How do you feel about this article? Help us to provide better content for you.
Thank you! Your feedback has been received.
There was a problem submitting your feedback, please try again later.
What do you think of this article?
It’s been almost 2 years since I started working on the Flood Alert project with electronics engineer Pete Milne and RS DesignSpark - building a device, with usability by vulnerable people in mind, to ‘alert’ to impending flooding, using Met Office Data. Specifically, the project is about using the U.K. Met Office API Data and combining it with the users’ own local data (collected using this alert device), to make a better predictive flood model. The reason, (or indeed, ‘story’), for it existing in the first place is to provoke the idea that as citizens we need to assist with data capture at a local level, if we want more accurate models for predicting events ranging from relatively minor, non-life-threatening (yet still highly inconvenient) floods, through to understanding full-blown natural disasters. Indeed, this model of combining ‘macro’ and ‘micro’/’local’ datasets might even have wider applications in areas outside of flood prediction, for example, in capturing the public's observation of wildlife to inform sustainability agendas, or traffic and air pollution lobbying, for instance.
Among the many learnings from the project, perhaps the most interesting outcome has been a realisation that although Climate Change is certainly a factor in recent years, it’s our urbanisation and governance of the land that has arguably been our downfall when it comes to flooding.
In a city environment, if we Concrete and Tarmac over green spaces - when rain falls it does not soak into the soil or become absorbed by trees at a sufficient rate to neutralise its effect. Similarly, if we build huge tower blocks without also expanding our wastewater treatment proportionally, the result is quite literally a tide of the proverbial filth heading for us in more ways than one.
In less densely populated areas, however, flooding has remained an on-going threat , as civilisations have formed around water, we have been building homes precipitously close to rivers, as evidently the rewards (farming, transport, drinking water, etc.) outweigh the flood which may or may not come in 10, 50 or 100 years. What has changed since the Industrial Revolution has been the density to which cities can grow, and to some extent, how we can control rivers better through irrigation, dams, reservoirs, etc. However, for all our technical prowess, even in developed countries, we are often still overwhelmed with an ‘unprecedented’ rainfall, and although climate change remains debatable for some, the devastation to densely populated areas is likely only to become more serious as our world changes. What is surely irrefutable is that we are not the master of the Earth’s weather system, so we need to better understand it to adapt to it.
For many years the Environmental Agency and UK Met Office have been keeping data on rainfall, and river levels (Fluvial data), which makes total sense, as for example Appleby, Cumbria, not far from where I grew up, you’ll see that areas which are not heavily urbanised, are nonetheless certainly at risk of flooding - when rivers burst their banks (see below).
Image: River Eden burst its banks in Appleby, a rural town in Cumbria. Image Credit: BBC News.
Another example is Carlisle, a city, in which it was estimated that in just one day, £1.3bn of damage was caused, according to BBC News. Although more urban than Appleby, Carlisle is still built along a river, and hence it sadly suffered from dramatic flooding in 2005, which was caused by record-breaking highs of rainfall.
Images: Differences between River Water (Fluvia), and Surface Water, unable to soak-away (Pluvial).
However, in addition to fluvial (river) water, Carlisle will also have issues dealing with surface water on hard surfaces, known as Pluvial water.
The combination of fluvial (river) and pluvial (surface) water overload is at the heart of why the Flood Alert was created - as there is scant data on the latter, and this likely calls for a mix of civic and citizen data to help improve our response to adverse weather, especially in urban densely populated areas.
To be sure, a Flood Alert cannot stop a raging torrent, but like a Smoke Alarm, you may react sooner with it, than without it. Furthermore, using localised predictive data models may mean you have more than ‘moments’ to react and might be able to deploy countermeasures like sandbags in time.
When Flood Defences Fail…Again.
Although this is not an exhaustive ‘one-size-fits-all’ example, the Carlisle situation, with its combined mix of river and surface water overload is useful to understand the multiplicity of why urban areas flood more than ever.
Images: Left: January 2015 Retrospective on 2005 Floods. Right: Flood Defences completed 2010.
Carlisle rightly improved its flood defences, investing a sizable £38m over subsequent years. With Embankments, Flood Walls and Pumping Stations expertly installed by 2010. Given the best data available at the time, this presumably looked reasonably sufficient to ward off the majority of weather fluctuations, and was apparently a credible upgrade to protect against all but the most infrequent ‘100+ year flooding’ occurrences.
And yet on Dec 5 2015, only 10 years later, Storm Desmond hit…
Images: Left: Dec 7 2015 BBC News. Right: Carlisle Flood Defences breached in 2015.
Storm Desmond left 43,000 homes in Cumbria and Lancashire without power and 5,200 affected by the water damage itself, according to one BBC Report. Another report said this was the largest in 600 years, which must have been disappointing for the Council’s flood prevention team, after spending £38m on ‘defences’, which failed after a mere decade, and of course devastating for families who’d only just got their lives back on track again.
The Increasing Use of ‘Unprecedented’, in regards to Natural Disasters.
The use of ‘unprecedented’ seems to be a word on the increase, when politicians and councillors need to publicly account for spending taxpayer's money on things that didn’t go as promised. Although I think there is less technical mystery in how banks have caused economic crises, (The Big Short being a scathing dramatisation of a reckless lack of accountability), clearly the fact is that our civic institutions have more data than ever - and yet can’t predict calamity, be it financial or environmental. Or perhaps they can, but if so, the relevant people are not getting the memo.
Image: Excerpt from the Council Report on Carlisle Flooding, and £38m Flood Defence Failure.
As the excerpt from the Carlisle Council Report shows, referring to Storm Desmond as simply ‘unprecedented’ only 10 years after a massive flood, must be hard to swallow for locals as justification, and does not offer much in the way of reassurance that they have learned from both 2005 and 2015 floods.
Furthermore, a recent BBC News Report on 6 August 2024 now warns that ‘Flood Planning Gaps Could Put “Lives At Risk”’ and heavily implies that not enough is being done as we head towards another decade since the last big flood. Indeed, I fear the next ‘Act of God’ is likely closer than 600 years away, and hope the defences hold, for my friends and family who live there.
The legal term for ‘Act of God’ is ‘force majeure’, and in many ways, I think this is a better phrase to use, as firstly I’m not looking to offend anyone religious, but attributing floods to a deity’s wrath feels a little dated, and rather problematically absolves humans of taking our share of proactive responsibility towards flood mitigation, even if outright prevention seems unrealistic. We can, in short, do a lot to affect change, and I think Technology and Science have important roles to play, but these are not secondary to our ethics and values which will ultimately define such actions. To this end, it’s worth zooming out to consider some more global and even historic perspectives on how we address such forces of nature…
The Cost of Flooding: Lives & Livelihoods.
I should preface this section by saying that part of the design process is to ‘expand’ things and look beyond the immediate brief of making an IoT Flood Alert piece of tech, to understand how it fits into the wider context.
There is a saying I find myself coming back to when trying to explore statistics of complex things: “Everything simple is false. Everything complex is unusable.” by Paul Valéry.
Another quote hints at our ability - or rather, a lack of ability - to process disaster, and prompted Kurt Tucholsky, a German journalist, satirist, and pacifist to observe that “a single death is a tragedy, a million deaths are a statistic”, later to be notably quoted by Stalin.
It is with some despondency, I hold these two maxims in mind when facing the uncomfortable task of asking how countries - and indeed individuals - evaluate what to do in so-called ‘natural disasters’. In 2015 Storm Desmond claimed the lives of 3 people in the UK, and caused £1.3bn of damage in Cumbria.
Image: For context, the top 10 most deadly floods, with many more listed besides.
Not to diminish the tragedy of any loss of life, but a cursory glance at the World's Deadliest Floods is nonetheless important to maintain perspective that overall the UK is, of course, a predominantly safe place to live, and it would be insensitive to not recognise the global view - and for example, the 2022 Floods in Pakistan that caused 1739 deaths, and £11.4bn of economic loss.
For reference, Pakistan's population is 235 million, and its GDP is around $375 billion. The UK has a population of 67 million, and GDP of $3 trillion.
Image: Flooding in Pakistan. Image Credit: Abdul Majeed/AFP via Getty Images.
It is telling, however, that these numbers are not proportional, and a close proxy to the disaster in Pakistan is Hurricane Katrina, 2005, in the US (Population 330 million) which caused 1392 deaths and yet almost 10x more economic loss at $186.3 billion. All things are not relative, and each country will ascribe its own set of economic and civic priorities.
Image: Flooding in New Orleans. Image Credits: CNN. Smiley N. Pool/The Dallas Morning News/AP
My juxtaposing these statistics beside one another is absolutely not intended to ‘weigh’ one human life against another, and certainly not against fiscal losses - but the sobering truth is that this sort of ‘accounting’ evidently happens in any government as they consider how to respond to any number of natural disasters, and crisis - and how they allocate preventative measures also. As much as politicians may get a bad rep, this is a thankless task which has to be done in some shape or form to determine who gets what support.
Image Credit: Aquanaomics
Although Aquanomics reports flooding to be a $5.6trillion global problem, I’m sure I’m not alone in saying that such numbers are so unfathomably large to us that I think they induce more ‘analysis paralysis’ in us than ‘action’... What does $5.6 trillion global GDP loss really mean?
If one report estimates Global GDP to be at $227 trillion by 2050, does this mean that Flooding is ‘no big deal’ if it’s only a 2% ‘dent’ in Global GDP, and as illustrated by the numbers above, the brunt of the financial loss and human loss will not be felt by countries like the UK, but almost without exception, much poorer countries and/or communities.
In short, these matters are more socio-political than we might want to openly admit, and it’s hard to say how any one country should better evaluate funds and initiate proactive measures for flooding or any other crisis, or future prevention of one. One thing that arguably has contributed in part to improve civic standards has been journalistic reporting and the resultant social pressure.
Making a Stink: The Politics of Prevention.
As many politicians will be aware, things need to get ‘local’ for action to be taken. Things have to get ‘personal’ for large numbers to galvanise and take action, more often than not.
The Great Stink as it was known in London, 1858, was a ‘break-point’ of where the untreated sewage that was dumped into the River Thames got so bad that Parliament famously doused its curtains with lime to try to quell the asphyxiating stench. It didn’t work!
Images: Illustrations of the literally deadly water of the River Thames, and a satirical newspaper cartoon depicting the dramatic rise of various horrific water borne diseases.
What is perhaps sobering about this, is that the UK Parliament had let sanitation slip for the masses since the Romans left England, and its steady decline had left a civilisation with an estimated 200,000 open cesspits (holes with all manner of filth in), which in a population of around 2.5million people, sounds vile. The wealthy will have had much more sanitary chamber pots, or perhaps a privy or early form of the modern toilet. In short, they were okay.
So it was only really when the stench of the river was so untenable to those in power, that they decided to act. This seems to be a pattern which turns up in living memory: Let us not forget that Covid-19 was treated with a spectacular global effort of the elite and powerful nations, as anyone from the average Joe Bloggs to the Prime Minister was not immune. Whilst Ebola (not a threat to the wealthy) remains very much a disease of poorer nations. My point being, our reaction to climate change is problematic, and is skewed by power and influence: yes, climate action will likely be determined partially by how much of a ‘stink’ the majority of citizens make (ie influencing politicians to an extent), but it will almost certainly be triggered by how much it impinges on the lives of those in power. Other parallels with Sustainability can likely also be observed, and we’ll perhaps save that for another day.
Flooding in Posh Places May Be Sooner Than We Think.
Getting back to flooding, Climate Central has an interesting simulation of various aspects of climate change, including Flooding through rising sea levels and rainfall estimates. Although one might debate the specifics of the data, the largely inescapable issue of sea level vs land elevation will hopefully focus the minds of respective country’s governors.
Image: Climate Central - UK and Netherlands.
Having just got back from visiting Amsterdam, in the Netherlands for work, it was very apparent that the beautiful canals and waterways, like Venice, are very obviously at risk from even a 1m rise in sea level. [Note the very red areas in the map above!]. This may be closer than we think, given that scientists estimate sea levels could rise by ~30cm in just 25 years. This again illustrates the issue, that those in power today will likely ‘get away’ with not changing their behaviour, but their grandchildren will not be so lucky.
The UK certainly has some alarming ‘red areas’ indicating serious water level increase in areas like Hull / the Humber Estuary, but also as far inland as Peterborough. London also has its concerns, but considering it already possesses The Thames Barrier to hold back flooding, it may be able to literally ‘stem the tide’, but it will have to actively manage surface and inland river volumes. In short, The Great Stink ‘fable’ may well apply again as even wealthy areas of London, and ‘home counties’ like Cambridge and Kent are not so easily ‘fixed’ with a Thames-style river barrier, and this may well impinge on the lives of the wealthy, and hence prompt action.
Image: Climate Central - India/Pakistan and US
Further afield, the aforementioned Pakistan also looks like it will have issues at the delta of its major rivers, which combined with any increases in rain will only make matters harder to manage. Similarly, the southern states, Louisiana was devastated in 2005, by Hurricane Katrina, and even as I write this, Hurricane Francine is now wreaking new havoc on the southern states.
Routes from Citizen Science to Journalism & Politics.
There is no doubt that Journalism has a huge effect on Politics. From uncovering corruption to ushering in new social values, it is an integral part of change, usually for the better.
Traditionally, ‘Data’ and ‘Hard Facts’ have been something only highly educated experts and organisations like the Media can collate, verify and publish with enough scale to influence social change.
However, for all the perils of Fake News, Social Media can of course be a force for good, if used appropriately. Combined with data, stories and compelling evidence from the general public, we all know how it can sway public opinion.
Citizen Science, a term coined in the mid-1990s by Rick Bonney, and is described as "developing concepts of scientific citizenship which foregrounds the necessity of opening up science and science policy processes to the public".
To my eyes this is about the general public being able to gather data, in sufficient quantity (now realistic with modern technology), and of enough credibility/validity that it can be taken seriously by institutions like academia and journalism, such that it makes news and reaches a wider audience - where it can have tangible impact - hopefully benefit that comes full-circle back to those of the public who want said change.
In short ‘People-Power plus Data’.
“In God We Trust, All Else Bring Data”
Indeed a quote I heard from NASA, but is attributed to W. Edwards Deming, and especially when hanging out in the US (when one is never entirely sure of atheist or religious affiliations) I’ve always loved the unifying nature of this quip when looking to put aside differences and get to the heart of the matter - to focus on the pragmatism of what you can control, and accept what you cannot.
With that said, NASA would also be the first to admit that John F. Kenndy’s famous speech containing, the often quoted - "We choose to go to the moon in this decade and do the other things, not because they are easy, but because they are hard" - that reminds even the most ardent scientist should appreciate that, like Kennedy, one needs a great sales pitch, as data alone does not galvanise a social movement, or indeed enable the economic boost to ultimately pay for it. NASA would not have gone to the Moon by ‘nuts and bolts’ alone, it needed an aspirational story or endeavour (plus a little rivalry with Russia) to make the majority of a nation feel this was a worthy use of taxes.
And so to finally circle back to the humble, yet plucky project that is Flood Alert, there is of course an element of ‘people power plus data’ - using the Met Office API Data and the users’ own local data for better predictive model. However, it also demonstrates the potential of citizens contributing to data capture in order to predict flooding both minor and major.
Big Data vs Hyper Local Data.
I am realistic that this modest, little ‘box of tricks’ is not going to out-perform the best minds at the Met Office, Government Agencies, or near-future advances in the International Space Station, and other Satellite meteorology - but it can fill the voids that considered are ‘too niche/local’ to be accurately covered by these institutions, by modelling the data that is frankly ‘too small’ or ‘too local’ to be economically compelling for tech companies, councils, or governments to have resource to focus and deliver on, but which still adversely affects a lot of people, destroying homes and livelihoods, even if thankfully not lives.
Image: Screenshot of Lee Valley Tributaries (near where I live), Note that the Map has many ‘River’ data points, but although the ‘Groundwater’ is toggled on - there are no data points in this map, making it more about rainfall in the borough and rivers only. (At time of writing).
I’ve dubbed this approach ‘Hyper-Local’, and arguably has a lot in common with the notion of ‘last mile’ delivery issues, where getting your parcel from London to say Glasgow is no problem for today’s mighty infrastructure, but getting it from the central depot to your house (the so-called ‘last mile’) is the most costly part of the journey in logistics.
Images: Environmental Sensor Development Kit (ESDK). The Good Air Canary and ESDK.
And so the same applies with Data, and indeed, has something in common with my creating a ‘localised’ air quality project previously for RS DesignSpark. Your borough might have okay air quality, but your street might be especially good or indeed, especially bad. It’s better to have your data at your fingertips! The same issue applies to flooding as it happens…
Image: ‘Zooming-in’ to Walthamstow and Wood Street, you’ll see there is now only 1 River datapoint for the whole area, and the ‘grey zone’ at the top of the map is not that relevant to the rest of the map, which may or may not be affected by heavy rain, and surface water issues.
To give a specific example for the problem with relatively un-localised data - the Environmental Agency can email me [in Walthamstow, London], that my nearest flood risk - the Lee Valley River and tributaries are experiencing high flow due to heavy rainfall. However, from my and other local resident’s experience this is often a ‘false alarm’ as it does not cause my particular street any major issue. Conversely, when my street had Surface Water issues, due to blocked road drains, these of course do not have any warning data, so it was a long and arduous process to gather enough data to get the Council to clean the Gullies, and the Water Authority to inspect the Drains, to resolve the matter (more on that later - I have some tips!).
In short, and in hindsight, we ideally needed a combination of 3 layers of data:
a. Met Office’s Weather Forecast Data.
Combined with…
b. Environmental Agency’s River (Fluvia) Data.
Plus…
c. ‘Hyper-Local’ Surface Water (Pluvial) Data.
This would mean less ‘false alarms’, (which is arguably no big deal), but when there is an issue it would hopefully create more predictive models of when to get the Council/Water Authority to service and clean Gullies/Drains. If for example, you know that after 7 days of hot weather (>25C), and then you have a storm coming, this is especially problematic - the hope is that with Flood Alert, you can get a ‘hyper local’ alert for this prediction, based on historical data.
Although I am very fortunate not to have been flooded, only 1 mile away from me, sharing the same area Environmental Agency ‘Alert’, a difference I would call ‘Hyper Local’, in Wood Street, there was significant flooding with homes ruined, boat evacuations, and cars written off. I’m presently unqualified to say irrefutably that this is only attributed to Surface Water, and/or blocked drains, but it is arguably not just about River Levels.
Without sufficient data, I do not want to speculate, but rather I want other citizens to gather data that is pertinent to their local areas, and is collectively powerful enough to make a compelling case for Councils and Water Authorities to act to improve the situation, whether it be ‘last-minute help’ like Sandbags, interim fixes like applying for Flood Defences, or longer-term strategies like rerouting water ways, flood gardens, and/or creating flood management systems in urban areas. These are all things evaluated on a case-by-case basis it seems, which is why helping citizens to bring forward data is key to them getting more support.
This is where the basis of the project started, and indeed, it is hoped to go further with support of the public, and other organisations, councils, sponsors and more.
At its simplest level, the Flood Alert project helps those who are less tech-savvy have a degree of warning about impending flood risk (even if imperfect as it is with river data). If we can create the ‘Hyper Local’ layer of data and combine it with Met Office forecasting, this would give greater clarity and better advice on how to react in advance.
Long term, but without wishing to get ahead of oneself, the project is a nice articulation of the way Machine Learning and predictive algorithms can be used on a relatable example of flood warnings. This tech Pilot costs around £100 per unit, so is more of a Beta Test, but if devices like this become more commonplace (the way Air Quality Units have) one wonders what ‘Edge IoT’ devices like this might enable in the UK, but in other countries where the severity is much worse. I would certainly be interested in the possibility of facilitating other flood alert projects being deployed in other locations, but given the importance of getting it to work reliably (and each country having its own Met Office / Environmental Agency data protocols which are hard to work on from the UK), it’s worth piloting this in the UK first and making it open source to see if others can build upon this to benefit others, and to contribute to this wider conversation around ‘big data’ and local data.
Indeed, just as I have waxed lyrical about the importance of ‘local data’, I would also extend this metaphor to imply that as a designer, one should create with people locally also, or as they say ‘nothing about us without us’. So please do get in touch if you’re interested after reading the rest of the build guide and evaluation of the project. Thanks!
Flood Forum: Who To Call If You’re Worried About, or Have Been Affected by Flooding
Early on in the project, I had the pleasure of speaking with Heather Stuart, from National Flood Forum - which is a UK charity that helps people who have been affected by flooding.
Heather made an important point about how flooding isn’t something that is easy to simply ‘get over’, and that because having so much of your personal belongings lost to water damage, and the safety of your home perpetually in doubt, this ends up taking a significant toll on an individual’s mental health. Having been lucky not to have been flooded, this was nonetheless something that reminded me of conversations with local residents and also friends and family from Cumbria.
Although UK floods rarely result in death, this does not diminish the need to help people, any more than relative poverty in a housing estate should not be ignored just because it’s not as bad as those in a famine or fleeing refugees of war. We all want a safe, secure and dependable place we can call home, and a combination of support after it happens is what Flood Forum is primarily focussed on - and it was heartening to hear that Heather liked the idea of an intervention that was proactive, and might give appropriate local warnings so individual residents / street take appropriate action in those crucial moments before flash flooding occurs.
Image: Heather Shepherd - Flood Forum. (Link)
The Impact of Flooding on Mental Health: Flooding can be a traumatic experience with long-lasting effects on mental well-being. Heather highlights the National Flood Forum's research in collaboration with Public Health England, which revealed a strong link between flooding and PTSD, anxiety, and depression. The emotional toll extends far beyond the immediate aftermath, impacting people's sense of safety, well-being, and future prospects.
The Importance of Community and Partnership in Flood Resilience: Heather emphasises the crucial role of community-level resilience and the need for strong partnerships between residents and risk management authorities (RMAs). She underscores the value of local knowledge and the effectiveness of collective action in advocating for change, accessing resources, and supporting each other during and after flood events.
The Potential of Technology in Flood Risk Management: Heather acknowledges the significant potential of technology, particularly accessible and user-friendly devices like the proposed so-called "flood warning Alexa," in enhancing flood preparedness and response. She stresses the need to cater to diverse technological literacy levels, including older generations less familiar with digital platforms.
"However, I think we are a society that is so focused on internet and modern technology, and there are so many people out there affected by flooding that this kind of technology is, you know, going onto a computer or trying to adjust with anything that's related to that, and it is just really beyond them."
The Need for Hyper-local Data Collection and Analysis: While acknowledging the progress made in flood forecasting and warning systems, Heather identifies the critical importance of hyperlocal data in improving accuracy and relevance. She sees significant value in enabling communities to contribute real-time observations on rainfall, drainage issues, and flooding events, empowering residents to become active participants in risk assessment and mitigation.
"Local people who are at flood risk are the experts of their own area. They go way, above the Environment Agency's alert when it comes to knowledge and what happens in an area."
Flood Re: The Flood Reinsurance scheme, set to end in 2039, highlights the ongoing challenges of securing affordable insurance for flood-prone properties.
Urban Flooding: Heather points to overdevelopment and inadequate infrastructure as primary drivers of increasingly frequent surface water flooding in urban areas.
Data Transparency: She dispels the myth that reporting flooding will negatively impact property values, as insurance companies already possess comprehensive flood risk data.
NB - Summary of Video was drafted by NotebookML [AI] from Youtube Transcript. Revised and Edited by Jude Pullen. (ie - really handy tool, but still needs oversight for editorial by humans!)
Blog Series Contents:
Prologue - The Case for 'Hyper-Localisation' of Civic Data
Research & Development:
Part 1: Filling the Local Data Gap
Part 2: Civic Services & User Experience Research
Part 3: Ideation of Flood Alert Concept
Part 4: Prototyping Back-Story
Part 5: Citizen Science Learnings
Open Source Build Guide:
Part 6: Build Guide for 3D Printed Assembly
Part 7: DIY Decals for 3D Prints
Part 8: Code & Data Guide
Future Ambitions:
Part 9: Project Reboot with Machine Learning
Comments