Theses Doctoral

Social and ecological insights across landscape, community, and household scales: Forest health, governance, and livelihoods in central India

Khanwilkar, Sarika Ann

Forests are embedded in diverse forest governance, resource use, and resource user settings which are linked as components of social-ecological systems. This dissertation examines forest health at a landscape scale, governance at a community scale, and livelihoods at a household scale within a social ecological system; I develop a measure of forest health, the Bare Ground Index, derived from satellite imagery and combine this with socioeconomic data to examine relationships between forest health and forest governance and livelihoods across central India. This body of work has identified livelihood and governance approaches that provide social benefits and maintain healthy forests in central India, a landscape with globally important biodiversity and socially and historically marginalized people. This context is reflected in additional human-dominated landscapes where identifying sustainable development solutions that provide social and environmental benefits is a priority.

As forests are lost, gained, and degraded around the world, satellite data has been a powerful tool in collecting estimates of forest cover change but less widely adopted to measure forest degradation, largely due to challenges in common interpretations of operational measures. In chapter 1, coauthors and I develop landscape-scale land cover and forest health datasets for central India. First, we identified land cover, including tree cover and bare ground, from Planet Labs Very High-Resolution satellite data using a Random Forest classifier, resulting in a 3-meter (m) thematic map with 83.00% overall accuracy. Second, we operationalize a measure of forest health and derived the Bare Ground Index (BGI), a normalized index that is a ratio of bare ground to tree cover at 90 m resolution. The BGI was mapped across forest (>10% tree cover). Although open areas occur naturally throughout the tropical dry forest of central India, results from field data indicated that the BGI served as a proxy for measuring the intensity of cattle presence in a landscape where grazing has changed forest composition. The BGI was developed as an indicator of forest health and now serves as a baseline to monitor future changes to a tropical dry forest landscape at an unprecedented spatial scale.

In chapter 2, coauthors and I integrated the BGI with socioeconomic data from surveys to households and locally elected leaders to assess forest health and governance patterns across 238 villages at the community-scale. We experimentally selected 80 total villages as treatment and control groups and used this dataset in various statistical analyses to assess the extent of exposed bare ground within forests around villages with and without local institutions involved in making decisions about the forest. Forest had less bare ground within forest where there was a local institution compared to villages without an institution at 3 and 5 kilometers (kms), distances that households traveled from the village to graze cattle or collect Non-Timber Forest Products, firewood, and fodder. Having a local forest institution was more strongly associated with bare ground within forest at 3 and 5 kms than measures of local forest use. In villages with institutions, the authority to modify rules about forest use was relatively more important than the length of time the institution had been established for bare ground within forest. Establishing formal institutions with authority over forest management is important to promote forest cover around forest-dependent communities but it is necessary to ensure that forest governance does not worsen existing socioeconomic disparities. Bare ground within forests near and far (1 and 10 kms) villages was not different in places with and without formal local institutions and was most strongly associated with local forest uses. Both formal forest institutions and forest uses like collecting firewood for cooking or wood for construction material impact forests in central India.

In my third and final chapter, coauthors and I examined firewood collection patterns and the adoption of Liquified Petroleum Gas (LPG) using surveys from 4,994 households in central India. Firewood collection is pervasive across central India’s rural communities and mainly used for cooking or heating. We adopted an energy justice approach, which emphasizes questions about who does and does not have access to alternative cooking fuels, because historically marginalized groups comprise a significant portion of central India’s total population. It was important to integrate social justice issues in a system where resource users experience multiple disparities, such as high levels of poverty. We found that despite overall growth in LPG use, disparities in access to clean cooking fuels remained and the probability of cooking with LPG was lowest for socially and historically marginalized households (i.e., Scheduled Tribe, Scheduled Caste, and Other Backward Caste). While 90% of LPG-using households continued to use firewood, households that have owned LPG for more years spent less time collecting firewood, indicating a waning reliance on firewood over time. This study found evidence that policies targeting communities with marginalized social groups living near forests can further accelerate LPG adoption and displace firewood use.

My thesis examined components of a social ecological system at landscape, community, and household scales. I integrated insights from across social and ecological disciplines to identify strategies for sustainable development in central India. First, I developed an operational measure of forest health. Following chapters identified characteristics of governance and livelihood interventions that present potential pathways towards achieving benefits for conservation and people. Environmental and development goals should be harmonized so that the central Indian landscape can continue to support biodiversity and people. My approach can be replicated across additional social ecological systems by linking a landscape-scale resource condition to community governance and household socioeconomic patterns.

Geographic Areas


  • thumnail for Khanwilkar_columbia_0054D_17795.pdf Khanwilkar_columbia_0054D_17795.pdf application/pdf 3.51 MB Download File

More About This Work

Academic Units
Ecology, Evolution, and Environmental Biology
Thesis Advisors
DeFries, Ruth S.
Ph.D., Columbia University
Published Here
May 3, 2023