Where does WISEN fit into the data commons?
Armed conflicts generate diverse forms of environmental harm, particularly when fighting is prolonged, intense, and geographically widespread. It manifests differently in urban, industrial, agricultural, and ecologically sensitive areas. Harms may be direct or indirect, temporary or permanent, localised or widespread, and often occur simultaneously in cumulative ways. Capturing their full scope therefore requires multiple, complementary data collection approaches.
Ground-based observations provide the most detailed and practically useful environmental data, but in conflict settings they are often difficult or impossible to conduct. Even after fighting subsides, such assessments are costly, require specialist expertise, are hard to scale, and may remain unsafe due to unexploded ordnance. Remotely collected data is therefore essential during both active conflict and post-war recovery.
Remote environmental data collection can be simplified into two approaches. The first approach is to track and investigate discrete incidents. The alternative approach we term ‘continuous class’ based, which is where objects or impacts are self-similar over a large geographic area. Typically data from satellites is assessed, often incorporating AI, to characterise phenomena like building damage, crop loss, deforestation, landscape fires, or cratering. A growing number of academic groups are producing such outputs for specific contexts, with systematic frameworks for urban and agricultural damage for all conflicts currently in the pipeline.
Both approaches are important and complementary; for example a health researcher might need acute air pollution exposure (i.e. from incidents) but also the chronic exposures from longer term changes in the average (i.e. continuous classes). However, depending on the conflict context, one approach may be more suitable. For example, where the destruction has been near-total, as in Gaza, it is more appropriate to use the continuous classes approach. So too, in contexts like Tigray, where there was an information blackout and no social/mass media footage/reporting. WISEN attempts to cover both approaches — Level 0 data being an example of continuous classes — though is more focused on incidents.
One future direction for WISEN as a knowledge platform would be a deeper integration of continuous class changes with incidents, together providing a comprehensive understanding of environmental risk and change in areas affected by conflicts.
