PREFACE

How the water bomb in Limburg led to adaptation of the Sellmeijer calculation rule

You are reading the nineteenth edition of Water Matters, the knowledge magazine of trade journal H₂O. You will find four studies on a variety of topics, written by water professionals based on solid research. The editorial board, consisting of experts from the sector, made a selection based on a clear relationship with daily practice in the water sector, which is the purpose of Water Matters. Research, results and findings form the basis for articles that describe new knowledge, insights and technologies with a view to practical application.

The first study focuses on the assessment of piping in Limburg using the Sellmeijer calculation rule. During the July 2021 high water, fewer cases of piping were observed than expected. This discrepancy raises questions about the assessment method for dykes in Limburg. Sewerage research received a lot of attention during the COVID crisis. But research on drug residues in sewage has been conducted for several decades. How can we use this to determine the extent and distribution of the use of (illegal) drugs? Aquatic ecosystems in densely populated deltas like the Netherlands are under pressure from overexploitation and pollution. Climate change adds to this. Better monitoring and modelling of climate effects is needed to make more reliable quality predictions. In anticipation, the researchers recommend measures to make the water system more climate-robust by improving sponge action and shading watercourses more.

For drinking water company PWN, the IJsselmeer, also known as the national rain barrel of the Netherlands, is the most important source of water. To safeguard the source also in the future, the drinking water company wants to construct additional storage basins and pre-purify the IJsselmeer water in a ‘purifying landscape’. What might such a landscape look like?

Water Matters, like the trade magazine H₂O, is an initiative of Royal Dutch Water Network (KNW), the independent knowledge network for and by Dutch water professionals. KNW members receive Water Matters twice a year free of charge as a supplement to their journal H₂O. The publication of Water Matters is made possible by leading players in the Dutch water sector. These Founding Partners are Deltares, KWR Water Research Institute, Royal HaskoningDHV and Stichting Toegepast Onderzoek Waterbeheer (STOWA). With the publication of Water Matters, the participating institutions aim to make new, applicable water knowledge accessible.

The English-language articles can be shared from the digital magazine on H₂O-online. Furthermore, articles from previous editions can be found on the site. Would you like to respond? Let us know via redactie@h2o-media.nl

Monique Bekkenutte Publisher (Royal Dutch Water Network)
Huib de Vriend Chairmanof the editorial board Water Matters

Assessment of piping in the northern Meuse Valley in Limburg, the Netherlands

Piping is a critical failure mechanism in dikes, whereby water flows under the dike, taking sand grains with it. It can lead to dike failure. In the Netherlands, piping is assessed using the Sellmeijer design rule. However, fewer cases of piping were observed during the Limburg floods of July 2021 than expected. This discrepancy raises questions about the assessment method for dikes in Limburg.

When piping occurs, a channel (pipe) is created under the dike by backward erosion. The water, containing sand particles, rises on the inside of the dike. Piping can lead to weakening and eventually failure of the dike (see Figure 1). The place where water and potentially sand particles flow out landside of the dike is called a sand boil.

Figure 1. Development of the piping failure mechanism ‎[1]

Using the Sellmeijer design rule, it is possible to calculate the gradient or water level critical to piping. The rule, developed in 1988 ‎[1], is based on a mathematical model. The current version (2011) was optimised based on a series of experiments, which improved accuracy ‎[3] but at the same time limited the scope of application. It is only applicable to dike sections with a grain size (d70) of the piping-sensitive soil layer between 150 and 430 μm.
Sellmeijer’s rule is set out in the Statutory Assessment Instruments for Primary Flood Defences [Wettelijk Beoordelingsinstrumentarium] (WBI2017) ‎[9] and is as follows:


: the critical gradient over the dike or the critical water level for piping
: a reduction factor that takes into account the resistance of the dike material
: a safety factor that provides additional security against unexpected circumstances
: a geometry factor describing the influence of the shape and dimensions of the dike.

Limburg as an exception

The problem with the current version of the Sellmeijer design rule for piping is that it does not hold for Limburg. The rule is conservative, meaning that the risks are overestimated. The soil in Limburg consists in many places of coarse sandy and gravelly layers, which have a higher resistance to piping. The water board therefore applies an additional multiplication factor of 1.8 to allow for a less conservative critical water level ‎[5]. This enables the water board to still estimate the risk of piping.

July 2021: less piping than expected

In July 2021, extreme rainfall flooded the floodplains in Limburg, with river water reaching the crest of the dike in some locations. The dikes were not designed to cope with such a high water level, leading to the expectation of many sand boils. However, only 13 boils were reported along the Meuse in Limburg and Brabant ‎[4]. This raises the question of whether the current assessment method is adequate in Limburg. Research has therefore been carried out to evaluate the applicability of the Sellmeijer design rule (including the additional factor of 1.8) in the northern Meuse Valley in Limburg and improve it if necessary.

Research sites

The study included four dike sections, at Well, Hout-Blerick, Buggenum and Thorn. These sites are potentially vulnerable to piping for two reasons: 1) the thin top layer that can rupture and 2) the presence of a thin fine-grained layer above a package of highly permeable soil.
We performed both analytical calculations and calculations with a numerical groundwater model. The groundwater models were developed based on the soil layer classification of the study sites using the COMSOL Multiphysics software ‎[6]. They were then calibrated using head measurement series from the July 2021 floods. By following this process, efforts were made to develop the best possible model for each site, in which the head calculated using COMSOL corresponds as closely as possible to water pressure and monitoring well data (Figure 2).

Figure 2. Calibration of the COMSOL groundwater model for the Buggenum site using water pressure (wsp-14) and monitoring well data (pb-14).

Piping in the groundwater model

The groundwater model ‎[7] assumes a steady state between two factors:
the horizontal pressure gradient on the sand grains at the bottom of the pipe (caused by the hydraulic gradient, i.e. the slope of the pipe)
the rolling resistance of the grains.
The model calculates the pipe height (expressed in number of sand grains) at which the pressure gradient exceeds the rolling resistance of the grains and the grains start moving, assuming piping occurs at this time.
Van Esch et al. (2013) ‎[8] have shown that the pipe height for dikes ranges from 3 to 30 grains. This study therefore assumes that piping can occur when the average pipe gradient exceeds the critical gradient for a realistic pipe height, where a realistic pipe height is at least 3 grains.

Results

For the Well, Hout-Blerick and Thorn sites, the outcome of the model calculations is that piping can only occur at a pipe height of one or two grains, or the smallest possible pipe height. According to both the literature and common sense, such theoretical pipe heights do not lead to piping.
The outcome for Buggenum is that piping can occur at a pipe height of 4.5 grains or smaller. A small sand boil was observed near Buggenum in July 2021. This indicates that the critical gradient, at which sand grains are carried away from the aquifer to ground level, can indeed occur here and so too, in theory, can backward erosion. The calculated pipe height is low compared to observations in field tests, which may also apply to the risk of piping.

Improving Sellmeijer’s rule for Limburg

The consistency between the observed water levels and sand boils, the calibrated Buggenum model and the literature leads to the conclusion that the Sellmeijer concept does essentially hold for the piping problem in Limburg. This means that an analytical calculation rule can be formulated to describe the problem. However, according to the Statutory Assessment Instruments [Wettelijk Beoordelingsinstrumentarium] (WBI2017) ‎[9], the grain sizes of the dikes at the study sites in Limburg fall outside the scope of application of Sellmeijer’s design rule (150 μm ≤ d70 ≤ 430 μm).
To expand the scope to include Limburg, a stochastic analysis was carried out. This involved assessing 1000 randomly generated dike sections for piping, using the proposed COMSOL model. These dike sections consist of collections of model parameters randomly selected from a realistic range. In terms of grain diameters, the range for this exercise was set at 100 to 900 μm. This range was selected to ensure that the grain size corresponding to the soil composition of the northern Meuse Valley in Limburg now falls within the scope of application. The scope is therefore larger than for the current Sellmeijer design rule.

Figure 3. Linear regression between Sellmeijer scale factor (Fs) of the stochastic analysis (FS,stoc) and associated with the critical gradient (FS,eq), leading to the derivation of factor 1.56

A linear regression analysis of the results of this stochastic simulation shows that the Sellmeijer design rule can be applied to dikes with parameter values within the scope of application broadened by a factor of 1.56. This factor must be added to the original 1998 Sellmeijer design rule ‎[2]. The reason for this is that the factor was derived solely on the basis of a mathematical model and is more broadly applicable than the 2011 revised version. The improved Sellmeijer design rule is then as follows:

New design rule in practice

The improved design rule can be applied not only in Limburg, but to all dikes whose parameter values fall within the adjusted scope (grain size 100-900 μm). The study shows that it is possible to assess piping analytically for dike sections outside the scope of application of the current Sellmeijer design rule. This addresses the need for a conventional assessment method for piping.
Further research should be carried out to determine the extent to which choices in schematic representation and the scarcity of measurement results and observations affect the predicted probability of piping.

Practical relevance

Around ten years ago, an inspection round of the dikes along the Meuse in Limburg indicated that piping was regularly a major failure mechanism [8]. This was a contributing factor in the decision to reject a significant number of flood defences. The most recent round in 2022 found that 20% of primary flood defences in Limburg are non-compliant, while compliance is uncertain for another 14%. Given that dike reinforcement is very costly, this means that a more accurate assessment of the probability of piping could result in considerable savings.

This article is based on final thesis research carried out at TU Delft on behalf of Arcadis‎[11]. To read the research in full, visit: http://resolver.tudelft.nl/uuid:b3f1fa37-480f-4cc6-9fc5-a6f6e3eed497.

Sanne van Dijk
(TU Delft)
Rimmer Koopmans
(Arcadis)
Juan Aguilar Lopez
(TU Delft)

Background picture:
Sandbags on the dike near Buggenum during the Limburg floods of July 2021

Summary

In the Netherlands, the probability of dike piping is assessed using the Sellmeijer design rule. A modified version is used for the Meuse Valley in Limburg due to a different soil composition. However, much less piping than expected was observed during the extreme floods in Limburg in July 2021. This study investigated this discrepancy, leading to a further adjustment of the design rule. This may lead to fewer dike reinforcements in Limburg, but further research is needed.

Literature

1. Deltares. (2013). Onderzoeksrapport Zandmeevoerende Wellen [Research report on sand boils]

2. Sellmeijer, J. (1988). On the Mechanism of Piping under Impervious Structures

3. Sellmeijer, J., J, L., S, L., van Beek, V., & Knoeff, H. (2011). Fine-tuning of the backward erosion piping model through small-scale, medium-scale and IJkdijk experiments. European Journal of Environmental and Civil Engineering, 15:8, 1139-1154. doi: 10.1080/19648189.2011.9714845

4. ENW. (2021). Hoogwater 2021 Feiten en Duiding [2021 Flood Facts and Interpretation] (Vol. 2). Retrieved from https://www.enwinfo.nl/publicaties

5. Van Beek, V. (2018). Grind en grindhoudende lagen in de Maasvallei [Gravel and gravel-bearing layers in the Meuse Valley]. Deltares memo 11202002-002- GEO-0002. Deltares [Unpublished].

6. COMSOL AB, 2019. COMSOL Multiphysics v. 5.4. reference manual

7. Aguilar-López, J., Warmink, J., Schielen, R., & Hulscher, S. (2016a). Piping erosion safety assessment of flood defences founded over sewer pipes. European Journal of Environmental and Civil Engineering, 22(6), 707-735. doi: 10.1080/19648189.2016.1217793

8. Van Esch, J., Sellmeijer, J., & Stolle, D. (2013). Modelling transient Groundwater Flow and Piping under Dikes and Dams. 3rd International Symposium on Computational Geomechanics (ComGeo III)

9. Deltares (2018). Basic report WBI 2017, v1.2

10. HKV (2024). Wettelijke beoordeling van de waterkeringen langs de Limburgse Maas [Statutory assessment of flood defences in the Limburg Meuse region]. Url: https://www.hkv.nl/actueel/wettelijke-beoordeling-van-de-waterkeringen-langs-de-limburgse-maas

11. Van Dijk, S.E. (2023). Sellmeijer in the northern Maasvallei. TU Delft Master Thesis. Civil Engineering & Geosciences.

^ Back to start

Wastewater: information source for the Dutch illicit drug market

Wastewater is a mirror of society. Wastewater analysis of traces of the COVID-19 virus was widely covered in the media. This revealed insight into geographical and temporal trends of infections in recent years. However, wastewater has been analysed for drug residues for two decades now. How can we use this information to determine the of illicit drug market?

Insight into illicit drug use is of social importance due to its impact on public health and on criminal activities related to production, trafficking and use. Population surveys and the interception of illicit drug trafficking provide valuable information on the consumption of illicit drugs in society as a whole. However, this data is less suitable for determining the local scale of consumption and market [1]. Drug residues in wastewater are more useful for this purpose. Wastewater analysis is now used on a structural basis to measure and compare drug use in regions or cities [2]. This study shows how illicit drug consumption in the Netherlands and the financial scale of the market can be estimated using wastewater data, population statistics, drug excretion and street prices.

Estimating the market in seven steps

Figure 1 illustrates the seven steps used to estimate the scale of the national illicit drugs market based on their residues in wastewater. In this study, 30 wastewater treatment plants, to which a total of 20% of the Dutch population is connected, were sampled daily for one or more weeks between 2015 and 2022 (step 1). This was done with the assistance of several water authorities and municipalities. The sampling weeks were deliberately scheduled outside public holidays, school holidays and major events because illicit drug use is expected to deviate from the 'baseline consumption' at those times. In step 2, the samples were tested to determine concentrations of residues of amphetamine (also known as speed), 3,4-methylenedioxymethamphetamine (MDMA, active ingredient of ecstasy), and benzoylecgonine (the human metabolite of cocaine). The measured concentrations were then multiplied by the daily wastewater volume, which was used to calculate the load per thousand residents at the respective wastewater treatment plant (step 3). The result was used to calculate consumption, by dividing the load by the excreted fraction by humans (step 4). The financial scale of the local market was then calculated based on the street quality (percentage of cocaine and amphetamine in powders) or dosage (MDMA in pills) and street price (step 5). Finally, a link was established between drug residues per 1,000 connected residents and the degree of urbanity (step 6). Taking into account the relationship between use and degree of urbanity, this was then extrapolated to the national market (step 7).

Figure 1. Seven steps to get from drug residues in wastewater to a picture of the national illicit drugs market [3]

Results

A total of 392, 364 and 357 unique 24-hour wastewater samples were selected for analysis of benzoylecgonine, amphetamine and MDMA, respectively. For MDMA and amphetamine, several measurement weeks were excluded due to suspected or identified residues of waste from illicit drug production in the wastewater.

Per capita amphetamine loads showed no significant relationship with urbanity, while for MDMA, 31% of the variation could be explained by urbanity and for cocaine as much as 64% (Figure 2). The monetary value of the three illicit drugs studied was estimated at €903 million (95% CL = 829-987 million) per year based on the 2022 price level (Figure 3).

Figure 2. Volumes of cocaine (left), amphetamine (centre) and MDMA (right) residues per 1,000 residents per day in relation to the urbanity of the wastewater treatment plant's treatment area. The dots are weekly averages, the lines the fit model. Solid lines show significant trends.

Figure 3. Estimated annual illicit drug use in the Netherlands in volume and euros

Discussion

About our estimate
This estimated scale of the illicit drug market is subject to uncertainty because the measurements do not cover the whole of the Netherlands and are based on random samples from just one or a small number of measurement weeks per location. However, long-term studies show little variation between weeks and the extrapolation to the whole of the Netherlands also takes into account correlations between use and degree of urbanity. This suggests that the sample is representative.

However, not all residues of consumed drugs may reach the wastewater treatment plant. Some wastewater might be lost along the way due to leaking sewers and not all drug residues might end up in the sewers due to public urination. In addition, the sampling weeks do not provide insight into potentially higher use during events and festivities. Together, this leads to a conservative estimate of usage and market size for the studied period from 2015 to 2022.

Comparison with other Dutch estimates
Illicit drug use can also be estimated from prevalence data (information on the proportion of Dutch people who use these substances occasionally, regularly or often). Based on this method, Baarsma et al. [4] estimated an annual national consumption of 8,750 kg of street-quality cocaine, 4,500 kg of street-grade amphetamine and 8 million
MDMA pills (around 1,300 kg of pure MDMA). This is just over half (57% for cocaine, 52% for amphetamine and 56% for MDMA) of our estimate of consumption between 2015 and 2022. This can be considered comparable given the uncertainties in both estimates.

Comparison with European estimates
If we compare the presumed market size of the Netherlands with international estimates based on customs, police and prevalence data, the estimated Dutch market accounts for 7.3% (cocaine), 6.3% (amphetamine, not including methamphetamine) and 11.6% of the total EU financial market size pre-Brexit. This would appear realistic, as the Netherlands generates 3.4% of the EU population (January 2020) and 6.0% of the gross European product, and the studied illicit drugs are also known to be relatively widely used in the Netherlands compared to other European member states [5].

Future application

Geographically defined quantitative illicit drug consumption data enables care providers, policymakers and law enforcers to focus municipal or regional policy on the most relevant illicit drugs. In addition, wastewater analysis can help evaluate drug policies, for example if measures are taken to discourage use and limit trafficking, or instead regulate trade as envisaged for cannabis in the ‘Experiment gesloten coffeeshopketen’. As part of this experiment, ten Dutch municipalities control and monitor the entire cannabis product distribution chain.

Wastewater analysis compliments other tools for mapping drug trafficking and consumption and their social and health effects.

Thomas ter Laak
(KWR; Institute for Biodiversity and Ecosystem Dynamics [IBED], University of Amsterdam)
Pim de Voogt
(KWR; Institute for Biodiversity and Ecosystem Dynamics [IBED], University of Amsterdam)
Erik Emke
(KWR)
Emiel van Loon
(IBED, University of Amsterdam)

Background picture:
Sampling for illicit drug analysis at Utrecht wastewater treatment plant

Summary

Wastewater is a mirror of society: it shows us, for example, where viruses are spreading or what medicines are being used. Wastewater analysis can sometimes be a solution when collecting data by other means is difficult. It is therefore used on a structural basis to measure and compare illicit drug use of regions or cities. Although accurately calculating the total market size still presents a complex challenge, this article describes how the market volume and its financial value can be estimated by combining local wastewater data with population statistics, human excretion rates and street prices of the drugs. The article also compares the results with previous estimates at national and European level.

The social impact of drug trafficking and use makes the results of this type of wastewater analysis highly relevant to public health and enforcement.

Literature

1. Zuccato E. et al. Cocaine in surface waters: a new evidence-based tool to monitor community drug abuse, Environmental health : a global access science source 2005: 4: 14.

2. Thomas K. V. et al. Comparing illicit drug use in 19 European cities through sewage analysis, Science of the Total Environment 2012: 432: 432-439.

3. ter Laak T. L. et al. Mapping consumptions and market size of cocaine, amphetamine and MDMA through wastewater analysis: A Dutch case study, Addiction 2024: 16649.

4. Baarsma B. et al. Drugs de baas, Hoe Nederland zijn drugsprobleem onder controle kan krijgen [Drugs in charge, How the Netherlands can take control of its drug problem], Netherlands: Denkwerk; 2022, p. 56.

5. EMCDDA. European Drug Report 2023; Trends and Developments. In: Publications Office of the European Union, editor, Lisbon, Portugal: European Monitoring Centre for Drugs and Drug Addiction 2023.

^ Back to start

Climate change: additional pressure on surface water quality

Aquatic ecosystems in densely populated deltas such as the Netherlands experience pressures from overexploitation and pollution. On top of this, meteorological variations are a major driving force behind water quality dynamics. We anticipate higher water temperatures and more extreme wet and dry periods in the future. How is this expected to affect water quality?

To start to answer this question, we looked at the trends in water quality over the past three decades for the Water Board Aa en Maas [1]. For this, we used the historical measurement series of this water board's water quality monitoring network. These data series consists mostly of measurements in free-draining watercourses in sandy areas with a lot of agricultural activity and some urban centres. At a large part of the monitoring sites, surface water consists of local runoff, mainly from agricultural areas. In addition, there are sites where inlet water and effluent from sewage treatment plants (STPs) also contribute a significant part of the discharge.

Setting up trend analyses

We looked for trends over the period 1990-2022 in various water quality parameters: nutrient concentrations and chemical and physical properties, such as electrical conductivity (measure of total dissolved salts), water temperature and acidity. For other substances, such as heavy metals and organic micropollutants, the time series were not long and/or consistent enough for trend analyses.

The trend analyses are based on two key principles. The first is that the trend analyses were first conducted per measurement site, and then aggregated to group and water board level [2, 3]. This ensures that the spatial variation in water quality does not result in greater uncertainty in the trend analyses.
The second is that multiple robust statistical methods were used, with little or no sensitivity to outliers and gaps in the datasets. We used three methods: (1) the Seasonal Mann-Kendall (SMK) trend test, which detects any significant upward or downward trend for the whole series; (2) the Theil-Sen slope estimator, which determines the median trend slope for the series (i.e. shows the strength of the trend); and (3) the LOWESS (Locally Weighted Scatterplot Smoothing) trend line, which shows changes in trends as a kind of running median.

General trends

Over the period considered, the various trend analyses showed a general improvement in water quality, particularly for the nutrients nitrogen (N-total) and phosphorus (P-total). Downward trends were also observed for all nitrogen and phosphorus sub-parameters (nitrate, ammonium, organic nitrogen, phosphate and particulate phosphorus). This picture was also maintained in separate analyses for summer and winter concentrations, for shorter series up to 2018 (excluding subsequent dry years) and from 2000 onwards, for separate subareas and water types, and for sites with and without the influence of STP effluent and intake water. The assumed main causes of improved water quality are: less leaching from agriculture, better water treatment and cleaner intake water.

Figure 1. Aggregate LOWESS trend line and the 25 and 75 percentile LOWESS trend lines (broken lines) for N total (Ntot), P total (Ptot), temperature (T) and chloride (Cl). The grey lines are the lines per site, the coloured lines are the aggregated LOWESS trend lines.

The contribution of climate change to the improvement in water quality cannot be clearly isolated from the above changes. However, climate change is probably mainly responsible for the observed increase in water temperature. The LOWESS trend analysis (Figure 1) also revealed a clear acceleration in surface water warming. Since 2010, this warming has been accelerating three times faster on average: since 1990, water temperatures have risen by up to 0.1°C per year and as much as 0.3°C per year since 2010. The result is an average water temperature that was 2.5°C higher in 2022 compared to 1990. Higher water temperatures make the water system more susceptible to algal blooms, proliferating aquatic plants and oxygen depletion, as warmer water can contain less oxygen.

Air temperature increased at a slower rate over the same period of 1990-2022 (approx. 0.04 oC per year) and showed no accelerated warming from 2010. Lower discharges and longer retention times due to climate change can cause additional warming of water, particularly in summer, as solar irradiation per litre of water increases. Transparency has also improved in many surface waters, allowing this irradiation to penetrate further into the water column.
The longer summer retention times may also help to explain the observed upward trend of chloride since 2012 (see Figure 1), due to upconcentration caused by more evaporation. Human activity may also cause additional high water temperatures, for example the additional intake of Meuse water in dry summers (Meuse water is often warmer than the local water).

Weather extremes

The data analysis also considered the impact of weather extremes. Heavy rainfall leads to higher concentrations and large loads of nutrients to receiving waters. However, the response of chemical water quality parameters to peak rainfall events is often short-lived (several hours to days). This is too short to be reflected in the monthly measurements, except for a few specific cases.

The effect of dry periods can be identified better from the monthly measurement data than the effect of short-term wet extremes. For example, the 2018-2020 drought led to lower than average nutrient concentrations in surface water in summer. This is probably due to less leaching of nutrients from agricultural plots and the increased self-cleaning capacity of watercourses (retention) due to longer retention times. Conversely, the subsequent wet winter season and the wetter summer of 2021 led to higher than average nutrient concentrations. The leaching water is then loaded with additional nutrients due to the long preceding dry period with hardly any leaching. Reduced crop uptake due to drought also played a role, allowing more nutrients to leach out with the same fertilisation regime.

To better understand the effect of weather extremes, a distinction was made between dry, average and wet sampling times for the summer months. In this analysis, a cumulative rainfall deficit of at least 40 mm counts as a dry period (10% of samples) and a 10-day running average of net rainfall of at least 4.5 mm per day counts as a wet period (3% of samples). During extremely wet conditions, nutrient concentrations are higher, while oxygen concentrations are lower (Figure 2). Low oxygen concentrations during wet conditions can cause immediate problems for local aquatic ecology. Peak nutrient loads threaten downstream waters and contribute to eutrophication (algal blooms, proliferating aquatic plants, oxygen depletion).

Figure 2. Boxplots showing mean and median concentrations of nitrate (NO3-N) and oxygen (O₂) for extremely dry and wet situations in the summer half-year and for the average situation.

Comparison with other studies

Increased concentrations of substances after heavy showers are also found in another study, with measurements from the Vinkenloop: an almost entirely agricultural catchment area near Westerbeek in North Brabant [4]. Water quality has been continuously monitored in this area since 2021, revealing short concentration peaks of substances such as total P and ammonium. The National Network for Monitoring the Effects of Manure Policy [Landelijk Meetnet effecten Mestbeleid] [2] and the trend study by Hallmann et al. [5] also show relatively low nitrogen concentrations in surface water during dry years, followed by increased leaching during a subsequent wet period.

A striking and potentially alarming finding was the three times faster increase in water temperature since around 2010. Similar accelerated warming of surface water was also previously observed in the Wetterskip Fryslân management area [3].

Recommendations

The analyses show that measuring on a monthly basis is too infrequent to determine the short-term effects of peak rainfall. To investigate the climate impact of more frequent and more intense peak showers on water quality, it would be better to start high-frequency measurements in some areas (such as the outflow points of the largest streams). For a wider range of parameters, it is also possible to set up autosamplers that automatically take samples during and after peak rainfall events. Such measurements also help interpret monthly measurement results in the rest of the monitoring network and are valuable for calibrating and validating models.

For correct modelling, it is important to identify which model adjustments in relation to the most relevant hydrological and biochemical processes are needed to properly simulate the impact of climate change. An accurate model will then make it easier to distinguish between the impact of climate and other influences, such as manure policy and improved water treatment. In addition, models can be used to explore the impact of measures to make the water system more climate resilient.

Measures that help buffer heavy precipitation in the soil, groundwater and surface water systems are also expected to have a positive impact on water quality. Restored water retention in the landscape prevents large nutrient losses during weather extremes, gives crops more time to absorb water and nutrients, and increases uptake and degradation in the surface water system. In addition, providing shade over watercourses can help regulate and reduce water temperatures in summer.

Kim Gommans
(Deltares)
Joachim Rozemeijer
(Deltares)
Luuk van Gerven
(Aa en Maas Water Board)

Summary

This study focuses on trends in surface water quality. Historical data from the Aa en Maas Water Board shows a general improvement in water quality since 1990, but also rising water temperatures. This makes the ecosystem more vulnerable to eutrophication and oxygen depletion. Thanks to climate change, the likelihood of weather extremes is also increasing. This has a demonstrably negative impact on water quality. One of the effects of dry summers is lower nutrient concentrations, while higher concentrations occur in subsequent wet periods. Precipitation extremes cause oxygen dips and large nutrient loads to downstream waters. Better monitoring and modelling of climate effects is needed to make more reliable quality predictions. In anticipation of this, we recommend measures to make the water system more climate resilient by improving water absorption and providing more shading over watercourses.

Literature

1. Rozemeijer, J., Gommans, K., Van Gerven, L. (2024). Waterkwaliteit van de toekomst. Data analyse naar de invloed van klimaatverandering en weersextremen in beheergebied waterschap Aa en Maas. [Water quality of the future. Data analysis of the impact of climate change and weather extremes in the area managed by the Aa en Maas Water Board.] Deltares report 11209190-024-ZWS-0001.

2. Buijs, S., Ouwerkerk, K. & Rozemeijer, J. (2020). Meetnet Nutriënten Landbouw Specifiek Oppervlaktewater: Toestand en trends tot en met 2018 [Nutrient monitoring network in agriculture specific surface water: Status and trends up to and including 2018]. Deltares report 1203728-005-BGS-0002, Utrecht.

3. Buijs, S., Ouwerkerk, K. & Rozemeijer, J., Hooijboer, A. (2021). Trends waterkwaliteit in het beheergebied van Wetterskip Fryslân in de periode van 2000 tot en met september 2020 [Trends in water quality in the Wetterskip Fryslân management area in the period 2000 to September 2020]. Deltares report 11206260-002-BGS-0003, Utrecht.

4. Schipper, P., Groenendijk, P., Van Gerwen, L., Van Loon, A., Lukács, S., Rozemeijer, J. (2022). Monitoring en modellering in twee pilotgebieden voor gebiedsgerichte aanpak [Monitoring and modelling in two pilot areas for an area-based approach]. STOWA report KIWK 2022-22.

5. Hallmann, C., Van der Pol, J., Brugmans, B. (2021). Trends en toestand ecologische, fysische en chemische parameters Aa en Maas: Effecten van inrichting en beheer & onderhoud [Trends and status of ecological, physical and chemical parameters in Aa en Maas: Effects of design and management & maintenance]. In WUR (no. 525641).

^ Back to start

A 'purifying landscape' in Lake IJssel (IJsselmeer) for drinking water production and biodiversity

For drinking water supplier and nature manager PWN, Lake IJssel (in Dutch: IJsselmeer) is the main source of freshwater. However, the availability of freshwater is under increasing pressure due to climate change. As a Nature-based solution, PWN wants to build additional water storage reservoirs and pre-treat Lake Lake IJssel water in a 'purifying landscape'. What could such a landscape look like?

Lake IJssel is affected by climate change: the rivers regularly carry less water, pollution and salinisation are increasing, and there has been a considerable reduction in ecological quality. At the same time, demand for water is rising. This leads to challenges regarding the availability of freshwater of sufficient quality for drinking water production. In response, PWN is working on a sustainable solution consisting of two elements: additional water storage reservoirs and a 'purifying landscape' in Lake IJssel. The reservoirs will allow more water to be stored, enough for 50 days, which is used when the regular intake water from Lake IJssel is of insufficient quality. In the purifying landscape, nature itself will pre-treat Lake IJssel water.
The initiative is part of the policy programme: ‘Programmatische Aanpak Grote Wateren' (PAGW) (in English: Systematic Approach River Lake and estuary Ecology) [1]. In an earlier stage an impact assessment was conducted that showed the added value for water quality and nature development in Lake IJssel [1]).

Purifying landscape

As most of the shorelines of Lake IJssel are made of basalt stones with a low ecological value [2], the construction of the purifying landscape will increase the surface of suitable habitats for fish and birds with 150 hectares. Reedbeds and submerged vegetation provide new fish spawning grounds, bird nesting sites, and a habitat for a biodiverse macroinvertebrate community. Important ecotopes, such as reed marshes, floodplain grassland and shallow water with vegetation only cover a small area of Lake IJssel. The PAGW therefore aims to increase the surface area of these ecotopes [1].
The objective of our research was to visualize the potential barriers and opportunities of a purifying landscape that contributes to both biodiversity and sustainable drinking water production. What does such a landscape look like?

This objective can be divided into multiple questions:
• What should be the most optimal size of each of the preferred ecotopes (reed marshes, submerged vegetation and floodplain grassland) and how should these ecotopes be connected to contribute to both biodiversity and pre-treatment of surface water?
• Which environmental conditions are needed for these ecotopes to develop?
• What physical, chemical and biological processes in these ecotopes can contribute to the purification function for drinking water production?
• What are the seasonal dynamics in ecology and water quality? In other words, what role can the purifying landscape play for birds and fish in each season, and how is this related to different water quality parameters throughout the year?
• How can the exchange of nutrients and organic matter take place between the landscape and Lake IJssel?

Research design

The barriers and opportunities of the purifying landscape were first explored by organizing workshops with experts and stakeholders and by an analysis of the available literature. The workshops used a landscape design approach to develop outlines from various angles, such as purification, fish, birds and management. Participants included experts and practitioners from various fields such as water management, technology, ecology. They explored the trade-offs and synergies between the different functions of the purifying landscape.

Once the various outlines had been evaluated, they were then combined into a functional schematic design. Finally, the purification potential of this design was explored.

Result: functional design

The design process revealed two key criteria. Firstly, the landscape should be a suitable habitat for fish and birds year-round. Secondly, the purification of the landscape should focus on nutrients, suspended solids and dissolved organic compounds. Thus, the landscape should be heterogeneous and accessible for fish and macroinvertebrates, and provide a place for passive and active filtration and conversion of suspended and dissolved compounds. The functional design outlines an area of roughly 150 hectares, which can be created using the sediment obtained from the proposed adjacent storage reservoirs. It consists of four zones: the inlet zone, the marsh zone, the purifying reed bed and the outflow zone (Figure 1). Three ecotopes are represented in the marsh zone: ‘Floodplain grassland’, ‘Reed marshes’, and ‘Aquatic plants’ [1].

In the relatively sheltered inlet zone, mussel beds filter suspended solids and algae from the incoming Lake IJssel water. Mussel faeces allow a macroinvertebrate community to develop here, which, along with the mussels, can serve as food for fish and birds. A fish-friendly pump then transfers the water to the marsh zone. This is a heterogeneous landscape, in terms of both space (differences in water depth, substrate and sediment) and time (natural water level, being low in summer and high in winter). Natural water level dynamics are necessary for reproduction and growth of characteristic fish and plant species. This zone will therefore have its own natural water level management independent of that in Lake IJssel, which encourages the germination of plants such as reeds, as well as ensuring a flux of organic material out of the landscape. This prevents succession towards a terrestrial ecosystem and reduces the need for management, in turn resulting in less disturbance to birds.

The marsh zone gradually flows into the purifying reed bed in the form of a free water surface (FWS) helophyte filter. More homogeneous vegetation with cattail and reed grows in this zone, which can be carefully harvested. After this helophyte filter, the pre-treated Lake IJssel water can flow into the water storage reservoirs or to the outflow zone back into Lake IJssel. Although less biodiverse than the marsh zone, this purifying reed bed also has ecological benefits.

The connection between the purifying landscape as a whole and Lake IJssel is very important. Organic matter can return from the purifying landscape through a fish passage via the outflow zone, into Lake IJssel. This will make the outflow zone an area rich in organic matter, which can serve as the basis of the food web for birds and fish. Parts of the outflow zone can be designed as a horizontal helophyte filter/bank filtration to achieve further purification.

Figure 1. Overview of the Purifying Landscape from above (top) and as a cross-section (bottom) showing the inlet zone, the marsh zone, the purifying reed bed and the outflow zone, and the direction of movement of water (white arrows) and fish (blue arrows).

Purification capacity

Purification in the landscape mainly takes place through biological processes, such as nutrient uptake and nutrient conversion by plants and micro-organisms [3] and filtration and conversion of organic matter by invertebrates [4]. Sedimentation and soil filtration also contribute to potential purification.

Exploratory calculations show that nutrients and BOD (biological oxygen demand) are only removed to a modest extent. Assuming that the entire landscape acts as a FWS helophyte filter, P, N and BOD removal is around 4%, 3% and 5% respectively, at a hydraulic retention time of 1.8 days. This retention time is relatively short, due to the small surface area of the helophyte filter. A FWS helophyte filter is usually used as the last purification step for wastewater (tertiary effluent), in which case the average retention time is 6.7 days [5]. A removal rate of more than 80% could then achieved [6].
However, the nutrient and BOD concentrations in Lake IJssel water are already relatively low: close to the estimated background concentrations of FWS helophyte filters. Moreover, the remaining inflow of suspended solids and algae towards the storage reservoirs is low, mainly due to filtration by mussels in the intake zone and root absorption in the purifying reed bed. The risk of them entering the reservoirs is thus greatly reduced. Total purification is not the goal - the purifying landscape serves as pretreatment for the deep reservoirs and subsequent technical treatment. Every small amount of pollution the purifying landscape removes is a win, as it reduces the burden on the technical purification system at PWN.

Seasonal dynamics

Natural seasonal dynamics play a major role in both the ecology and purification of the landscape (Figure 2). Over the year, the landscape functions as breeding and spawning grounds and resting places for birds and fish. Purifying processes are also affected, through nutrient uptake in spring and summer, and decomposition of organic matter in winter. While mussels filter the water for much of the year, too much mussel larvae can cause disruption of the technical treatment. As a result, the purifying landscape will have varying purification efficiencies throughout the year.

Figure 2. Seasonal dynamics of the purifying landscape. With a schematic example of (a) pollutant concentrations in Lake IJssel water including inflow and outflow zone, (b) the purifying and ecological functions, and (c) the expected concentrations of pollutants in water outflowing to storage reservoirs or the outflow zone.

Discussion

This research shows how a multidisciplinary team can design a purifying landscape that can benefit both ecology and drinking water purification. It seems that nature development and purification can be combined in a landscape if we think in terms of both space and time: combining functions and making use of interactions between natural processes. Throughout the entire 1.5-year design phase, ecologists, engineers and water managers, working both separately and together, maintained a focus on all aspects of the landscape and incorporated these aspects into the design. For instance, on the advice of the engineers, the filtering mussels were moved to the start of the system, and the water in the marsh zone was given a braided structure on the advice of the ecologists.

Multiple questions still need to be answered to further refine the design. For instance, the purification efficiency of the whole system, merits further examination, as does the ratio of surface areas between different zones and ecotopes. Aspects such as how the connectivity between the purifying landscape and the water storage reservoirs, the outer dike area, the inlet zone and the outlet zone should be designed, also need to be further investigated in practice. Proper monitoring and management are also important to ensure that the system can be flushed if, for example, an excessive accumulation of mussel larvae or dead plant material occurs.
We hope to answer these questions in the LIFE WATERSOURCE project, which encompasses a 1-hectare pilot landscape to study how natural and seasonal variations can affect ecology and purification in practice.

With thanks to TKI Deltatechnologie, which co-funded the research.

Tom van der Meer
(Wageningen Environmental Research)
Jeroen Veraart
(Wageningen Environmental Research)
Koen Zuurbier
(PWN)
Myrthe Fonck
(PWN)
Joost Lankester
(Netherlands Enterprise Agency [Rijksdienst voor Ondernemend Nederland], RVO)
Martine van Mourik
(Wageningen Environmenntal Research)
Thomas Wagner
(Wageningen University)
Ilse Voskamp
(Wageningen University)

Summary

Drinking water supplier and nature manager PWN wants to better respond to climate change as well as improve the sustainability of its drinking water supply. To do so, it plans to construct additional water storage reservoirs and a 'purifying landscape' in Lake IJssel. The purifying landscape will pre-treat Lake IJssel water based on natural processes, while enhancing the natural value of Lake IJssel. This article describes the creation of a functional design, with input from technical knowledge from the drinking water industry and expertise from ecology and water management. Design workshops with experts and stakeholders led to a design for an ecologically rich and heterogeneous purifying landscape with a natural water level.

Literature

1. Heins et al., (2020) Werkdocument Programma Aanpak Grote Wateren. Rijkswaterstaat , Lelystad.

2. Tack et al. (2024). Food webs in isolation: The food-web structure of a freshwater reservoir with armoured shores in a former coastal bay area. Science of the Total Environment.

3. Wagner et al. (2020). Pilot-scale hybrid constructed wetlands for the treatment of cooling tower water prior to its desalination and reuse. Journal of Environmental Management, 271.

4. van der Meer et al. (2023). Removal of nutrients from WWTP effluent by an algae-mussel trophic cascade. Ecological Engineering, 190.

5. Kadlec & Wallace (2009). Treatment Wetlands 2^nd ed., CRC Press.

6. Dotro et al., (2017). Treatment Wetlands: Volume 7, IWA Publishing.

^ Back to start