Frequently Asked Questions

The Oxford Flood Network is still in its beta phase. This means we're still learning about things like reliability issues, wireless range, community engagement and data trust. As we develop the network we learn more about the most effective ways to work, but it does mean that the system may not be available 100% of the time.

Q: Who runs this?

Oxford Flood Network was a project set up by Ben Ward and Andrew Lindsay to experiment with the concept of a 'guerilla sensor network' - one which is set up by the community, rather than by an authority. Flood Network Ltd was set up in June 2015 to develop the idea and make it a commercially sustainable model. 

In 2014 we partnered with Nominet UK's R&D department who developed the backend systems which store and analyse the data and manage the devices and funded early development of the sensors.

Q: doesn't the environment agency do this already?

Their sensors produce similar data. However, we decided to develop Flood Network to fill in the gaps between their sensors whilst living in Oxford and seeing the lack of detail on the kinds of waterway that cause flooding. This has many useful side effects, such as street-level observations of flooding and the ability to make decisions based on the data. We like to think of Flood Network as an enhanced layer of detail on top of their data.

Q: Why bother? I know when it floods

The flood sensors constantly collect data, but they also disseminate it.  This means they can give lots of up-to-date readings from different locations at the same time, and your flood sensor could inform hundreds of people about conditions. Thresholds can be set for levels where flooding has historically occurred. If you're not in the area you can still look at the flood map to see if your home is likely to be at risk, or if your business is accessible by staff and customers.

Q: Are they ever stolen?

Yes, a couple of sensors have been stolen or lost in the past year. In certain installations we used very temporary mountings such as magnets and velcro, and this may have contributed. Where possible we try to fix the sensor to a structure or bracket.

Q: Is this only for rivers?

Not all sensors are measuring river heights. The sensor can easily be used to detect water underneath floorboards in flood cellars. Many of the older houses in Oxford have been built with flooding in mind and have these flood cellars to manage ground water variations.

Technical FAQ


Approximately a year. We have not seen any significant problems due to premature battery drain. Some of the older sensors were out in the field for over a year before being replaced with a new device.


You can see from the data on the left hand side of the map, we get small and big gaps in the reporting. In general, the sensors report back every 15 minutes. These gaps are due to a number of reasons:

  • Sensor Firmware Bugs: The sensor occasionally fails to read the range to the water level. Raw data can be processed to deal with many things, such as ducks passing underneath, and changes in temperature which affect the ultrasonic sensor. (If you want to help with the hardware development, much of the source code and designs for this sensor are on Github)
  • Wireless Range: Between the sensor and the gateway device attached to broadband there is a wireless radio link. Sometimes the signal doesn't get through to the gateway. This can be often caused by environmental conditions such as a new object in the way, or wet foliage. When a reading is missed we can fill in the gaps using 'interpolation', but this can be misleading if no data is received for extended periods, so the graph stops plotting under these conditions.
  • Gateway Devices: The gateway devices are mains powered and attached to volunteers' broadband. Sometimes a knocked cable, or a broadband problem can cause the gateway to go offline. Occasionally a crash can stop the gateway operating, but recent versions of the gateway have mechanisms to recover from this. The older gateways have to be manually rebooted.
  • Repositioning of sensor: Sometimes we move the physical location of a sensor. We can keep the original data from the old location, but it wouldn't make sense to move it to a new graph representing a different location. 

Because there are many complex parts involved in this process we're always learning ways to retain data continuity without affecting the reliability of the readings.