A flood sensor network for Calderdale
Photo by Twitter user Brittany Forbe
Hyperlocal water level sensors for improved monitoring
Boxing Day 2015 saw parts of Northern England subject to devastating flooding, with many homes and businesses affected. Our own industrial estate in the Calder valley was amongst the sites hit, with the water high enough to reach the roofs of cars and cause shipping containers to float around the car park!
With similar events in 2012 still fresh in the mind, it is hard to believe that this sort of disaster is unusal, or unlikely to happen again in the future.
Photo of Hebden Bridge from HebWeb on Twitter
Though flood defences are in place and other prevention schemes are being reviewed, improved advance warning of such flooding may help to reduce the damage caused. This is a complex challenge, with many factors involved in understanding how, when and where such flooding may occur.
A great example of a grassroots initiative tackling this problem is Flood Network, with distributed sensors sharing water level data to improve resilience and response. In this post we will look at the project before detailing how we added our own water level sensor to the network.
Born in Oxford
Oxford was heavily hit by flooding after storms in late 2013 / early 2014, prompting Oxford local Ben Ward and others to think about how to build a water level monitoring solution with the Internet of Things.
This led to the Oxford Flood Network, deploying it's first sensor in February 2014, monitoring the water level of a small stream in Oxford. Soon gaining support from Nominet UK, the project grew, with more sensors added to the network as time progressed.
Because of the community driven nature of the project, people were encouraged to contribute however they could. For example, in August 2014 a workshop was held at technology festival Wuthering Bytes. Participants helped to assemble sensors, work on software and give their input on the project as a whole.
Ben giving a presentation to delegates at Wuthering Bytes 2014
As the Oxford Flood Network grew it was clear that the project could be implemented elsewhere, and so Flood Network was launched, with the aim of building the UK's biggest network of flood sensors.
With the river Calder running just outside our offices, we felt ideally positioned to add our own node to the Flood Network.
Installing our first sensor
It is now possible to sign up to the Flood Network, purchase a monitor and, provided you are within 40 metres of a river, stream or other suitable location, easily contribute.
Included in the kit is the following:
The Flood Monitor sensor
A Raspberry Pi powered gateway
Spare battery for the sensor
Power supply and network cable for the gateway
Instructions are provided to help you get your sensor up and running as quickly as possible. Due to the developing nature of the project these are best considered as a 'beta' or draft version, with feedback requested by the Flood Network.
Image from Oxford Flood Network
Following the supplied instructions we soon had our sensor installed above the river. The sensor was calibrated, with measurements taken and supplied to Ben Ward. With our sensor installed and gateway connected to our network, we soon started to see data being transferred between the two.
Within hours our sensor had appeared on the Flood Network map. Ben then asked us for some further information on the surrounding area, for example what was upstream and downstream of the location, and how these areas may be affected by changing water levels.
We also now had login details for access to our data – and the data from other nodes – via a browser-based console. This allows for remote monitoring of the connected sensors and is the perfect starting point for further work.
It is important to note that this project, including the map, is under a heavy state of development and as such should not be treated as concrete evidence for the likelihood of flooding or as a fail-safe warning system.
However, we have shown that adding a sensor to the Flood Network is relatively simple to do, a real testament to the hard work of those involved on the project thus far.
Our initial installation has allowed us to prove that the technology works and motivated us to build on it. In order to get our first sensor working it was put up in the most convenient location to hand, a tree above the river. This allowed us to begin collecting data and test performance of our local environment.
Next we will relocate the sensor to a more permananent place, fabricating a bracket mounted to a building alongside the river. This will mean re-callibration is required, as well as considering just how high the water can get before there is a significant danger of flooding. This needs to be as accurate as possible if we are to be able to rely on the readings from our sensor.
A selection of sensors from Oxford Flood Network
We also want to look at adding more sensors to our local network. There is an update on the horizon that will introduce LoRaWAN connectivity to the flood sensors. This will give long range, low power connectivity that is both low cost and license exempt. This will allow sensors to be located a great deal further than 40 metres away from a broadband router, with a long battery life, meaning reduced deployment and maintenance costs.
We will be at the upcoming FloodHack event that will be held at the Open Data Institute in Leeds, with some hardware, ideas and further details of our plans. We look forward to seeing others there and collaborating on solutions that can make a real and immediate difference!