Upmanu Lall

Project: "Collaborative development of machine learning tools for spatio-temporal prediction of dry spell distributions to predict water deficit risk and its impacts across Chile using climate models and historical information"

Global Center: Santiago

Project Description: Climate variability and change are posing increasing operational risks to a number of sectors in Chile. Following a recent mega-drought specific risk that has been identified as a concern is that of persistent dry spells. Chile is a significant producer and exporter of fruits and vegetables, that rely on optimal water availability. It is also a leading producer of copper and lithium, both of which are sensitive to water availability, and it has a growing reliance on hydropower and solar/wind energy. Co-variations in precipitation, cloudiness, solar, wind and streamflow resources consequently affect all the critical sectors of the economy.

The proposed research will focus on one critical metric - the duration and frequency of dry spells in the upcoming 1 to 4 months -- with the goal of predicting this using a combination of information from existing climate forecast models and historical climate information. The forecasts will tie into the potential applicability for each of the economic sectors identified in a way that is relevant to the users in a geographically specific manner.

This initial work will set the stage for a future expansion of the effort to cover other climate related variables with the intention of developing a web based platform that could provide forecasts of these climate risks and their sectoral impacts as a step towards the mitigation of the emergent risks. If successful, the project would engage public and private sector sponsors and subscribers so that a self-sustaining resource center is developed for Chile. The proposed planning grant will facilitate the development of a proposal to APEC on a related topic, targeted for June 2022.