have overgrown by urbanization, this data offer unique insights into places that may be vulnerable to natural disasters, especially areas where settlements have now grown into former riverbeds (Fekete, 2020). Coastal dynamics and spatial change in the mangrove cover of a region can be investigated using such satellite imagery. The need to delineate the mangrove cover and shoreline positions and assess its variation through space and time is particularly crucial, given the long history of human settlements and human-induced activities within or near the coastal zones (Burningham & Fernandez-Nunez, 2020). Thus, the study focuses on a multi-temporal approach where shoreline change dynamics and mangrove changes along the coastal areas of the Godavari Estuary has been assessed using different remote sensing datasets of over a span of 82 years. Factors influencing the shoreline equilibrium and its impacts on coastal urban areas have also been investigated.

Study Area
The Godavari mangroves are located between latitudes 16°30 0 and 17°N and longitudes 82°10 0 and 82°25 0 E in the East Godavari District of Andhra Pradesh, India. The Godavari River emerges in the eastern slopes of the Western Ghats (close to the west coast of India) and flows eastward across the Indian peninsula for 1,465 km, before being decanted into the Bay of Bengal on the eastern coast of India (Praveenkumar & Jothiprakash, 2021). The second largest river basin (about 3,12,812 km 2 ) in peninsular India after Ganga accounts for about one-tenth of the country's total geographical area, out of which the delta takes up nearly 5,100 km 2 (GRMB, 2016). In the pre-dam scenario, the Godavari River ranked ninth among the largest sediment transporting rivers in the world (Malini & Rao, 2004).
The lower Godavari River bifurcates into two distributaries -Gautami and Vasishta. Vainateyam, the third distributary branches out of the Vasishta farther downstream, while Gautami's terminal branching forms the fourth distributary, Nilarevu. Both these distributaries enter the sea through two mouths, one at Bhairavapalem in the north and the other near Kottapalem at the southern limit of the mangrove-bay complex (Fig. 2). The older tributaries of Gautami have been retained as Corangi and Gaderu canals that are about 11 and 21 km long, respectively (Rao, 2009a). In July 1978, the state government declared a portion of the Godavari mangroves as CWLS under Sect. 18 of the Wildlife Protection Act, 1972, extending in an area of about 235.7 km 2 in order to rehabilitate saltwater crocodiles that were on the verge of extinction, to protect other endangered species such as Olive Ridley turtles and Indian Otters, and to conserve the mangrove vegetation of The region has a semi-arid climate, with temperatures ranging from 22°C in January to more than 35°C in April and receives an average rainfall of about 1100 mm annually (Selvam, 2003). The region is also prone to severe cyclonic activities accompanied by stronger waves and occasional storm surges which often reach several kilometers inland in this extremely low-lying region sloping seaward at an average gradient of just about 0.011 percent (Rao et al., 2010). Tidal data from Kakinada at the northeastern part of the Godavari delta indicate that the annual mean tidal range is * 1.0 m and that the mean annual high tide is 1.38 m and the low is 0.36 m with the highest average monthly tides occurring in October and November and the lowest in February and March (Rao et al., 2013).

Remote Sensing Datasets
The present study was carried out using several sets of remotely-sensed, multi-temporal satellite images and topographic maps, spread over the past eight decades. Topographical map (1:250,000 scale) prepared by U.S. Army Map Service from 1937 (T 1 ), high-resolution declassified CORONA and KeyHole-9 images from 1965 (T 2 ) and 1980 (T 3 ), respectively, Landsat 7 TM images from 2000 (T 4 ) and Resourcesat-2A LISS IV MX images from 2019 (T 5 ) covering the intended region were digitally processed. Statistics for T 1 data should be deemed only as rough approximations due to its arguable mapping accuracy. Image to image rectifications for the 1980 data was carried out using Landsat 2 MSS from 1977 because of cloud cover, with an overlap error less than 1 pixel. The geometric projection for both images was set to UTM (Standard LANDSAT projection for the Indian region). The details of the data used in the present study have been shown in Table 1.
The topographical map and the satellite images for T 2 , T 3, and T 4 were downloaded from the University of Texas site (http://www.lib.utexas.edu/maps/ams/india/#) and the United States Geological Survey (https://earthexplorer. usgs.gov/), respectively. Satellite data for T 5 were procured from NRSC Data Centre (NDC), National Remote Sensing Centre (NRSC), Hyderabad. The CWLS boundary was extracted from the shape file of protected areas of India available from the Wildlife Institute of India (WII), Dehradun.

Data Processing and Base Preparation
Three CORONA strips with dimensions 3 9 30 inches each (* 3 m resolution) were provided in four split components (a, b, c, and d), whereas one strip of Keyhole image (* 4 m resolution) with dimensions 9 9 18 inches was provided in two image components (a and b). They all were processed in Image Composite Editor to stitch the components of each image together, which faced errors related to projection type as ends of the strip were more skewed due to the panoramic effect. Thus, over-warped edges in the overlap area were also trimmed before georeferencing to achieve good matching. For this study, the first two components were stitched for each CORONA strip.
Following the completion of this process for all the images, KeyHole (1980) and CORONA (1965) images were georeferenced with Geographic Coordinate System World Geodetic System (GCS WGS) 1984 datum in QGIS 10.0, where Bing Map was effectively utilized as the reference image to provide more than 200 accurate Ground Control Points (GCP). Since the study present study employed only one strip of KeyHole imagery, no mosaicking was required. However, in order to obtain a uniform spectral intensity within and across all the COR-ONA strips, a number of color correction techniques were used iteratively in Erdas Imagine Software 2014. The detailed methodology for the registration of declassified imageries is given in Fig. 3.
Landsat 5 TM data for 2000 and Resourcesat-2A LISS IV MX data with 70 km swath for 2019 were in their original 30 m and 5 m resolution, respectively. Landsat 5 TM images were pan-sharpened where a high-resolution panchromatic band (B4) was merged with the lower resolution multispectral bands to create a single high-resolution (15 m) color image for better visual interpretation. These datasets were then subjected to a series of pre-processing techniques to create a seamless mosaic using the Erdas Imagine 2014 image processing software. Both the images were displayed in False Color Composite (FCC) using the band combination of 4, 3, and 2.

Classification and Shoreline Analysis
A multi-temporal image comparison was conducted by manual visual image interpretation through which a tenclass suitable, site-specific classification scheme was developed (Table 2), which has been adopted from the coastal feature classification proposed by Nayak et al. (1991).
Satellite images of all the years were subjected to multiresolution segmentation using eCognition Developer software. It is a method which allows an image to be segmented into a network of homogeneous image regions at any required resolution (Tiwari et al., 2009). These image objects are abstract representations of image information that serve as building blocks for ensuing classification. It was tested out using various combinations of compactness and shape parameters at different resolutions. To obtain the image objects for all the time periods, multilevel resolution segmentation was carried out at different scales for each image since they had varying resolutions, with a constant compactness of 0.3 and shape factor of 0.7. All the features were obtained in separate image objects, which could then be further refined and classified in the ArcGIS environment.
The land-water interface is where the shoreline change analysis was done, leaving aside creeks, stream mouths and its inner parts and rivers. The classified maps of each year were used for delineating the shoreline by merging the polygons of all features except water. The scale selected for the wetland categorization was 1:250,000 as suggested by Nayak et al. (1991) for a better understanding of the areas, their inter-comparison and change detection. The overall configuration of the delta-front shoreline in the respective time periods was compared using overlay analysis in ArcGIS from which the geomorphological changes between successive datasets have been interpreted. Furthermore, a field survey was also carried out and geotagged field photographs were collected for use in the classification procedure (Fig. 1).

Change Detection
After the images of the five time periods (T 1 , T 2 , T 3, T 4 and T 5 ) were classified independently, the post classification change detection method was applied in ArcGIS 10.8, where they were compared to assess the quantitative and qualitative aspects of the changes in the spatial extent of mangroves and other land cover types. From the change map, the change area matrix table was generated which provided ''from-to'' information.

Accuracy Assessment
The quality of the classified images was tested separately for T 2 , T 3, T 4 and T 5 through a rigorous accuracy assessment technique carried out in ArcGIS 10.8. Ground verification of 110-115 points was done for the land use and land cover classes of each time period and Bing Map was used as a high-resolution reference image. After this, the confusion matrix approach was adopted to calculate the overall accuracy and kappa statistics.

Results and Discussion
The long-term changes in the mangrove vegetation in the study area can be attributed to various factors such as natural regeneration and restoration, accretion and erosion, extensive land use changes particularly shrimp farms and aquaculture, which has been discussed in the following sections in detail.

Mangrove Change due to Natural Regeneration and Restoration
The interpretation of the satellite images reveal that the mangrove vegetation is concentrated more on the northern side (CWLS) of the Godavari River than on the southern Red NIR side (Non-Sanctuary Area) as can be seen in Fig. 4. In 2019, the CWLS constituted almost 60 percent of the total mangrove cover present in the study area. This is because of the prolonged tidal flushing of mangroves in the sanctuary area which takes place through the creeks of Matlapalem kalava (canal), Corangi and Gaderu canals, bringing in dissolved nutrients like phosphates and nitrates to the mangroves (Ramasubramanian et al., 2006;Ravishankar et al., 2004). The canal system is highly intricate, and the flow nature is complex due to stratification that occurs throughout the year. Seawater enters from the north at Kakinada Bay and the south at Bhairavapalem. Long and widely branching mangrove creeks, where significant lateral water trapping occurs, provide an excellent habitat and nursery for a variety of fish (Carangidae, Sciaenidae, Clupeidae and Leiognathidae) and invertebrate species (P. monodon, Metapenaeus affinis, Penaeus indicus and M. monoceros) many of which are commercially important (Satyanarayana et al., 2002). The most dominant mangrove species found in the region are Rhizophora mucronata, Rhizophora conjugata, Acanthus ilicifolius, Bruguera gymnorrhiza, Suaeda nudiflora, Lumnitzera racemosa, Excoecaria agallocha and Excoecaria marina (Satapathy et al., 2007). This region is also protected from high energy water currents and velocity winds by the Hope Island sand spit which is an island located off the coast of Kakinada that acts as wave breakers (Bagaria et al., 2021a;Ramasubramanian et al., 2006). It is noted for its natural breakwater spread over 1,000 ha that has protected the coast during cyclones (Shankar, 2021). During 1933 to the early 1970s, the Andhra Pradesh Forest Department (APFD) took over the management of   mangroves and systematically felled mangrove trees . This unscientific management practice resulted in the development of hypersaline condition, which was one of the primary causes for the degradation of Godavari mangroves Selvam et al., 2004). Apart from this, other major causes included anthropogenic factors such as grazing by cattle, and collection of mangroves for firewood and fencing materials by nearby villages. Therefore, it can be examined from Table 3 that the mangrove cover was reduced from 10,639.25 ha to 9,862.06 ha in CWLS during 1937CWLS during -1965 In the 1980s, the degraded areas might have been left undisturbed and recovered/healed on its own due to local conditions as in the 1980 classified image, the restoration of mangrove vegetation in the degraded area is clearly noticed. Table 3 shows that the degraded area reduced from 1,519.91 ha in 1965 to 1,048.86 ha in 1980 in the sanctuary area after the CWLS was established with an overall reduction of 30.5 percent (1,356.60 ha) in the whole study area during that period. Suaeda maritima and S. mudiflora are commonly found in degraded and partially degraded areas .
The change analysis of 2000 and 2019 indicates a significant reduction of degraded areas (1,075.98 ha), where the mangrove cover has also shown a positive increment of 1,123.03 ha in the CWLS. This shows the efficacy of various government interventions and restoration strategies in the area. Few studies (Bagaria et al., 2021b;Reddy & Roy, 2008) observed that the majority of positive changes in CWLS during this period was due to the active involvement of the government that has made efforts towards protecting the natural areas and undertaken various plantation and afforestation initiatives.
The APFD initiated various mangrove restoration programs in the Godavari Estuary in 1991 by digging canals in the degraded areas of CWLS . The M. S. Swaminathan Research Foundation, a Chennai based non-government organization (NGO), had initiated reforestation projects in the area during 1997 to 2004 funded by the India-Canada Environment Facility (Rao, 2009a). Around 165 ha of degraded mangroves had been restored in five demonstration villages in the delta region during this period . As part of the shelterbelt program, the state forest department also planted Casuarina equisetifolia along the coast Reddy & Roy, 2008), as a result, the area under coastal vegetation increased significantly from 763.41 ha in 1980 to 1,449.25 ha in 2000 (Table 4).

Mangrove Change due to Accretion and Erosion
Between 1980-2000, the results indicate pronounced erosion (1,812.98 ha), more than accretion (1,380.79 ha), along the delta front shoreline and estuarine banks of Godavari, putting the mangrove forests at higher risk. This is due to the fact that this area had been barricaded by several barrages and dam constructions which had picked up momentum during that time period in the Godavari River basin starting with the magnificent Sriramsagar in 1972, which was the first large-scale dam across the Godavari River followed by Jayakwadi in 1976, Balimela in 1977, Upper Indravati in 1994, Upper Wainganga in 1995 and many others built across the Godavari River and its multiple tributaries (Rao et al., 2010). Godavari delta has been classified as a delta that is receding faster than sea-level rise (Syvitski et al., 2009). According to Gupta et al. (2012), the drop in historic sediments of Godavari post damming was as high as 74 percent.
Though between 2000 and 2019, the mangrove forest increased by 15.95 percent (2,416.09 ha) as a result of increased protection, consequent regeneration and accretion. During this period, the extent of mangroves increased by 1,583.03 hectares owing to newly accreted areas ( Table 5). The accretion was most likely caused by alluvium sedimentation from the Godavari River's flow through the Corangi and Gaderu canals. Until now, this is one of the very few Indian studies that has reported an increase in the extent of mangroves since 2000 (Singh et al., 2004). Erosion also reduced the mangroves by 1,028.21 ha during the same period. Floods during the southwest monsoons have been gradually eroding the mangrove vegetation along the southern bank of the Nilarevu River (Ramasubramanian et al., 2006). The small mangrove islands near the Bhairavapalem mouth seen in the 2000 image have also been eroded in the 2019 image. In 2000, the Hope Island sand spit had grown drastically and now, the northern part of it bears plantations and new mangrove formations.

Mangrove Change due to Aquaculture Farms and other Land Uses
Aquaculture and shrimp farming are the second major activity spread in the area after agriculture. The classified data reveal that aquaculture in the region covered a very scarce area of 58.43 ha and 68.01 ha during 1965 and 1980s, respectively. The map of 2000 and 2019 indicates that the major changes have been taking place in the vicinity of the mangrove-abutting areas as a result of increased anthropogenic pressure. These shrimp farms are mostly located in the private lands of the rural settlements fringing the mangroves. It can be seen from the analysis, that the rate of land use conversions of agricultural fields, degraded areas, mudflats and mangrove vegetation into shrimp farms during the late 1990s has increased substantially. The overall area of aquaculture was 6,388.62 ha in 2000 that almost doubled to 11,924.67 ha in 2019 (Table 4). Analysis of the satellite data of 2000 and 2019 also shows that a total area of 1,227.11 ha of mangrove vegetation has been converted into aquaculture farms in the last four decades. About 38 percent (4,519.44 ha) of the aquaculture farms have been built on mangrove areas in over a span of 40 years out of which 277.58 ha is in the sanctuary area. The mangrove vegetation underwent a negative change of 3,437.50 ha mainly due to this reason during the same period. Satapathy et al. (2007) suggested that this has contributed to the production of 12,000 tonnes of shrimps from converted mangrove wetlands annually. Now catfish, carp, and a few species of shrimps as well as finned fish are commonly cultured in the area. The tiger  1937-1965 1965-1980 1980-2000 2000-2019 Restoration prawn spawner (Penaeus monodon) is the main species cultivated in the region (Muralidhar et al., 2013). According to a study, the mangroves in this area have the potential to support an estimated catch of around 50,000 tiger prawn spawners annually, with a net worth of 6 million USD (Rönnbäck et al., 2003). It is also noticeable that the rural settlements and urban areas have increased significantly during the years from 1, 824.42 ha and 5,248.15 ha in 1965 to 4,508.87 ha and 7,948.48 ha in 2019, respectively. Analysis shows that 24.4 ha of mangrove cover, 207.2 ha of mudflats and 93.6 ha of degraded areas have been encroached upon by agriculture and built-up area during 2000-2019. It is evident that the mangrove vegetation cover can vary greatly based on the area's dynamics, natural events, and development activities (Reddy & Roy, 2008).
Research confirmed that shrimp farming and aquaculture have been an appealing and profitable choice for many farmers that had been begun in 1988 in the eastern coast of India who had to regularly deal with crops being flooded, intrusion of salt into irrigation water, and Bay of Bengal cyclones (Jayanthi et al., 2010). Despite this, a recent study done in 2019 by Animal Equality indicated that fertile lands across the Godavari delta region that are being furiously converted into aquaculture for shrimp and fish farming have shown an increasing salination in the quality of groundwater. Their putrefaction outside the ponds has also caused turbidity of water, foul odors and siltation with detrimental effects on local flora and fauna. As a result, it has raised several adverse effects in the region both in the physical environment and socio-economic environment (Satapathy et al., 2007). Gupta et al. (2001) observed that conversion of agricultural land to aquaculture has longterm environmental impacts such as permanent loss of agricultural land, overall groundwater contamination, etc. Thus, local aquaculture farmers must be educated about the detrimental consequences of unregulated and unmonitored aqua-farming. The government should provide special grants/incentives to projects that focus on expanding and improving the way the fish is farmed sustainably.

Geomorphology and Shoreline Changes
The geomorphic process of erosion and sedimentation, flooding and sea-level changes continuously modify the shoreline. The accurate demarcation of shoreline is very important for planning conservation measures (Nayak et al., 1991). Rao et al. (2003) found CORONA photographs to be very accurate for mapping of geomorphological changes in the Godavari estuary region which could not have been interpreted through modern day remote sensing imageries as these coastal linear features representing the former shoreline positions have been mostly changed due to recent intensive land use conversions. Analysis suggests that the high erosion trend of the coast has transformed to less and stable condition by 2019 as shown in Fig. 5. This observation can be supported by a similar study conducted by Murali et al. (2020) in the same region. Significant changes are noticeable in the Godavari shoreline configuration due to deposition (D) at the sand spits on the southern side of the Bhairavapalem mouth and erosion (E) at the coastline along the Bay of Bengal on the northern side of the Nilarevu River, i.e., from the river mouth to the tip of Hope Island as well as at the barrier islands (Fig. 6). Deposition of sediments on the southern bank of the River Gautami has been recorded by a study (Ramkumar, 2003), which has resulted in a gradual shift of the Gautami estuary since past eight decades. This observation is backed up by the results of the present study where the Gautami estuary has moved to the west from 1965-2019 (Fig. 6). Enlargement and elongation of Hope Island in the north and northwest directions is also seen over the years. A recent study (Bagaria et al., 2017) found that there was a disconnection between the estuary lobe and  (Table 6). According to Rajawat et al. (2015), due to dam construction this region saw significant coastal erosion brought on by changes in the equilibrium of sediment transport. The results from the shoreline analysis are a major contribution towards the proper understanding of shoreline dynamics across the Godavari Estuary.
During the course of 82 years, the river span of Nilarevu has also widened and that of Gautami has narrowed indicating a shift in the direction of river discharge. Rao et al. (2005) also observed such instances of increased erosion in the delta because of the widening of one of the draining channels. The river course has changed much near Bhairavapalem mouth, which was earlier towards the east in 1965, has eventually shifted towards the northeast in 2019. In 1965, the river span was about 0.6 km while in 2019 it has increased to 3.12 km owing to erosion (Table 6).

Classification Accuracy
The overall classification accuracies for T 2 , T 3, T 4 and T 5 came out to be 93.91, 93.85, 92.85 and 95.49 percent, with kappa statistics of 0.93, 0.93, 0.92, and 0.94, respectively (Table 7). Accuracy assessments for each time period are shown in Table 8. The minimum level for accuracy assessment in the identification of wetland categories in remotely sensed data should be at least 85 percent (Nayak et al., 1991), indicating that the analysis produced very good results. The best results were of sand, followed by mangroves. While cropland and urban area showed relatively lower accuracies. Overall, the land cover classes could be distinguished well.

Conclusions and Recommendations
The present study attempts to understand the long-term changes in the mangrove vegetation and shoreline dynamics that have occurred along the eastern coast of the Godavari region and evaluating the cause of the same. Time series data for five time periods during a span of 82 years (1937-2019) reveal that the period from 1980 to 2000 was the most significant with reference to mangrove cover changes, as there was a drastic reduction in mangrove vegetation by 2000, which was estimated to be 3,437.50 ha. Overexploitation, conversion of mangrove into other land uses especially aquaculture and coastal erosion were found to be the three major threats in the region. Despite many causes of mangrove cover depletion, it has been found that there was an overall increase of 2,417.41 ha in the mangrove cover during the period 2000-2019 as compared to 1980-2000. The CWLS area had a positive increment of 1,123.03 ha in the mangrove  vegetation during 2000-2019 which indicates the effectiveness of various government interventions for conservation in the area. Accretion and restoration have played a significant role in the formation and improvement of the mangrove cover in the area during the last eight decades.
The study also revealed that the Godavari front shoreline has shifted significantly during the past eight decades. CORONA and KeyHole photoreconnaissance satellite imagery has been found very useful for this study. There is a huge potential of using these easily available, high-resolution declassified data for long-term change-detection studies. This data can also help in delineating the various geomorphic changes in different time periods and also provide inputs to study the coastal erosion and accretion. Therefore, it is suggested that improved tools of remote sensing (high-resolution data) and GIS can enable a better understanding of coastal areas, so that a sound strategy can be devised for the conservation and management of such ecosystems.
The present research pertaining to the study of shoreline changes and mangrove mapping will help the decision makers and planning practitioners during the formulation of Coastal Regulation Management and Conservation Plans in identifying regions prone to mangrove degradation. This data is expected to also support the local authorities in adopting multi-disciplinary measurement strategies for the protection of eroding stretches of the mangrove shoreline as well as identifying the load of human activities on the sediment stability within the coastal zone.
To ensure proper maintenance of mangrove ecosystems, it is imperative to assess, sustain and further improve the extremely precious mangroves of the Godavari Region.

Joint Mangrove Management
It is recommended to engage various sectors and stakeholders involved in coastal areas through joint mangrove management and place development needs in a wider framework of planning and management. This will help address the underlying causes of the vulnerability of mangrove ecosystems, such as ensuring sustainable use of adjacent land, marine and river ecosystems to avoid impacts to soil surface, freshwater, nutrient and sediment supplies. Local fishermen communities can be stewards for mangrove management, and their support is likely to boost a project's chances of success. Such communities can volunteer in projects for restoration, regulate the use and access of mangroves, and directly benefit from sustainable management as well.

Local Awareness
Improving local awareness about the impacts of human activities in coastal areas is very important for the protection of mangrove ecosystems. The East Godavari River Estuarine Ecosystem (EGREE) Foundation in partnership with various Village Level Institutions (VLIs) and Non-Government Organizations (NGOs) have the authorization to execute such awareness among the stakeholders at various levels in the Godavari Region.

Alternative Sustainable Livelihood Models
Promoting sustainable livelihood models that do not rely on the consumption or pollution of mangrove resources could be another way to reduce the pressure on mangroves. Training and skill acquisition programmes on alternative

Alternative Energy Sources
Encouraging the use of alternative energy sources to meet the growing demand without cutting mangroves for fuel or engaging in oil and gas exploration activities that damage mangrove ecosystems. The abundance of palm and casuarina trees in the region can be used as an alternative for mangrove wood for fuelwood and construction purposes.

Sustainable Aquaculture
Since, aquaculture is one of the main drivers of mangrove loss in the area, it needs to be developed in a more sustainable way. The local fishermen communities need to be sensitized about the negative impacts of unbridled, unmonitored aqua-farming. Standards for sustainable aquaculture take mangroves into account, prohibiting siting and encroachment in mangrove areas and requiring mangrove restoration and rehabilitation.

Integrating Mangrove Conservation in Sectoral Legal Frameworks
Provisions of sectoral laws such as forest management regimes, development permitting regulations and environment laws should be taken advantage of, in order to promote the conservation of mangrove ecosystems. It should be ensured that mangroves are included into emergency and natural disaster frameworks and that pollution rules such as Andhra Pradesh Pollution Control Norms take into account the particular impact of activities on mangroves.

Allocation of Financial Resources
Appropriate funds should be allocated to institutions that are involved in mangrove conservation and management for equipment, human resources and technical tools to undertake, inter alia, inventories, economic valuation, capacity building, public engagement, monitoring and management activities, recognizing the high value of mangrove ecosystems and the importance of their prioritization.
guidance and encouragement. We thank the staff of RRSC-S for providing overall support during the course of our work. We also thank Kaladi Yaspal Varma, M. Tech (Remote Sensing), JNTUK Institute of Science and Technology, for supporting the geotagged field information collection for the study area.
Author Contributions Both the authors contributed to the conceptualization and design of the study. Material preparation, data collection and analysis were performed by SDR, with technical guidance from ST. The first draft of the manuscript was written by SDR and ST reviewed and commented on previous versions of the manuscript. Both the authors read and approved the final version of the manuscript.
Funding No funding was received for conducting this study.

Data Availability
The datasets generated during the present study are available on reasonable request.

Declarations
Conflict of interest The authors declare no conflicts of interest.
Consent for publication Authors have read and agreed to the published version of the manuscript.
Consent to participate Not Applicable.
Ethics Approval Not Applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.