Category: Uncategorized

Having just finished the first phase of our WEG-funded project, I’m taking stock of how it unfolded, to better inform the next, much bigger phase of works scheduled for next year. I divided this 1600 metre re-meandering project into two partly because it felt like too much to deliver in one go and partly because I felt we would learn a lot from doing a short phase first, with time to gather thoughts over the winter.

It was touch and go in the summer whether we would get started at all in 2019, but I’m glad we did because we learned a lot.

Ground conditions were a challenge: we were working on a peaty floodplain, where the gravel was up to 80 cm below the surface and the groundwater half that. That kind of ground can take very little traffic before it breaks up. The site is a SSSI, so keeping it in good condition was paramount.

Groundwater was also a challenge. Although we cut ‘in the dry’, the channel would fill with groundwater overnight: even in late September, with groundwater levels at their annual low-point.

In some ways it might have been better to start at the downstream end, in order to cut the channel in such a way that it drained as we carried on upstream. But that would have presented another set of problems, not least that (for now) we were returning the new channel into an impounded and raised section of old channel: this backed up water level would have flooded back up the new channel and given us even more of a water ingress headache.

So we started at the upstream end. For the first week the weather remained bright and the dryish ground held up very well … so long as the tracked dumper took straight lines back and forth with the spoil. This first section of spoil was placed along the edge of the existing river channel, ready to be pushed in at the end when we made the cut through.

But then as we worked downstream we came into a more fragile, peaty part of the floodplain. In addition from a certain point on we had to start taking the spoil off site and this across and off the floodplain.

We worked out a methodology so that the digger always worked within the confines of the channel it was cutting. This meant that we could keep the surrounding ground intact, but it also meant we got just the one pass at cutting the channel to the exactly correct levels for the pools and riffles.

We brought the dumper up the line of the new as yet undug channel and filled it by turning the digger a slightly laborious 180 degrees, one bucket at a time.

This worked okay for about a day or two. But where the dumper tried to follow the meandering course of the as yet undug channel it began to cut the ground in the tighter turns: tracked vehicles turn by going faster on one side than the other which creates a shearing effect on the ground. Eventually the floodplain surface broke apart and after that the dumper started to churn the ground up , to the point that it almost got stuck on a few occasions.

So we decided to take the peat away first and excavate down to the hard gravel which we could use as a roadway. This meant the digger had to “hay-make” the peat into accumulating piles and roll it on out of the site to a spot where the dumper could run in straight lines to and from the floodplain.

This seemed to proceed reasonably, if long-windedly well, but groundwater would seep in up to the surface of the gravel, so what was exposed as a dry gravel road on the day we cut down to it, had become a soggy, muddy road by the following morning.

When we then cut through that gravel to the desired bed level, the channel immediately began to fill with water. Consequently we had to leave coffer dams of gravel when we started each morning to keep the overnight infill out of the new day’s dig.

And thus we proceeded, day by day, until the weather broke and things got really tricky. Once it started to rain the ground because even more hazardous. After one, long and very wet weekend we returned to find our new channel absolutely brimful, like an infinity pool.

We tried pumping the water out, but this was a hiding to nothing really. The pump kept blocking, or sucking in air and stopping altogether. Somehow, by hook and by crook we managed to get to the bottom end and to hay-make the material back out to the one spot the dumper could reach without having to turn.

This main drag to and from the river became a real mess, but it was, at least, the only mess we made.

After all that the cut through was the simplest job of all. We cut a small channel from the existing river in to the new one, waited for the existing channel to drain down a bit and then carefully laid a large tree across the old channel. Building against this edge we created a land bridge across the old channel and then rolled on down from here filling in and levelling as we went.

On the way out we placed a goodly number of branches and ranks from trees that had been cleared a couple of years before. These will be pinned in place to add some diversity and grit to the new and still somewhat bare river channel.

I’m happy to report that trout, being curious creatures, didn’t take long to move in. Macrophytes, shards of starwort and ranunculus particularly, have already caught up on some of the stones. And I watched a kingfisher follow the new channel as we dug it.

I am now planning the fine details on the next phase.

If anyone would like to come and see the works do drop me a line. If a few people are interested we could arrange a field day and maybe a small workshop looking at some of the planning and delivery issues.

A few more photos of the unfolding works on the new channel in the first phase of our Water Environment Grant project.

 

A few photos as we begin work on Phase 1 of our Water Environment Grant (WEG) funded project – see previous post.

I am working with the expert help of Five Rivers, Stew on the digger and Jari (aka Gary) pedalling the tracked dumper. I’m on supervisory duties with the laser level and a satchel of drawings.

We have a brand new wide-track (90cm) excavator to play with. This machine has unbelievably low ground pressure: “like riding around on bog mats,” says Stew. This is good news as the ground here is very delicate and relatively saturated even at this time of year. The new channel fills with groundwater overnight.

So far so good: the peat floodplain is holding up, so long as we run the machines in straight lines, while the gravel is down there pretty much at the depths my unscientific road-pin probing survey indicated.

The real-challenge is trying to recreate the distinctive shape of a natural, spring-fed pastoral river – those undercut banks and the terraces on the insides of meanders, the gentle rise and fall of the river bed, the beautiful and almost inimitable randomness of a natural river  – while every other part of the process is about precise details and planning.

That’s where I hope my bursting-at-the-seams photo album of unmodified spring-fed streams in New Zealand will come in handy!

Three phases of a 1.75km re-wilding project – blue, light-blue and purple – in which the River Nar will be taken out of its diverted and impounded course and restored to a meandering channel in the centre of the floodplain.

In early September we will be starting work on the first phase of our most ambitious River Nar project to date, what will become almost 2 km of re-created “natural” chalk-stream: with funds from the Water Environment Grant we are moving the river out of a diverted, impounded and now dredged channel and putting it back where it belongs – in the bottom of the floodplain.

An awful lot of forensic work, sifting through old maps and aerial photos, of thinking, discussions, thinking again, drawing, measuring, and drawing again – not to say form filling and fund-raising – have lead to this. Probably rather more than ever preceded the diversion (several hundred years ago) and the dredging (twenty to fifty years ago). It takes more work to fix something than bust it.

Like most chalk-streams, the River Nar has been diverted along much of its course. The diversions have been driven by putting the river to use in one way or another over the centuries. It was widened and straightened in the early Middle Ages in order to float stone in barges to build the abbeys and castles along its course. It was diverted to feed water to those abbey kitchens, and latrines, and to power mills (some of which are recorded in the Domesday book). And in the late 18th / early 19th centuries it was used to create “floated” water-meadows, an agricultural innovation which trebled the crop of hay that could be taken off the floodplain meadows: the improved crop of hay fed greater numbers of sheep, which were “folded” at night onto arable fields on the surrounding hills, and thus fertilised those hills with the goodness of all that floodplain grass. Quite a neat form of cyclical land-use.

An aerial photo from about 1946 showing the diverted channel (dark and straight) and feint remnants of the original channel to the north.

Whenever you take a river away from its original course and divert it towards and then along the edge of the valley, you inevitably lessen the gradient for the duration of that diversion, you turn a swift-flowing river into a linear pond (hence the word “impound”) and in so doing build up “a head” of potential water-power than can be released at key places, at a mill or across a water-meadow, for example. In a way these functional structures are all part of the rich history of the valley: but they have big consequences for wildlife. The Domesday mills shut salmon and sturgeon out of the River Nar over one thousand years ago, for example. And in more recent decades dredging and the sheer volume of silt that runs in to the river from surrounding farmland and roads, has greatly exacerbated the problems caused by diversions and impoundments.

A typical reach of diverted, impounded and dredged channel.

The diverted and impounded and now dredged and silt-filled channels are not nearly so good for fish, or invertebrates, or the plant-life one would normally associate with a free-flowing chalk-stream.

This project, therefore, is an attempt to “re-wild” the river, to give it back its gradient, meanders, deep pools and gravel shallows, and to improve conditions for the fish and the plants and insects which thrive in a natural, free-flowing chalk-stream habitat.

Outline plans of Phase 1.

We are going to cut a new channel – although in fact it’s our best guess at a re-creation of the original channel – which will meander back and forth along the bottom of the floodplain. The gradient that is currently expended over a very short part of the existing channel will be spaced evenly along the entire length of the new channel, creating a swift moving river that falls from one riffle and pool to the next.

We’ll be cutting the channel with diggers … but in fact cutting a new channel is something natural rivers do all the time: even spring-fed rivers such as the River Nar. Have a look at the aerial image (below) of a spring-fed river in New Zealand, one that has never been straightened or modified and where the floodplain either side is open, uncultivated grassland. There are relic channels and ox-bows all over the place. In places the river runs through several channels side by side.

The River Nar is doing the same thing in one section of our new project site: a breach in the impounded leat is allowing some of the flow to escape through a willow carr and thus find its way back to the remains of the original channel. If we left it alone for a thousand years, it would probably finish the job on its own. This project puts those natural self-repair processes on fast forward.

In planning a project like this, an awful lot of work goes into investigating the history of the river and how it got to be how it is. Old maps and aerial photos suggest that the river here was diverted first to the north and then to the south to create a crude water-meadow or perhaps simply to shift it out of the way of the floodplain meadows.

There is evidence of basic catch-work drains along the pathway of what would have been the original channel, and of smaller drains running perpendicular to the original river. Towards the lower end of the project site the existing channel is suspended higher and higher above the floodplain, held behind a leat-like earth wall, until suddenly the river passes a sill in the river bed and over about 300 yards all the accumulated gradient is released in a headlong rush towards a mill. It is the strangest mill-leat I’ve ever seen, but maybe it worked.

Luckily some of what is almost certainly the original river channel remains in the valley floor towards the lower end of the project site: this grown-in channel meanders along an old parish boundary.

The first phase – 400 meters – will be cut this September. We will cut the channel by digging down through the floodplain peat until we get to gravel. But the heights of each riffle have to very carefully worked out in order to ensure we build a river that has a gravel base and flows downhill! You can’t just excavate until you hit gravel: you have to excavate to a very specific set of heights and try to coincide this with the gravel that is down there. Since you also don’t want to be digging and removing tons of gravel, you have to run the course along a pathway that coincides with where the gravel is at the correct depth under the floodplain. There’s no easy way to do this: one day someone will invent a cheap underground radar. In the meanwhile it’s me walking back and forth with a Trimble levelling device and a metal road-pin which I drive into the ground until I hit gravel. I then try to build a map of the floodplain surface and subsurface, the heights of the gravel under the surface, in order to arrive at the best course for the new river, trying to coincide this with a best estimate of where that original course would have been. It’s easier when there is an old channel, or evidence of one, to follow.

The peat that comes out of the ground will be used either to repair any compression of the floodplain caused by the machinery or it will be taken off the floodplain and used to fill an old gravel pit near the site. Some of the spoil will be used to plug the existing channel at the cut off point downstream as far as some freshwater springs: these will continue to feed water to the existing channel which will then become a fen-like oxbow channel and add diversity to the range of habitats in the valley.

I will post pictures as the project unfolds. In the meanwhile here are some before and after mock-ups of what I hope it will look like:

 

The photograph above of the River Nar near Castle Acre was taken in 2015, just after we had strimmed the banks ahead of rehabilitation works. The issues here were a dredged channel and a steeply domed riverbank piled high with the arisings. The channel is too wide and too deep. There is little hydrological connectivity except on the far bank where a berm has formed within the overwide channel. The ‘dredging’ which was more the result of mechanised weed-management than a concerted dredging effort had taken place annually, and always from the same bank. The programme was haphazard too, in that no clearing was done in inaccessible parts of the river: in the far distance a line of trees had protected the river and here the bed is undamaged. In an ideal world we would have built up the river bed with gravel before doing any other work, but gravel was not easily available here and our budget precluded us from bringing any in. We did fill two short sections in the reach as photographed. Otherwise the work consisted of installing LWD and then re-grading the bank to create a narrow, sinuous channel and a much lower bank profile which the river can – and does – spill over in high flows. The progression of photos below shows the evolution of the work over the following years: 2016, 2017 and 2019.

Now that the project is two years old I thought I’d share a few before and after shots of the work. Newton Common is the next reach of river downstream of Emmanuel’s Common. You can see from the before photos that river here had been canalised and deepened and was as a result pretty much overrun with bur-reed.

I made a long-section survey of the river-bed which revealed the places where the bed had been lowered below the natural fall-line of the river. Each of these was effectively a sump. The resultant siltation was the river’s attempt over time to repair itself: we gave the stream a helping hand with as complete an infill of gravel as we could manage within the budget and the time.

A bed-level survey revealed the sump-like reaches where the river had been lowered beneath the fall line of the valley. The Newton Common project is marked by points 1 to 29.

The top image was taken in October 2016: the channel is covered with burr reed. The second image was taken before the project began and shows how deep and wide the dredged channel was. The third image was taken just after the gravel went in, April 2016 and the fourth was taken two years later in April 2019

The top image was taken in October 2016, the middle image during the project and the bottom image in April 2019: note the presence of ranunculus and berula!

Learn from nature: LWD is rarely better than when an old tree falls in.

Large Woody Debris (or in plain English ‘Trees in Rivers’) – if done well – is a very effective river restoration / rehabilitation technique.

Until very recently we tended to tidy rivers up. In fact we still do. I often see pictures posted on Instagram and Twitter of tidying up exercises on chalk-streams masquerading as river restoration: “clearing out” a stream, or uniform and tidy constructions – faggot hurdles, bank edges – that would be more at home in a garden or golf course. 

Traditionally river-keepers tended to pull out trees and branches that had fallen into their streams: the aesthetic was one of a ‘tidy’ river, the practical intention was to make the river easier to fish, with fewer obstructions to casting, and no tangly obstructions to lose fish in. It was about creating an environment that suited anglers more than it suited the fish they were seeking.

State bodies like the Environment Agency have colluded with this mindset. When a tree falls in they remove it, motivated by the need to avoid ‘flood-risk’.

But slowly rivers-keepers and the Agency are changing how they look at this issue. That’s because in both fishing and habitat terms and in terms of flood risk, removing trees and large branches from rivers is self-defeating, the very opposite of what we should be doing.

Trout love cover: remove the cover, remove the trout.

From the fish and habitat point of view it’s simple: fish love cover. A tree in a river provides fabulous refuge from predation. In New Zealand’s spring creeks which are not tidied up at all, the largest trout will always lie within bolting distance of a submerged tree, or some form of log-jam or snaggy overhead cover. This makes catching them tricky, but on the other hand they wouldn’t be there if the LWD wasn’t there too.

Research carried out by Dr Murray Thompson has also shown that LWD is also fabulous habitat for invertebrates, that invert numbers are far higher in rivers where there is plenty of submerged LWD.

LWD is also great from a morphological point of view, especially in chalk-streams.

Chalk-streams are very low energy rivers. The forces which shaped them (melting glaciers) have long since retreated from our landscape. We have since that time heavily modified their channels by straightening, dredging and impounding them. Chalk-streams very rarely achieve the flows needed to overcome these modifications and re-shape themselves to a more natural planform. Thus their shape and their habitat is locked in a man-made straight jacket and the rivers are effectively imprisoned.

But a fallen tree can make a huge difference, releasing gravel and stones into a system that lacks material (quick morphology lesson: when a river is modified, by dredging, say, the natural processes of the river work to erase that modification, depositing material into the enlarged space until the river reaches the correct dimensions again. To repair itself, therefore, a river needs material, but chalk-streams are too gentle to supply it, except in the form of fine particles like silt. When a tree falls in, however, this creates the material and energy that enables self repair). A single tree can kick-start a whole sequence of channel repair and re-meandering.

I own a fishery at Frampton in Dorset where the river once flowed through a stately park, impounded by ornamental weirs. All the bends had been taken out. When I bought the fishery the then owner removed fallen trees all the time. He even removed one “for me” after he had sold me the river. I had to explain that I really didn’t want him to.

One large filled tree is visible in the distance, the other to the left of the shot. Ten years ago this reach was dead straight. The entire green bank on the left is new and the bend to the right is new. The same has happened in the opposite direction downstream of the lower tree.

Looking upstream from the lower tree-fall. The vegetated island in the middle of the shot and the meander to the left were created by a single fallen tree upstream.

2013 when the lower tree fell in.

2014, one year later: we’re starting to see a braided channel, a deep pool, a meander.

The third big fallen tree lies upstream of the island and gravel riffles which it created.

Since then three really big willows have fallen in. I have left them where they fell and watched what happened. Those willows have catalysed a whole sequence of channel repair, far better than any river restoration guru could ever have designed. The Frome is quite a powerful chalk-stream, so these changes have been on fast forward compared to gentler rivers, but even so they are extraordinary: hundreds of yards of re-meandered stream, of pool and riffle sequence, all catalysed by three fallen trees.

Those willows and many other instances where I have observed the aftermath of a natural tree fall, suggest to me that even now with LWD “restoration” we choreograph our work far too carefully. I’ve done it myself. In fact on the 7km LWD project I did with Simon Cain on the river Nar the work got more cost-effective, more natural and far, far more efficient as we progressed. To me good LWD imitates the impact of a fallen tree: it begins and ends with that.

Early phase of LWD work on the Nar: this was pretty good. It worked well but was quite intensive.

This was better! Less intensive, more natural, more cost-effective.

Another important learning curve has been where and when to use LWD. Basically there is a depth and gradient beyond which LWD doesn’t really work very well, at least from a morphological point of view.

LWD makes it best impact in relatively shallow, wide reaches, where the bed is more or less intact and where the river is not impounded. It is perfect for repairing an overwide river, but is not nearly so good at repairing a dredged river or an impounded river. In the latter two cases you need to resolve the river-bed and gradient issues first.

Do’s and Dont’s of LWD in chalk streams

• Don’t overly choreograph your LWD. Resist the desire to “build” something tidy or uniform.

• Look at and learn from natural tree fall and try to copy it.

• Use big timber, whole trees or very big branches.

• Work with the natural meander pattern of the river. Work out what the natural meander wavelength should be (see my other notes on this) and use LWD in sync with that, on the insides of bends or what will become bends.

• Think about daylight: the LWD will allow sediment to accrete in the slackened flow downstream. This sediment will only consolidate with vegetation and vegetation needs daylight.

• But don’t overdo the above. Rivers need shade too. Try to create a mosaic of light and shade.

• Be judicious about which trees you cut down. Alder is best. Willow can be great but is difficult to control later. Ideally you will have multi-stand alders and you can use some and leave others.

• Don’t expect LWD to do much in an impounded or overly deep river. Really you should remove the impoundment (weir or mill hatches) first. When you do you will have a fast flowing but overly wide and homogenous channel. That’s when LWD can work it’s magic.

Natural LWD in a New Zealand spring creek: spot the trout within a fin-flick of tangled cover. Note in the pictures below how the really big LWD has shaped the channel. It’s like this the whole way up the river!

Minn’s Meadow was our attempt at finding a way of properly fixing the dredged parts of the river. Because dredging is the most fundamental and widespread way in which chalk-streams have been damaged, I’m really interested in what we can do about it.

Here we have taken the gravel from borrow-pits beside the river and restored long riffles in sync with the meander pattern of the river. The borrow pits were filled back in and have settled into flood-plain hollows that imitate old oxbows and have a nice side-benefit of wetland habitat. There were quite a few snipe hanging around them in the winter.

It was also nice to see that the riffles were used by spawning trout (and sea-trout) within a month of going in. I feel the fact that the gravel was ungraded and from the floodplain beside the river might well have influenced how readily it was adopted by spawning trout.

Here’s a few before, during and after photos. Before we started all the gradient in this meadow was lost over the first few yards, whereafter the river was deep and sluggish. Now we have spaced the gradient out of a series of riffles with natural shapes and spacings. It is great to see patches of ranunculus growing where there was only eel-grass.