Rise of the Machines: Algorithmic Zoning for Climate Risk

Stephanie M. Stern & Cherie Metcalf

Fires, floods, storms, and other extreme events related to climate change are responsible for billions of dollars in property damage and thousands of deaths in recent years across North America. In this Article, we propose a new model of local zoning to address the escalating threat of climate harm: climate algorithmic zoning (“CAZ”). This model draws on the emerging potential to harness large, independent data sets to generate more granular, location-specific climate risk predictions and regulations. These predictions can guide a range of traditional zoning functions, including building codes and permits, restrictions on existing development, and local planning and infrastructure. We anticipate that CAZ will arise in the highestrisk localities where climate threats and damage outweigh political opposition from the development lobby and homeowners. Within those localities, CAZ will tilt toward regulating new construction or reconstruction, for both political and constitutional reasons.

There is no future in which algorithmic climate risk prediction does not affect residential investment, property valuation, insurance, and protection. Insurers and lenders have already embraced climate risk algorithms. Rather than shying from CAZ, we contend that local governments should adopt a complementary approach that integrates CAZ into land use regulation proactively and leverages local governments’ distinct institutional and democratic advantages. These competencies include the ability to zone and plan to prevent climate damage, knowledge of local conditions and culture, and the ability to engage communities democratically.

Algorithmic, big data approaches carry risks and shortcomings; the approach can entrench existing biases that disadvantage vulnerable communities, create and amplify risk assessment errors, produce rapid shifts in decision-making that upset expectations, and reduce transparency. CAZ at the local level can leverage traditional tools of local governance to mitigate these challenges, for example, through more democratic design of algorithms, use of CAZ as non-exclusive inputs into broader decision frameworks that incorporate important local values, and retooling existing legal doctrines of zoning, such as substantive due process and amortization.