Living by the Sea: Range and Habitat Distribution of Asiatic Lions


Study zone

Located in the southwestern part of western India in the state of Gujarat, the region of Saurashtra generally represents the semi-arid province of Gujarat-Rajputana 4B23, which covers 11 of the state’s 33 districts. The region forms a rocky plateau (elevation 300-600 m) bordered by coastal plains with an undulating central plain broken by hills and dissected by various rivers flowing in all directions24. With the longest coastline (~1600 km) in India, Gujarat is blessed with rich coastal biodiversity25.26. The coast of Saurashtra in Gujarat is encircled by the open sea between two gulfs (68°58′–71°30′ N and 22°15′–20°50′ E) and divided into two segments, namely. the southwest coast from Dwarka to Diu (~300 km stretch) and the southeast coast from Diu to Bhavnagar (~250 km stretch)26.

The Asiatic lion landscape covers an area of ​​approximately 30,000 km2 (permanent lion range: ~16,000 km2; record driving range: ~ 14,000 km2) of various habitat types within Saurashtra. The landscape includes five protected areas (Gir National Park, Gir Wildlife Sanctuary, Paniya Wildlife Sanctuary, Mitiyala Wildlife Sanctuary and Girnar Wildlife Sanctuary) and other forest classes (Reserved Forests, Protected Forests and Forests unclassified).

The coastal habitats extend over the districts of Bhavnagar, Amreli, Gir-Somnath and Junagadh (Fig. 1). Within these districts (Fig. 1), the tehsils (sub-divisions/taluka) of Mangrol, Malia, Patan-Veraval, Sutrapada, Kodinar and Una are categorized under the South West Coast (hereafter Coastal Habitat western), Jafrabad, Rajula, forms the south-eastern coast and Mahuva and Talaja constitute the Bhavnagar coast and represent separate units of lion range (Fig. 1). The total area covered by the study is 2843 km2 on the eastern coast and 1413 km2 on the west coast (Fig. 1).

The Saurashtra region is endowed with three distinct seasons, namely. dry and hot summer (March-June), monsoon (July-October) and mainly dry winter (November-February). It receives an average annual rainfall of around 600 mm, with most precipitation during the southwest monsoon27. Average maximum and minimum temperatures are 34°C and 19°C respectively28. 110 km ago2 expanse of forests along the coast. The rest of the zones are multi-use composed of private, industrial, pastoral wasteland and various properties. The natural vegetation consists mainly of Prosopis juliflora and Casuarina equistsetifolia. On the beach and dune areas, vegetation such as Ipomea pescaprae, Sporobolus Trinules, fimrystylis sp., Rattlesnake sp.and Spurge nivuleria29. Mudflats along the coast are limited to Talaja, Mahuva, Pipavav Port, Jafrabad Creek and Porbandar, which are sparsely covered by the Avicenna marina29. Fishing, agriculture, horticulture, animal husbandry and some large and small scale industries are the main economies of the coastal belt.

The coastal segments are characterized by the variety of vegetation, sandy beaches, small cliffs, wave-cut platforms, open and submerged dunes, minor estuaries, embankments and the transition from open sea to gulf environment with tidal mud.26.29 and also support a diverse assemblage of biodiversity25. This biodiversity is further enriched by several perennial/ephemeral rivers originating from the Gir AP (Shetrunji, Machundari, Raval, Ardak, Bhuvatirth, Shinghoda, Hiran, Saraswati, etc.)12. These rivers meet the sea at different sections of the coast, forming important coastal ecosystems25. River stretches act as important corridors for wildlife movement9,12,30. By dispersing through these corridors, lions began to inhabit these coastal habitats30.31.


All research activities involved in this study on Asiatic lions were conducted after obtaining permission from the Ministry of Environment, Forestry and Climate Change (MoEF&CC), Government of India (Letter No: F No. 1-50/2018 WL) and Senior Chief Conservator of Forests (Wildlife) and Chief Wildlife Warden, State of Gujarat, Gandhinagar (Letter No: WLP 26B 781-83/2019-20). Procedures and protocols were followed in accordance with Standard Operating Procedures of Gujarat Forestry Department, Government of Gujarat, regarding the handling of wild animals. Qualified and experienced veterinarians and their team performed all procedures related to radio collaring. Additionally, the study is reported in accordance with the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines where applicable.

A long-term lion monitoring project was initiated in 2019 by the Gujarat Forest Department to understand the movement patterns and ecology of lions in the Asian lion landscape. In view of the heterogeneity and vastness of coastal areas, ten individuals were carefully selected for satellite radio collaring based on their frequent movements in different coastal habitats and monitored from 2019 to 2021.

The lions were deployed with Vertex Plus GPS collars (Vectronics Aerospace GmbH, Berlin, Germany) that weighed less than three percent of the individual’s body weight, regardless of age and sex. Lions were immobilized using a combination of ketamine hydrochloride (2.2 mg per kg body weight; ketamine, Biowet, Pulawy) and xylazine hydrochloride (1.1 mg per kg body weight; Xylaxil, Brilliant Bio Pharma Pvt. Ltd., Telangana)32 administered intramuscularly using a Telinject™ GUT 50 gas dart delivery system (Telinject Inc., Dudenhofen, Germany). A blindfold has been placed to protect the eyes and reduce visual stimuli33.34. Each sedated individual was sexed, aged, and measured according to standard operating procedure (SOP) of Gujarat Forestry Department, Government of Gujarat, and recorded the data in the trapping data sheet. The radio collars were deployed taking into account the circumference of the individual’s neck, ensuring freedom of movement of the individual so as not to interfere with the individual’s routine activities. After deploying the radio collar, we used the specific antidote for xylazine, i.e. yohimbine hydrochloride (0.1 to 0.15 mg per kg body weight; Yohimbe, Equimed, USA Unis) intravenously, which allowed the total recovery of the immobilized individuals.32 within 5 to 10 minutes. Individuals were intensively monitored for 72 h and then monitored regularly throughout the period of radio collar operation. The entire radio-collaring exercise was carried out by trained and experienced veterinary officers and their teams which comprised wildlife health and field staff.

Each collar had a unique VHF and UHF frequency. The radio collars were equipped with a programmable GPS program and configured to record hourly location cues and provided the data via the Iridium Satellite Data Service Constellation in Low Earth Orbit (Iridium Communications Inc., Virginia, USA United) at four o’clock. – hour intervals after being activated. Data logs included location marks in decimal format in degrees (latitude/longitude), speed (km/hour), altitude (meters above mean sea level), UTC timestamp (dd -mm-yyyy h:m:s), direction (degrees) and temperature (Celsius). The radio collars were equipped with mortality sensors and a programmable drop-off activation system. Gir Hi-Tech Monitoring Unit, Sasan-Gir, Gujarat monitored and coordinated these activities. Location data for each radio collar was uploaded using GPS Plus X software (Vectronics Aerospace GmbH, Berlin, Germany) into the Gir Hi-Tech monitoring unit (a technology-driven scientific monitoring initiative in landscape established in 2019 in Sasan-Gir, Gujarat).

Data analysis

In this study, we calculated the home range of lions residing in the coastal region using the fixed kernel method. We expressed them as 90% and 50% fixed core (FK) to summarize the overall home range and core area, respectively.35,36,37. In addition, the home range of lions classified as ‘linking lions’ and protected area lions were summarized for comparison (Table 1).

MaxEnt standalone software (version 3.4.1)38 has been applied for fine-scale modeling of lion distribution39.40. Logistic output format has been set for MaxEnt output. 30% random lion occurrence points were used as test data to assess model performance. The area under the receiver operating characteristic curve (AUC) was used to assess the discriminative ability of the model as a function of the values ​​of sensitivity (correct discrimination of true positive localization points) and specificity (correct discrimination of true positive localization points). no true negatives)41. The Jackknife regularized training gain for the species was used to understand the effect of each variable in building the model. The logic output of the MaxEnt was presented in tabular form as “percentage contribution” and “swing magnitude” values ​​(from 0 to 100%). Spatial inputs were prepared on the GIS platform using ArcMap (version 10.8.1, ESRI, Redlands, USA)42. Input data for MaxEnt was classified as (i) lion occurrence data, (ii) model variables were prepared as described below:

  1. I.

    Occurrence data

    At the first level, inconsistent location fixes (records with missing coordinates, timestamps and elevation) and outliers were filtered out. Then, hourly GPS location markers of each lion obtained from remotely monitored radiotelemetry data were randomized to overcome spatial and temporal biases. Data has been reduced by taking every three hours of location43.44. The data was then categorized by season, i.e.. summer, monsoon and winter. These consolidated data were then spatially thinned by one kilometer using SDMtoolbox (version 2.0)45.46.

  2. ii.

    Model variables

The variables used for distribution modeling broadly included different land use categories, including natural habitats and anthropogenic factors, namely roads and areas of human settlement. All variables were rasterized at 10 m spatial resolution.

Land Use Land Cover (LULC) data for Saurashtra was obtained from Bhaskaracharya National Institute for Spatial Applications and Geoinformatics (BISAG-N), Gandhinagar, Gujarat. The data was then categorized into 18 subclasses: forest, sandy areas, salt flats, salt marshes, open brush, dense brush (wasteland), waterlogged, river/stream/drain, lakes and ponds, mining areas /industrial, reservoir /Reservoirs, Mangrove/marsh area, Cultivated land, Agricultural plantation (horticulture and agroforestry), Urban core, Mixed habitat, Peri-urban, Village (Fig. 2).

Roads and highways were also analyzed as separate variables in the model. The roads were categorized into village roads, major district roads, and state and national highways and digitized individually to further estimate the Euclidean distance (Table 2). Euclidean distance from human settlement (urban core, peri-urban, villages and mixed settlement) was analyzed and taken as a separate input variable for the model.


Comments are closed.