Webinar: Developing and deploying low-cost, distributed monitoring to evaluate NFM

Today I attended a NERC webinar on developing low-cost monitoring to evaluate Natural Flood Management (NFM). Marc Mulligan from Kings College London talked about how each NFM technique is different and will vary in design and scale. It’s important to consider geographical context, therefore Mark’s approach aims to monitor a wide range of interventions extensively rather than a single site intensively.

Monitoring is expensive in specialist equipment and sensors, associated infrastructure, as well as time to visit sites, collect and analyse data. Mark explained how to overcome this, they developed a network-connected, low-cost, DIY-build sensors and automatic web-based analysis techniques suitable for replication over large number of sites and NFM measures. This involved using FreeStation loggers to collect data.

Mark mentioned how ground based monitoring infrastructure is declining globally and we cannot remotely sense all the variables we need to measure. Therefore, FreeStation offers an environmental monitoring system with open source DIY hardware. The FreeStation solution includes a variety of instruments, all good enough for measuring what we want. It is designed with consumer, off the shelf, low-cost components, easy to build and install. The project started in 2014, and lots of data has been collected so far. Online tools have also been developed to help us better manage and analyse the data streams.

Low cost does not mean low accuracy, precision or quality. Low cost means non-specialist, consumer, mass produced and using your own labour.

A few examples of where these systems have been deployed include at Blackwater, on the Colne, Mole, Stour, Leck, Mar Dyke and Lannock Manor Farm.

Mark outlined there are 2 different types of NFM – ‘at a point’ and those interventions which retain water over much larger areas. Examples of at a point interventions include stage sensors (similar to low-cost car parking sensor, using sonar sensors) either side of dam. This can then determine the exact distance between the sonar and the water level.

Problems with sensors include build up of rubbish on the sensor, vegetation growth, needing to avoid heavy shade in order to get sufficient solar power, signal (2G and 3G signal not great in wooded ditches), force of water and high flows, and vandalism of sensors, although this is rare. Of 110 loggers only 2 instances of theft. 1 logger was stolen and 1 instance where someone tried to steal a solar panel which caused some damage. To avoid this, it’s best to label devices to avoid suspicion, work with local people, keep out of sight, don’t put next to paths and disguise where possible. Also need to protect from wildlife, for example hide and cover cables with metal.

Mark outlined some findings for several NFM interventions. He mentioned leaky dams were found to work if they are designed and implemented well. Small volumes retained behind these dams, and can activate flood meadows locally. Similarly, beaver dams can make a contribution. They are self-building and self-maintaining, and offer many co-benefits for biodiversity and habitat through creating wetlands.

However, there are some challenges associated with at a point interventions. Leaky dams and retention ponds can be tricky and expensive to build. They also require some maintenance, particularly after extreme events. If rubbish builds up against their structure, they can become vulnerable to failure.

Monitoring is difficult and can be inconclusive. Estimating discharge is difficult without building flow control structures. NFM interventions can also store water below ground, which needs to be measured. As well as this, comparing upstream and downstream flows within a network of dams is subject to the impact of side flows and inputs. Comparing either side of a single dam and stage locally, doesn’t help understand the impact of a network of dams. Plus, a single dam at a specific location appears to behave and respond differently, so extrapolation is challenging.

Assessing the impacts of NFM is challenging as different types have varying impacts which depend on their design and local context.