Water Risk: A Tale of Two Cities
In the data center industry, there is a lot of focus on carbon. Global warming is an urgent concern, and due to the huge amounts of electricity our facilities consume, our industry has a lot of responsibility to mitigate our carbon emissions. One of the ways this is often accomplished is by evaporating water to cool our data halls. But is this really an acceptable trade-off, environmentally-speaking? In our opinion, that depends very much on location.
While the impact of carbon emissions is global, the impact of water use is most definitely not. Rather, water use makes itself felt locally, in the particular watershed from which a facility draws its water, and the consequences of this vary greatly by the water supply and demand of a given region. Additionally, as both climate and populations change over time, the supply and demand on a given watershed will change as well.
This is why it is important to make water decisions at the facility level. First, this requires understanding the current and future water situations at each site. We use the WRI’s Aqueduct Water Risk Atlas to analyze water stress at each of our facilities now and into the future, with WRI’s predictions for 2030 and 2040. By comparing the regional water risk to the current water usage of our facilities, we can make informed decisions about the carbon/water tradeoff.
While most of our facilities consume zero water for cooling, let’s look at the water stress of two of our US Midwestern facilities that happen to be among our top water consumers: Cincinnati – Lebanon and Chicago – Aurora I. These sites currently receive all of their electricity from the grid. We know from the EPA’s EGRID stats that the RFCW grid subregion that supplies both Cincinnati and Chicago has a relatively high carbon output, averaging 0.487 MTCO2/MWh in 2019, with a correspondingly low 6.1% of energy from renewable sources in that year (the latest year for which data is available). Therefore, these sites are high on our target list for energy efficiency upgrades that could include continuing to evaporate large amounts of water.
What does our water risk assessment tell us about this strategy?
|Location||Current Water Stress||Water Stress 2030||Water Stress 2040|
|Cincinnati – Lebanon||Low (<10%)||Low (<10%)||Low (<10%)|
|Chicago – Aurora I||Low (<10%)||Extremely high (>80%)||Extremely high (>80%)|
If we look only at the current water conditions, these two facilities seem to be in the same situation – both are in regions where the current water stress is low, meaning that water supplies are enough to meet demand. However, the outlook for the future is very different! While Cincinnati is expected to continue to see low water stress, WRI predicts the situation in Chicago to change drastically in just the next decade, to the point where water stress is projected to be extremely high by 2030!
This suggests the need for very different strategies for these two facilities. The water risk assessment lets us feel comfortable continuing to consume water for cooling in Cincinnati because we can expect the water situation to remain stable, at least for a couple of decades.
However, Chicago is a different story. If we continue burning water at the current rate in Chicago, not only are we contributing to increasing regional water stress, but we risk business disruption as water becomes less available, or at least reputational damage as the local community becomes concerned about water use. Therefore, we are looking at ways to decrease the water use of our Chicago – Aurora I facility, along with increasing energy efficiency at both sites and a move to renewable power to avoid grid emissions.
And that is one way we use climate risk tools like water risk assessment to decide how to plan our tale of two cities.