IIT Bombay Study Shows Fragmented Forest Accounting Is Crucial For Effective Afforestation Planning
Large, continuous forests are ecologically healthy. They foster rich biodiversity, are resilient to natural and man-made disturbances and can self-regenerate. These unfragmented forests deliver long-term socio-economic benefits. In contrast, fragmented forests disrupt the movement and survival of plants and animals; for example, tigers need large, connected forests to hunt, breed, and survive without coming into conflict with humans. While the Forest Survey of India (FSI) and other independent studies regularly report on India’s gross forest cover, there has so far been no systematic framework to understand structural connectivity and monitor forest fragmentation across the country.
In a recent study, Prof. RAAJ Ramsankaran of IIT Bombay and his collaborators, Dr. Vasu Sathyakumar and Mr. Sridharan Gowtham of SASTRA Deemed University, have proposed a framework that uses remote sensing data and open-source digital tools to map forest connectivity at both the state and national levels. In addition to providing insights into how connected the forests are, the framework can also be used to analyse the impact of afforestation efforts, determine the resilience of different forest types to deforestation, and identify the states undergoing severe changes in forest cover.
Read also: IIT Madras Develops Agriculture Waste-Based Packaging To Replace Plastic
A key aspect of this approach is its classification of forest landscapes into seven types, each with distinct ecological implications: cores, which are relatively large and intact forest habitats; bridges, which connect different cores; loops, which connect parts of the same core; branches, which are narrow extensions from cores; perforations, which are non-forest clearings within cores; edges, which form the outer boundary of cores; and islets, which are small, isolated forest patches. The study finds that cores are the most resilient to deforestation, while islets are the most vulnerable, often undergoing further fragmentation or loss within a short span of time. In this sense, afforestation activities that primarily result in the creation of islets may not meaningfully contribute to forest health or connectivity.
“Our resilience-based ranking offers a practical tool for policymakers,” says Prof. Ramsankaran. “Rather than treating all forest areas the same, it helps identify which morphologies are most vulnerable (like islets) and which offer long-term ecological value (like cores).” He adds that afforestation programmes such as CAMPA or the National Mission for a Green India can benefit by focusing on strengthening existing cores and building bridges between them, which could potentially yield better-connected, more resilient, and ecologically sustainable forests. The framework also has the potential to inform infrastructure planning by helping identify areas where connectivity is most at risk, thus supporting more scientifically informed decisions and reducing ecological disruption.
The framework relies on an image processing technique called Morphological Spatial Pattern Analysis (MSPA) to detect and classify the structure of forest landscapes. As part of the study, the researchers applied the analysis to digital forest cover maps of India for the years 2015 to 2019, obtained from the Copernicus Global Land Service (CGLS) Land Cover Map. Unlike most previous studies on forest cover, which report only net gains or losses, this study mapped forest loss and gain separately.
The results show that from 2015 to 2019, all states in India experienced a net loss in forest cover. Overall, India lost 18 square kilometres of forest for every 1 square kilometre gained. Nearly half of the 56.3 sq. km. of gross forest gain occurred in Andhra Pradesh, Tamil Nadu, Karnataka, and Rajasthan, while Tamil Nadu and West Bengal together accounted for almost half of the 1,032.89 sq. km. of gross forest loss.
More significantly, over half of the newly added forest covers are islets, which do not substantially improve structural connectivity. This suggests that even where forest cover is increasing on paper, the ecological value and resilience of those forests may be limited. “Our results clearly show that most of the newly added forests during 2015–2019 were islets, highly fragmented and ecologically vulnerable patches. There is a need to move beyond the current quantity-based afforestation approach and explicitly incorporate structural connectivity into forest planning,” explains Prof. Sathyakumar regarding the implications of the study.
While the findings appear to differ to those of FSI, which often indicate an overall increase in forest cover, the results from FSI and this study are not directly comparable. FSI uses different criteria from the CGLS to identify forests and does not distinguish between fragmented and continuous forests.
Read also: IIT Bombay, ICICI Foundation Partner For Expansion Of Biomedical Innovation Hub
FSI defines forested areas as those with a minimum of 10% tree canopy cover and relies on satellite imagery with a 23.5 m resolution. In contrast, the CGLS dataset used in this study applies a 15% canopy threshold and a 100 m resolution. The researchers also had to rely on the internationally accepted CGLS dataset, as FSI data are not publicly available for similar analyses. “Since FSI reports do not include forest connectivity assessments, direct comparisons aren’t possible. However, our data source has a globally validated accuracy of over 85%, making our connectivity results reliable. If FSI’s data were made available in GIS-compatible format, our methodology could be readily applied to it,” says Prof. Sathyakumar.
One limitation of the current study is that at 100 m resolution, narrow linear features such as roads and railways may not be fully detected, and forest fragments smaller than 100 m may be missed. However, the strength of the framework lies in its scalability, cost-effectiveness, and use of open-source tools. It can be expected to give consistent results with similar datasets at finer resolutions and can be applied at different spatial and temporal scales.
“Our framework is fully extensible to finer scales, such as districts or protected areas, and can be used to analyse the impacts of linear infrastructure like roads and rail lines on forest connectivity in a more focused manner,” explains Prof. Ramsankaran. This makes it a valuable tool for long-term forest monitoring, planning and informed infrastructure development in and around forested areas, both in India and in similar contexts globally.
