Chile's compliance data show projects that comply while loads persist and baselines that are never revisited. What 50 years of IA teach about monitoring planetary drivers.
Since NEPA (1970), impact assessment (IA) has become universal, yet monitored wildlife populations fell 73% (1970-2020) and seven of nine planetary boundaries are now crossed. Has IA managed project symptoms rather than drivers? We test three propositions using Chile's open compliance data: 9 million biodiversity records (15,947 reports, 1989-2026) and 4.3 million effluent results (~530 discharges, 2017-2025). (1) Drivers vs symptoms: effluent exceedance rates halved (0.9% to 0.4%; p<0.001), yet nitrogen loads from a stable panel of 265 discharges did not decline (p=0.71); over 95% of the load came from compliant samples, 63% reached water bodies with multiple dischargers, and only 28% of discharges are linked to an IA permit. (2) Adaptiveness: 62% of long biodiversity series changed method and limits are set per discharge rather than per catchment load, so monitoring cannot feed back into decisions. (3) Changing baselines: 81% of sites with over a decade of history were sampled once and only 13% of baseline sites were resurveyed; baselines are snapshots, not references. Project-level compliance is not pressure control. We propose permanent monitoring networks with standard protocols, catchment load targets embedded in IA, and interoperable open data so that follow-up becomes learning for the next 50 years.
Gino Araya is Director Project at ECOS-Chile, working on environmental impact assessment, monitoring and compliance for mining, energy and infrastructure projects in Chile.