Floods on the Somerset Levels: a sad tale of ignorance and neglect

When King Alfred hid in the Somerset marshes from the Danes in 878, he could scarcely have imagined that the marshes of Athelney and the lands of Wessex to the North East would be drained and occupied during both winter and summer, since beforehand, these lands could only be used during the summer: thus the name of my home county - the land of the summer pastures. The Saete of Somersaete, comes from the summer pastures of Norway which are called Saeters or Seters.

In showing our respect for the freedom of others from serious floods, I would like to describe some of the ignorance and neglect that has led to the recent flooding of the Somerset levels, and to then shine a beam of hope for the future, which quantifies a solution to the problem.

Rainfall

No one disputes that the past three months have been very wet. But how does it compare with the past? Using the monthly series of rainfall for England and Wales we find the following results for November and December only.

There was serious flooding in 1854, 1872-3 and 1929-30. During the latter flood at Taunton, from November 1929 to January 1930 537mm were recorded, which is over 70% of the annual average. The floods lasted from December to February. For the same time period in 2013-14 the rainfall in the upper Brue, which drains into the Levels, was 434mm.

An analysis of the highest consecutive three monthly rainfall since 1766 shows that this is the fifth highest, giving it a return period of about 1 in 50 years. This result excludes 1960, when Bedlamgreen in the upper Brue recorded 556mm. Before 1766 possibly the worst flood in historical times was that of 1607: I will mention this event at the end of this paper, since it was not caused by prolonged rainfall.

It would therefore not be correct to state that the present floods are in anyway linked to changes in air temperature, both local, or global.

Drainage

Several Acts of Parliament in the early 19th century led to improvements in the drainage of the Levels. In the uplands, field boundaries often consisted of a hedge and ditch. Both were maintained, often on an annual basis. Where they did not lead directly or not at all into a watercourse, ditches retained a considerable volume of water.

During the late 1970's I calculated the likely volume of potential storage of field ditches in the upper Brue catchment. The Brue is one of three main rivers that drains into the Levels. Figure 1 shows the extent of field ditches, excluding those by highways. Assuming a cross section area of 0.7 m2 gives a storage volume of about 77000m3. Many ditches have a bigger cross section. This is big enough to have cancelled out nearly all of the overflow volume during the flood at Bruton in May 1979, when 36 houses were flooded. Applying the same density of ditches over the upland catchments of the Tone, Parrett, and Brue and assuming a cross section area of 5m2 gives a storage volume of about 16Mm3. I will describe the way in which this could improve the situation in the penultimate section of this paper.

Rate of pumping

From the 17th century pumps were used to remove excess water from the land. The earliest pumps were driven by windmills, but steam power, followed by diesel and electricity enabled much greater volumes of water to be carried away. In the scholarly account of the draining of the Somerset levels by Dr Michael Williams, published in 1970, Williams never gave a single figure for the capacity of the pumping stations. In 2014 the situation is more complicated as Table 2 shows.

Not all of these volumes can represent the truth. None of these sources give an estimate of the effect of the removal of these volumes of water. However, all sources are united in giving the area of flooded land as 65km2. Figure 2 shows the extent of the present flooded area.

Taking a value of 1.5Mm3 from the Environment Agency website, over an area of 65km2, gives a reduction in water level of 2.3cm per day. This means that a house with 0.6m of floodwater can expect, all else being equal, to have to wait 26 days for the pumps to remove all the flood water from indoors. This rather sorry situation must be amended because as with some of the estimates of pumping rates, it too is in error. From the evidence in Figure 2, the likely area that is now flooded in February 2014 is about 125km2. Clearly the Environment Agency needs to check its basic data via a peer review process before it is published. Given a flooded area in excess of 100km2, yields a decrease in water level of about 1.5cm per day, and an evacuation time of 40 days.

Dredging of the river channels

During the mid 1980's considerable sums of money were spent on dredging the main rivers (Basil Tinkler, personal communication). The effects of an increased channel width and depth, possibly with a compound channel to reduce the effects of silting, can be calculated. If the capacity of the rivers Tone, Parrett, and Brue in their lower reaches were to be increased by 10m3 s-1 then allowing for the effects of the daily tidal regime, would be expected to evacuate an additional 1.29Mm3 per day. This is probably more effective because the pumping in February 2014 is to remove water from the flood plain, whereas an increase of channel capacity helps to prevent floodwater from overflowing into people's houses and farmland in the first place.

Wildlife

Another aspect of the problem is wildlife. No one disputes the need to conserve wildlife and wholesale drainage without Nature Reserves does not respect the freedom of those who wish to enjoy its presence. Both the needs of farmland, civil life, and wildlife can be accommodated through sensible agricultural and drainage practices.

For example the present floods have, in many cases been destructive and fatal to many animals. In the summer it is especially damaging to plant life. Dredging of rivers can proceed on one side of the river only. Numerous wildlife niches can be made with improved hedgerow ditches. At present most of the ditches are filled in through neglect, or are in a poor state of repair. As a result the lower ends of fields remain waterlogged, roads are flooded unnecessarily and most economic activities are adversely affected.

The way ahead

The Somerset levels were created as a result of people wanting to produce food and a way of life that was not perpetually hampered by stagnant water and ruined crops. This can be maintained in years to come. Indeed it must be continued otherwise lives and valuable crops and wildlife will be destroyed.

The maintenance of lowland rivers, adequate pumping, and the reinstatement of hedgerow ditches and ditches alongside roads can, potentially provide considerable storage of water. It is the legal responsibility of land owners to maintain their ditches in good order. The enforcing authority ought to act accordingly. To help store and release floodwater, each ditch would have a thin plate weir installed at its lower end with an outlet near the base to allow water to drain slowly away over a period of a few days. The details can be calculated on a case by case basis. Proper maintenance would be needed.

The additional capacity of the drainage ditches and the 0.5Mm3 per day of general pump capacity would produce an estimated 1.29 (channel enlargement) + 0.5 (pumping) + 16.1 (ditches draining out over 1 day at a rate of 200 l/s per km2) = 17.89Mm3 capacity for excess floodwater. In this way the worst effects of the present floods would be avoided.

If the hedgerow ditches were increased to a cross sectional area of 10m2 then capacity for flood water would increase to 8.19Mm3. If the true area of flooding is about 125km2 and the average depth is 0.5m then the volume of floodwater is 62.5Mm3.

Using the lower volume of flood capacity this gives a time period of 13 days for the flood to take place without any adverse effects. Any shorter time period would cause the system to be overloaded. For the four days from 1-4 January 2014 the average daily rainfall in the upper Brue was 15mm. This produces a volume of rainfall over 908km2 of upland area of 13.62Mm3 which is less than 17.89Mm3 provided by storage, channel improvements and pumping. In the present case the floods have built up since the start of December. The first reports of serious flooding were around Christmas time. Therefore the present proposal should cope. Without the use of the extra storage capacity of ditches a critical time period of 35 days would be needed for the floods to be contained. Clearly pumping and dredging would not have coped in the present case.

A warning

Very rarely a flood may occur that is not caused by heavy and prolonged rain, but a tidal surge and storm surge as well. It may also have been caused by a tsunami. This probably took place in January 1607 which led to flood waters up to 4m deep. A contemporary report by W. Barley and I.Bayly tells us:

" that the cuntrey all along to Bridge-water was greatly distressed thereby, and much hurt there done it is a most pitifull sight to beholde what numbers of fat Oxen, were there Drowned: what flocks of sheepe....

It has been estimated that 500-2000 people were drowned during this flood. It affected much of Somerset and also South Wales. It is hoped that such a catastrophic event does not find the land and its people on the Levels as ill prepared as they were in 2013-14.

Dr Colin Clark is a research and consultant hydrologist who has over 60 published papers. He has been advisory editor to eight scientific journals and is in charge of Charldon Hill Research Station in Somerset, where the longest running measurements of evaporation in England are currently being made. Contact: colin4chrs@hotmail.com