As the amount of data stored by our global civilization soars, concerns about the energy consumption of the required infrastructure, in particular data centers and the associated carbon footprint are being raised. And even though several cloud suppliers are striving for carbon neutrality, this is not yet common practice. And even green energy does not have a completely neutral footprint, because of the fossil energy consumed in the production of the solar panels and windmills, the effects of mining for their source materials, etc. What does this mean for the digitalization of the water sector? Are we increasing the industry’s carbon footprint by going digital? As both the climate and biodiversity crises are steadily intensifying, how does digital water help?
This is actually quite a complex and comprehensive question to answer. Therefore, I will only try to answer it partially, in order to illustrate the direction of the complete answer.
The (environmental) benefits of digital water
Digital water is not a goal in itself but a means to an end, or actually multiple ends with multiple benefits. One of these is to increase insight in and improve the performance of water systems. IWA’s 2019 report on Digital Water lists three major environmental benefits of digital water: reduced risk of sewage overflows into the environment, reduced greenhouse gas emissions from utility operations, and improved conservation and management of critical water resources.
I am not a specialist on the first and third topic but am happy to accept this claim. Both seem to be aimed more towards mitigating the biodiversity crisis and water resource management also towards adapting to changing climatic conditions. Reduction of greenhouse gas emissions from utility operations may contribute to mitigate the climate crisis and some aspects of this at least are quantifiable. I will discuss the particular case for water-loss reduction in the following paragraphs.
The reduction of water losses through sensors and data analysis is a very good example of how digital water may help to reduce GHG emissions. Every cubic meter of water that is lost corresponds to the loss of the energy that was used in the treatment and transmission of this water. How much? The table below gives an overview of avoidable energy consumption as a function of leakage rate and the energy cost of water treatment and distribution, assuming that 4% of losses are unavoidable (which is, of course, a simplification). The numbers have been calculated by multiplying the energy consumption by one million and then by the leakage rate minus 4%.
So any digital water solution that helps to bring back water losses that can operate within this energy consumption bound results in a positive result in terms of carbon footprint. But what amount of “digital water” do these energy bounds allow?
Let’s first look at the energy consumption of the data centers mentioned above. It’s hard to find convincing numbers on the energy consumption of data storage and processing, so I’ll use a very simple approach that aims to include both storage and data processing and analysis, by dividing the total energy consumption of data centers worldwide by the world’s total amount of data stored. The latter can be estimated at 3.5 ZB (3.8×1012 GB) for 2018b,(not all of which will be in data centers) and the former at 205 TWhc. This renders a rough estimate of the energy intensity of data storage and analysis: 0.053 kWh/GB/year. Note that the energy consumption of data centers is increasing much less rapidly than the amount of data stored, so the energy intensity continues to drop.
The amounts of data you can store and process for the energy saved in water loss reduction are really staggering, see the following table, which shows the amount of data you can store per million cubic meters of water produced before the data storage energy consumption becomes more that the energy loss from water loss, as a function of the leakage rate and emission intensity of the power source:
Note that these are gigabytes per year per million m3 of water produced. To put this number in perspective, let’s assume that we have DMA’s serving 2500 connections each with an annual consumption per connection of 150 m3. This corresponds to 375,000 m3 per year. Let’s assume you need 10 flow and pressure sensors per DMA to adequately detect and localize any leakages and that these produce 1 data record per second each. That’s about 315 million records per year, and assuming 10 bytes of storage per data item (including some overhead), this adds up to 3 GB of data per year for the DMA, or 8 GB per million m3 of water. About a thousandth to a millionth of what we ‘have available’ in terms of energy savings. And that’s not even considering data compression, which would easily reduce the amount of data by a factor of 10. That is if you wish to store it all and keep it for posterity. Our estimate above of the energy intensity can be off by a factor 10 or 100; we can still conclude that when looking at water loss and data storage, digital water pays off massively and the energy cost of data storage is negligible in this regard.
Of course, these sensors also consume power. The above example amounts to 26 sensors per million m3 per year. Let’s assume these sensors consume 20 W (probably less), this corresponds to about 4500 kWh of electricity per year. This is close to the break-even point at a 5% water loss rate, but significantly below at higher loss rates. And if we underestimate the sensor power consumption by a factor 10? Then it still pays off at leakage rates of 10% and higher.
There is a use case for digital water which is bright green: water loss reduction. It is likely that other use cases for digital water lead to the same conclusion. Worth mentioning is the wastewater treatment demo case in the Fiware4Water project, in which digital water reduces energy consumption and N2O emissions. It would be great to see more of these calculations and let’s continue to move ahead with digital water.
PS: Just as I submitted this blog for publication, a link to another blog that is very relevant in this context entered my mailbox.
d https://www.eea.europa.eu/data-and-maps/daviz/co2-emission-intensity-5#tab-googlechartid_chart_11, https://www.epa.gov/energy/greenhouse-gases-equivalencies-calculator-calculations-and-references, https://www.mdpi.com/1996-1073/9/12/988/pdf