Paulina Concha Larrauri of the Columbia Water Center

The RainMiner app assists cities looking to augment their centralized water distribution systems with more sustainable localized networks

Interview by Grace Chan
Image: Fraction of wet days (a) pre-deconvolution and (b) post-deconvolution [before and after data has been cleaned up]. Note that most of the regional patterns were preserved (Source: Paulina Concha Larrauri)

Today I am interviewing Paulina Concha Larrauri, Staff Associate at Columbia University’s Water Center. Paulina is here with us to discuss the RainMiner app that she co-created, which assists cities looking to augment their centralized water distribution systems with more sustainable localized networks. RainMiner recently took first place at the Vaisala Open Weather Data Competition.

Paulina talks with us about how her app can help water-stressed cities better manage this precious resource.

Grace Chan: Hi Paulina, thank you for speaking with me today. Why don’t we start by you telling me a little about your role in the Columbia Water Center — what does your research focus on?

Paulina Concha Larrauri

Pauline Concha Larrauri: I’ve been a Staff Associate at the Water Center for the past three years covering different projects, all related to water of course.

Some of them are agriculture applications; for example, climate risk management for supply chains that are dependent on agriculture commodities that get hit by climate. That was one of the first projects I did, which was for PepsiCo, and I have had other projects related to assessing climate risk for urban water supplies in some areas of the Northeast United States.

Because I am from a very water stressed area in Mexico I care a lot about how water is managed. More recently I’ve been working on studying the potential of rainwater harvesting in large cities as a way to mitigate water stress. It’s part of a trend to seek new ways to get water to cities, moving from the huge centralized water infrastructures we have now to more decentralized systems.

GC: What are your interests, and how did they shape your idea for the RainMiner Concept?

PCL: One of my main interests is to find ways to use water more efficiently, for humans and for our ecosystems. So for me, RainMiner was about thinking how rainwater is usually not utilized in cities; we have this precious resource, even in water-stressed areas that have problems with their aquifers like Mexico City, and they’re still not using all of the precipitation that they have. So the RainMiner idea was like, “Okay, how can we know if rainwater harvesting is a good idea in a certain place?” So that’s how it came to life.

GC: Recently you and your partner Mounir Enennbach won the Vaisala Open Weather Data Challenge with RainMiner. Could you tell me what the purpose of the challenge was and go a little deeper into what RainMiner is?

PCL: The Vaisala Challenge was to create an application using open source weather data in an innovative way. There weren’t any more restrictions than that. The RainMiner concept we developed is a tool for decision makers, and even for homeowners and people like you and me, to see what the potential is for rainwater harvesting locally.

When we started RainMiner, we were looking at the potential of the whole United States by county based on available roof area, population, water demand, and of course local precipitation. We first had to work on the precipitation dataset; because it was gridded we had to “clean it” to make it more precise for our purposes.

The RainMiner app showing the average yearly supply rate in the United States derived from rooftop area approximations based on a GLM model 

The goal was to have an idea of how reliable a system would be for collecting rainwater in a specific area in the United Sates, and to get maps of different variables related to rainwater harvesting potential across the country. We could then find the answers to questions such as: What percentage of the demand would it be able to cover? How does seasonality impacts reliability during the year? How can water reuse increase the reliability? And things like that.

For decision makers, politicians and policymakers this could be useful because you can actually see in which part of the country rainwater harvesting is a good alternative or at least they can assess what percentage of the demand could be covered.

And it’s not only about water stressed areas. For example, here in New York it rains a lot. We have a lot of water anyway, but rainwater harvesting could become an option, for example, for reducing combined sewer overflows. Because you are not releasing all that water into the drainage immediately, you are keeping it, that can reduce a major source of urban pollution. Of course a little analysis would have to be performed but that is a potential plan.

GC: Are there any other tasks RainMiner can be used for?

PCL: The RainMiner app, as I mentioned, is for decision makers because it provides a macro-view for the entire United States. But now we have come up with other adaptations of the tool that are more targeted for local areas. For another competition that we were finalists for, an urban challenge for Rio de Janeiro, we focused on a specific part of Rio with the RainMiner concept. In Moros dos Cabritos, the favela we used as test case, there are constant interruptions in the water supply, but it also rains and floods a lot. So based on the amount of rooftop area and cisterns [water storage tanks] that could be sized to houses there, we could calculate the potential demand of water that could be covered during the year.

We are constantly exploring new ways to make analyses and to present them. In my projects I’ve always relied on climate data…This type of analysis is part of my daily job.

Locally it can be a selling point for homeowners since a rainwater system can save a lot on your water bill depending on where you are. For example, you could have 50% of your demand covered by harvested rainwater, and in areas where water is expensive this can translate into savings. In areas with no connection to the municipal water supply, this can be a life changer. So it can be a tool for decision makers and a selling point for these systems.

GC: How does RainMiner work?

PCL: In the app that we are still developing, you start with a menu that has two tabs — one to compute reliability and another that calculates costs, NPV [net present value] and payback time. If you are assessing an individual system, you would enter things like: What is your rooftop area? Where are you located? What are the intended uses you have? How many people live in the house? It then calculates the total daily demand and pulls information from the rain data for that specific area.

So it creates a report, telling you what is the reliability in the summer, in the winter, in the fall, etc. It also tells you what percentage of your demand is covered as well as the mean demand covered, and it provides a chart with the reliability that different tank sizes would achieve in that area.

Average demand satisfied in (a) winter, (b) spring, (c) summer, and (d) fall with 0% recycling; (e) winter, (f) spring, (g) summer, and (h) fall with 50% recycling 

The second feature takes costs into consideration, so based on different rainwater harvesting configurations — buildings, roof types, if it is a new or an old house or building, treatment level, etcetera — you can see what the costs for your system are and what is the payback time and NPV considering local water tariffs. For example, if you are just using the water for gardening you would not use a lot of filters. If you use it for drinking you would have a lot of filters. On top of that, maintenance is also considered in the cost analysis. This way you can decide what configuration makes more sense in a particular place.

Another feature is to provide a complete view of a city or of an entire country, depending what data is available, as we did with the United States. So for example, depending on what borough of New York City you are in, you could see what is the harvesting potential based on the type of buildings, roof areas and demand.

GC: Do you think you will continue to create similar environmental data analysis projects like RainMiner in the future?

PCL: Yes, we are constantly exploring new ways to make analyses and to present them. In my projects I’ve always relied on climate data. In another project I created an app that helps you visualize extreme weather data in a particular location and its correlation with crop yield, in my case oranges and potatoes. That tool also has a forecasting feature. So definitely, this type of analysis is part of my daily job.

GC: What do you think you will do in the future with your research at the Columbia Water Center?

PCL: Well, I would like to explore more the potential of water reuse technologies in combination with rainwater harvesting because I think they are part of the water solutions for the future, just like Singapore has demonstrated.

And something else that I am very curious in exploring is actually managed aquifer recharge (MAR), because where I am in Mexico has big issues with aquifer stress. So I would like to also learn more about that.

GC: Is there anything you would like to add?

PCL: I mean, there are so many interesting aspects about water that I just want to keep learning more about it!

GC: Thanks so much!

PCL: Okay!

* All data visualizations included are works in progress while the app is still being developed.

Leave a Reply

Your email address will not be published. Required fields are marked *