| Home » Fisheries Research and Development » 13th Salmonid Habitat Enhancement Workshop »

Agricultural Land Use and Salmonid Habitat Restoration: The River Calder, Cumbria.

Dr Keith Hendry & Dr David Cragg-Hine

APEM Ltd, Enterprise House, Manchester Science Park, Lloyd Street North, Manchester M15 6SE, U.K.

Abstract

The River Calder, a short high-energy stream running off the Cumbrian fells, has experienced a deterioration in its migratory salmonid population over recent years. Intensive investigations have indicated that spawning and juvenile habitat have deteriorated. A once productive nursery tributary, Worm Gill, has been subject to extreme erosion, the river channel having re-aligned itself in some places. Examination of hydrograph records indicates that high flow events, which appear to be increasingly erosive, are becoming more common. Anecdotal evidence also supports the claim that the river is rising and falling more rapidly than previously. As such the evidence points towards changing patterns of both run-off and stream discharge. Initial observations suggested that the catchment, including a large area of common land known as Kinniside Common, was subject to overgrazing. A thorough evaluation of stock levels was complimented by a detailed vegetation survey that confirmed the catchment was indeed heavily overgrazed. Although there is no direct empirical link between cause and effect in this instance, overgrazing is considered by the authors to be a major contributory factor to the destabilising of the Worm Gill channel that has in turn led to a dramatic decline in the juvenile fish populations of the tributary. In-stream habitat restoration, although regarded as desirable by most parties involved, cannot be initiated until channel and bankside stability has been established. The paper explores the often-disputed link between overgrazing, stream run-off and erosion, and in particular centres on the difficulties experienced in the UK when trying to reach consensus for remedial management action, both with respect to treating the perceived cause of the deterioration and the symptoms.

Introduction

The River Calder is a small salmonid spate river in Cumbria, North West England. The main angling interest is salmon, although the river does support sea trout. Over recent years anglers have noticed a decline in catches, raising concern over the status of the stock. As a consequence of these concerns the Calder Conservation Committee (CCC) was formed, a group comprising local fisheries interests and owners, landowners and graziers, industrialists and the Environment Agency. The remit of the Committee was to investigate and understand the reasons for the decline in salmon stocks, with a view to arresting the decline and returning the stock to a healthy condition. With this objective in mind, APEM were commissioned to collate all existing information and undertake whatever field investigations were considered necessary. In addition, several other studies have been undertaken, including a River Habitat Survey (RHS), an overgrazing investigation and an Environment Agency hydrological and geomorphological review.

Figure 1. River Calder and tributaries, Cumbria, North West England.

The river, 17.2 km in length, has two major headwater tributaries, the Little Calder and Worm Gill, which rise in the upland fells near the Cumbrian coast north east of Sellafield (Fig. 1). The Calder flows across a variety of solid geology. The lower course of the river is on Sherwood Sandstone, the most recent solid geology and a major aquifer. Moving upstream, the main Calder channel encounters pre-carboniferous rocks. Before the upper confluence of the Little Calder and Worm Gill, the solid geology is dominated by the Borrowdale Volcanic series. Above the confluence, Little Calder flows entirely across Skiddaw Slates. However, Worm Gill flows along the boundary of the Borrowdale Volcanics and Skiddaw Slates from the confluence with the Little Calder to its upper reach which is found on granitic intrusions. In addition to the solid geology, the catchment also has a variety of drift deposits, mainly dominated by peat alluvium in the upper reaches and glacial tills (e.g. boulder clay) in the lower reaches; pockets of sand and gravel are also present. The Calder is used as a source of water for industry and for public water supply by United Utilities.

The standard rainfall average is less than 1500 mm for the majority of the Calder catchment (1152 mm); however, the uppermost reaches of Worm Gill have a standard rainfall average of between 1500 – 2000 mm (NRA, 1994a). The mean daily flow measured about 1Km from head of tide at Calder Hall is 1.794 m³/s, dry weather flow (Q₉₅) being 0.299 m³/s and mean annual flood flow 30.029 m³/s.

Land use in the upper Calder and two main tributaries is dominated by livestock farming on the open fells. Improved pasture is encountered on the main stem interspaced with mixed woodland down to the industrialised site that dominates the final kilometre of the river to the estuary. This section has been the subject of major modification to alleviate flooding over recent years.

With respect to water quality, the river has been classified under the Environment Agency in England & Wales scheme ‘General Quality Assessment (GQA)’ which utilises chemical (levels of dissolved oxygen, BOD, and total ammonia) and biological (the diversity and abundance of benthic invertebrate fauna) data to determine the position of a water body within a six tiered grading system. The majority of Calder has been given the highest GQA grade of A (Very Good)[1] for both biology and chemistry.

Salmon Stock Status

Adult Stocks

Catch return data is provided by the local angling club (Calder Angling Association) that accounts for over 90% of the declared rod catch on the river (Fig. 2). Examination of the data reveals that catches have remained low during most the 1990’s, although 1998 was a good year but was possibly augmented as a result of stocking. However, over the 25-year period for which reliable data are available, rod catches have fluctuated wildly, ranging from a maximum of 126 salmon in 1988, to a low of 12 in 1983 and 1999. Anecdotal evidence from Calder Angling Association suggests that catches prior to the period represented by the data set were much higher historically, with reports of annual rod catches of between 200 and 300 fish in the 1950’s and 1960’s.

Figure 2. River Calder Rod Catch – Calder Angling Association Data

It is also useful to compare the Calder with other North West rivers to determine whether the catches reflect regional trends or if factors are operating within the Calder that are impacting upon the stock independently. The procedure adopted was to combine the catches of the major North West salmon rivers ^[2]and convert them to logarithms. The logged annual catches from the Calder are then subtracted from the North West rivers annual average, providing an indication of any deviations from the trends in the regions rivers (the control group) as a whole. Fitting a trend line to the data (dotted line) helps to identify any long-term tendencies and reduces inter-annual variation^[3].

Figure 3. River Calder vs Combined Rod Catch of North West Control Group Rivers

The results are illustrated in Figure 3. The trend line shows significant divergence away from the other rivers in the region (p<0.05), indicating that factor(s) are operating specifically within the Calder catchment that have a detrimental impact on fish stocks that are not evident in other North West rivers. It would appear that this tendency has become more pronounced since the mid 1980’s.

This change is also well illustrated by comparing the Calder rod catch as a percentage of the total rod catch of the North West rivers for the two five year periods at the beginning and at the end of the 25 year data set. Over the period 1976 to 1980, the Calder comprised on average 2.4% of the total of the Northwest rivers (. For the corresponding period between 1996 and 2000, the average was 1.0%. This type of analysis does not confirm the cause of the decline in catches but merely that the fishery in the Calder is subject to additional pressures not present in other rivers of the region.

The same exercise was performed on the River Ehen, which shares the same ‘estuary’ and hence might be expected to show a similar trend if a common factor was affecting both rivers (e.g. illegal estuarine or coastal netting). Examination of Figure 4 shows that the Ehen is behaving in a very different way to the Calder. The trend line moving towards zero and then going negative indicates that the Ehen is improving relative to North West rivers in general (p<0.05). Put another way, the five-year average for Ehen rod catch as a percentage of North West Rivers over the period 1976 to 1980 was 2.6%. The five-year average for the corresponding period between 1996 and 2000 increased to 4.4%.

Figure 4. River Ehen vs Combined Rod Catch of North West Control Group Rivers

In addition spawning activity over a 23-year period was also examined using redd count data. It is acknowledged that caution must always be exercised when using redd count data because of the highly variable quality of counts from year to year, caused primarily by high river flows, obscuring redds. In addition larger sea trout redds can be mistaken for salmon. Nevertheless, provided these limitations are borne in mind, redd count data can provide a useful indication of long-term trends together with information on distribution of spawners within the system.

Looking at the five year averages, for the period 1974 to 1978, the average annual redd count was 148 salmon and 186 sea trout. The five year averages for the most recent corresponding period for which data are available (1992 to 1996) are 42 for salmon and 10 for sea trout, representing quite dramatic reductions of 72 and 95 % in the redd count. It is apparent that there has been a marked decline in spawning effort as would be anticipated from the decrease in rod catches discussed above.

Juvenile densities

With respect to juvenile salmonid densities one site on the main stem of the River Calder at Pelham House, provides the longest available data set for the unmodified main stem dating back to 1986. In 1986, although the trout population was poor, salmon were abundant; particularly parr that were present in high numbers, 25 per 100 m² representing a class A fish population (NRA, 1994b see Table 1). By 1993 both fry and parr had declined, and have remained at a comparatively poor level (class D and C respectively) during the most recent survey, 1998 (Environment Agency data in Hendry & Cragg-Hine, 1998a).

	CLASS
Species Group	A		B		C		D		E		F
0+ Salmon		86		45		23		9		0
>0+ Salmon		19		10		5		3		0

Table 1. Atlantic salmon abundance (fish per 100 m^-2)associated with absolute classifications in the National Fisheries Classification Scheme (National Rivers Authority 1994b). Grades run from A to F (e.g. grade A >86 and grade B 45-86 0+ salmon per 100 m^-2).

For the Little Calder, a significant decline in salmon population densities is apparent. Fry densities have virtually collapsed to a class E from a C, whilst the previously abundant parr are now present at only moderate to poor levels class C/D. Trout populations have also declined to extremely poor densities (E).

Data for Worm Gill is available for a single site from 1982 just upstream of the confluence with the Little Calder. Initially this part of the river supported an excellent salmon parr density (class A). Fry were poorly represented whilst trout were only present at moderate to poor densities (D/E). Over the intervening years the salmon parr population has deteriorated to what is now a class C/D population. Interestingly, salmon fry densities were moderate to good at the remaining two sites below the UU intake in 1993. However, the recent results show that salmon are all but absent from Worm Gill in 1998 upstream of Swarth Beck, whilst juvenile trout densities are for the most part poor.

Restocking Programmes

During the mid-1990s the Environment Agency embarked upon a stocking programme in an attempt to reverse the decline in salmon in the river. In June of 1994 the river system was extensively stocked throughout with 40,500 fed salmon fry. Areas stocked included both the major tributaries Worm Gill and the Little Calder together with their becks. Monitoring data to aid interpretation of the success of stocking is available for Worm Gill(EA data in Hendry & Cragg-Hine 1998a). Several months after stocking an electric fishing survey was carried out (October 1994). The density of stocked fish in the area upstream of Scalderscrew Beck to the UU intake[4] should have been of the order of 128 fry per 100 m², a rather high Class A density, in June immediately after stocking. However, the actual number recorded in October of that year was zero. The area immediately upstream of the Little Calder confluence contained a small number of fry but at a low density representing a class E juvenile population. During the summer of 1994, several large flows were experience, one in excess of 17 m³/s (10 times average daily flow of 1.66 m³/s) following 30 mm of rainfall in one day. In the absence of evidence top the contrary[5], it would appear that the fry were simply washed out of Worm Gill by the flood waters.

Analysis of water quantity, quality and ecological data excluded causes such as over-abstraction, acid rain, and sheep dip pollution, all of which have been cause for concern in one river or another in the North West of England in recent times. In-river poaching is not considered significant, partly due to the presence of the industrial complex in the catchment and partly due to the almost single ownership of the fishing rights by the local angling club. Estuarine poaching, if substantial would have been anticipated to also affect the adjacent River Ehen, as they both share the same estuary. The fact that the long-term trend in salmon rod catches in the Calder appears to be markedly different to the adjacent Ehen (i.e. downwards and erratic) when compared to other North West rivers combined, indicates that a river specific issue was the cause. This prompted an evaluation of the physical river habitat of the Calder.

Habitat Investigations

Availability of Habitat Types

In order to assess the quality and quantity of salmonid habitat throughout the river Calder, a detailed habitat assessment was undertaken. This involved a ‘walkover’ survey of the entire catchment, from the estuary downstream of Sellafield to the headwaters of the Little Calder and Worm Gill following the detailed methodology as described in Hendry & Cragg-Hine[6] (1997).

The survey was originally undertaken in spring1998. However, after a severe 1 in 100 year flood event on 3^rd August 1998 (108 m³/s), it was decided to repeat the survey following significant bank damage and obvious gravel movement within the river. The post flood survey was undertaken in mid September 1998.

Habitat	Length of habitat (m)	Area (m²) Mean width 9.25 m
Parr	3,567	32,993
Parr and Fry	1,433	13,259
Fry	680	6,290
Pool	2,333	21,584
Glide	453	4,194
Silted spawning habitat	20	185
Spawning habitat	163	1,607
Bedrock	40	370
Table 2. River Calder (main channel) – Habitat Types
Habitat	Length of habitat (m)	Area (m²) Mean width 4.9 m
Parr	1,860	9,114
Parr and Fry	1,680	8,232
Fry	426	2,091
Pool	566	2,777
Glide	126	620
Silted spawning habitat	27	131
Spawning habitat	77	436
Table 3. Little Calder – Habitat Types
Habitat	Length of habitat (m)	Area (m²) Mean width 5.9 m
Parr	960	5,664
Parr and Fry	2,133	12,587
Fry	913	5,389
Pool	633	3,737
Glide	140	826
Spawning habitat	80	491
Cascade	13	79
Table 4. Worm Gill – Habitat Types

Throughout the main stem of the Calder, once upstream of Sellafield, there are extensive areas of good quality juvenile habitat, particularly for parr but with adequate areas for fry (Table 2). In-river cover from boulders and cobble is excellent, with extensive areas of woodland providing overhead cover without the effects of tunnel vegetation. When comparing the habitat surveys before and after the large flood event of summer 1998, it was evident that many new areas of spawning gravel had been deposited, presumabley as a result of upstream erosion. The latter 1,607 m² of spawning gravel should be sufficient for around 169 redds, assuming 9.5 m² per spawning female (Beall & Marty, 1987). Several areas of quite extensive erosion were also encountered in the upper part of the Calder main stem, in one case over two metres depth of riverbank being removed.

Similarly, the Little Calder (Table 3), one of the two main tributary streams, offers good quality juvenile habitat interspersed with numerous small holding pools. Historically, during the 1950’s and 1960’s the upper reaches, which meander through an upland plain, were regarded as an important spawning area (S. Payne, pers. comm.). However, gravel beds encountered during the survey were less extensive than those present previously, amounting to some 1,607 m², although this is still sufficient to support 46 redds.

In contrast to the Little Calder, Worm Gill (Table 4) is a much more unstable channel. The 1998 flood had a significant effect, with bank erosion evident throughout much of its length and the river being substantially widened as a result. In some cases channel migration occurred, creating a new river bed. Worm Gill is dominated by juvenile habitat, although much of it appears to be unstable. Spawning habitat area is similar to Little Calder, with sufficient for 51 redds, but again it appears very unstable.

The survey concluded that spawning habitat was not limiting in the catchment with respect to quantity, particularly after the increase in available gravel following the flood event of August 1998. In total, when including the minor tributaries not discussed above, some 288 redds could be accommodated in terms of area available. This is far in excess of the 1990’s 5-year average redd count of 42 redds and easily accommodates the 1970’s 5 year average of 144 redds. Indeed, if the figure of 288 redds is worked through a basic life stage survival model (Kennedy, 1988 and Kennedy & Crozier, 1991) the total available juvenile habitat from all tributaries, equivalent to 35,695 m² would theoretically support parr densities well in excess of 120 per 100 m². This clearly indicates abundant if not excessive parr for the available habitat (see Hendry & Cragg-Hine, 1998).

Interestingly prior to the August 1998 flood event, the total amount of spawning gravel was 1,360 m², indicating that the quantity of gravel regarded as suitable for spawning had increased by just over 100% as a consequence of the flood. This pre flood area of spawning gravel would just about be sufficient to support the 1970’s 5-year average count of 144 redds, possibly indicating that during this period the river’s capacity to support salmonids would be around saturation point. This assertion would be borne out by the Class A juvenile densities recorded in the parts of the system for which historic juvenile data is available (e.g. Worm Gill, 1982).

Clearly the quantity of both spawning habitat and juvenile habitat are not limiting. However, as mentioned earlier, there are concerns about the stability of juvenile habitat in Worm Gill, riverbed mobility and the invariably coarse nature of the substrate being the main indicators combined with the recent poor electric fishing results.

Siltation of Spawning Habitat

Perhaps of more significance are concerns regarding the fines content of spawning gravels in the river. Analysis of gravels deposited in the main stem of the Calder, indicate that over 50% of the material comprises of fines <2 mm in size. It is commonly accepted that fines content should be less than 20% if not lower. For example Peterson (1978) and Roche (1994) have demonstrated that the minimum permeability (1000 cm/hr) for successful emergence of fry corresponds to a sand (< 2 mm) content of no more than 12 to 15%.

Recent work undertaken for the Environment Agency has used a fingerprinting approach to assemble information on the relative importance of surface and channel/subsurface sources of the interstitial sediments retrieved from salmonid spawning gravels in England and Wales (Walling et al 2001). The study concludes that in upland catchments surface soil erosion is the primary source of silt where overgrazing promotes the efficient delivery of fine sediment particles from slope to channel.

In addition, during the autumn of 1997 a study-visit was carried out by Dr. F.Theurer of the United States Department of Agriculture, Natural Resources Conservation Service, with the purpose of collating evidence on whether fine sediment from rural land use is affecting salmon and trout spawning habitats in England and Wales (Theurer et al 1998). As a result of this study-visit, the authors concluded that sediment pollution is widespread in England and Wales and is having a deleterious effect on salmonid fisheries, particularly through siltation of spawning gravels. It was acknowledged that Environment Agency staff had identified bank erosion as a source of sediment, and had several projects in hand to deal with this problem. However, there was a less clear appreciation of the substantial source of sediment that originates away from the river corridor.

The authors discussed the connection between rural land use and decline of salmonids, and described the potential sources of fine sediment from both agricultural and non-agricultural activities. Agriculture-induced erosion was believed to be the primary source of sediment entering redds during the incubation period for salmonids.

The high fines content of Calder spawning gravels coupled with the work referred to above clearly casts doubt on the quality of spawning gravels in the Calder catchment. This may provide an indication as to the causes behind the recent poor and somewhat erratic performance of the salmon stock.

Hydrology

The flood event of 3^rd August 1998 focused attention on the problem of high flows in the Calder and revealed the inherent instability of some parts of the catchment. The flood was a 1 in 100 year event peaking at 108.171 m3/s at 10.45 am on August 3^rd. The two surveys revealed that significant quantities of new gravel were deposited after having been eroded from other areas of the catchment. This increased the available gravel by over 100%. However, although more gravels have become available for spawning, the fines content and the instability of spawning areas which arise from this magnitude of flood induced gravel transport, raises questions with respect to both egg survival and redd washout during the incubation period.

It is thought that these mechanism may be responsible for the reduced egg survival which may then in turn be the bottleneck on production. If this hypothesis is indeed the case, then both the magnitude and frequency of flood flows in any given year will affect juvenile survival and subsequent smolt output. Given that some years are wetter than others this may be the reason behind the erratic performance of the stock.

An examination of maximum daily flows on the Calder (Figure 5) for two ten year periods commencing in 1974 and 1989 respectively, reveals an increased incidence of flood flows exceeding 30 m³/s and those exceeding 50 m³/s.

Analysis of rainfall data reveals a trend towards increasing peak flows during summer months, although the trend just fails to prove significant at the 95% level (Maas, pers. comm.). The latter is possibly a function of the comparatively small amount and duration of information for this type of analysis and data, but nevertheless, the ’trend’ does mirror experience from elsewhere. Recent information from SNH (1999) has indicated that wetter conditions (up to 10% by 2050) are likely in North West Scotland and that there will be increasingly strong seasonal differences. In addition, recent examination of flow records for the Lower River Bann in Northern Ireland illustrate a similar trend, high summer peak flows becoming much more common (D. Cragg-Hine, pers. comm.).

There is also anecdotal evidence of a change in the duration of the hydrograph resulting in increases to the speed and magnitude of flood events in the Calder. In other words the ‘flashiness’ of the river, appears to be increasing. Local landowners established in the area for over 50 years, are convinced that the river rises and falls much more quickly now than in previous years (D. Halliday, pers. com.). Thirty years ago when the river was in a normal flood, the ford at the foot of the Little Calder was impassable for livestock when returning sheep to the fells for at least a day, if not two days after it stopped raining. Now, following a winter flood the river is passable in a matter of hours on the same day after the cessation of rain, the river both rising and falling much more quickly.

It should be stressed that the evidence for this change in the hydrograph of the Calder is not conclusive, and is the subject of much debate. However, considerable evidence thus far points to a change having occurred such that it is worthy of far more detailed investigation than has been possible with the existing study.

This phenomenon has been observed in other NW rivers, notably the River Wyre draining the Trough of Bowland and has been investigated in some detail on the River Lune (Orr, 2000). The latter PhD study concluded that there had indeed been an increase in the frequency of intermediate magnitude floods (since 1950) largely as a result of land drainage and local climatic variability after 1970. In addition more rapid runoff observed over the past 25 years was attributed to both increased rainfall intensity and heavy grazing within the catchment. The intermediate magnitude floods correspond to discharge considered to be significant in sediment transport. The extensive bank erosion observed was related to three factors; overgrazing of banks and loss of stabilising vegetation, transfer of instability from elsewhere and an increase in stream power.

With respect to the Calder catchment, moor gripping to improve drainage does not appear to be relevant and forestry, although present in areas surrounding Worm Gill and the headwaters of the Little Calder, is not extensive. However, sheep grazing is widespread and is though to be relatively intensive throughout the upper Calder Catchment.

Overgrazing

Following on from the issues raised in this study and increasing concerns from within the farming community in and around the Calder valley, the Department for Environment, Food & Rural Affairs (DEFRA) commissioned a study to investigate grazing in the Calder catchment (National Grazing Management Team; 2001). The study involved a detailed survey of vegetation within Kinniside Common (Calder headwaters) together with an assessment of the livestock numbers and grazing management within the common.

It was evident from the survey that grazing pressure was undoubtedly affecting vegetation types. For example, Heather Heath, abundant in adjacent well-managed sites, was present only at low levels in Kinniside, a factor attributed to high grazing pressure. Similarly, the presence of areas of Wet Heath, with dominant grasses such as purple moor grass (Molinia caerylea) and deer grass (Trichophorum cespitosum) also appear limited due to high grazing pressure. Conversely, bilberry (Vaccinium myrtillus), a species thought to be associated with high grazing pressure, was common. It was also apparent that heavier grazing pressure was concentrated in the valley bottoms, albeit still in upland fells.

Based on the relative area of the different habitat types on Kinniside common, a sustainable stock carrying capacity was calculated at 3425 ewes in the 2143 hectares of fell. With a 10% margin for error this provides a sustainable carrying capacity figure of 1.75 ewes per hectare. The current grazing demand is equivalent to 2.2 ewes per hectare. The study concludes that there is an excess average grazing demand of at least 0.45 ewes per hectare (964 ewes) throughout the year. However, the study also makes the point that there are large numbers of livestock (sometimes equivalent to more than 3 ewes per hectare) on the common in late summer and autumn.

This has a number of additional implications. First of all, the winter reserves of plants may become depleted, leading to a reduction in sustainable productivity in future years and promoting a downward spiral of overgrazing. Secondly, dwarf shrubs are particularly prone to overgrazing at this time of year because they are flowering and setting seed and may be selectively grazed as the more palatable grasses die -back. Finally, the practice of winter foddering of livestock was also highlighted as being potentially detrimental, in that vegetation and soil can be mechanically damaged due to tracking and rutting to such an extent that recovery does not take place the following summer.

However, it is important to realise that overgrazing per se does not necessarily equate to broader environmental damage. There are many differing interpretations as to what constitutes overgrazing ranging from ecological, agricultural, through to geological and erosional viewpoints. Figure 6 illustrates three states of overgrazing, which occur at different densities of stock.

		- ECOLOGICAL
0.25 SHEEP per ha		- AGRICULTURAL
2.0 SHEEP per ha		- EROSIONAL

>10 SHEEP per ha HAS BEEN RECORDED

Figure 6. Upland sheep densities and levels of overgrazing (modified from Hendry et al., 1998b).

Evans (1977) found that erosion was created in Hey Clough (Peak District) during the 1960s with year round stocking densities as low as 2.0 sheep per ha. This level of grazing density was also sufficient to initially cause a decline of heather moorland, and then to initiate erosion in the form of sheep scars within the better quality acid-grasslands, which often had wavy hair grass as the dominant species (Evans, 1993). Hence, at densities somewhere above 2.0 sheep per ha, problems with erosion may be anticipated.

Furthermore, a study conducted by Anderson and Radford (1994) on the impacts of shepherding on the Kinder Scout area in the Peak District showed that active shepherding had the effect of reducing grazing intensities by a factor of up to 14, from 2.5 ewes ha^-1 to 0.18-0.43 ewes ha^-1. This encouraged the revegetation of previously bare and eroding ground, illustrating the benefits of shepherding. Historically the role of the shepherd was to keep the stock moving and force the sheep to graze in specific areas. This had the effect of extracting the maximum sustainable stock presence from the fell without irreparably damaging the vegetation or soils. In the absence of shepherding, the sheep densities on Kinniside Common (2.2 ewes ha^-1) are probably much higher in the valley bottoms than is apparent from the analysis undertaken above. Therefore, it should be concluded that within the Calder catchment, localised sheep densities are likely to be sufficiently high to cause overgrazing at a level where erosion is likely.

Overgrazing and Runoff

The link between sheep overgrazing and runoff has recently been examined under an Environment Agency R&D project “Impacts of Grazing and Upland Management on Erosion and Runoff” (APEM, 1998). Increasing stock density, primarily as a result of European Agricultural Policy, has been demonstrated to result in reduced vegetative cover, increasing the speed of runoff during rainfall resulting in a more rapid rise in riverine water levels and discharge. Consequently stream power is increased, resulting in elevated levels of erosion. This in turn has the combined effects of destabilisation of river substrates and increased sediment transport from bank erosion. The net impact on a river system is a combination of gravel instability and siltation. Both mechanisms can impact seriously upon salmonid egg survival.

In response to the National Grazing Management Team Report, DEFRA is currently proposing that Kinniside common is designated as an Environmentally Sensitive Area (ESA) under the Common Agricultural Policy (CAP), which will have the effect of reducing sheep numbers and with concomitant compensation payments to graziers. However, it is important to realise that reductions in sheep by themselves are not necessarily the whole answer to the problem.

It is well established that changes in farming economics have lead to widespread changes in hill-farming practice. Not only are more sheep kept in upland areas, encouraged by CAP subsidy; but also they are kept there over the winter with fodder to supplement the diet. In by-gone days, hill-sheep would be transported to lowland areas with less severe climate, sometimes hundreds of miles away, to over winter before returning in the spring. Finally, and perhaps most importantly stock is no longer shepherded. In the absence of shepherding sheep are free to roam where they choose on the open hillside, and can over-graze and damage preferred areas (APEM, 1998).

Sheep become adapted to localised conditions very quickly, and if constrained within areas such as nutrient deficient peat soil based grazing, will fare well after a period of adjustment (Halliday pers. comm.). If however, they are allowed to select the more palatable and nutritious mineral soil based grasses in the valley bottoms, they will lose their ability to cope with the harsher environment on the fellside. Hence, they will stay in the valley bottoms generating an uneven grazing pressure in these most sensitive areas adjacent to streams and riverbanks.

Furthermore, with increasing leisure activity on the high fell, particularly hill walkers and mountain biking, sheep will have a tendency to move away from this disturbance. Again, they will migrate towards the valley bottoms. In absence of shepherding these displaced sheep, particularly pregnant ewes, will stay contributing even further to much higher localised levels of overgrazing than might be apparent than when taking the whole fell into consideration.

Modern land management practices, such as high sheep densities and upland drainage have been implicated in influencing the hydraulics and hydrology of rivers, resulting in increases in peak flows and increased velocities (Hendry & Cragg-Hine 1995, Robinson, 1986; Howe et al., 1967; Burt, 1995). Evans (1996) provides evidence that as sheep numbers increased in the north Derwent catchment, so too did the rate of run-off, an observation supported by a study in Shetland by Birnie and Hulme (1990).

The influence of land management on hydrology has however, been the subject of much debate over the last century. In a detailed review of the subject Robinson (1990) concluded that a number of complex processes are involved and hence there are dangers in over-simplification of cause and effect; soil type, gradient, nature of drainage and scale all being important.

Nevertheless studies from other countries have established a link between overgrazing and changes in runoff . For example, Owens, Edwards and Van Keuren (1997) have studied the runoff and sediment losses from a small pastured catchment in eastern Ohio (US). Using a variety of grazing regimes they demonstrated that the increased runoff and erosion in the initial 12 year period of their study resulted from the non-rotational winter feeding on the pastures. In Australia, Greene et al (1998) investigated the effects of grazing regimes on the surface soil properties of a dunefield land system. At high intensity grazing (4 animals per hectare) there was a rapid depletion of perennial grasses, removal of most of the shrubs and a conversion of the soil structure to one that was either easily erodable or, formed a strong, physical crust. They conclude that this crust may cause a change in the hydrology of the land system and limit recovery of palatable sward, thereby propagating grazing pressure elsewhere. Mwendera and Saleem (1997) working in Ethiopia, assessed the hydrological response to cattle grazing in the Ethiopian Highland using study plots and multiple grazing regimes. They determined that heavy to very heavy grazing pressure (3.0 animal unit months (AUM) ha^-1 and 4.2 AUM ha^-1 accordingly) significantly increased surface runoff and soil loss, as well as reducing the infiltrability of the soil.

In the Calder catchment there is some circumstantial evidence of the effects of overgrazing on stream bank erosion high up on Worm Gill, immediately above the UU abstraction point where the river splits into two similar small tributaries, both headwater streams about 1.5 m wide high on the fell. The left-hand tributary, although flanked by relatively steep banks was accessible to sheep for grazing, whereas the right bank was too steep, and formed a small gorge at its entrance which excluded sheep. During the April survey, both streams were intact, with good mixed juvenile habitat throughout. However, following the flood of August that year, the more open left hand stream which was accessible and subject to grazing suffered major bank collapse into the hillside with large quantities of soil and sand entering the river. The stream was badly braided and littered with large cobble and boulder, becoming so shallow that the juvenile habitat was no longer present. In contrast, the ungrazed stream, even though only 20 or so metres away over a flat plateau of fell, was untouched by the flood, it’s banks and habitat intact. Although this evidence must be regarded as anecdotal, in that there was no detailed measurement of the stream, gradients, the surrounding riparian zone or flows, it does provide a degree of corroboration that overgrazing in the riparian area adjacent to the stream is occurring in the Calder catchment and that this may have implications with regard to the susceptibility of the banks to withstand flood waters.

Conclusions

The investigations conducted into the fish population of the River Calder have confirmed that the salmon stock has been subject to a decline over recent years. The rod catch remains erratic, whilst redd counts and juvenile densities are significantly lower than recorded previously. In particular Worm Gill, a once prolific nursery stream, now contains few if any juvenile fish for much of its length. Spawning gravels, although present in abundance, are unstable and contain unacceptably high concentrations of fines.

It has been suggested (N. Haycock, pers. comm.) that this is a natural phenomenon, and that the river is merely beginning an active eroding geomorphological phase that may last several hundred years. Certainly there is evidence of high levels of activity in the past with historic exposed banks visible in the floodplain (G. Maas, pers. comm.). The decline in fisheries could therefore be a by-product of this natural process.

From the data gathered during the course of this study it cannot be claimed that cause and effect have been demonstrated. However, in addition to the decline in the salmon population, there is evidence of overgrazing. Whether this overgrazing is at a level sufficient to promote erosion, either across the whole fell or indeed just within the valley bottoms is debatable, although observations from the walkover survey are compelling. Furthermore, the question of a link between overgrazing and increased stream power and erosion is also inconclusive, both within this study and arguably further afield, although again, examination of the data available such as it is, is persuasive. Finally, the presence of large quantities of fines, presumably as a consequence of sediment run-off is a worrying development. Although again, at this stage siltation cannot be attributed to overgrazing or changing hydrological conditions, sufficient circumstantial evidence, together with experience from other studies is available to give cause for concern.

In conclusion therefore, the study has provided sufficient support for the hypothesis that there is a link between declining salmon in the Calder, changing river discharge patterns and overgrazing, to suggest that this issue is investigated in a much more thorough and controlled manner than has been possible here. Detailed hydrological studies in relation to stream salmonid productivity are required for a range of vegetation states at various levels of grazing pressure. If, as is anticipated, global warming precipitates a change in rainfall patterns, then the issue requires even more urgent investigation.

References

Anderson, P., Radford, E. 1994. Changes in Vegetation Following Reduction in Grazing Pressure on the National Trust’s Kinder Estate, Peak District, Derbyshire England. Biological Conservation, 69.

APEM, 1998. Impacts of Grazing and Upland Management on Erosion and Runoff. R&D Technical Report P123, Environment Agency, Bristol. ISBN 1 873 160 95 X.

Beall, E., and Marty, C. (1987). Optimisation of Natural Reproduction of Atlantic Salmon in a Spawning Channel: Effect of Female Density. Restouration des Rivieres a Saumon. INRA, Paris, 231-238.

Birnie, R.V. Hulme, P.D. 1990. Overgrazing of peatland vegetation in Shetland. Scottish Geographical Magazine 106.

Burt, T.P. 1995. The role of wetlands in runoff generation from headwater catchments. In Hughes, J. Heathwaite, L., (Eds). Hydrology and Hydrochemistry of British Wetlands. John Wiley & Sons, Chichester, p 21-38.

Environment Agency (1998) West Cumbria local environment agency plan, Environment Agency, Penrith.

Evans, R. 1977. Overgrazing and Soil Erosion on Hill Pastures with Particular Reference to the Peak District. Journal British Grassland Society. 32, 65-76.

Evans, R. 1996. “Soil Erosion and its Impacts in England and Wales” Friends of the Earth Trust. London.

Hendry, K. and Cragg-Hine, D., (1995) River Wyre: Salmon & Sea Trout Spawning Habitat Restoration/Creation. NRA NW Region Internal Report.

Hendry, K. and Cragg-Hine, D. (1997) Restoration of riverine salmon habitats. Fisheries Technical Manual 4, Environment Agency, Bristol.

Hendry, K. & Cragg-Hine, D., (1998a). The River Calder Survey. Calder Conservation Committee. BW1/2202 (BNFL). pp 71.

Hendry. K., Johns, M & Fisher E., (1998b). Upland Management: The Implications for Fisheries. IFM Annual Study Course Conference Proceedings, Cambridge. p80-83.

Hendry, K., Cragg-Hine, D., O’Grady, M., Sambrook, H. and Stephen, A. (In press) Management of habitat for rehabilitation and enhancement of salmonid stocks. Fisheries Research. Proceedings of Salmonid 21C, Penrith, 2000.

Howe, G.M., Slaymaker, H.O., Harding, D.M., 1967. Some aspects of the flood hydrology of the upper catchments of the Severn and Wye. Transactions of the Institute of British Geographers, 41, 33-58.

Kennedy, G.J.A. 1988. Stock enhancement. In:.Mills, D.H & Piggins, D. (eds.) Proceedings 3^rd International Atlantic Salmon Symposium, Biarritz, October 1986. p345-372.

Kennedy, G.J.A. and W.W. Crozier, 1991. Strategies for the rehabilitation of salmon rivers - post project appraisal. In: Mills, D.H. (Ed.) Strategies for the rehabilitation of salmon rivers. Proc. Joint Conf. Atl. Salm. Trust/IFM Linnean Soc.: pp 44 - 62.

National Grazing Management Team; (2001) Kinniside Common, Cumbria. Vegetation and grazing management evaluation, 2000. Rural Development Service, Leeds, WRS: MHP117.

NRA, (1994), Derwent & Cumbrian Coast Catchment Management Plan, Environment Agency, Penrith.

NRA (1994) The NRA national fisheries classification scheme: a users guide. R and D Note 206, NRA, Bristol.

Orr, H. 2000. The impact of recent changes in land use and climate on the River Lune: Implications for catchment management. Unpublished PhD Thesis. University of Lancaster.

Peterson, R.H. (1978) Physical characteristics of Atlantic spawning salmon gravel in some New Brunswick streams.Fisheries and Marine Service Technical Report, 785. St Andrews, Canada: Fisheries and Environment.

Robinson, M., 1986. Changes in catchment runoff following drainage and afforestation. Journal of Hydrology. 86, 71-84.

Robinson, M., 1990. Impact of improved land drainage on river flows. Institute of Hydrology, Report No. 113, Wallingford, pp226.

Roche, P. (1994) Habitat availability and carrying capacity in the French part of the Rhine for Atlantic salmon (Salmo salar). Water Science Technology, 29, 257-265.

SNH (1999). Impacts of Climate Change on Plants, Animals and Ecosystems in Scotland. Information & Advisory Note 111, 8pp.

Theurer, F. D., Harrod, T. R. and Theurer, M., 1998 Sedimentation and salmonids in England and Wales. Environment Agency, Bristol, R&D Technical Report P194, 63 pp.

Walling, D. E., Collins, A. I. and McMellin, G., 2001 The provenance of interstitial sediment retrieval from salmonid spawning gravels in England and Wales: a reconnaissance survey based on the fingerprinting approach. Exeter University/Environment Agency Report, October 2001, 39 pp.

[1] A small proportion (<10%) of the river adjacent to the estuary was classified as C (fair) until recently due to intermittent industrial pollution, although these problems have now been resolved.

[2]. Border Esk, Eden, Derwent, Cocker, Ellen, Ehen, Irt, Esk, Duddon, Leven, Kent, Lune, Ribble, and Wyre.

[3]. A horizontal trend line would indicate that the target river was no different to the control group. A negative asymptote towards zero would indicate that catches were improving relative to the control group. A positive asymptote would indicate deteriorating catches relative to the control group.

[4] 10,019 m² of Juvenile habitat – fry & parr combined

[5] Invertebrate & water quality data reveal no deterioration from Grade A GQA status over this period.

[6] EA Fisheries Technical Manual 4 - Restoration of riverine salmon habitats.

[Return to the top of the page]