- India’s dry forest region gained 2.10 million hectares between 2014 and 2024, but this increase masks substantial deforestation, much of it on government-managed lands.
- Restoration monitoring should move beyond tree-cover statistics to separately track native forests, plantations, grasslands, and other open natural ecosystems.
- Beyond mapping restoration, it is important to identify land tenure in which restoration occurs to understand the social and ecological outcomes.
- The views in this commentary are those of the authors.
Since its launch in 2011, the Bonn Challenge has mobilised commitments to restore more than 350 million hectares of degraded and deforested land by 2030, including India’s pledge to restore 26 million hectares.
The International Union for Conservation of Nature (IUCN) defines Forest Landscape Restoration (FLR) as a process that seeks to regain ecological integrity while enhancing human well-being. In practice, this means that ecosystem restoration should improve biodiversity, wildlife habitats and corridors, watersheds, and other ecosystem services. It should also strengthen local livelihoods, particularly through the provision of non-timber forest products and support for pastoral communities.
Yet in practice, restoration is often narrowly equated with tree planting. This is partly because plantations can sequester large amounts of carbon, making them attractive for climate mitigation targets. India, for example, has committed under the Paris Agreement to create an additional carbon sink of 3.5 to 4 billion tonnes through afforestation and the expansion of tree cover.
Many tropical countries use globally available forest-cover products to monitor and report restoration progress. However, these global products come with limitations because they are generated using different satellite imagery, training data, classification algorithms, and forest definitions. For example, the Global Forest Watch dataset, derived primarily from Landsat imagery, can produce substantially different estimates of forest extent and change compared with GlobeLand30, which is generated using Landsat and China’s HJ-1 satellite imagery. A recent global study comparing ten global forest datasets found that these maps agreed on only 26% of the globally mapped forest area. The disagreement was particularly high in fragmented or sparsely covered biomes such as dry forests. Global forest maps also do not specifically identify plantations such as acacia, bamboo, and casuarina, which may be misclassified as natural forests in remote sensing analyses.
India has about 193 million hectares of land, or nearly three-fifths of its geographical area, under the dry forest biome. This biome extends from the lower regions of Himachal Pradesh in the north to southern Tamil Nadu (excluding the Western Ghats), and from Rajasthan and Gujarat in the west to Odisha and Jharkhand in the east. Here, the annual average rainfall is less than 1,000 mm, and the ecosystems include open grasslands, woody savannahs, deciduous forests, and dry evergreen forests. Work by Bengaluru-based researchers M.D. Madhusudan and Pradeep Koulgi shows that several thousand hectares of land in this region are open natural ecosystems, including grasslands and scrublands.
This dryland region has been a major target of afforestation and plantation activities over the last decade. As a result, large parts of the dry forest biome have experienced substantial tree-cover gain. When India reports its overall progress on forest restoration, the outcome is likely to show a net gain in forest cover. However, our recent research uncovers a different story.
When gains conceal degradation
In our recent study published in the journal Environmental Research Letters, we find that between 2014 and 2024, the dry forest region of India gained about 2.10 million hectares of forest but also lost 0.39 million hectares. Here, we define forest as a land parcel with at least 10% tree canopy cover on at least half a hectare, a criterion used by the Food and Agriculture Organisation to define forest. Our study uses Landsat satellite imagery and machine-learning techniques to map forest cover. Our study mapped about 14.11 million hectares of forest in 2014 with 89% accuracy. In contrast, the Hansen global forest cover data, one of the most widely used global forest maps, mapped only 9.2 million hectares of forest in the same region with 71% accuracy.
However, we found that it was not possible to differentiate native forests on government land from block plantations of trees such as palm, eucalyptus, and casuarina, generally grown on farmland. Hence, we generated forest boundaries for government-managed forests from the Survey of India’s toposheets.
Using this information, we found that about 0.17 million hectares, or 58% of the forest loss we observed, was within government land, where much of the native forests remain, while about 0.78 million hectares, or 37% of the gain, was outside government land. For example, dry forest loss on government lands was about 17,900 ha in Andhra Pradesh and 32,600 ha in Telangana. But gains outside government land were about 166,200 ha and 54,100 ha for Andhra Pradesh and Telangana, respectively, thereby masking the loss of forests in these States.
Similarly, a few states gained substantial amounts of dry forest in 2024 relative to the baseline year 2014: Andhra Pradesh (44%), Haryana (181%), Gujarat (30%), Karnataka (123%), Rajasthan (66%) and Telangana (54%). Tree cover gain on government lands is likely due to government initiatives such as Compensatory Afforestation Plantations and the National Afforestation Programme – Green India Mission. At the same time, India’s growing timber needs could drive gains outside forest land and in the use of timber as raw material for the paper and pulp industries.
Missing the woods for the trees
In India and many other tropical countries, the story of FLR is increasingly becoming a case of “missing the woods for the trees.” The Forest Survey of India (FSI) publishes biennial reports on the state of India’s forests. For the public, the message is often that forest cover has been increasing by several thousand hectares every year and that India is well on track to achieve its target of restoring 26 million hectares by 2030. However, the missing question is how much of this forest-cover gain has occurred outside recorded forest land, and what the status of grasslands and other open natural ecosystems is.
The Government of India defines forest cover as any land parcel of one hectare or more with at least 10% tree canopy cover, including private land. This definition used by FSI remains important because many private lands in the Western Ghats and northeast India still contain ecologically important forests. However, it is important to monitor and report the loss of forest within government land, where much of the native forest remains and the gain from outside government land, which is likely plantations of fast-growing trees like eucalyptus, poplar and casuarina, especially in the dry forest biome.
Grasslands and other open natural ecosystems in India have evolved over several thousand years and support unique biodiversity, including species such as the Great Indian bustard, blackbuck, wolves, and hyenas. However, India currently lacks systematic monitoring of grasslands and open natural ecosystems. The FSI publishes little information on these ecosystems, and there is no dedicated monitoring of whether degraded lands have naturally regenerated into grasslands or savannah woodlands, or of whether these ecosystems are being lost to agricultural expansion, fire, grazing pressure, or climate change. By excluding grasslands and open natural ecosystems from restoration planning and monitoring, there is a growing risk that tree plantations in these habitats will be promoted primarily to increase forest-cover statistics. Such interventions could negatively affect native biodiversity and pastoral communities that depend on these ecosystems for their livelihoods. Although programmes such as the Green India Mission recommend the protection and restoration of grasslands, in practice there remains a strong bias towards tree plantations.
At the same time, India has important examples of successful grassland restoration, including the Banni Grasslands in Gujarat and the Kalpavalli Grasslands in Andhra Pradesh. Grassland ecosystems can often be restored through assisted natural regeneration by regulating grazing pressure and reducing excessive fuelwood and fodder extraction. Such restoration successes should also be recognised and reported under the Bonn Challenge.
Ways to achieve ecological integrity
The Bonn Challenge (2011), the New York Declaration on Forests (2014), and the United Nations Decade on Ecosystem Restoration are important global commitments that encourage countries to protect and restore forests and other ecosystems. However, the monitoring and reporting of restoration progress remain weak, primarily because there are no universally accepted guidelines, indicators, datasets, or reporting mechanisms for these commitments. Countries participating in restoration initiatives should therefore be encouraged to look beyond tree planting as a measure of FLR and to also report on the restoration of grasslands and other non-forest ecosystems.
Given the current limitations of global forest and land-cover products, monitoring grassland restoration and distinguishing tree-cover gains in plantations and private farmland from ecological restoration will remain challenging. However, as demonstrated in our research, researchers, organisations and state and national governments can use locally collected training data and open-source satellite imagery, such as Landsat, which is readily accessible through the Google Earth Engine platform, to map and monitor restoration progress independently. And with current advancements in remote sensing, monitoring grassland restoration should also be possible. Furthermore, integrating land ownership and tenure records can help clearly distinguish restoration occurring on government forest lands from tree-cover gain on private agricultural lands.
Ultimately, the goal of restoration should be to enhance ecological integrity and human well-being, and monitoring frameworks should extend beyond tree-cover gain and carbon sequestration. Monitoring restoration across diverse ecosystems, while also identifying the land-tenure contexts in which it occurs, is essential for understanding the social, ecological, and economic factors that shape restoration outcomes and for designing more effective interventions.
Banner image: Labourers plant saplings as part of a tree plantation project. Representative image. (AP Photo/Rajesh Kumar Singh)
Dhanapal Govindarajulu is a senior research fellow at the Bharti Institute of Public Policy, Indian School of Business, Hyderabad and Johan Oldekop is a professor in environment and development at the Global Development Institute, University of Manchester, UK.
Credits
