Not All Water Consumption Is Equal: A Water Stress Weighted Metric for Sustainable Computing

TL;DR

This paper introduces SCARF, a framework that assesses water impact of computing by considering spatial and temporal water stress variations, enabling more sustainable location and timing decisions.

Contribution

SCARF is the first framework to evaluate water impact by integrating water stress variations over space and time in computing sustainability assessments.

Findings

Optimizing location and timing can significantly reduce water impact.

Case studies demonstrate the effectiveness of the AWI metric.

Water stress-aware decisions improve sustainability in AI and manufacturing.

Abstract

Water consumption is an increasingly critical dimension of computing sustainability, especially as AI workloads rapidly scale. However, current water impact assessment often overlooks where and when water stress is more severe. To fill in this gap, we present SCARF, the first general framework that evaluates water impact of computing by factoring in both spatial and temporal variations in water stress. SCARF calculates an Adjusted Water Impact (AWI) metric that considers both consumption volume and local water stress over time. Through three case studies on LLM serving, datacenters, and semiconductor fabrication plants, we show the hidden opportunities for reducing water impact by optimizing location and time choices, paving the way for water-sustainable computing. The code is available at https://github.com/jojacola/SCARF.