Future cost of climate change for humanitarian crises

Humanitarian crises are the tip of the iceberg in climate change adaptation, yet their future is rarely quantified in human and economic terms. We use machine learning to simulate future estimates of people in need of humanitarian aid and required funding under the business-as-usual scenario (SSP2-RCP4.5) with warming of 2.1–2.4°C by 2100. Humanitarian needs rise to a baseline of 410±22 million people and USD2024 64±8 billion annually by 2050 worldwide, increases of 28% and 30% respectively compared to the current (320 million people and USD 49 billion). A lightly optimistic simulation holds needs near the current, while a medium pessimistic simulation leads to 614±68 million people and USD2024 96±19 billion by 2050, increases of 92% and 96% respectively. Our results show empirical vulnerabilities and an opportunity cost, as resources for crisis response displace funding for adaptation and mitigation. Yet, sustained investment could curb the impacts even with climate inertia.