Indian savanna project

The dry grasslands of peninsular India are unique habitats that support a vast proportion of India's agro-pastoralist community. They are also home to critically endangered species such as the Indian bustard (Ardeotis nigriceps) and lesser florican (Sypheotides indica) and other endangered and endemic species such as the Indian wolf (Canis lupus pallipes), Indian fox (Vulpes bengalensis), and blackbuck (Antilope cervicapra).

Unfortunately, dry grasslands have not received the level of attention from conservationists or policy makers that is necessary, resulting in a lack of protection for endangered and endemic wildlife which occupy this habitat. On the contrary, Government policy officially declares most of these grasslands, scrub and thorn forests as “waste” or “unproductive land” ( For example, in one of the largest Indian states, Maharashtra, over 15% of the state's land area of scrub, grassland and grazing land is categorized as “wasteland”.

The prioritisation of landscapes for conservation of multiple species in human-dominated areas is recognized as a key global challenge. Identification of priority landscapes for conservation is dependent on the biological values associated with those landscape types. The presence of endangered and critically endangered species in any particular landscape increases its value, even if the threats and associated mitigation requirements are high.

Our key objectives are to

  1. create a countrywide map of dry grassland biomes of India, with fine scale maps at the district level for the states of Andhra Pradesh, Karnataka, Madhya Pradesh and Maharashtra;
  2. create a classification for conservation prioritisation based on objective criteria in a quantitative framework and
  3. design landscape specific conservation management plans for the protection and sustainable use of grassland habitats based on a participatory consultative approach involving all stakeholders.

Remote sensing

For the purpose of identifying and mapping grasslands areas in India, remote sensing data was analysed using various RS & GIS software such as ERDAS IMAGINE 2011, ArcGIS 9.3 and an extensive ground truth survey. A brief summary of the remote sensing data processing is given below:

Computing grassland occurrence probability

MODIS time series data of 2011 was used to make a country-wide map of grassland occurrence probability. Unsupervised classification was performed on the data, which displayed NDVI values. The data covered the whole country, and had a resolution of 250m.

Medium resolution landcover maps

From the grassland occurrence probability maps, certain areas in Maharashtra, Andhra Pradesh, Karnataka and Madhya Pradesh were selected for further analysis. LANDSAT ETM+ data was used for this purpose, since it has a resolution of 30m and is in 7 different spectral bands, thus broadening the scope of information available from the data. Supervised classification in ERDAS IMAGINE was performed on this data using ground control points to create land cover maps of these selected areas.

Detecting large grassland areas

From the land cover maps, the pixels corresponding to grasslands were extracted, and subjected to threshold based filtering algorithms such as “clump”, “sieve” and “filter”. The result of this process was a map of large, well-connected grassland areas from the LANDSAT scenes.

Fine scale land cover mapping

10 large areas across the 4 selected states were selected for finer scale land cover mapping. LISS IV imagery for these areas was used for this purpose. LISS IV data is in 3 spectral bands and has a spatial resolution of 5.8m. Supervised classification was performed on this data to produce land cover maps of the selected areas in 5.8m resolution.

Field surveys

Ground truth surveys

The field surveys were two pronged in approach. Our aim was to record locations of various land cover classes such that they can be used as ground control points for accurate mapping of land cover using 30 m data from LANDSAT. The various land cover types recorded were: grasslands, scrub, forest, bare rock, various types of agriculture, plantation and fallow. Close to 500+ ground control points were collected from all four target states over a span of 7 months across 15000+ kilometers of road travel. Our second aim during the preliminary field survey was to survey grassland areas that were large and contiguous based on our unsupervised land cover classification maps and grassland probability maps. Track plots were deployed and mud trails were scanned for mammalian track impressions and scat/pellets, photographs and coordinates were recorded for each of the signs encountered. This equipped us with preliminary presence data of grassland species from across the four states. Interviews with pastoralists and villagers allowed us to determine presence of species along with incorporation of any past sightings of target species (Indian Bustard, Lesser Florican, Blackbuck, Wolf etc.). Images of grassland areas can be viewed in the gallery section and a key to identifying track impressions and scat/pellets of grassland mammals is also available for reference.

Occupancy survey

This phase of the project commenced in May 2013. The objective of the occupancy survey is to establish occurrence probabilities of key species of large mammals in the grasslands and adjoining areas. The survey is ongoing and a few images are viewable in the gallery section.

Conflict surveys

In collaboration with Dr. Krithi Karanth, Centre for Wildlife Studies, surveys on human-wildlife conflict were conducted at three selected sites in the three states. These surveys will determine the presence of wildlife in the selected areas, as well as crop damage, livestock loss and any other incidences of conflict that may be detrimental to the conservation of biodiversity in these areas.

Team savanna

  • Dr. Abi Tamim Vanak (Principal Investigator)
    The conservation of species, landscapes and biodiversity is the central theme that drives my research interests. However, given the increasing lack of natural areas free from human-impact, I focus not just on conservation in natural systems, but also in human-modified and human-dominated landscapes.

    Much of my work is along two streams: the study of movement ecology of mammals ranging from Indian foxes to African elephant; and the conservation prioritization of semi-arid savanna grasslands that involves the study of mammalian carnivores, and in particular, meso-carnivores. In general, carnivores exist at relatively low numbers and yet can dramatically influence animal and plant communities. Thus, by studying carnivores, we can gain insight into the importance of the individual within natural systems. However, I am by no means restricted to carnivores or even mammals in my research program.

  • Ameya Gode (Research Associate): Ameya is the GIS analyst of the team. Primary tasks include data acquisition and processing. He classified Landsat imagery and LISS IV data into high accuracy land cover maps. Also, he analyses the remote sensing data, and uses it in sample design, change detection, area quantification and habitat analysis. He completed his M.Sc in GIS and Earth Observation for Natural Resource Management from ITC, Universiteit Twente.

  • Abhijeet Kulkarni (Research Associate): As a part of this project Abhijeet’s role is to identify the potential grassland areas using RS, GIS and help in creating fine scale maps of dry grassland biomes. Conduct field surveys to collect ground truth points for accurate classification, sign – based surveys to identify key grassland sites. Design and implement occupancy surveys using combination of camera trap survey and sign surveys to identify presence of key faunal species in these areas as well as to create species distribution maps. He is also responsible for the field logistics, development of a rapport with locals, communicate with forest personnel and media support. In the past, he assisted in Dr. Vanak’s PhD field research. His role was to assist in capturing foxes for radio collaring, collect movement data using radio telemetry, dietary analysis and organize dog vaccinations and disease testing camps in the study area. Abhijeet has vast experience in semi-arid and forest landscapes; leopards in sugarcane fields, large-carnivores and dogs have been his key themes.

  • Chintan Sheth (Research Associate): Chintan is part of the field survey team along with Abhijeet Kulkarni and has surveyed, mapped and digitally recorded most of the grassland areas of the four states. Chintan also takes keen interest in reptile systematics and distribution and has published in scientific journals. He is currently working on the geographic distribution of savanna species for his master's dissertation.


  • Dr. Enrico Di Minn (University of Helsinki, Finland): Dr. Di Minn's research focuses on evaluating the economic benefits derived from biodiversity conservation, spatial conservation prioritization, and trade-offs between biodiversity conservation and alternative land uses to reduce policy conflicts. He addresses these topics using interdisciplinary research and involving conservation practitioners. He is particularly interested in developing conservation solutions that can better inform real-world decision-making.

  • Dr. Krithi K. Karanth (Centre for Wildlife Studies, India): Dr. Karanth has been working on conservation issues in India with a prime focus on mammalian extinctions, human-wildlife conflicts and tourism trends to name a few. Her analyses on the controversial issue of resettling villagers outside a national park in India revealed some promising ways of tackling this conservation problem elsewhere in the country.

Email PI Abi Tamim Vanak at

Or write to
Royal Enclave, Sriramapura,
Jakkur Post
Bangalore 560 064
Telephone +91-80-23635555 (EPABX)
Fax +91-80-23530070