words and pictures by charles rangeley-wilson

River Nar at Castle Acre

A restoration project sponsored by the Norfolk Rivers Trust via the Catchment Restoration Fund.


An obviously entrenched reach of the River Nar: note the difference the between former flood-plain height and the newer emergent-berm on the right bank (this difference equates to the depth of dredging) and also the mound of spoil under the nettles along the left bank.

The River Nar in Norfolk, like most English chalk-streams, is a highly modified river: tiers of channel modification over the centuries – mill and priory diversions, culverts, fords, water-meadows and land-drainage dredging – have built one after the other to create a chalk-stream that is very much corralled, impounded and channelised.

But the River Nar is a low-energy chalk-stream. The modifications, particularly the entrenchment and impoundment of the stream, have so disabled the riverine processes that the river will take many decades if not centuries to regain its natural morphology. 

The aim of the project, therefore, is to fast-forward what would otherwise be a very slow process of self-repair and ‘re-wild’ the river: but working within the context of the heritage of the surrounding landscape – the water-meadows, mills and priories.

Planning and Strategy

The river was divided into sections according to the basic underlying morphological limitations. Very roughly speaking the differences between reaches related to:

• relative gradient (low versus moderate)

• degree of impoundment

• degree of entrenchment

We found that reaches of low gradient (longitudinal gradient of the river bed) tended also to be the more impounded and entrenched. This is because channel modifications relating to the priory, to mills and later to the water-meadow system, almost invariably involved moving the river from the centre to the edge of its flood-plain: either to create a head of water, or to get the river out of the way, or to create channels used to flood the water-meadows – or all three reasons together.


The map above shows some of the alternating impounded (red) and free-flowing (blue) reaches of the River Nar near Castle Acre.

• When a river is diverted to the edge of the flood-plain it loses its gradient: it has to run along a lesser degree of incline from the centre to the edge of the valley, and is thus effectively impounded. This loss of gradient alters the dynamic equilibrium of the stream and causes the river to drop sediment in the impounded channel.

• Accumulated sediment and slow flows alter the riverine habitat and the modified river inevitably fills with plants like burr-reed. Burr-reed has the effect of slowing flows yet further and compounding the problem. In fact, a map of burr-reed exactly coincides with a map of the impounded, modified gradients.


A typically impounded and entrenched reach of the river, just upstream of a culvert. Here the river is over one meter deep and the water-surface is over half-a-meter below the level of the flood plain.

On certain reaches of the river the compounding impacts of a loss of gradient, sedimentation and the encroachment of burr-reed amounted to a perfect storm of habitat degradation: to make it all worse, these reaches – being inundated with burr-reed and thus prone to backing-up and flooding – were heavily dredged in the post-war years. Long-section levelling-surveys along these reaches invariably showed the river bed to have a ‘sagging belly’ profile upstream of a given impoundment. See link below:


These impounded, diverted and entrenched reaches are expensive to fix. The best way to restore morphological stability to these reaches is to reunite them with the natural fall of the valley. However, where the basic diversion relates to something like a 900 year-old priory this is easier said than done.

Other possibilities might include: leaving the diversion in place as a relic still-water channel and creating a new replica natural channel in the hollow of the valley; or second best, restoring the bed-level of the diverted channel, filling in the “sagging belly” to at least make the best of the diverted channel.

Prioritising Phases of Work

Because these impounded and dredged reaches are relatively difficult and expensive to fix they were not prioritised for the first phase of work. Instead, the first phase was dedicated to the higher gradient reaches where Large Woody Debris berms might be sufficient to catalyse self-recovery. These reaches were, on the whole, relatively straight, only moderately entrenched or impounded and often excessively wide.


A typical reach suitable for LWD: the drifting dunes of sand and silt on the river bed indicate a channel that is too big for its flow. However, this reach is not excessively deep, impounded or entrenched.

Reference Reaches

There are a (very) few places on the River Nar where the channel morphology is close to natural  – a ‘letter-box’ cross-section (wide and relatively shallow), with a firm gravel bed, shallow, shelving banks and with a natural meander pattern. These are ‘reference’ sites.

There are also (more commonly) places on the river which can be described as ‘naturalising’ reference sites. These, typically, have been modified at some stage in the past, but have reverted back towards a natural state. These places show evolving channel dimensions that are similar to the very pure reference reaches.

Both sites are vital bench-marks on which to base plans for the sites targeted for restoration. If anything, for the LWD type of project, the ‘naturalising’ reaches are the most useful, because the targeted reaches in this project cannot (for various heritage reasons) be restored to a fully natural state (this would involve re-meandering the channel by taking it beyond the confines of its existing course).


A ‘naturalising’ reach: a channel which was straightened and widened in the past, but not impounded or dredged. These reaches can self-repair quite effectively if left alone and if there is the right mix of trees (to fall in the channel) and sunlight (to stimulate riparian growth).

Large Woody Debris

The idea was to use berms of Large Woody Debris felled from riparian trees in order to re-create pseudo-sinuosity within the confines of the relatively modified, over-wide and straightened channels. This would, at the same time, open canopies in the tree-cover and allow sunlight to stimulate riparian growth.

Reference reaches were analysed to get an idea of the desired natural channel dimensions. Gentle curves in the stream, characterised by depositional benches on the insides of the bend give a good idea of natural channel cross-section dimensions.

The ‘natural’ width was used to calculate the correct meander wavelength and this, in turn, was used to give an approximate placement frequency for the LWD berms. 

This is not the same as restoring natural meanders, but it is perhaps the best that could be achieved within the constraints of the site and budget.


An excerpt from the site plan showing how the LWD berms were spaced to maximise available capacity for sinuosity within the confines of the relatively modified channel

The total length for this phase of work was calculated at approx, 2750 meters.

The delivery team comprised three trained foresters (Acorn Tree Services) and one supervisor: either Simon Cain as the project leader or myself as project supervisor on behalf of the Norfolk Rivers Trust.


The budget, including materials, equipment hire, welfare units, etc. covered approx 8 weeks work. At a rate of 100 meters a day the full length could be treated within seven weeks, allowing one week leeway. The project was completed on time and within budget.

Materials: chestnut posts, galvanised wire and staples.

Equipment: chainsaws, post-pounder, dinghy, hammers, fencing pliers.

The following before-and-after photographs give a good idea of the way in which the LWD was used. Where possible whole trees were pinned in place: these tend to create the most natural-looking structures. However, it is not always possible to fell the trees close enough to the stream to pull them in by hand. Often, therefore, the trees were cut into sections and arranged in the channel.

In practice we found the best types of structure represent something like the fingers of an outstretched hand where the base of the hand is the centre of the arc of the meander: these recreate an emergent point-bar in a meandering river.

On straighter sections the limbs could be used as a series of spars, the outer edges of which mark the desired bank line. It is not necessary to describe the new bank line with a line of timber running in-line with the channel.

For Phase2 of the project, the re-grading of leveed banks see here:








%d bloggers like this: