The Great Stour in Kent (above), along with a number of other iconic chalk streams like the River Nar in Norfolk or the Great Eau in Lincolnshire, is classified in the least sensitive to abstraction band ASB1, even though the original 2005 groupings of river types by sensitivity to abstraction placed all chalk streams in the most sensitive group ASB3.

In December 2020 I wrote a post pointing out the anomalies that exist in the banding of chalk streams according to their sensitivity to water abstraction. I had discovered, for example, that my local chalk stream, the River Nar, was in the least sensitive band ASB1, along with the Great Eau, Lincolnshire’s most iconic chalk stream, while numerous other classic streams, like the Ver and Chess were in the middle band ASB2.

In fact, when I looked at the banding of all the chalk streams, frequently I couldn’t make much sense out of what seemed inconsistent assignments. Neighbouring and very similar streams like the Cerne and Sydling were – inexplicably – in different bands. Even allowing for the fact that a mixed geology chalk stream, or a very highly modified chalk stream in an intensely developed setting, or the lower reaches of a very big chalk stream, could all arguably be placed in ASB2, there seemed to be many inconsistencies: Wylye tributaries the Swan, Heytesbury Stream and Chitterne are all ASB3 but its SAC tributary the Till is ASB2, for example. Why?

The abstraction sensitivity banding is designed to provide a level of protection against the pressures of abstraction by assigning to a stream a “sensitivity to abstraction” associated with a maximum allowable % reduction from natural flow. Thus at very low flows, for example, the most sensitive of streams in ASB3 are deemed to tolerate a 10% reduction from natural flows, while for the least sensitive in ASB1 that reduction can be 20%. At higher flows these reductions can be greater: 24% at Q30 for ASB3 and 30% for ASB1.

The ASB banding is an evolution of the flow standards for river types developed by the UK Technical Advisory Group (Acreman at al 2005 and UK TAG 2008) and is based on an assessment of three components: a physical score which takes into account including gradient, catchment size, rainfall, and base-flow; a LIFE (lotic index for flow evaluation) score for the expected macroinvertebrate community; and a fish-guild based score for the expected fish community. In the Acreman / UKTAG banding chalk streams and especially headwater chalk streams were deemed to be amongst the most sensitive river types, so one would expect most chalk streams to now find themselves in ASB3.

Indeed, the key word here taken from EA literature is “expected”. This word suggests to me that the banding should be set according to the river’s potential or type, not necessarily its current condition. First of all, let’s look at the score of a classic chalk stream in excellent condition, the Alre:

WFD no.	Name	Phys	LIFE	Fish	ASB
GB107042022610	Alre	31	23	33	3

The Alre is one of the three tributaries that come together to make the Itchen. The others, the Candover and Cheriton score more or less exactly the same, so this gives a good idea of a classic chalk stream score for rivers in good condition.

Now let’s take an anomaly I highlighted above, the fact that the Sydling is ASB3 and its neighbour the Cerne is ASB2. Their scores are:

WFD no.	Name	Phys	LIFE	Fish	ASB
GB108044009700	Sydling Water	33	23	32	3
GB108044009710	Cerne	32	23	21	2

The Sydling’s score is more or less the same as the Alre’s and the difference between it and the Cerne is, clearly, the (expected) fish community. But the Cerne is a headwater, pure chalk tributary just like the Sydling. It is dominated by trout. Salmon might occasionally spawn in the Cerne, although they can’t ascend the Sydling. Of course there will be eels, bullheads and minnows too. The two streams should – to anyone who knows them – score exactly the same for expected fish community, albeit the Sydling might score better according to its actual fish community because it is in better condition than the Cerne: the valley is dominated by organic farms, and there is only one small village on the Sydling, whereas the Cerne valley is more developed, and the stream more heavily impacted by abstraction and sewage discharges.

But as said, the scoring is supposed to be according to type not condition.

Now, let’s look at some chalk streams that should be just like the Alre and Sydling but aren’t, not least because of how much they are abstracted: the classic chalk stream tributaries of the Colne.

WFD no.	Name	Phys	LIFE	Fish	ASB
GB106039029830	Misbourne	11	23	21	2
GB106039029870	Chess	11	23	32	2
GB106039029890	Bulbourne	11	23	32	2
GB106039029900	Gade (upper)	12	23	32	2
GB106039029920	Ver	12	23	31	2

Apart from the lower score for the Misbourne (why?) the expected fish community scores are the same as for the Sydling and Alre. The expected LIFE scores are the same too. That’s good: it implies that expected is indeed the operative word, given that the actual fish and invertebrate communities of the Chilterns streams will not as good as the Wessex streams.

But the physical score which – note – is based on gradient, baseflow index, catchment size and rainfall, is drastically different. This makes no sense: these are not exactly mutable qualities and I doubt there are very many differences at all between these physical parameters of the River Chess and any number of Wessex chalk streams whose physical scores are around 32/33. You could possibly argue that these Chilterns chalk streams are more heavily modified than those in Wessex, but then how come the River Wye, the most heavily modified and urbanised of all the Chilterns chalk streams scores as follows:

WFD no.	Name	Phys	LIFE	Fish	ASB
GB106039023880	Wye (High Wycombe fire station to Thames)	21	23	33	3
GB106039023890	Wye (Source to High Wycombe fire station)	31	23	33	3

To me this suggests a basic inconsistency of application. Maybe different area teams interpreted the scoring system in different ways, or maybe the assessment has been done in a rather mechanical way by personnel who aren’t that familiar with the streams they are assessing and within parameters that just don’t fit the physical reality.

For example, the scores for chalk streams local to me in Norfolk suggests the same inconsistency complicated by unhelpful WFD waterbody boundaries.

WFD no.	Name	Phys	LIFE	Fish	ASB
GB105033047791	Nar upstream of Abbey Farm	22	13	12	1
GB105033047792	Nar downstream of Abbey Farm	11	13	12	1
GB105033047620	Babingley River	22	13	12	1
GB105033053480	Heacham River	22	13	12	1

In these East Anglian rivers it is common for the waterbody to include chalk headwaters and reclaimed fenland drain. Often the waterbody boundaries don’t always fall in a place that allows these changing characteristics to be properly assessed and fitted to the waterbody in question. But really there should no difference at all between the expected invertebrate and fish communities of the East Anglian rivers Nar, Heacham and Babingley and the Wessex rivers Alre, Sydling or upper Test.

The second recommendation of the Water Quantity section of the CaBA chalk stream restoration strategy launched in October last year was for a review of this abstraction sensitivity banding:

CaBA CSRG recommends a review of the Abstraction Sensitivity Banding. All chalk streams should be banded ASB3, unless there is evidence to support a lower band. ASB3 may not be appropriate on the lower reaches of very big chalk catchments or highly modified systems, for example the lower Colne or Lea, the lower Wey, Gade, Stort etc.

In the table below I have indicated the chalk streams which should – unarguably, in my view – be reassigned into ASB3: coloured blue these are classic chalk streams, whose morphology and expected or actual fish and invertebrate communities are exactly the same as all the other chalk streams already in ASB3. In some cases it might make sense to split the waterbody in order to do this with logical consistency . In my view, the stand-out anomalies are:

Frome headwaters and Wraxall Brook

River Cerne

River Piddle (middle reaches)

North Winterbourne (trib of Stour)

River Wylye

River Till

River Bourne (Wilts)

River Ebble

Sweatford’s Water

Allen River

River Dever

River Test (middle reaches)

Wallop Brook

River Dun

River Itchen (upper and middle reaches)

River Meon

River Loddon (headwaters)

River Whitewater

River Ver

River Gade (upper reaches)

River Bulbourne

River Chess

River Misbourne

River Beane

River Quin

River Ash (upper reaches)

River Stort

Great Stour (middle reaches)

Nailbourne and Little Stour

River Dour (headwaters)

River Nar (upper river)

River Babingley

Heacham River

River Burn

River Stiffkey & Binham Stream

River Glaven

River Bure (upper reaches)

River Wensum (upper and middle reaches)

Great Eau (upper reaches) & Burwell Beck

River Lud

West Beck incl Elmswell beck and Little Driffield Beck

Gypsey Race

River Dever

If I haven’t included a given chalk stream that is currently ASB1 or 2 in the list above that doesn’t mean it shouldn’t be moved to ASB3. It just means it’s not as stand-out obvious as the streams in the list above are.

North of London, for example, the lower-lying topography creates shorter chalk streams that segue into different kinds of lowland stream, all lumped into one waterbody. To protect the chalk stream reaches of many streams like this it’s probably going to be necessary to split the waterbodies between the chalk upper reaches (which could then be classed ASB3) and the lower reaches. Across the piece, the lower reaches of larger chalk streams, or very heavily urbanised streams could reasonably be assigned ASB2. Re-assessing them all will, of course, require good local knowledge, applied consistently across all chalk regions.

In my view though, the EA could make a good start by moving those stand-out streams listed above en bloc into ASB3 whilst committing to a review of the rest.