Robust optimisation for sequential investment problems: an application to climate adaptation in the Mississippi River Basin

Abstract

Effectively adapting to climate change requires long-term investment strategies informed by climate forecasts. This paper presents methods and a case study assessing how an investor would approach sequential land investments in the Mississippi River Basin (MRB) using 32 climate models under two emission scenarios (RCP4.5 and RCP8.5). Each model produces projected farmland values, which are used in a robust optimisation framework to identify optimal investment policies under varying levels of conservatism-reflecting the degree to which worst-case outcomes are considered. The model is linearised and scalable across long horizons and asset sets. The case study spans 2023-2090 and uses regression-based projections of land values. Robust investment strategies are derived for each climate model and scenario. Results show that as conservatism increases, investment becomes more geographically constrained, with significant variation in optimal regions. While all climate forecasts predict warming, the regions benefiting from such changes vary by model and scenario. Investment patterns range from concentration in specific areas to broad diversification or complete withdrawal. The analysis underscores substantial disagreements among climate models, which critically affect spatial investment decisions and highlight the importance of robust planning under climate uncertainty.