Q&A with David Fortune
Why do we need flood forecasting? Sounds like something nice to have. But is it imperative?
Flooding can be an essential part of the economy – so much agriculture depends on the nutrients that flood brings. However, most of us know flooding for the damage, disruption, and of course loss of life, that it brings.
By giving warning of impending flood, people can take action to reduce damage, reduce disruption, be informed to get out of the way, take refuge and thus save lives.
A forecasting system can also be an essential part of a flood management system. Perhaps by choosing to flood agricultural land rather than city centre, to some extent we can control flooding and reduce the risk of more damage.
It can take a long time to recover from flood damage. Particularly the toll it takes on people’s lives, health and wellbeing. Flood forecasting systems help.
And maybe you wonder why weather forecasts aren’t enough? They are certainly getting better and better. But what’s missing is the estimates of how much floodwater there will be; who it will affect; to what extent; and for how long. Flood forecasting systems include the vital elements of flood estimation, prediction and mapping.
Tell us about your background in flood forecasting?
About twenty years ago, we were designing and developing what was formally known as InfoWorks RS (now InfoWorks ICM) for river modelling. We had the opportunity to join a bid for a project sponsored by the World Bank to develop flood forecasting for the State of Andhra Pradesh in India. I have many years of experience of designing and delivering real-time operational systems so I jumped at the chance to be involved. I was appointed as the real-time systems expert on the project – but to the coastal flood forecasting team that was managed by Deltares.
Over the next three years I went backwards and forwards to India and spent a total of five months on the project. We built an effective design and development team and built relationships with meteorologists and hydrologists.
We were faced with huge technical problems. The coastal inundation was the result of storm surges by cyclones as they cross the ocean and come onshore. But the track of a cyclone is incredibly hard to predict. They stop and start, go backwards and forwards, and go around in circles so the exact landfall of the storm surge is unknown until just before it happens. The real-time system needed to give answers very quickly.
And we needed to come up with an innovative interface between the wind forecast and the 3D water model. Inevitably, we delivered a system that worked really well.
That must have been an interesting, and challenging, experience. So where did we go from there in Innovyze?
I learned a lot during that project. One thing that particularly struck me was that I would never want to go through that experience again starting from scratch.
What we needed was a product: pre-built software tools that would allow us to quickly and confidently configure flood forecasting systems anywhere in the world. And, of course, cost-effectively!
So that’s what we did – we built what was formally known as FloodWorks – a proper product, used at the heart of many flood forecasting systems around the world. And we learned more from those implementations – mainly how to make installation simpler and faster.
And that resulted in InfoWorks ICM Live. And there, I would say we perfected how to make integrated 1D/2D flood models run in real-time quickly enough to push out relevant predicted flood maps to those that need them operationally.
That’s flood forecasting. Last time we met, you introduced the subject of digital twins when we were talking about digital transformation. What’s the link between flood forecasting and digital twins?
Flood forecasting is a real-time operational process or system so I will compare that with general real-time, operational digital twins.
There has to be a point to an operational digital twin, for people to invest their time and money – there have to be intended benefits. So that could be information about performance of a water distribution network in the day ahead. Could be alarms to warn of potential service failures. And it could be optimization of the way the network is operated in the day ahead. And “optimization” covers a lot of different aspects, it’s a broad subject – we will talk about that. And it could be all three – multiple benefits.
It’s clear that flood forecasting, almost by definition, easily fits the purpose of an operational digital twin.
Now let’s consider system characteristics or components.
First of all, the operational digital twin needs a constant supply of real-time, telemetered monitoring data. And in the case of flood forecasting, that’s going to be flows and levels. If the digital twin is forecasting performance, it will often have other forecasts as an input. For flood forecasting that is weather forecasts: rain, wind, and temperature.
At the heart of a predictive operational twin is a model to predict performance. And that model should be an up-to-date representation of the performance of real world. We can talk about what kind of model would suit another time. But it is clear that the hydrological models, and the integrated one-dimensional and two-dimensional hydraulic models in a flood forecasting system fit well.
Then consider visualization. Just like other digital twins, flood forecasting systems need to present managers with clear pictures: that usually means 2D and even 3D flood maps. And those maps may well be presented to the public on open websites. That’s what happens in Flanders in Belgium where ICM Live, and previously FloodWorks, drives the national flood forecasting system.
So, what does this mean for Innovyze? And, more importantly, what does this mean for our customers?
It’s all about experience. Through building and implementing flood forecasting systems worldwide over the last fifteen years or so, we at Innovyze have gained a lot of knowledge:
How to work with real-time data – it’s never as reliable as you would wish but there is a lot you can do to ensure you get the best out of it.
How to make sure real-time systems stay working operationally, 24 hours a day, 365 days a year – that’s all down to how you design, write, test and implement your code.
How to push the forecasts out to the managers that need them using mobile and web technology - wherever they are at any time of the day.
The importance of collaboration and teamwork between Operations, Modelling, IT – and of course solutions suppliers such as Innovyze.
So, we have grown that experience in partnership with water managers and their engineering consultants and we are able to transfer it to other water managers. Whether they are in flood management, stormwater and wastewater, or water distribution. So much of the expertise is transferrable.
The success of real-time operational systems depends on so much more than technology. We get that at Innovyze. And I say that makes us the ideal technological partner – we understand the way that technology supports successful solutions.
Nevertheless, we are a technology company. And I am excited by the way that technology is advancing and can be adapted by operational systems. Internet of Things, Cloud Computing, solid-state monitors, and lower cost weather radar, just for example. What I am looking forward to is that this new technology can simplify the complexity of operational systems such as flood forecasting systems. And that’s making it accessible to many more environmental and water managers than ever before.
Let’s talk about that another time.
So, if you are out there listening and recognise you may have an issue with flooding, or sewer overflows. Or you want visibility of which of your customers you are in danger of failing to deliver water to at the right pressure. Then talk to us. Chances are that digital twins can help. And we have the experience to help you take advantage.
Stay tuned for our next Q&A and in the meantime, check out the Innovyze Podcast page to hear more: All Things Water Podcast