Keywords

FormalPara Highlights

  • It is vital that Natural Flood Management interventions are carefully designed and fully tested for their effectiveness

  • We outline newly developed methods for assessing the effectiveness of NFM with a focus on lowcost and open-access solutions.

  • We show how modelling is best suited to large scale strategic assessment of the optimal type, magnitude and locations of interventions

  • Assessment of the effectiveness of specific interventions is best achieved through a field measurement approach

  • Where modelling or measurement are not possible, space-for-time substitution with comparable sites for which the intervention has already been applied and tested should be considered

4.1 Introduction: Understanding the Biophysical Effectiveness of Nature Based Solutions

This chapter defines what is meant by effectiveness of nature based solutions (NBS) for flood mitigation, highlights the opportunities and challenges provided by a range of methods for assessing the effectiveness of NBS and introduces new methods for low-cost assessment of NBS effectiveness in a range of flood mitigation contexts. The methods discussed include monitoring, modelling, and space for time substitution. In the latter an intervention in a different location but similar context can be used in substitution for the potential impact of an intervention not yet built, see Pickett (1989). We discuss methods that can be used to assess the effectiveness of different flood-focused NBS in different biophysical contexts (i.e to understand what type of NBS works where), and to understand the scale required for effectiveness in flood mitigation, including the generation of co-benefits which, for nature based solutions, often by far outweigh the direct benefits of flood risk reduction (Chap. 6).

The methods outlined here can be used to indicate the biophysical effectiveness of the investment made in terms of reducing flood risk and, where NBS are shown to be effective, to advocate for greater investment. To measure NBS effectiveness in mitigation of floods one needs to know:

  • the size or capacity of the NBS, for example the volume of upstream flood storage captured in relation to that needed for downstream damage loss mitigation,

  • where in the catchment the NBS should be located to be most effective in reducing flows to downstream assets at risk,

  • how effective the NBS are for different events, ie to what extent does the intervention mitigate the hazard for different storm sizes, durations and antecedent conditions,

  • the magnitude of losses mitigated by NBS in relation to other (grey infrastructure) mitigation methods.

Here we describe new methods for assessing the effectiveness of the NBS before discussing challenges that still exist for assessing the effectiveness of NBS. We focus on no or low-data situations, which is the norm for most NBS investments, in contract to grey infrastructure investments where detailed feasibility studies are required ahead of any major investment. Our focus is a subset of NBS: natural flood management (NFM) and it is the low-data situations associated with NFM investments for which we have developed the Eco:Actuary toolkit, a suite of tools to support risk analysis, NFM investment planning and NFM effectiveness monitoring for installed interventions. These methods have been implemented in the Thames case study (Chap. 12). For NAIAD case studies that cover data-rich situations where extensive and detailed background hydrological data are available, high resolution ‘engineering’ models can potentially be adapted and used to assess NBS, for example, the MIKE-SHE/HYDRO model application in Copenhagen (Chap. 17), and HEC-HMS and HEC-RAS applications in Slovenia (Chap. 15) and the Lower Danube (Chap. 10).

4.2 The Eco:Actuary Toolkit

4.2.1 Introduction

The Eco:Actuary toolkit was designed to assess different types of NFM at scales from small to very large, and is designed for locations where there is little or no pre-existing data available to support assessment, which is often the case for non-engineered solutions, like NFM (Dadson et al. 2017). The toolkit comprises three components:

  1. 1.

    The //Smart:River and //Smart:Soil systems for in-field monitoring and web-based analysis of the effectiveness of specific interventions made. //Smart: is intended as a turnkey system for monitoring the effectiveness of specific in-place NFM investments accurately, at low cost and with ease of construction, deployment and analysis. It combines self-build open source hardware (FreeStations, www.freestation.org) alongside custom on-device firmware for data collection and pre-processing and server-side software (serverware) to provide near real-time visualisation and analysis of NFM performance. FreeStation internet connecting environmental monitoring designs are around 1/30th the cost of commercial devices, make use of open source technology, but must be self built according to the instructions provided. //Smart: includes configurations for use on point source in-line storage, such as leaky woody debris dams (Thomas and Nisbet 2012), and retention ponds (Wilkinson et al. 2019) and non-point source land management options such as conservation (regenerative) agriculture (Boardman and Vandaele 2020). Example live data are available at www.policysupport.org/smart

  2. 2.

    The Eco:Actuary web-based Investment planner, which is an empirically based spreadsheet model capable of connection to data APIs (Application Programming Interfaces) such as the UK Environment Agency Real Time Flood Monitoring API, for understanding the total magnitude and type of upstream interventions necessary to make a significant difference to observed flood peaks at a monitored location. This tool is further described in Chap. 12.

  3. 3.

    The Eco:Actuary web-based spatial policy support system (PSS) – which uses a physically based raster spatial model and globally available data to map flood risk, asset exposure, damage, mitigation and avoided loss by current green infrastructure and under scenarios for climate change and NFM investments. This coupled physically-based, probabilistic model can be used to generate a large population of rainfall total, spatial distribution and intensity events that are applied to a mapped human and physical landscape. As with all policysupport.org tools, all data are provided for application of the EA PSS anywhere, globally.

4.2.2 The //Smart: Tools for Monitoring NFM Effectiveness

4.2.2.1 The //Smart:River System

The //Smart:River System is for monitoring the effectiveness of NFM measures used in small rivers and streams such as an in-line storage intervention (eg leaky woody debris dams, Fig. 4.1). There are three methods for monitoring effectiveness using //Smart:. The first focuses on a single intervention and involves a FreeStation water level (stage) and soil moisture device either side of the intervention (Fig. 4.1) alongside local measurement of rainfall, temperature, humidity and air pressure using a FreeStation weather station. The data can then be analysed using the //Smart: visualisation and analysis interface (Fig. 4.2) to assess the differences in soil moisture and stage either side of the intervention which to calculate the operation of the intervention. A typical visualisation for FreeStage is shown in Fig. 4.2 with measured stage shown alongside live rainfall for the nearest station as captured from the Environment Agency Rainfall API in this case.

Fig. 4.1
A photograph of a water level and soil moisture monitoring device placed at the wood barriers of a leaky dam.

Live FreeStations monitoring water level and soil moisture upstream and downstream of a Leaky Dam (left) and close-up of a water level FreeStation (right). Each has a pole-mounted aerial to improve mobile connectivity

Full size image
Fig. 4.2
A photograph of a control panel placed between small plants to monitor river depth and hourly rainfall.

The //Smart: visualisation and control panel interface of hourly measured river depth and live hourly rainfall from the Environment Agency Rainfall API

Full size image

The second method for monitoring the effectiveness of in-line storage interventions calculates the storage behind the dam (or in the retention pond) either as a volume or as a proportion of measured flow downstream at the asset at risk from another FreeStation stage gauge. To calculate the storage, the stage data are analysed by the //Smart: serverware in relation to the cross sectional area of the store (and any downstream asset at risk station location). Cross sectional area is measured using a low-cost hand-held FreeStation LIDAR (the ‘FreeDAR’). The storage is calculated using the //Smart: serverware, by calculating the change in storage behind the dam in periods when water is rising (“the rising limb”) as an absolute volume stored or as a proportion of a downstream flow. Volumes are calculated from surface areas assuming a unit velocity (1 m/s) to provide unit discharges. This assumption holds for small streams and rivers in similar geomorphological settings as the assets at risk or comparator dam, as is the case for most leaky debris dam and retention pond NFM since they address small scale flooding. When the water level is receding, the dam is draining and these “falling limbs” represent drainage of the stores. For example, Fig. 4.3 shows that during a week-long period some ~3 m3 in total are captured during the rising limbs by a dam on this small experimental stream. This can be compared in the system to the corresponding discharge over the same period at the asset at risk downstream.

Fig. 4.3
A line graph of a leaky dam inline water storage of depth filtered, and rainfall for E A station. Two y-axes depict depth filtered, reversed L 10, and rainfall for E A station E 10130. The x-axis reads the time from July 18, 18 hours to July 25, 12 hours. The curve for depth filtered peaks around 0.1 on July 20, zero hours.

Cumulative inline water storage behind a leaky dam calculated by //Smart:River

Full size image

The third method is usually applied to networks of multiple interventions and involves analysis of unit discharge upstream of the interventions in comparison to downstream. Differences in the magnitude of flow peaks and the duration of recession limbs reflect the impact of the combined interventions. Figure 4.4 shows how a network of 31 leaky debris dams decreases the flow peaks downstream (red) compared with upstream (blue) and extends the recession of flow.

Fig. 4.4
A line graph of a cumulative inline storage L 44 D 1, excluding drainage. The x-axis reads the time from August 11, 18 hours to August 18, 12 hours. The curve starts at 0 and steadily increases reaching 2.2 on August 18, 12 hours. Values are approximate.

Flow peaks downstream (red) compared with upstream (blue) of 31 leaky dams as monitored by FreeStation //Smart

Full size image

We find that, usually the total storage behind a single leaky dam is a very small fraction of the flow downstream, and so – unless the asset at risk is very close downstream or the dam raises flood waters to spill over onto a low asset value river floodplain – many hundreds or even thousands of these leaky dam interventions are needed in order to make a significant impact on downstream flood peaks for anything other than the smallest rivers. Chapter 12 discusses this further and demonstrates how this has been applied in the Thames.

4.2.2.2 The //Smart:Soil System

//SmartSoil: continually assesses how much water enters and is stored in the soil, in absolute volumes or in proportion to downstream flow. It combines FreeStation soil moisture sensors with specific serverware for analysis of storage by soil over a given area (as absolute volume or relative to downstream flow) and can also be used to measure the impact of soil management interventions aimed at NFM (such as low-till or other regenerative agriculture methods) relative to a counterfactual in a neighbouring field (eg conventional tillage). When used to assess the effectiveness of soil management interventions, one sensor is deployed within the land use where the soil management intervention is practiced (eg the regenerative agriculture field) and another in the counterfactual (land use without intervention, eg under conventional tillage, but all other hydrological factors the same, such as slope, soil type). Each device measures soil moisture at one or two depths (e.g. 30 and 60 cm). With information of the field areas, soil bulk density and soil depth, //Smart: can calculate the difference in soil water storage and drainage between the intervention and the counterfactual. There is also a local measurement of rainfall, temperature, humidity and air pressure. The result can be calculated as an absolute volume of water retained or percentage of downstream flow (with downstream flow volume data).

These methods have been applied in the Thames demo during the NAIAD project and are discussed in Chap. 12. //Smart:Soil was used to assess the contribution of regenerative agriculture to increased flood storage, by increasing infiltration and thus reducing runoff over farmland. Since these techniques can be applied over large areas of agricultural land, they can be effective NFM for evening large rivers. The benefit of stream based NFM over farm based NFM as that the former, whilst requiring an initial investment and maintenance, are not dependent on the landholder changing their land use/management practice. Farm based NFM require buy-in and/or incentivisation of the landowner.

4.2.3 The Eco:Actuary Investment Planner

The Eco:Actuary Investment Planner (EIP) is designed as a simple, online spreadsheet-based tool that allows assessment of the scale and approximate cost of upstream interventions of different types that would be required to reduce peak flows at a monitored station, by a given proportion. It is useful to investors considering the type and scale of NBS investment that will be effective at mitigating flooding for a river for which long term measured river flow data is available, and thus is only relevant for gauged catchments.

4.2.3.1 Using the EIP to Assess the Effectiveness of NBS Investment

The EIP is available at www.policysupport.org/ecoactuary. The user enters the location and the proportion of the discharge (flood peak) they wish to mitigate and the tool calculates the volume of flood storage that NFM must hold back to achieve that level of flow mitigation. For applications in England, this tool uses data supplied by Eco:Actuary (EA) through DEFRA Opendata (data.gov.uk) in particular the UK Environment Agency Real Time Flood Monitoring API and connects to this directly through a real time API (application protocol interface). The non-UK spreadsheet tab allows users to paste in their own discharge time-series data, which the tool will then use. Users can update the default per-unit volume construction cost for different Natural Flood Management (NFM) techniques (leaky debris dam, retention pond, regenerative agriculture) and the tool will estimate the budget required to mitigate flow at the level required. The user sets the flow mitigation goal as a percentage of flow. An example application of this tool is given in Chap. 12. EIP takes no account of how the placement of NBS in specific catchment positions can make a difference: a spatial model like Eco:Actuary is necessary for that (see below).

4.2.4 The Eco:Actuary Spatial Policy Support System (PSS)

4.2.4.1 Purpose of Eco:Actuary

Eco:Actuary (E:A) is a spatial Policy Support System (PSS) focused on assessing fluvial flood risk, exposure of multiple asset types and their values, estimation of baseline damage losses and better understanding the mitigation of potential damage losses by existing and proposed green infrastructure and NFM interventions. As well as assessing baseline risk and mitigation E:A can also be used to understand the impact of scenarios of climate, land use, asset value and asset adaptation (damage function modification) change. Like all policysupport.org tools, Eco:Actuary is applicable anywhere globally based on datasets provided with the model. As a coupled physically based, probabilistic model the EcoActuary PSS can be used to assess risk and mitigation in unmonitored catchments. A detailed user guide to Eco:Actuary and the PSS itself are available at www.policysupport.org/ecoactuary.

4.2.4.2 Using Eco:Actuary to Assess NBS Effectiveness under Different Scenarios

The impact of climate change or proposed land management interventions can be achieved by comparing modelled outputs for flood damage losses under the applied scenario compared with those in the baseline simulation (representing climate and land use at 2015). Within Eco:Actuary the following types of scenarios can be assessed:

  • Climate change: Eco:Actuary provides access to downscaled monthly CMIP5 (Emori et al. 2016) and AR4 (IPCC 2007) climate change scenarios for 17 climate models (GCMs) and a range of Intergovernmental panel for climate change (IPCC) scenarios which, combined with user-specified changes to the rainfall intensity and storm size distributions can be used to examine impacts of likely climate changes, spatially.

  • Impact of extreme events: Changes to rainfall extremes can be simulated by changing the maximum intensity and maximum monthly rainfall parameters for the distribution of rainfall event intensities and event volumes generated in the “Hazard Ensemble” of the catastrophe model. This is in addition to any changes in monthly rainfall volume under climate change, which in turn determine the number of events.

  • Land Use and Cover Change: Eco:Actuary allows users to generate large-scale land use and cover change scenarios to represent landscape scale NFM interventions such as afforestation or rewilding. To understand impacts of observed deforestation or afforestation, users can also use the land use modeller QUICKLUC (Mulligan 2015) within Eco:Actuary. QUICKLUC projects recent rates of afforestation/deforestation forward on the basis of past observations and other relevant factors.

  • Asset adaptation: scenarios with different flood damage curves (representing damage losses for different depths of flood water) can be used in Eco:Actuary to represent asset adaptation (such as lifting assets from the ground).

  • Natural flood mitigation infrastructure: Eco:Actuary allows users to assess the impact of existing natural flood, mitigating infrastructure (wetlands, canopies, soils, floodplains, water bodies) on reducing flood risk and can also change the volume and distribution of these infrastructures.

We have outlined the Eco:Actuary suite of tools to assess NFM. These tools can be used as part of the NFM planning and preparation stage to assess the current location and value of assets at risk and of mitigation by natural infrastructure (Eco:Actuary PSS), or for exploratory Cost Benefit Analysis for proposed intervention types and magnitudes using available discharge records (EIP). The toolkit can also be used post-investment to assess the effectiveness of NFM. A detailed review of the application of the toolkit in the UK Thames Demo is given in Chap. 12, whilst the following section demonstrates an application of the Eco:Actuary PSS to the Lower Danube.

4.3 Challenges in Assessment of NFM Effectiveness

4.3.1 Data Availability and Uncertainty

One of the key limits to using modelling to assess effectiveness is data availability and uncertainty: even for key datasets such as rainfall and terrain. The availability and quality of these data are fundamental to the accuracy of modelling of flood risk. Many smaller streams that are relevant to NFM are not gauged and, even in well monitored countries, the distribution of rainfall stations is much broader than the scale of spatial variability of rainfall.

4.3.1.1 The Complex Drivers of Flooding

A further challenge is that flooding can result from poorly known conditions antecedent to the rainfall event or poorly known characteristics of the event (e.g. the detailed spatio-temporal rainfall intensities, blockage of channels and culverts, subsurface flow, snowmelt). Even with recent advances in meteorological monitoring and prediction, the location and timing, size and duration of large rainfall events is still poorly known, and therefore the prediction of the hyper-local outcomes that characterise flooding and generate flood damage profiles is thus challenging. Soil and subsurface hydrology is also poorly known at policy relevant scales, and there are many green (natural) and grey (engineered) flood water stores in the landscape to take account of, some of which are also poorly known.

4.3.1.2 Model Uncertainty

Model uncertainty is also a challenge In the case of Eco:Actuary the applied models are physically based, i.e. representations of physical processes that generate discharge, rather than empirical relationships between rainfall and discharge. Complex, physically-based models are more useful in scenario studies (since the physical principles can be applied to scenarios as well as baseline conditions, whilst empirical relationships usually cannot). Nevertheless, physically-based models make many assumptions which are difficult to test in real landscapes. The advantage of low cost, DIY (self build) approaches to monitoring is that more dense data collection networks can be developed that can provide data to help reduce model uncertainties. A hydrological model is usually calibrated, e.g. against discharge data of rivers and streams in order to achieve greater correlation between modelled and observed flows but calibration is a ‘sticking plaster’ that does not help understand the reasons for model disagreement. A key to effective use of models in NFM is to make them easy to apply, such that a number can be compared, alongside field measurements and a ‘weights of evidence’ approach to understanding outcomes can be developed.

4.3.1.3 Where the Assets at Risk Are

It is important to locate NFM where they have greatest influence on the most downstream assets at risk. Good asset maps for different classes of asset type are necessary but are not usually available; Eco:Actuary uses Open Street Map data for a variety of asset types combined with EO products representing nighttime lights and building height in order to map asset types and assess the scale of asset value within an asset class (with brighter, taller assets being of higher value, than shorter, darker ones). These are coupled with locally specified parameters representing the absolute max and min monetary value per asset type. For example, the asset type of greatest value is shown for each sq km. the Danube catchment in Fig. 4.5.

Fig. 4.5
A satellite map of asset of greatest value of river Danube. A color legend reads, none, airport, bridge, commercial, cropland, highway or road, hospital, hydropower plant, industrial, pasture, pipeline, power utility, railway, residential, retail, school, telecoms, university, waste water plant, water treatment, and plant.

Eco:Actuary output: Asset type of greatest value, River Danube catchment. (Base Map: Map data ©2019 Google)

Full size image

4.3.1.4 Asset Valuation Uncertainty

To calculate avoided damage, accurate asset values and damage functions are crucial. Eco:Actuary replacement value and loss of function value have to be estimated on the basis of proxies (nighttime light, building height and asset type) since no other data are usually available in the public domain at the scales that NFM analyses are applied. This creates uncertainty in the absolute values produced by the valuation component, but the results will nevertheless be of the correct order of magnitude and scaled appropriately relative to each other. More readily available data are necessary to reduce uncertainty.

4.3.1.5 The Scale of NBS

The impact of all interventions decays with distance downstream of the intervention. Thus, small scale NFM interventions in large catchments, such as leaky log dams will only have local impact and very significant interventions are required to have significant impacts downstream, particularly given data and model uncertainties. Figure 4.6 shows output from the Eco:Actuary PSS indicating the rapid decline in influence of a key wetlands in the Danube with distance downstream: within only 40 km downstream, the wetland storage represents only 0.4% of the discharge. The Danube is 2860 km long so for most of it, the wetland has virtually no flood-relevant influence or at the least any influence is of a smaller magnitude than the errors in the data that we have available to assess it.

Fig. 4.6
A zoomed-in satellite map of downstream influence of wetlands including lakes, rivers, and reservoirs. A color legend reads values equals class base from 0.000e plus 00 to 100.000.

Rapid decay of downstream influence (%) of wetlands using Eco:Actuary Policy Support System. (Base Map: Map data ©2019 Google)

Full size image

It is also important to consider the effectiveness of the sum of all interventions upstream, so interventions cannot be considered in isolation from each other. The location of NFM and networks of NFM are also important: analysis using the Eco:Actuary PSS shows that to be effective interventions have to match the scale of rivers. Small interventions will only have impact on small streams and on large rivers interventions must be substantial to make a positive hydrological impact. Substantial mean few large or many small interventions.

4.3.2 Ways Forward

Assessing the effectiveness of NFM is challenging, especially for the civil society organisation that lead on many of these investments. By providing easily used monitoring and modelling tools based on open-science and open data we have removed some of the barriers to more effective deployment of NFM as part of the flood mitigation toolkit. For large scale strategic analyses we provide the Eco:Actuary PSS and the Eco:Actuary Investment Planner to better understand the scales and magnitudes of intervention required to achieve specific levels of benefit in particular contexts.

For post-deployment assessment of the effectiveness of NFM, we provide the //Smart: system which allows users to cheaply and effectively measure water stores relative to volumes of flow downstream near assets at risk, for a range of intervention types, intervention settings and storm conditions. This has the potential to measure the real-time contribution of NFM to flow and flood mitigation. For smaller interventions this is a much more direct measure of their efficacy than modelling could ever be. The //Smart: system facilitates this through Internet-of-Things connected electronic monitoring of storage and its influence downstream using self-build FreeStation technology. However this is not possible for interventions that have not been installed yet so some combination of measuring and modelling may be necessary.

An alternative to monitoring and modelling is space-for-time substitution. A large number of NFM interventions have been installed around the world in similar settings and thus, in many cases, finding an analogous NBS in a similar setting to that proposed can provide a good indication of the likely efficacy of the proposed NBS. However this requires similar (and well known) conditions between the analogue and the proposed intervention as well as a number of years of data on the effectiveness of the existing intervention at times of flood. Some ‘scaling’ may be needed to account for differences in scale or context between the analogues. Databases of NBS such as the Oppla platform can facilitate this process.

4.4 Conclusion

This chapter has introduced newly developed methods for assessing the effectiveness of nature based solutions for flood mitigation, with a focus on low cost and open-access solutions. It highlights the opportunities and challenges of different monitoring and modelling methods, from physically based scenario models such as the Eco:Actuary PSS for assessing impacts of climate change and green infrastructure on flood risk to assets around the world, to using networks of low cost environmental monitoring stations for direct measurement of effectiveness of specific interventions. The latter has been used by a number of citizen-led Flood Action Groups in the UK to assess the effectiveness of the NFM interventions. More detailed applications of these methods are described in Chap. 12.

Empirical and physically based modelling is best suited to large scale strategic assessment of the optimal type, magnitude and locations of interventions whereas assessment of the effectiveness of specific interventions is best achieved through a measurement approach or – where that is not possible – space-for-time substitution with comparable sites for which the intervention has already been applied and its effectiveness tested.

NFM will be an important component of the global toolkit for future flood management in order to address persistent small-scale flooding on smaller rivers that are beyond the reach of governmental grey infrastructure flood protection schemes and also contribute to ‘green-grey’ solutions for larger rivers, recognising the many co-benefits of NFM for nature and society. It is, however, vital that NFM interventions are carefully designed and fully tested for their effectiveness and cost-effectiveness so as not to waste valuable flood mitigation resources and to ensure that these solutions do not themselves generate further problems. Only through accessible, low cost approaches can this be achieved since each NFM project is usually rather small scale, low budget and unique. Generic or expensive engineering assessments such as those carried out for some grey (engineered) infrastructure investments may not be suitable.