Deforestation is a significant environmental concern, but the impact of human modification on remaining forests is often overlooked. This modification reduces ecosystem integrity and diminishes forest benefits, potentially rivaling outright deforestation in environmental consequences. Global science bodies increasingly recognize this issue, highlighting the need for improved tools and data to support policy-making and achieve Sustainable Development Goals (SDGs) and commitments under various international conventions (UNFCCC, UNCCD, CBD). Ecosystem integrity, fundamental to these conventions, is defined as a system's capacity to remain free from anthropogenic modifications to structure, composition, and function. Such modifications reduce ecosystem benefits and often precede deforestation. Intact forests typically provide a wider range of benefits, including climate regulation and biodiversity support. Various human activities like logging, infrastructure development, farming, urbanization, over-hunting, and resource extraction degrade forest integrity. While satellite imagery and analytical approaches have improved forest extent monitoring, consistent monitoring of modification levels remains a challenge. This study aims to develop a mechanism to measure and map the overall degree of modification resulting from landscape-level anthropogenic influences.

Previous research has individually quantified the effects of human activities on forest integrity at various scales. Studies have investigated intense, localized modifications (e.g., road building, canopy loss), diffuse changes associated with increased accessibility, and landscape-level connectivity alterations. This study aims to integrate these findings into a comprehensive, globally consistent measure of forest integrity. Existing studies also utilize binary measures, such as Intact Forest Landscapes (IFLs), which do not capture the degree of modification within remaining forests. This study addresses this gap by creating a continuous index reflecting varying levels of modification.

The study constructed a Forest Landscape Integrity Index (FLII) by integrating three data inputs: (1) observed pressures (infrastructure, agriculture, tree cover loss); (2) inferred pressures (modeled based on proximity to observed pressures, considering factors like edge effects and reduced connectivity); and (3) loss of forest connectivity. The FLII score ranges from 1 (lowest) to 10 (highest). Observed pressures were quantified using existing global datasets. Inferred pressures were modeled based on spatial association with observed pressures, incorporating distance-decay functions to reflect decreasing impacts with distance. Loss of forest connectivity was assessed using a method that quantifies the reduction in connectivity between forested areas compared to a potential configuration. The FLII was then calculated by combining the observed, inferred, and connectivity metrics, with the final score scaled from 0 to 10. The global map of FLII scores were categorized into three illustrative classes: low (≤6.0), medium (≥6.0 and <9.6), and high integrity (≥9.6).

The analysis reveals a substantial loss of forest integrity globally. Only 40.5% (17.4 million km²) of the remaining forest has high integrity, primarily located in Canada, Russia, the Amazon, Central Africa, and New Guinea. Even high-integrity forests show some degree of human modification (36%). The remaining 59% of forests exhibit medium or low integrity, with 25.6% having low integrity. Biogeographic realms display varying levels of forest integrity. For example, the Neotropic realm (Amazon region) has significant high-integrity areas. In contrast, the Palearctic and Neotropic realms have the largest extent of low-integrity forests. The Indo-Malayan realm has the highest percentage of low-integrity forest. National mean FLII scores vary widely, ranging from 9 in Guyana and Gabon to 3 in Sierra Leone. A substantial portion (26.1%) of high-integrity forests lies within protected areas, but even within these areas, a significant portion shows low or medium integrity (47.1%).

This study's findings highlight the significant and widespread impacts of human activities on global forests. The FLII provides a more nuanced assessment of forest condition than simpler measures based solely on forest extent or binary categories. The results emphasize the need for conservation efforts to prioritize the retention of remaining high-integrity forests. Preventing further degradation is more effective and cost-efficient than restoration efforts. The identification of areas with high integrity forests and their incorporation into effective management strategies are vital. The spatial resolution of the FLII enables targeted policy and management planning at various scales.

This study presents a novel global assessment of forest landscape integrity, revealing that only a small fraction of remaining forests maintain high ecological integrity. The FLII provides a robust framework for evaluating and monitoring forest condition, informing conservation strategies and policy decisions. Future research should focus on incorporating temporal dimensions, climate change impacts, governance effectiveness, and improved data on other drivers of forest integrity loss. Adapting the FLII for finer-scale applications will increase its practical utility for local-scale conservation planning.

The study's definition of forest integrity doesn't explicitly address past, current, and future climate change impacts or invasive species. The modeling of inferred pressures relies on distance-decay functions, which may not perfectly capture the complex spatial relationships between human activities and forest integrity. The categorization of FLII scores into illustrative classes is a simplification of a continuous index. Future research should address these limitations by improving data availability and refining the modeling techniques.