What is the lifetime of drought-related water savings? Measuring rebound with survival models

Abstract:

Water saved during drought is an invaluable buffer for utilities facing water supply shortages. Although it has received less attention, the “lifetime” of those water savings is also a key consideration for long-term planning. Savings that evaporate once a drought is over can be called on in the future, while persistent reductions in demand may reflect a transition to a new and more efficient water use baseline, brought about by structural changes. In this study, we use a novel consumption change detection procedure to identify residence-level conservation and rebound in Costa Mesa, California after the severe 2012-2016 drought. We follow residences for 3-5 years after their first detected savings event in 2014-2015, using an accelerated failure time model to analyze savings “lifetimes”. We find that less than 30% of conserving residences rebounded to within 25% of prior consumption levels after 5 years, implying savings lifetimes of 10-12 years. Conservation survival was higher in younger and politically progressive neighborhoods and lower in residences with a customer change between 2014 and 2019. Although these savings lifetimes are long, our comparison of savings events during California’s severe 2012-2016 drought and the milder 2007-2009 drought does not show that drought severity produced more long-lasting savings; instead, heightened public awareness may have driven an additional but temporary savings boost. Still, our analysis suggests that droughts can produce deep and permanent water use reductions; better customer data are necessary to understand whether these are driven by behavior or structural change through permanent upgrades.