Figure 2 Dorpspolders surrounded by minor dikes.

River floods (Figure 1) have been common in the Netherlands, ever since the rivers were embanked. River embankments started approximately 1100 AD, and were completed approximately 1350 AD. The first dikes were not the present river dikes, but dikes that enclosed the area of a village (Figure 2), so-called 'dorpspolders'. The dorpspolders had their own drainage systems, some until the 1950's. In the second half of the 20th century, water management areas increased in size, and became more efficient.

Figure 1 High water of the River Waal. (photo: P.C. Beukenkamp). Seepage occurs in the low lying areas protected by dikes. This could eventually lead to a dike breach if the dike is undermined by the water flow.
Figure 3 Ice dams can lead to a rapid rise of the water level.
Figure 4 Ice creeping on top of the dike between Lelystad and Enkhuizen (picture by Karel Tomei).
Figure 5 Topographic map of the Meidijk, a dike between the Waal and the Meuse (Maas). Dike breaches occurred where a sandy channel belt crosses the dike (see Figure 6).
Figure 6 Geological map (Berendsen 1986) of the Meidijk area, showing the Meidijkse wielen (dike breaches; in blue) in the Gameren channel belt (yellow).

Some of these low dikes have been preserved, but unfortunately, most have been dug away. They should be preserved as much as possible, not only because of their cultural-historic value, but also because they slow the inundation of a polder in the case of a dike breach (see animation below). In the beginning, dikes were low, and flooding occurred because the river water simply overtopped the dikes. Gradually dikes became higher and higher, and often the base of the dike was not widened. This resulted in narrow, unstable dikes, that often led to catastrophic dike breaches. Especially after rivers froze over, and the ice started to move downstream, ice jams could lead to a rapid rise of the water level (Figure 3). Subsequently, seepage increased, and dikes became undermined, or the floodwaters simply overtopped the dikes, causing headward erosion and subsequent dike collapse. In later years, ice jams were removed using explosives. Sometimes the ice crept on top of the dikes (Figure 4).

Figure 7 Aerial photograph of the Meidijkse wielen (dike breach scour holes). Picture: H.J.A. Berendsen.

The Bommelerwaard is one of the areas where dike breaches were common, because the Bommelerwaard is situated in between the rivers Waal and Maas, and floodwaters sometimes came from both rivers. For example, the N-S running Meidijk (Figure 5) that was meant to protect the area from marine floods, was breached 26 times. These breaches occurred at the location where a sandy channel belt crosses the dike (Figure 6), leading to seepage and undermining of the dike, finally resulting in the formation of dike-breach scour holes (wielen in Dutch, Figure 7). Because similar circumstances led to many dike breaches, geological-geomorphological maps can be very important to detect potentially hazardous areas, and a 'sand depth map' (Berendsen et al. 2002) can be very valuable for engineering and planning.

Figure 8 Dike breach scour hole (wiel). Picture: Province of Utrecht.
Figure 9 Dike breach scour hole (wiel) in the embanked floodplain, that became filled with sediments. Picture: H.J.A. Berendsen.
Figure 10 Artificial mound in Delwijnen, built after the 1861 flood of the Bommelerwaard. Picture: H.J.A. Berendsen.
Figure 11 Man-induced dike breach in the Betuwe, during the second World War. The dike-breach deposits are clearly recognizable.
Figure 12 High water of the River Waal. Seepage occurs in the low lying areas protected by dikes. This could eventually lead to a dike breach if the dike is undermined by the water flow. Picture: Rijkswaterstaat.
Figure 14 Three dike levels can be seen in the city of Zaltbommel. Picture: H.J.A. Berendsen.
Figure 17 Widening of the dike and adding a berm. Picture: H.J.A. Berendsen.

If a dike breach occurs, an almost circular scour hole is formed (Figure 8), that is filled with water. Dike repair may leave the scour hole either within the embanked floodplain, or within the polder (the areas protected by dikes). If the scour hole is in the embanked floodplain, it may eventually be filled up with sediments (Figure 9).

The last river flood that inundated the Bommelerwaard occurred in 1861. After this flood, mounds were built in several villages for the protection of people and cattle (Figure 10).

In 1926 the Land van Maas en Waal was flooded. This was the last river flood that led to a dike breach (with the exception of an artificial dike breach inflicted during the Second World War, Figure 11).

However, during the flood of 1995 (Figure 12) the dike near Ochten (Betuwe) almost burst, during one of the worst floods of the 20th century, and 250,000 people had to be evacuated. Figure 13 shows a near-collapse of a dike near Rotterdam. In this case, the dike was built on peat, and partly collapsed because it had not been properly stabilised.

Figure 13 Near-collapse of the Lek dike near Rotterdam. Picture: P.C. Beukenkamp, ca. 1992.

At many places, different dike levels can be seen, for example in the city of Zaltbommel (Figure 14). The city was built at level 1. Level 2 is the dike level of 1861, and level 3 is the present dike level, established after recent raising. Now the dikes are almost everywhere at the desired 'delta' level, leaving the front door of old houses along the dike at a level approximately 1 m below the present level (Figure 15). This is particularly common in the Bommelerwaard area.

Figure 15 The front door of houses built on the dikes is often 1 m below the present dike level. Picture: H.J.A. Berendsen.

After the 1995 high water (Figure 12), dikes were reinforced. This involved only a moderate raising of the dike level, but especially a widening of the dikes (Figure 16 and 17).

Figure 16 Recent reinforcements of the Lek dike.
Figure 18 The 1995 flood almost resulted in disaster (picture by Rijkswaterstaat).
Figure 20 Increase of precipitation in the Netherlands in the 20th century (KNMI).
Figure 22 Inundation of the Land van Maas en Waal during the 1805 flood (Hesselink 1992). February 13, 11 AM.
Figure 23 Inundation of the Land van Maas en Waal during the 1805 flood (Hesselink 1992). February 13, 7 PM.
Figure 24 Inundation of the Land van Maas en Waal during the 1805 flood (Hesselink 1992). February 14, 9 AM.
Figure 25 Inundation of the Land van Maas en Waal during the 1805 flood (Hesselink 1992). February 15, 5 AM.

Dike reinforcement in the Netherlands has for long been a controversial subject, because the original reinforcements involved removing houses that were built on top of the dikes, straightening and widening of the dikes, removal of trees, ponds, etc. In many places, this resulted in a dull landscape. In the 1980's it was agreed to maintain the original situation whenever possible, and to use 'keen designs' at places where old farmhouses, churches etc. have to be preserved. The 1993 and 1995 floods of the Rhine and Meuse (which almost resulted in disaster, Figure 18) helped to diminish public opposition to reinforcements.

The dikes are now designed for a chance of a water level exceeding the dike level of 1 in 1250 years. The inportance of a massive dike along the river Lek, can be deduced from Figure 19: a dike breach near Wijk bij Duurstede could potentially inundate all the major cities in the western Nehterlands.

Figure 19 Area potentially inundated by a dike breach near Wijk bij Duurstede.

As a result of the increased greenhouse effect, discharge of the Rhine is expected to increase by approximately 15 % during the winter, and to decrease during the summer by the same amount (Kwadijk 1993). So far, we have seen an increase of precipitation all over the year (Figure 20).

To accommodate higher floods, new plans have been developed. These do not necessarily involve the construction of higher dikes, because that would also increase the risks of dike bursts. Instead, the following measures are planned or being carried out (Figure 21):

Figure 21 Measures being carried out or contemplated to increase the storage capacity of the embanked floodplains and enhance drainage to the sea (Middelkoop 1998).

In this way a peak discharge of 17,000 m3/s at Lobith can be accommodated (the peak discharge measured so far is 12,000 m3/s at Lobith). It may be necessary to take additional measures, and increase the capacity to a flood of 18,000 to 20,000 m3/s at Lobith. In this densely populated area these measures will meet great resistance from the local population. However, disasters like that in New Orleans after hurricane Katrina in 2005 (Berendsen 2005), are a lesson that should not be forgotten.

Hesselink (2002) modeled the effect of a dike breach, with the help of the Delft FLS model (Hesselink et al. (2003). The model was calibrated with historic data of the moderate 1805 flood of the Land van Maas en Waal (Figures 22, 23, 24, 25). The pictures and the animation (zipped AVI, 1,72 MB) show, that minor dikes significantly slow the inundation of the polder.

A CD-ROM with more animations, made by Alkema (2003), under the supervision of Dr. Hans Middelkoop, is available from Projectbureau Belvedere. Distribution of the CD-ROM for educational purposes is free.


  1. Alkema, D. (2003), Waterstaatkundige inrichting van rivierpolders: een hernieuwde rol voor cultuurhistorische elementen. CD-ROM with animations and analyses of 28 flood simulations of the Land van Maas en Waal. Ministerie van Volkshuisvesting, Ruimtelijke Ordening en Milieubeheer, Projectbureau Belvedere and Utrecht University. (www.belvedere.nu)
  2. Berendsen, H.J.A., E.L.J.H. Faessen, A.W. Hesselink & H. Kempen (2002), Zand in banen - Zanddiepte-kaarten van het Gelders Riverengebied met inbegrip van de uiterwaarden. Sand-depth maps of the eastern part of the Rhine-Meuse delta (with a summary in English). Arnhem: Provincie Gelderland, in samenwerking met Rijkswaterstaat, Waterbedrijf Gelderland en Universiteit Utrecht, 53 p. Coloured maps.
  3. Berendsen, H.J.A. (2004), Landschap in delen. Overzicht van de geofactoren. Assen, Koninklijke Van Gorcum.
  4. Berendsen, H.J.A. (2005), Gone with the wind and the water: New Orleans. Geografie, November 2005.
  5. Hesselink, A.W. (2002), Morphological development of the embanked floodplains of the Rhine and Meuse in the Netherlands in historical time. Netherlands Geographical Studies 292, 190 pp. KNAG/Faculteit Ruimtelijke Wetenschappen Universiteit Utrecht.
  6. Hesselink, A.W., G. Stelling, J.C.J. Kwadijk & H. Middelkoop (2003), Inundation of a Dutch river polder, sensitivity analysis of a physically based inundation model using historic data. Water Resources Research Vol. 39, no 9.
  7. Kwadijk, J. (1993), The impact of climate change on the discharge of the River Rhine. Netherlands Geographical Studies 171, 208 pp. KNAG/Faculteit Ruimtelijke Wetenschappen Universiteit Utrecht.
  8. Middelkoop, H. (1998), Twice a river. Rhine and Meuse in The Netherlands). RIZA report 98.041 Arnhem: RIZA.
  9. Silva, W., F. Klijn & J. Dijkman (2001), Room for the Rhine branches in The Netherlands. Irma Sponge, Rijkswaterstaat, Delft Hydraulics Report.