Synthesis of the Upper Thurne Research and Recommendations for Management

Report to Authority

Holman IP and White SM

July 2008

Department of Natural Resources Cranfield University Cranfield Bedfordshire MK43 0AL Telephone: +44 (0) 1234 750111 Ext. 2764 Fax: +44 (0) 1234 752970

Executive Summary

The importance of the Upper Thurne (including , Horsey Mere, North and South Broads) for biodiversity is recognised under national and international conservation legislation. Appropriate water management (water resources, quality and flood defence) is fundamental to conservation in the broads. However, the incomplete understanding of the surface water system and their interactions with the wider catchment, particularly with respect to nutrient cycling, was recognised by the Appropriate Assessment Team (Broads Authority, 1999). The resulting workshop, held in in November 2001, developed a framework for a research and monitoring programme which intends to inform the ongoing activities which contribute towards the published 20 year Vision for the Upper Thurne water space.

The ecology of the Upper Thurne broads has gone through a number of Phases in response to changing environmental conditions:

 Phase 1 – until the early 20th century – dominated by stoneworts (charophytes) and low- growing waterweeds, due to the low nutrient levels;  Phase 2 – early to mid 20th century– luxuriant aquatic plant growth dominated by taller growing species better able to take advantage of enhanced nutrient levels;  Phase 3 – mid 20th century to present – phytoplankton dominated system due to high nutrient and salinity levels.

The Upper Thurne research has aimed to help understand how a range of hydrological, chemical and ecological factors contribute to achieving ‘clear water’ (Phase 1) conditions in Hickling Broad. The research has included catchment modelling of water and nutrient movement, groundwater modelling, hydrodynamic modelling of water and salt movement from the sea to the Upper Thurne, mesocosm (small experimental ponds) experiments of salinity effects, the use of remote sensing, lake sediment analyses, laboratory experiments of stonewort response to a range of factors including water temperature, cutting, establishment, pollutant concentrations.

The report has 4 sections:

1. How the Upper Thurne water spaces have changed; 2. A description of the current status of the Upper Thurne waterways, and how these compare to the Favourable Condition criteria under the EC Habitats Directive; 3. A synthesis of the activities to identify the significant catchment water management issues, which is focussed around salinity and ochre, biocides and heavy metals within the sediment and water column, point and diffuse sources of nutrients (nitrate and phosphate) under current and future climate, sea level rise and coastal protection, monitoring and the population biology of charophytes; 4. Recommendations for management actions to address the significant issues previously identified and thereby achieve Favourable Conservation Status.

The recommendations for management actions to achieve Favourable Conservation Status centre around:  reducing salinity and ochre discharges from the land drainage pumps, principally the Brograve pump. An approach to identifying a solution is suggested based on principles of no significant change in current flood risk; compatibility with a range of farm systems; being consistent with existing agri-environment schemes; having a means of removing any seawater from a future coastal breach without discharging it through the Special Area of Conservation; and that there is recognition that the Brograve sub-catchment is a system that is ‘naturally’ brackish and which produces limited ochre;  reducing diffuse source losses of nutrients from agriculture. Assuming that the farming community are following Codes of Good Agricultural Practice and Good Agricultural and Environment Condition requirements, a non-exhaustive list of practical measures to reduce nutrient losses from agricultural activities are suggested.

i Key findings from the Upper Thurne studies

The importance of land drainage to salinity management  Much of the salinity in the Brograve drainage system enters via the coastal marshes, especially Hempstead marshes;  Changes to the management of the land drainage systems have the potential to reduce the salinity entering the rivers and broads from the pumps e.g. raising water levels in the Hempstead Marshes by ~ 1 metre might lead to a 15% reduction in the amount of salt being discharged by drainage pumps;  Such changes need to be considered in conjunction with flood risk and ‘knock-on’ effects to neighbouring drainage systems and the .

Sources and transport of nutrients from the catchment to the rivers and broads  Total Phosphorous concentrations in the Upper Thurne broads regularly exceed the target limit for favourable conditions, but point sources appear to contribute little;  Agricultural drains and water from the drainage pumps have the highest nitrogen (N) concentrations, but N concentrations in the broads are reduced through biological uptake or sedimentation;  Increased rainfall and higher temperatures through climate change will increase nutrient (N and phosphorus (P) and sediment losses from the land;  Erosion control measures, on susceptible soils and slopes, should be employed as part of good agricultural practice to reduce sediment and nutrient losses.

The movement of water and salt within the rivers and broads  Water and salt being discharged from the land drainage pumps form the main source of water and salt entering the River Thurne from its catchment;  Constrictions within the river system, principally at old bridge, where the narrow openings within the bridge impede both the downstream and upstream movement of water (depending upon tidal conditions);  The role of the land drainage pumps changes with the tides – reducing salinities during extreme tides but increasing background salinity during normal tides.

Environmental needs of the Stoneworts  Increased water temperature leads to considerably higher growth rates and seed production of Stoneworts, suggesting that climate change may influence future growth patterns;  While laboratory plant cutting experiments showed that cut stems had the ability to re-grow and branch, uprooting of Stoneworts, which is a risk of weed harvesting on the broad, leads to a high rate of plant mortality;  Stoneworts are affected by a range of chemicals, particularly metals and boat antifouling paints, which have been found in sediment and water in Hickling Broad;  Stoneworts in the broads are likely to be able to withstand a salinity increase of up to around 12.5 % of typical recent values for Hickling Broad, but such increases are likely to cause changes in community structure;  Replicated small ‘pond’ experiments have shown that modest reductions in salinity, to around 1600–1800 mg Cl L-1 may have a substantial effect on Total Phosphorus and chlorophyll and hence on the potential for plant growth.

Looking down from above (‘remote sensing’)  ‘Remote sensing’ data can map the development of algal blooms in the broads, potentially providing early warnings for water users such as sailing clubs;  It is shown to distinguish the presence of potentially toxic blue-green algae from other non- toxic species;  The distribution and health of aquatic plants, both around the margins of the broads and, if the water is clear, those submerged beneath the waters surface can be mapped.

ii Table of Contents

Executive Summary ...... i

Key findings from the Upper Thurne studies...... ii

Table of Contents...... iii

Introduction ...... 1

How the Upper Thurne water spaces have changed ...... 2

A description of the current status of the Upper Thurne waterways; ...... 4 Favourable Condition criteria and current condition...... 4

Synthesis of the research activities to identify the significant catchment water management issues...... 4 Introduction...... 4 Aquatic plant monitoring programme ...... 6 Salinity and ochre...... 6 Introduction...... 6 Ecological effects of the salinity ...... 7 Causes of surface water salinity ...... 9 Solutions to the ochre and salinity problems...... 10 Hydrodynamic effects of changes to drainage management...... 12 Biocides and heavy metals...... 13 Point and diffuse sources of nutrients...... 15 Ecological effects of nutrients ...... 15 Point sources of P ...... 16 Sediment sources...... 16 Diffuse sources of sediment, phosphorus (P) and nitrogen (N) ...... 17 Effects of climate change on water quality...... 18 Sea level rise and coastal protection ...... 19 Monitoring ...... 19 Population biology of charophytes...... 19

Recommendations for management actions to address the significant issues previously identified and thereby achieve Favourable Conservation Status...... 21 Reduction of salinity and ochre...... 21 Reduction of diffuse source pollution (N, P, sediment and crop protection products)...... 22

References...... 22

iii

Introduction The low-lying River Thurne catchment is located in northeast (TG 4020), adjacent to the North Sea coast (Figure 1). The upper portion of the river system incorporates large, shallow ‘broads’ connected directly, or via channels, to the main river. This includes Hickling Broad, Horsey Mere, Martham North and South Broads. The importance of the Upper Thurne for biodiversity is recognised under national and international conservation legislation. The Upper Thurne Broads and Marshes is designated nationally as a Site of Special Scientific Interest (SSSI) and internationally as part of the Broads Special Area of Conservation (SAC) and Special Protection Area (SPA) under the EU Habitat and Birds Directives, respectively. Hickling Broad and Horsey Mere were also designated in 1976 as Wetlands of International Importance especially as a Waterfowl Habitat under the Ramsar Convention, with the whole Upper Thurne SSSI subsequently designated as a Ramsar site in 1994.

Figure 1 The Thurne catchment, showing the (black) broads and watercourses and the (grey) main designated sites (The Broads SAC/Broadland SPA/Broadland Ramsar, Upper Thurne Broads & Marshes SSSI, Calthorpe Broad SSSI, Priory Meadows SSSI, Ludham & Potter Heigham Marshes SSSI, Shallam Dyke Marshes SSSI and the Winterton-Horsey Dunes SSSI and SAC.)

According to Hails (1996), the broads are “an example of the rich biological diversity and productivity which has arisen in an essentially cultural landscape. The wetland habitats and species now found in the region are the result of centuries of manipulation by local communities for fuel production, wetland plant products, and extensive summer grazing. Maintenance of the area's biological value depends on the continuation of traditional management practices, combined with measures to restore damaged habitats and to counter eutrophication from sewage and agricultural run-off, as well as the effects of mass tourism, and, in the longer term, rising sea levels.”

Appropriate water management (water resources, water quality and flood defence) is fundamental to conservation in the broads. However, the incomplete understanding of the surface water system and their interactions with the wider catchment, particularly with respect to nutrient cycling, was recognised by the Appropriate Assessment Team (Broads Authority, 1999). The resulting workshop, held in Norwich in November 2001, developed a framework for a research and monitoring programme.

The 20 year Vision for the Upper Thurne water space given in Broads Authority (2006) provides a consensus-based view, produced by the Upper Thurne Working Group, of how the Upper Thurne should operate in 2026. It provides a ‘target’ for ongoing activities which should be informed by the recent and ongoing research. The 20 year Vision encapsulates:  A reversion of drained marshland to low intensity agriculture, leading to a reduction in salinity and ochre loads into the broads;  ‘Gin clear’ waterways with an aquatic plant dominated community, leading to an increase in the richness and diversity of birds (including wintering waterfowl), fish and invertebrates;

1  Water-based recreation, particularly sailing/boating and fishing;  A vibrant local economy providing high quality facilities, goods and services, with many local people earning a livelihood in the protected landscape;  Management of the water space based on mutual understanding and agreement, providing a highly regarded example of what can be achieved in integrating different interests and uses into a special landscape.

The overall aims of this current Synthesis proposal are to condense and 'translate' the understandings gained from the research carried out since the November 2001 workshop on aspects of the Upper Thurne catchment, as a basis for informing Partner organizations in the development and application of restoration options and/or revised management practices within the catchment(s). How the Upper Thurne water spaces have changed The ecology of the Upper Thurne broads has gone through a number of Phases (summarised in Table 1) in response to changing environmental conditions:  Phase 1 – until the early 20th century – dominated by stoneworts and low-growing waterweeds, due to the low nutrient levels  Phase 2 – early to mid 20th century (Photo 1 and 2)– luxuriant aquatic plant growth dominated by taller growing species better able to take advantage of enhanced nutrient levels  Phase 3 – mid 20th century to present – phytoplankton dominated system due to high nutrient and salinity levels

Table 1 Aquatic plant community phases in the Upper Thurne (from Broads Authority, 2006) Phase 1 Rivers and broads in their  Low levels of phosphorous and moderate nitrogen with pristine state. high calcium carbonate.  Water plant communities include Stoneworts (Chara Elements still present in sp), Holly-leaved naiad (Najas marina) and low growing Martham Broads. waterweeds such as Reddish pondweed (Potamogeton alpinus) and bladderwort (Utricularia intermedia) Hickling Broad moved into  Phytoplankton virtually absent this Phase in the late 1990s  Wealth of invertebrate life into the early 2000s  Water crystal clear Phase 2 Usual Phase for Horsey  Gradual rise in nutrient loading Mere as at 2005/06.  Robust, taller, nutrient demanding plants colonise with competitive advantage over Phase 1 communities. . Plants such as Horned Pondweed (Zanichellia palustris), Fan-leaved Water-crowfoot (Ranunculus circinatus), Hornwort (Ceratophyllum demersum), Greater bladderwort (Utricularis vulgaris), Yellow Water- lily (Nuphar lutea), White water-lily (Nymphaea alba), Spiked and Whorled Water-milfoils (Myriophyllum spicatum and M verticillatum) and Fennel-leaved pondweed (Potamogeton pectinatus)  Increase in periphyton (epiphytic algae) growing on plants, which Phase 2 species are able to withstand better than Phase 1  High biological productivity, exemplified by large populations of diverse fish species, a wealth of invertebrate species Phase 3 Hickling Broad during the  Phytoplankton dominance 1970s and 1980s and  Major reduction in the biomass and diversity of the currently (2008) aquatic flora  Increased water turbidity  Phosphorus levels greater than 100μg l-1  Accelerated rate of sediment deposition  Loss of aquatic invertebrate diversity

2 Photo 1 “Gathering water lilies” (1886) by PH Emerson, showing Phase 2 emergent aquatic vegetation in a Norfolk Broad

Photo 2 The Brograve level, 1949, under traditional high water level management (Photo: Jocelyn Gardiner)

In the 1980s and 1990s, vegetation in Hickling Broad started to recover after the closure of the Martham landfill site and decline of the gull roost. Neomysis integer recovered, and helped to reduce algal crops (Moss, 2001). In 1998, the lake came back to the clear water Phase 2 state dominated by charophytes, including the rare Chara intermedia, although salinity remained high. The charophytes growth caused problems in navigation, and the Broads Authority, who are responsible for navigation and conservation, implemented a series of cutting trials. However, in 2000 it was not necessary to cut, as the condition of the plants declined, which continued in 2001. The vegetation showed a slight recovery to 2003, but has since declined (Table 2).

Table 2 Area and annual increase in dense C. Intermedia lawn coverage in Hickling Broad (source: Broads Authority) Year Area (ha) Annual Increase (ha) % of Broad 1994 13.6 11.7 1995 17.1 3.5 14.7 1996 25.8 8.7 22.2 1997 33.3 7.5 28.7 1998 39.0 5.7 33.6 1999 48.5 9.5 41.8 2000 18.8 -29.7 16.2 2001 No data - - 2002 No data - - 2003 31.2 12.4 (since 2000) 26.9 2004 20.5 -10.7 17.7 2005 11.0 -9.5 9.5 2006 0 -11.0 0

3 A description of the current status of the Upper Thurne waterways;

Favourable Condition criteria and current condition To achieve Favourable Condition in the Upper Thurne broads, the following criteria need to be met:  Good water quality: o annual mean concentration of Total Phosphorus < 30 μg/l (Broads Authority , 2006); o No targets currently set for nitrogen; o Upper limits for chloride set at 600 mg/l for Hickling Broad, and 1000 mg/l for Horsey Mere and Martham North and South Broads  Clear water - Secchi disc visible to bottom of water column throughout the water bodies  Aquatic plant beds present across the whole of the water bodies, except in marked channels (e.g. main channel across Hickling Broad and side channel connecting it to Catfield Dyke)  Actively growing margins  Disturbance free winter bird refuges (for feeding as well as resting/sleeping) over 50% of the total area of the Upper Thurne

The current condition of the lake features are:  Hickling Broad - Unfavourable declining  Heigham Sound - Unfavourable recovering  Horsey Mere - Unfavourable – no change  Martham North and South Broads - Favourable

The above "Favourable Condition" criteria are taken from the Upper Thurne Water Space management plan (Broads Authority, 2006). The full Condition Tables produced by Natural England contain the criteria for the terrestrial features of the SSSIs, which are not dealt with here. Synthesis of the research activities to identify the significant catchment water management issues

Introduction The Upper Thurne research has aimed to help understand how a range of hydrological, chemical and ecological factors contribute to achieving ‘clear water’ conditions in Hickling Broad (Figure 2). These factors are affected by a range of pressures exerted by activities by the human users within the catchment and the wider environment which include drainage, pollution (point and diffuse sources), abstraction, coastal breaches The pressures exerted by these activities have been ordered by the perceived significance of their effect on the achievement of Favourable Conservation Status. This section is completed by a synthesis of activities which provide underpinning knowledge of the ecological requirements and population biology of charophyte species.

Figure 2 Schematic of factors contribute to achieving ‘clear water’ conditions in Hickling Broad.

A conceptual framework of the factors affecting chara growth in Hickling Broad was developed by Servera-Martinez (2005) from literature review (Figure 3).

4 Figure 3 Conceptual framework linking factors affecting charophytes in Hickling Broad, Norfolk (from Servera-Martinez, 2005).

5 Aquatic plant monitoring programme The Chara intermedia lawns in Hickling Broad have been monitored since the mid 1990s (Table 2). The Chara lawns developed from the 1994 ‘foci’ in shallower water distant from the navigation channel, and have expanded only onto silt sediment. Expansion of the lawns appeared to be largely by vegetative spread, as no germinating oospores or seedlings were observed. Colonisation of bare sediment also occured by growth from plant fragments which readily produce new rhizoids and shoots from the branchlet nodes, and were constantly being generated by bird grazing and recreational activity in the broad. In the later 1990s the dense lawns increased annually not only in extent, but also in vertical height (Harris, 2000) through a cycle of lawn overwintering and new summer growth at the top of the lawn. Despite increasing grazing pressure, especially due to the gradual build-up in coot numbers, the dense lawns continued tom increase in height annually, reaching greatest heights in 1999 (Harris 2000). However, following the dramatic decline in the Chara lawns from 2000 and subsequent poor growth vigour attributable to other factors, new growth was not able to keep pace with removal by localised grazing by coot in Sailing Club Bay (Harris 2004). Several factors may have been involved in the 1999/2000 die-back, the most obvious of which were the release of nutrients available for algal growth via bird faeces during the period of intense coot grazing in autumn 1999 and the observed widespread anoxia in the tall, dense lawns which was inimical to plant growth. The gradual loss of the dense mono-dominant lawns of C. intermedia which covered 42% of the broad, resulted in the exposure of 48ha of unvegetated, highly mobile sediment. The organic-rich sediment is easily disturbed, especially in shallow water, and lifted into the water column by wave action from where it is subsequently deposited onto aquatic plants (Harris, 2006). In the deeper parts of the broad, plant propagules are continually covered by a ‘rain’ of organic sediment which also reduces light penetration through the already turbid water, and no species have been able to successfully colonise the sediment (Harris, 2006). Poor water clarity and sediment disturbance by the strong winds experienced in late summer 2004 are both likely to have contributed to a lack of establishment and further loss of Chara. The winds were strong enough to uproot large amounts of milfoil, and sediment disturbance may have been sufficient to dislodge or smother small plants of Chara in areas with patchy cover (Harris, 2004). In the early 1990s, when milfoil was already abundant in Hickling, plants of C. intermedia colonised the sediment surface below the milfoil canopy, often using the stems for support to grow vertically. It is highly likely that milfoil acted as a ‘nurse’ at this stage, protecting the stonewort from sediment disturbance. In the absence of milfoil, there has been no recolonisation of the mobile sediment by C. intermedia (Harris, 2006). If this scenario is correct, with M. spicatum the primary and C. intermedia the secondary coloniser, Harris (2006) suggests that it may be many years until there is sufficient stonewort biomass in Hickling to bring about a return to clear water.

Salinity and ochre

Introduction The unusual saline nature of the surface waters in the River Thurne catchment has been recorded since at least the 1892 High Court (Chancery Division) case of Micklethwait v Vincent (1892), the 'Hickling Broad Case' (Innes, 1911). The source of the salinity was thought to be "probably due to salt springs" within Hickling Broad and Horsey Mere (Gurney, 1904), a view supported by Innes (1911). Following the work of Pallis (1911), it is generally recognised that the source of the salinity within the Brograve sub-catchment is by direct underground communication between the sea and the dykes. The Thurne valley has the highest proportion of acidified soils in Broadland; occurring mainly on drained land, which produce both acidic drainage water and the brown, orange brown or yellow `ochre', Fe(OH)3. The lowering of water levels as a consequence of the drainage improvement to the Brograve and Somerton Level, which lead to increased salinity, also increased the formation of ochre significantly. White et al. (2005) estimated an ochre-derived sediment load of 800 t year-1 from the Brograve and Somerton New Pumps.

6 Ecological effects of the salinity Replicated laboratory experiments by Lambert (2007) to investigate the effects on growth of Chara connivens, C. intermedia and N. obtusa to salinity variations in Hickling Broad showed that Chara connivens and Nitellopsis obtusa growth was significantly inhibited (p< 0.01) at median salinities between 1996 and 2006 of 6.1 mS cm-1. Chara intermedia growth rate was significantly higher than these two species at this salinity range and appeared to show accelerated growth rate at (as shown by mean relative growth rate) at 12.5 – 50% addition of seawater to Hickling water. However, the increased growth response of the Chara intermedia was not statistically significant across the range due to wide variation in individual plant response. The experiments also demonstrated that all three species would suffer stress, as indicated by a reduction of photosynthetic efficiency at 25% addition of seawater to recent median Hickling Broad salinities. The experiment results suggest that changes in salinity within the Broad are likely to change the charophyte species assemblage.

Lambert also carried out a three-year survey (2004-06) of 26 environmental variables at 124 historical chara-holding water bodies which supported 18 charophyte species. The purpose of which was to evaluate the environmental ranges of charophytes in the field. The electrical conductivity data appears to have an almost trimodal distribution with the Hickling Broad data exhibiting a distribution with a lower conductivity limit of around 3600 µS cm-1, although another statistical test (Waller-Duncan ‘k’ test) placed the charophyte species recorded in Hickling Broad into a brackish range of 5456 - 6112µS cm-1. Certain chara species show clear salinity tolerance with Chara hispida being recorded growing over the widest chloride range, and Chara intermedia at the highest mean and median concentration. However, the chloride ion concentration clearly shows the Thurne sites (Hickling Broad and Horsey Mere) as outliers within all charophyte species’ chloride data.

Barker et al (2007) describes a two-year mesocosm experiment, based on 48 tanks of 3 m diameter containing around 3m3 of water when full. The experiment had four salinity treatments, of around 600 mg Cl L-1 (as in the early years of the 20th century), 1000 mg L-1 (which is about half the current value), 1600 mg L-1 and around 2500mg L-1, the latter two spanning current salinity values in Hickling Broad. To each pond were added three packages each (25-50 g wet weight) of Chara intermedia, Chara hispida L, Chara globularis Thuill. Myriophyllum spicatum, Potamogeton pectinatus, Hippuris vulgaris, Callitriche sp., Elodea canadensis Michx., Ranunculus circinatus Sibth., Lemna trisulca L. and Ceratophyllum demersum L.; mixed innocula of zooplankton from Hickling Broad and Daphnia species and two male stickleback. Nitrate and Phosphate were added in excess of current loadings in Hickling Broad to ensure that responses seen were not due to nutrient shortage. In the second year, two further males and two female sticklebacks were added and reproduction allowed the fish community to rise to the carrying capacity.

Although Barker et al. report statistical analyses from the first year of data, they acknowledge that the first year was not strictly intended to be an experimental year, and that the effects of low versus normal populations of fish in the two years are statistically not comparable. Therefore focusing on the results from the second year of the experiment, a number of findings were made: 1. In the presence of fish predation, cladocera declined greatly and copepods increased in proportion as salinities increased. Daphnia generally disappeared in the mesocosms above Low (600 mg Cl L-1) salinities, despite laboratory tests showing greater resilience to higher salinities; 2. Salinity was significantly positively correlated with both Total Phosphorus (TP) and phytoplankton chlorophyll a (Figure 4). The highest salinity level was associated with a marked increase in the TP, due to the reduction of sulphate to sulphides which remove iron, allowing phosphate to move from the sediment into the overlying water. Due to the loss of cladocera at this highest salinity, the remaining grazing community (which was dominated by copepods) could not effectively control the increase in algal populations, measured as chlorophyll a. The soluble phosphorus released from sediment was taken up rapidly by phytoplankton, measured as chlorophyll a, keeping SRP low, but allowing TP, which includes the P incorporated into living [algal]

7 cells to increase. The uptake of SRP was greater with increasing salinity as salinity decreased the production of the more efficient zooplankton grazers (daphnids first, then other Cladocera, followed by calanoid copepods, which are relatively inefficient grazers, but more tolerant of salinity)- reduced grazing resulted in more algae and more efficient use of available SRP released from the sediment; 3. Salinity did not influence periphyton chlorophyll a (which would affect the light climate at the plant surface) but did have some effects on the invertebrates likely to graze periphyton, as Gammarus duebeni became very abundant, suggesting that potential increased periphyton growth (due to higher P) was counterbalanced by a salinity-induced increased grazing pressure. 4. In addition to effects of salinity manifested through phosphorus availability and effects on the zooplankton community, it was also associated with reduced submerged plant species richness and a reduction in macrophyte PVI (Figure 5);

Figure 4 Changes in (left) Total Phosphorus and (right) phytoplankton chlorophyll a in relation to salinity treatment in a mesocosm experiment. Light dotted line, low salinity; light dashed line, moderate salinity; heavy dashed line, sub-present salinity; heavy line, high salinity [Aquatic conservation: marine and freshwater ecosystems, Control of ecosystem state in a shallow, brackish lake: implications for the conservation of stonewort communities, Barker T, Hatton K, O’Connor M, Connor L, Bagnell L and Moss B, Copyright © 2007. John Wiley & Sons Limited. Reproduced with permission]

Figure 5 Total plant abundance in relation to salinity in a mesocosm experiment. Dotted line, low salinity; light continuous line, moderate salinity; heavy dashed line, sub-present salinity; heavy line, high salinity. [Aquatic conservation: marine and freshwater ecosystems, Control of ecosystem state in a shallow, brackish lake: implications for the conservation of stonewort communities, Barker T, Hatton K, O’Connor M, Connor L, Bagnell L and Moss B, Copyright © 2007. John Wiley & Sons Limited. Reproduced with permission]

Barker et al propose a framework for the relationships linking salinity to plant performance in the mesocosms (Figure 6). Increasing salinity inhibits reproduction of different zooplankters at different concentrations, but the threshold of salinity on reproduction depends on the predation by fish - more predation, leading to greater vulnerability to salinity. A normal fish

8 population will reduce the numbers of the bigger zooplankters first, such that the large, visible and vulnerable Daphnia magna cannot coexist with fish. Fish will thus eliminate D. magna, the most efficient potential grazer, leaving the smaller, less efficient species. Although these smaller daphnids can co-exist with fish, given the availability of refuges such as plant structures or dark water, they are much more susceptible to increased salinity. Hence the combination of salinity plus predation leads to lower zooplankton biomass, which coupled with the chemical effect of salinity on phosphorus release from the sediment (which enables increased potential phytoplankton growth) produces more actual phytoplankton growth. Whilst the effect of salinity-induced increased phytoplanktyon chlorophyll a causing reduced light availability is clear, the case for increased periphyton shading is less apparent in the experimental data, but the overall effect of salinity on macrophyte dry weight is a statistically significant reduction.

Figure 6 Summary of main relationships linking salinity to plant performance in a mesocosm experiment on the Hickling Broad ecosystem [Aquatic conservation: marine and freshwater ecosystems, Control of ecosystem state in a shallow, brackish lake: implications for the conservation of stonewort communities, Barker T, Hatton K, O’Connor M, Connor L, Bagnell L and Moss B, Copyright © 2007. John Wiley & Sons Limited. Reproduced with permission]

Causes of surface water salinity Pallis (1911) mapped salinities in parts of the Upper Thurne long before the drainage improvements of the 1950s to 1980s. The Brograve Level is shown as having a high salinity of 15-24% that of seawater, as the freshwater from the Lessingham Valley was not providing dilution to the Brograve Level at this time (as occurs now) but was being discharged by Ingham Mill into the head of the Waxham Cut. Driscoll (1984) showed that drainage improvement in the West Somerton Level in the 1980's lead to a significant increase in dyke water salinity, as dyke deepening penetrated the underlying clay, removing the impediment to saline seepage (Holman, 1994; Holman & Hiscock, 1998). Within all of the brackish drainage systems (e.g. Brograve, Somerton, Horsey, Eastfield and Stubb), the distribution of salinities is not uniform. Holman (1994) showed that the salinity within individual ditches varied from essentially zero to over 60% seawater-equivalent (Figure 7), dependent on a number of topographical, geological and hydrogeological factors. The progressive drainage improvements and associated lowering of water levels have caused an increase in the salt load being discharged into the Thurne river and Broad system (Moss, 2001).

9 Figure 7 Distribution of surface water chloride concentration in spring 1991 (from Holman 1994)

Solutions to the ochre and salinity problems Because ochre production and dyke water salinity are both a consequence of land drainage and low water levels, it is appropriate to consider the studies which have looked at solutions together. Harding and Smith (2002) assessed solutions to the generation of ochre in the Brograve system and recommended:  Comprehensive remedial solutions– based on: o raised regional water levels to within 30 cm or 45 cm of the marsh level; o re-sealing the bed of the main drains within the coastal marshes (Hempstead Marshes, Great Moss Fen etc) with clay; o Construction of a coastal interceptor Spine Drain, running from Eccles-on- Sea to near Horsey Corner.  Local schemes or piecemeal remedial solution- based on: o Conifer bark filters, installed in pairs within the drains, to remove ochre; o A settling lagoon near the Brograve pump; o Clay lining of the coastal main drains; o Discharging drainage water from the Brograve system directly to sea; o Raising water levels locally; o Dyke widening to provide improved aquatic vegetation habitat.

Subsequently, ELP and Cranfield University (2005) reviewed the results from Harding and Smith (2002) in the context of delivering a partnership vision for the Brograve catchment to: “Identify a preferred, sustainable solution to address Ochre and salinity problems in the Brograve catchment, within the context of all the existing formats of farming in this area. Such a solution needs as far as possible to be balanced to meet the environmental requirements of the European Wildlife sites, and aims to be acceptable, and where possible beneficial, to people living and working in the catchment. Opportunities for additional environmental enhancement will be considered with rate payers and other stakeholders”

ELP and Cranfield University (2005) identified a range of scenarios of measures that aimed to provide comprehensive catchment wide solutions. Given the potential impacts on land, property and people, the selection of a preferred option was not possible. In addition, it was recognised that a solution to the high priority objective of solving the water quality problems

10 for the international sites, could be resolved with solutions which did not meet the full vision but which had lesser impacts on the catchment. The final scenarios investigated were:  Scenario 1-1 – included: o regional water level increases, whereby the water level in the drains would be maintained within 35cm of ground level during the summer, April-October inclusive, o construction of new shallower ditches in coastal marshes, and infilling of original deep ditches; o restoration of past aquatic and floodplain habitats  Scenario 1-2 – included: o regional water level increases, whereby water levels in the drains would be maintained at whatever levels are needed to maintain average water tables within 35cm of ground level o construction of new shallower ditches in coastal marshes, and infilling of original deep ditches; o restoration of past aquatic and floodplain habitats  Scenario 2 – included: o Splitting the Brograve drainage system, such that: . a new drain is installed to collect the saline water from the coastal marshes (Hempstead, Great Moss Fen, Waxham, Poplar Farm and Fir Tree Farm sub-catchements) and to discharge it directly to sea through the Hempstead marshes . The good quality water from the Lessingham Valley enters the Brograve Level, to provide dilution, to be discharged by the existing pump o Water level increases in the Brograve, Lessingham and Calthorpe sub- catchments, as in either Scenario 1-1 or 1-2  Scenario 3 – similar to Scenario 2, except that the new coastal drain is deepened to act as a (uncertain) sacrificial salinity line.

The appraisal of the above scenarios provides the following conclusions:

 The only scenario which is likely to fulfil the Vision is Scenario 1-2, but this scenario also has maximum impacts upon land use, livelihoods and potentially, on property.  If the primary objective is resolving water quality issues for the SACs, then Scenario 2 combined with the higher water level in the Brograve, Lessingham and Calthorpe sub-catchments, provides the maximum benefit as it removes the water with highest salinity and ochre while retaining the fresh water for discharge to Horsey Mere.  Scenarios 2 and 3 provide considerable flexibility to vary water level management within the non-coastal sub-catchments, depending on the balance between water quality benefit compared to impact on land use and flooding.  There is considerable uncertainty with Scenario 3 and very high and increasing running costs over time.  Both Scenarios 2 and 3 require discharge to sea, for which a discharge consent may be required from the Agency. In addition, the design of the outfall will need careful engineering design given the dynamic nature of the coast.  It is imperative that any proposed solution to the water quality problems in the Brograve catchment does not lead to increased flood risks. In taking any proposed solution forward, it was recommended by ELP and Cranfield University (2005) that, in addition to creating additional flood storage capacity within the Main Drains and floodplain, alternative control mechanisms to better manage Lessingham water levels are investigated  For Scenarios 2 or 3, it is suggested that there is a managed hydraulic connection between the Lessingham Valley and Hempstead Marshes to provide a means of evacuating seawater following a coastal breach from the Brograve catchment, rather than via the Brograve pump.

11 Simpson (2007) used a numerical groundwater model to simulate three previously proposed management or engineering remedial measures to assess their affect on saline inflows. He found that:  raising the water levels in the drains of the Hempstead Marshes from around -1.1 to - 1.8 mOD to -0.4 mOD will reduce the saline inflow into the Brograve sub-catchment by around 15%. This will decrease the overall saline inflow into the whole Thurne catchment from 3081 m3/day to 2822 m3/day;  (ii) lining the main drain of the Hempstead Marshes with low permeability material produces a reduction in the saline inflow into the Brograve sub-catchment by around 7%. The saline inflow into the whole catchment decreases from 3,081 m3/day to 2,958 m3/day;  (iii) The construction of a coastal interceptor drain could in theory prevent the inflow of saline groundwater into the Brograve system. However, such a drain would increase the saline inflow across the coastal boundary by around six times (from 3,081 m3/day to 19,750 m3/day), remove large quantities of fresh groundwater from the Pleistocene Crag aquifer and lead to high energy and pumping costs.

Simpson (2007) has shown that there are partial solutions to reducing the saline inflow into the drainage systems in this lowland coastal catchment, but stresses that any intended alterations must consider other potential impacts, such as changes to flood risk, land management restrictions or hydrodynamic effects on the Thurne river.

Hydrodynamic effects of changes to drainage management White et al. (2008) applied the ISIS hydrodynamic and water quality model to the Thurne system. They improved upon the original model set-up developed by Halcrows for the Broadland Flood Alleviation Project. It is apparent that the difficulty of calibrating the simulated river levels and salinities at the two locations within the Lower Bure and Thurne with good data (Acle Bridge and Repps) indicates that the factors controlling the hydrodynamics of this low hydraulic gradient system are variable in space and time. Nevertheless a reasonable match has been achieved at Repps and Acle Bridge across a diverse range of tide and river flow events which provides a degree of confidence in the robustness of the model.

Analysis of measured tide and salinity data, and model results show that:  In a high tide and low flow situation, the saline water comes up to Acle Bridge and sometimes further upstream to Repps and beyond. If the pumps are operating the salinity is reduced at Acle Bridge, Repps and Hickling as the salinity is diluted by the extra flow. so therefore: If salinity is from tidal influence, then turning off pumps has impact of raising salinity as the dilution of the tidal saline water from downstream is less.  In the situation of normal tides and with a variety of river flows, higher saline values at Repps can be caused by the water being pumped from upstream. The impact is diluted by higher river flows. The impact has generally disappated by the time it reaches Acle Bridge. If the pumps are tuned off the salinity levels drop therefore: If salinity is from pumps then turning off pumps causes a drop in salinity as pumps are no longer pumping saline water into the system.  If there is a sustained period of high tide levels with a low tidal range, the saline water gets trapped and salinity levels at Acle Bridge can remain high until the tide levels drop or the tidal range increases. The sustained levels of salinity seen at Acle Bridge under these circumstances are not repeated at Repps as the saline water does not get trapped as high up the system. The peak levels of salinity are seen at Repps for very high tides of over 2.5m ODN (at ) but they are not sustained in a period of higher tide levels.

The extreme scenarios of high tides (1 in 200 year tide) and high (1 in 100 year flow) and low river show similar patterns in Hickling Broad (Figure 8) and Horsey Mere:  The high tide levels cause high levels of salinity even with a high flow

12  At around 50 hours the impact of the high tide levels causes high levels of salinity especially when the river flow is low. This reduces as the tidal levels reduce.  For the option of the1 in 100 year flow, at 50 hours the salinity impact from the high tide is less than the options with the lows flow as there is more dilution of the saline water. As the flow subsides but the tide levels are still high, the salinity rises as the dilution effect from the freshwater is less.  When the pumps are on, the peak salinity levels from the tidal saline intrusion are reduced as the water from the pumps dilutes the impact of the tidal salinity. Switching the pumps off has the effect of reducing dilution and increases peak salinity concentrations

12000 4

3.5 10000 3

8000 2.5

2 l / 6000 D g 1.5 O m A

y m t

i l n e i l 1 v a e S 4000 L

0.5

2000 0

-0.5 0 0 10 20 30 40 50 60 70 80 -1

-2000 -1.5 Time (hrs)

200yr tide 100yr flow pumps on 200yr tide 100yr flow pumps off 200year tide low flow pumps full 200year tide low flow pumps off Tide levels Figure 8 Simulated salinity levels at Hickling Broad for the tide, flow and pump scenarios (from White et al., 2008)

Overall it is apparent that under normal tidal conditions, the input of saline drainage water by the pumps leads to an increase in salinity in Hickling Broad, which is consistent with previous studies into the origin of the salinity of the Broad. However, the modelling has demonstrated that under more extreme tidal situation in which saline water moves up the Thurne, the pumps act so as to dilute salinities. The drainage pumps therefore have a role in the salinity ‘story’ of the Upper Thurne which varies with the tides which must be taken into account in their future management.

Biocides and heavy metals

The use of tri-butyl tin (TBT) as an active ingredient in boat antifouling paints was banned in 1987 owing to widespread evidence of negative ecological effects, but has been replaced by copper based compounds (which act as a molluscicide). As copper is not as effective in deterring algal growth herbicidal “booster biocides” have been added to improve performance. A number of workers have looked at the potential threat to charophytes posed by boat antifouling paints and associated compounds.

Sayer et al. (2006) investigated TBT concentrations within a core extracted from Hickling Broad and relationships with biostratigraphical data. TBT was present in the core down to 12 cm depth (dated 1971 ± 4 years), which is coincident with the shift from a plant-dominated to a phytoplankton-dominated ecosystem. Stratigraphic changes in the diatom, charophyte and

13 zooplankton data and sediment lithology coincided with the TBT contamination. As a result, Sayer et al. (2006) postulate that TBT caused a chronic reduction of periphyton- and phytoplankton-grazing invertebrates, particularly molluscs and zooplankton, breaking down a strong feedback loop reinforcing plant dominance, thereby precipitating (with other environmental stressors) a regime shift and the substantial loss of aquatic plants.

Smith (2003) looked at the effect of the newer biocides directly in field and laboratory studies, and indirectly through investigating the presence of copper in sediment within Hickling Broad. Smith (2003) found that total copper content of the top 5 cm of sediment in Hickling Broad exponentially increased towards the boat channel, reaching around 20 mg/kg and a porewater concentration of 100 µg/l nearest to the channel. Evans (2004), cited by Lambert (2007) recorded maximum copper concentrations of filtered water drained from sediment samples taken near the boat channel of 27 mg/l, sufficiently high to affect molluscs (LaBreche et al., 2002). The chlorophyll fluorescence ratio (Fv/Fm), which is used as a measure of photosynthetic function and therefore stress, had a negative relationship with the total copper content of the sediment and was also depressed by Irgarol 1051 and Diuron in field and laboratory tests. The temperature effects of Irgarol 1051 on C. vulgaris in laboratory tests were significant as freshly painted boats would have been entering the broads around the time of greatest apparent temperature sensitivity to the active ingredient;

Lambert et al. (2006) studied the effects of Irgarol 1051 and Diuron on UK freshwater macrophytes. Within a survey of rivers and broads of in May, June and September 2001, Irgarol 1051 was detected in the Hickling area at concentrations from below detectable limits to 2430 ng/l; GS26575 (the principle metabolite of Irgarol) from below detectable limits to 36 ng/l, and Diurone from below detectable limits to 86 ng/l. The concentrations detected by Lambert in the Hickling area, with the exception of Irgarol (for which the maximum observed concentration was double the previous highest recorded level in UK freshwaters) were generally in the middle range of the data range. Laboratory-based toxicity tests on Apium. nodiflorum, Myriophyllum spicatum and Chara. vulgaris showed that Measured Environmental Concentrations were consistently significantly greater than the calculated No Observed Effect Concentration and that Chara vulgaris was the most sensitive of the species tested.

Barker et al. (2007) dismiss the effects of boat antifouling paints on the grounds of the “innocuous use of the same presumably TBT-contaminated sediments in the mesocosms”. However, while the previous studies have demonstrated the high concentrations of copper and TBT-contaminated sediment at shallow depth, the methodology used by Barker et al. (2007) to collect and prepare the sediment had the potential to significantly affect the sediment chemistry, mixing sediment of different depth/ages, changing the redox condition of the sediment (by introducing oxygen) and removing contaminants that might have been mobilised from the sediment or porewater.

Lambert (2007) surveyed 70 sample points throughout Hickling Broad in August 2005 investigating the relationship between macrophyte distribution and abundance and the chemical properties of the interstitial water from the top 2 cm of sediment and the shallow and slightly deeper sediment (5 cm depth). A key finding was that copper was ubiquitous throughout the samples of interstitial water. Only one site contained an MEC of less than 50 µgl-1, 50 sites contained an MEC of between 50 µgl-1 and 90 µgl-1, 13 sites contained an MEC of between 90 and 200 µgl-1, four sites contained an MEC of between 200 µgl-1 and 300 µgl-1, and one site contained an MEC in excess of 300 µgl-1. The two highest concentrations of copper were ascribed by Lambert (2007) to locations in the proximity of the two inflow points to the Broad at the Hickling drainage mill (>300 µgl-1) and the Catfield dyke (>200 µgl-1), although there were three samples near Catfield dyke and one sample within it which had much lower MEC of between 50 µgl-1 and 90 µgl-1, while other samples from around the Hickling drainage mill had MEC of less than 150 µgl-1. Two further high concentrations were found in the vicinity of the boat house (>150 µgl-1), in addition to a larger number of lower concentrations. Lambert suggests that the highest copper values located close to Catfield Dyke, Hickling Mill and the boat house might be caused by either bioaccumulation and deposition by algae from the water column or anthropogenic sources, and states that further

14 research of a similar fashion is required of other broads for comparison before firm conclusions may be made.

Further observations from the survey data and a time series (1996-2006) analysis of Environment Agency water quality data were

1. A statistical analysis (PCA of element associations) indicated that copper, chromium, cadmium and zinc were distributed in similar patterns and concentrations within the Broad, suggestive of similar sources. 2. The canonical RDA analysis by Lambert indicated that while water hardness was the strongest correlate with charophyte abundance (both total and C. intermedia), nitrite, copper and sulphate were lesser positive correlates. 3. There is a highly significant (p<0.01) positive relationship between mean copper concentration in Hickling Broad (based upon Environment Agency sampling) and charophyte canopy and percentage coverage between 1999 and 2006.

Lambert followed this survey work with extensive replicated laboratory experiments demonstrating that in an artificial Hickling growth media, the mean concentration of filterable copper of approximately 100μgl-1 as recorded in the interstitial waters of Hickling Broad in August 2006 caused rhizoid loss and inhibited growth of photosynthetic shoots of Chara intermedia plants grown from apical cuttings.

TBT is banned and Irgarol 1051 and Diuron are no longer licensed for use on boats under 25 m in inland waters. The historically high recorded concentrations and the long term persistence of these compounds and/or their metabolites in the environment provides some cause for concern. Whilst their use may have contributed to the original regime shift of the Upper Thurne system, it seems unlikely that their continued presence is a serious factor in the failure to achieve clear water, given the recovery of the chara lawns in the later 1990’s and early 2000’s. However, the presence of these contaminants (including copper) in the sediment, porewater and water column (Figure 9) may make the chara less resilient to other stressors.

14 ) l /

g 12

μ Cu Filtered (

s Copper - Cu n 10 o i t a

r 8 t n e

c 6 n o c

r 4 e p

p 2 o C 0 7 8 9 0 1 2 3 4 5 6 7 9 9 9 0 0 0 0 0 0 0 0 ------y y y y y y y y y y y a a a a a a a a a a a M M M M M M M M M M M ------0 0 0 9 9 9 9 8 8 8 8 1 1 1 0 0 0 0 0 0 0 0

Figure 9 Monthly open water copper concentrations measured within Hickling Broad from May 1997- February 2008 (source: Environment Agency)

Point and diffuse sources of nutrients

Ecological effects of nutrients Analysis of the Lambert (2007) dataset of 124 water bodies supporting 18 charophyte species showed that there was a statistically significant difference (p<0.01) between the mean nitrate ion concentrations at sites where charophytes were present and absent, with the upper

15 confidence limit for the presence of charophytes being < 3 mgl-1 nitrate although outlying data points showed chara presence at individual sites at concentrations up to 19 mgl-1.

All chara species growing at concentrations of inorganic phosphorus of > 80 µg/l were in either Hickling Broad or Horsey Mere, whilst those at >150µg l-1 were in Hickling Broad and died within in 12 months following recording, making interpretation of this data uncertain! There was a statistically significant difference (p<0.01) between the mean phosphate MECs of sites where charophytes were present and absent, with the upper confidence limit for the presence of charophytes being < 60 µg l-1 phosphate. There is a wide variation within the MECs of TP of many of the species recorded, with several being found at extremely high MECs occurring at a variety of sites. However, subsequent laboratory studies on Chara connivens, C. virgata and Nitella opaca in an artificial Hickling Broad growth media showed that growth of all three species was significantly reduced above 100 µg/l filterable inorganic phosphate.

Point sources of P Holman and Deeks (2007) updated the Soil and Water Assessment Tool (or SWAT) (Gassman et al., 2007) model of Whitehead (2006) to assess the contribution of point source discharges to TP concentrations in the Upper Thurne. Results suggest that:  Point source discharges have a very small (<5 μg l-1 Pl) effect on annual average TP concentration in Horsey Mere, Heigham Sound and Hickling Broad;  Point source discharges may have a significant (10-30 μg l-1 P) effect on annual average TP concentration in , although this may partly reflect an under-prediction of summer discharge from the Somerton pumps;  Reduced flows from the Somerton South pump, to represent the effects of groundwater abstraction, lead to a small simulated increase (<5 μg l-1 P) in annual average TP concentration in Martham Broad;  Diffuse source control of P losses from agriculture may be more effective in reducing surface water P concentrations.

Sediment sources Understanding sediment dynamics is important, due to the role of sediment as a transport media for nutrients (principally phosphorus) and for changing channel dimensions, with consequent effects on navigation and floods. The most comprehensive investigation of sediment sources and transport processes is given in the desk-based study of White et al. (2005), although the availability of data on sediment sources and sediment supply processes are patchy.

The headwater catchment upstream of the Brograve Mill (Management Unit T1) was estimated to contribute 6-10 t year-1 to the Thurne system, which is equivalent to an erosion rate across the headwater catchment of 0.11 – 0.19 mm/year. This estimate was based on modelled erosion rates for a 1 in 10 years return period event (McHugh et al., 2002) and represents the lowest rate in the Broadland headwater catchments. The Internal lower Thurne catchment unit from Brograve Mill to Thurne Mouth (T2) was estimated by White et al. (2005) to contribute between 29-39 t year-1, equivalent to an erosion rate across the Management Unit of 0.24 – 0.33 mm/year. The sediment supply associated with ditch management was not quantified (although this is the subject of a current MSc project at Cranfield University)

Sediment inputs from bank erosion in T2 were estimated as a function of water level range, flow velocity, boat pressure, channel curvature and bank protection at 40.33 t/year, although this may be an under-estimate, as it does not include erosion of the whole bank profile by slumping or undercutting. The only significant industrial sediment source is the Ludham Sewage Treatment Works, which provides an estimated annual sediment load of 1.4 t/year.

16 Plants and algae provide a substantial load of organic detritus to the broads system. Using data from Moss (1977), the phytoplankton numbers in Unit T2 were estimate at around 53000 per ml in 1973, which has been classified as Excessive. White et al. (2005) found it impossible to produce a sediment budget or to relate sediment inputs to sediment accumulation, but:  Hickling Broad was reported to have reduced in depth between the 1930s and 1990s by approximately 25-30 cm (Defra, 2001)  Rose and Appleby (2005) report sedimentation rates in Hickling Broad of 0.014g/cm2/yr in 1922 and 0.055 g/cm2/yr in 2002  Sayer et al. (2006) reported that sedimentation rates since 1950, within a core taken in Hickling Broad, have fluctuated between 0.031 and 0.055 g/cm2/yr, with a mean value of 0.041 ± 0.009 g/cm2/yr, equivalent to 0.31 cm/yr

Diffuse sources of sediment, P and N ADAS have applied both the Environment Agency’s spatial toolkit modeling and the PSYCHIC model to the much of Thurne catchment (as part of the Catchment Sensitive Farming initiative for the Bure, Ant and Muckfleet catchments) to assess erosion and P delivery (Figure 10). Neither model takes account of the presence of the pumped drainage system. The outputs from the two models of sediment risk are quite different, with the highest risk areas from the spatial toolkit modeling being usually in the lower risk categories of the PSYCHIC model. The PSYCHIC model output shows high risk of surface sediment and P along the Waxham Cut, Hickling Marshes. The Environment Agency’s spatial toolkit modeling shows high sediment risk on the Brograve Level, Eastfield marshes and along the River Thurne. So while Figure10 provides useful information on potential source areas of sediment of P, there are significant limitations in how that can be translated into a delivery into the surface water bodies. The Environment Agency’s spatial toolkit modeling for nitrogen shows that most areas of the Thurne catchment have a uniformly high risk.

Figure 10 Extract of (left) EA spatial toolkit modelling and (right) Psychic modelling for the Thurne showing surface sediment risk [darker hue = greater risk] from the CSF in the Bure and Muckfleet

A combination of simple modelling techniques, field work and laboratory analysis is being used in an ongoing research project by Faye Horne at UEA to understand the dynamics of nitrogen (N) in the Upper Thurne catchment in terms of sources, fluxes and seasonal - - variations. All types of dissolved N are being measured including nitrate (NO3 ), nitrite (NO2 ), + ammonium (NH4 ), dissolved organic nitrogen (DON) and total nitrogen (TN). Nitrous oxide (N2O), a potent greenhouse gas, is also being studied to determine whether there are any significant sources from the waterbodies or field surfaces within the catchment.

17 A simple GIS model has been used by Faye Horne to calculate the amount of N potentially available for run-off from different types of land use, according to cropping, fertiliser application rates, stock numbers and export coefficients from Johnes (1996). Cereal production is the biggest source of N from land use, which potentially contributes 81904 kg N year-1, followed by cattle and then sugar beet which contribute 49106 and 33394 kg N year-1 respectively. This technique can be used to identify the areas most at risk from N enrichment.

- The field work has shown that in winter NO3 - is the predominant form of N within the river and broads due to higher run-off during this season, although maximum concentrations found - -1 - -1 were less than 30 mg NO3 l which is below the 50 mg NO3 l limit set by the EC Nitrates Directive. In contrast in summer, DON, such as urea and amino acids, predominates as this is when decomposers of organic matter are most active, although proportionally concentrations of DON remain fairly constant throughout the year, as this is largely a biologically unavailable pool of N. All of the sample sites chosen throughout the catchment were found to be sources of N2O; concentrations were lowest in the broads and highest in the - + pump water. N2O concentrations correlated well with both NO3 and NH4 concentrations. This suggests that both denitrification and nitrification are sources of N2O in the Upper Thurne, so that if N loadings to the water were to increase so too would the N2O flux. The concentration data collected for each type of N will be used to calculate flux values for the whole catchment, allowing a mass balance to be created.

Effects of climate change on water quality Whitehead (2006) used the SWAT model to help understand current and future nutrient dynamics within the Thurne, Bure and Ant catchments. SWAT was used to assess the affects of climate change (using the UKCIP02 scenarios) with and without future socio- economic (landuse) change (Holman et al., 2005). The simulation results showed that both changes in climate and land use affect future diffuse source pollutant losses and TP concentrations in Hickling Broad (Table 3). However, the distribution of nutrient and sediment losses within the catchment is not spatially uniform, demonstrating the effects of soil type, land use and crop rotation. Whitehead (2006) simulated the effects of a limited range of soil conservation practices (cover crops, no till, no till with cover crops and conversion of arable to pasture) on average TP and nitrate-N concentrations in Hickling Broad (Table 3). These demonstrated the potential efficacy of soil conservation practices in reducing sediment losses (and thus P losses) and N leaching in the catchment.

Table 3: SWAT results for management practices in the Thurne model (from Whitehead, 2006) [GS – Global Sustainability; RE – Regional Enterprise] Average TP (mg l-1) Hickling Broad Current Scenario Management Cover Cover & No till No Till Pasture Baseline 0.1 0.03 0.03 0.03 0.02 2050s Low 0.11 0.05 0.05 0.05 0.04 2050s High 0.12 0.06 0.06 0.06 0.05 2050s Low GS 0.11 0.04 0.04 0.04 0.03 2050s High RE 0.13 0.07 0.07 0.07 0.06 -1 Average NO3-N (mg l ) Hickling Broad Current Scenario Management Cover Cover & No till No Till Pasture Baseline 0.85 0.8 0.49 0.54 0.54 2050s Low 0.98 0.51 0.76 0.77 0.49 2050s High 1.01 0.96 0.59 0.81 0.59 2050s Low GS 0.83 0.64 0.64 0.67 0.43 2050s High RE 1.32 0.96 0.79 0.83 0.6

18 Sea level rise and coastal protection The Thurne catchment lies within the Kelling to Lowestoft Shoreline Management Plan (sub-cell 3b). A Shoreline Management Plan (SMP) is a non-statutory, policy document for coastal defence management which provides a large-scale assessment for addressing the risks to people and environment (historic, natural and developed) in a sustainable way. A consultation period was held during 2005, but was not fully adopted by April 2007. This coastline has a rich diversity of features including soft cliffs at the north of the catchment and low-lying plains fronted by dunes and beaches. The structures currently providing protection for the dunes, sand banks and beaches against erosion, are 14km of concrete sea wall, as well as timber and rock groynes, and a series of rock reefs at . In addition the beach is recharged periodically. The issue of coastal protection is already emotive in the catchment, given the significant erosion near Happisburgh following the damage to the coastal defenses. Sea level rise due to climate change and isostatic rebound of the land from the last ice age poses a particular risk to the Thurne catchment, given the soft coastline and the low-lying land. Although there are no published studies regarding this area, there has been significant research activity in sub-cell 3b as it is the test case for the Tyndall Centre’s Coastal Simulator. Given the current uncertainty over future coastal defence in the catchment, this issue represents a concern of over-riding importance to many of the local inhabitants.

Monitoring Shallow lakes and wetlands are, by their very nature, complex environments. This can often result in conventional field-based approaches, which are based on monitoring conditions at a location at a point in time, proving ineffective when attempts are made to monitor the ecological status of these habitats. The collection of data by sensors mounted on aircraft and satellites, in combination with conventional approaches, offers an alternative means of monitoring important wetland habitats such as the Norfolk Broads. The research carried out by Hunter (2007) investigated the potential contribution that data from remote sensing instruments (mounted on aircraft during the project) may make to monitoring programs in shallow lake and wetland environments such as the Norfolk Broads. In particular for the assessment of (i) phytoplankton abundance and species composition and (ii) aquatic vegetation distribution and ecophysiological status in shallow lakes. Using high resolution in-situ and airborne remote sensing data, Hunter demonstrates that semi-empirical algorithms could be formulated and used to provide accurate and robust estimations of the concentration of chlorophyll-a. It was further shown that it was possible to differentiate and quantify the abundance of potentially toxic blue-green algae (cyanobacteria) using the biomarker pigment C-phycocyanin, such that diurnal and seasonal regional-scale time-series of phytoplankton dynamics in the Norfolk Broads could be constructed. It was further shown that remote sensing can be used to map the distribution of aquatic plants in shallow lakes, both around the margins of the broads and, perhaps more importantly, those submerged beneath the waters surface that are of high conservation interest. Hunter also shows that remote sensing metrics could be constructed for the quantification of plant vigor, in particular related to the ecophysiological response of Common Reed (Phragmites australis) to lake nutrient enrichment.

Population biology of charophytes Smith (2003) mostly used highly replicated laboratory-based studies to investigate the population biology of charophytes. In the context of the Upper Thurne his important findings mostly relate to restoration and climate change impacts. In particular, Smith (2003) found that: 1. C. intermedia propagation in experimental laboratory conditions could only be achieved by vegetative growth and not from germination of oospores (i.e. seeds). The vegetative source of early shoots indicates that over wintered vegetative material could have a competitive advantage over growth from germinating oospores. There was a higher density of shoots in samples where the source of growth was from

19 oospores than that in samples where the source of growth was from nodes. However, germination did not occur in oospores from those samples until the vegetative matter was removed. 2. Water temperature is critical for Chara growth and oospore production, thus climate change may influence future growth patterns. Temperatures of c. 10oC, which are generally found in Hickling Broad between March-April, result in little growth; whilst 20oC, which is characteristic of July and August, results in considerably higher growth rates. 3. Analysis of grouped data from a number of broads (including Upper Thurne broads) showed that the percentage cover of charophytes was related to degree days > 15 oC and 20 oC in the previous year, possibly due to the development of either the oospore bank, or a higher density of vegetative individuals. Analysis of Hickling Broad data after 1993 showed that day-degrees > 15 ºC prior to the end of June is negatively associated with charophyte % cover, which appears to conflict with the experimental results 4. Plant cutting experiments showed that cutting encouraged branching and increased overall production, which has obvious implications for future chara management in Hickling Broad. Mortality was low- no plants of C. intermedia or C. hispida were lost and only 2/95 plants of C. aspera died. Cutting of single stems showed that C. intermedia established 25% more branches after the first cut, while C. hispida formed more branches after the second cut. There was some evidence suggesting that C. intermedia may suffer when cut early in the season, as early cutting made a significant negative difference to dry mass, but a second cut appeared to increase dry mass formation. However these experiments must be interpreted with caution as cutting was done cleanly without the uprooting force typical of cutting by machinery in field conditions. 5. Establishment experiments showed that charophyte stem sections with as few as two nodes (which are dispersed naturally through physical disturbance by herbivores and mechanical cutting) are equally capable of growing into independent plants as longer stem sections. Stem sections were shown to be capable of re-growth whether they are buried, planted, or merely resting on the substrate surface, suggesting that they would be tolerant of planting in, or being covered by a range of sediment depths if used for re-introduction during restoration of shallow lakes. 6. Chara vulgaris would be the best species to re-introduce into a lake with soft sediments, as it produced shoots that were longer and had greater plan-form area than C. intermedia or C. aspera within establishment experiments. 7. Based on a survey of 29 chara sites in Norfolk (although none were in the Thurne system) Smith (2003) suggests that sediments with higher shear strength are more likely to be colonised by charophytes, rather than implying that charophytes increase the shear strength of sediments that are unstable. Smith also considers that the experimental evidence shows that charophytes are not physiologically better adapted to develop in very firm sediments than in soft ones, which suggests that there may be an environmental factor, not present in his controlled experiments that precludes colonization of softer sediments. This may relate to oospore germination as, in sediments with a low shear strength of 0.0098 kPa (typical of eutrophic broads with no re-growth of charophytes), the majority of oospores sank below 3 cm (Smith 2003) with important implications for germination and emergence.

Based on his water body survey data, Lambert used 95% confidence intervals to summarise predicted ranges for key variables for which there were found to be a significant difference in range between where charophytes were found and where they were absent. The data were be grouped into:  Water quality variables which are required at minimum concentrations – calcium, chloride, magnesium, sodium, sulphate and water redox;  Water quality variables which are predicted to limit the probability of charophytes being present above predicted concentrations- cobalt, copper, manganese, nitrate, inorganic phosphate and silica;  Physical factors which are predicted to limit the probability of charophyte existence – sediment shear strength of the top 2 cm, canopy % cover and light extinction coefficient.

20 The Hickling data were frequently outliers in the distribution of sites in Lambert’s study. In addition the charophytes died within in 12 months following data collection, hence the Hickling data was not incorporated into the calculation of general tolerance limits for charophytes. However, the fact that Hickling when included was frequently an outlier in predicting for charophyte presence, combined with the complete vegetative loss of the charophyte community in Hickling Broad in 2006, tends to support the conclusion by Lambert that under current water and sediment chemistry conditions, we should not expect charophytes to successfully re-germinate from the oospore bank.

Lambert suggests that the environmental variables which emerge from his data as significant predictors for charophyte presence or absence may represent unexplored factors affecting charophyte communities, directly as toxic agents (in the case of cobalt, copper and manganese) or as limiting macro- and micro-nutrients which remedial measures such as nutrient reduction, lake bio-manipulation of fish communities, or sediment removal have not addressed.

Recommendations for management actions to address the significant issues previously identified and thereby achieve Favourable Conservation Status

It is apparent from the previous sections that there are a number of issues that need to be addressed in order to achieve Favourable Conservation Status. Given the many diverse interests in this catchment, the following recommendations are based on generic principles and activities, rather than ‘on the ground measures’, but provide the framework for future stakeholder engagement activities. However, where possible, they should be linked to existing schemes (including the current agri-environmental schemes) to maximise credibility and increase the potential for future compliance: Examples of relevant schemes might include:  Proposed capital grant scheme for farm-scale mitigation under the Catchment Sensitive Farming scheme  DEFRA codes of Good Agricultural Practice;  Defra / ADAS Fertilizer recommendations;  DEFRA / HGCA guide to Arable cropping and the environment;  NSRI Guide to Good Soil Structure  Environment Agency guide to waterway bank protection  DEFRA soil strategy

Reduction of salinity and ochre

Given the evidence that increased salinity has had a detrimental impact on the ecology of the Upper Thurne broads, and is hindering the stable recovery of the clear water conditions, there is a need to reduce chloride concentrations. However, Barker et al. suggest that a quite modest reduction in salinity, to around 1600–1800 mg Cl L-1 may have a substantial effect on TP and chlorophyll and hence on the potential for plant growth, as the sub-present salinity mesocosm treatment were not greatly different from the lower salinities. Significant ecological benefits may be obtainable without reaching the Favourable Condition chloride concentration criteria of 600 mg/l for Hickling Broad.

As the largest pump in the catchment, the Brograve pump has been a focus of attention. One of the Nature Conservation Objectives of the Water Level Management Plan (WLMP) produced by the Kings Lynn Consortium of Drainage Boards (now the Water Management Alliance) in March 2001 to cover the Brograve drainage district was to ensure that discharges from the Brograve pump do not compromise the water quality of the receiving waters. A solution to the salinity and ochre issues in the Brograve catchment will not be easily gained through consensus, given the range of potential impacts on individuals and businesses. However, as a starting point and building upon the work of ELP and Cranfield University (2006), it is recommended that such a solution is designed on the following principles:

21 1. No significant change in current flood risk (though recognising that flood risk is likely to increase regardless as a result of climate change) – which might be achieved by re-profiling drains, providing additional flood storage and better water level management control than the current water level trigger at the Brograve Pump; 2. A range of farm systems should be able to continue within the marshes of the catchment – this may be based on the splitting of the drainage system or the installation of private pumps (as are present in other parts of the catchment) to allow localised lower drainage levels; 3. Opportunities should be afforded for farm businesses to voluntarily implement measures consistent with existing agri-environment schemes (Entry Level and Higher Level, as appropriate) or to enter into habitat creation partnerships with conservation NGOs or charities; 4. In the event of a coastal breach that effects the rivers and broads, as well as the marshes, (as in 1937 or 1953), there is a means of removing the seawater from the catchment without discharging it through the SAC with the minimum delay. This might be achieved by the installation of an auxiliary coastal pumping station (as suggested in ELP and Cranfield University, 2006, for near Hempstead) or the Broads IDB having immediate access to a mobile pump and generator. 5. That there is recognition that the Brograve sub-catchment is a ‘naturally’ brackish system, and that management changes should not aim to achieve freshwater conditions. Pallis (1911) showed that dyke salinities were brackish or saline (10,000-16,000 μS/cm) when the Brograve Level would mostly have been under high water level management. 6. Similarly, that ochre production in the Brograve sub-catchment is unlikely to be stopped. It should be recognised that limited production of ochre may be beneficial to surface water quality, given its propensity for removing P from solution.

Reduction of diffuse source pollution (N, P, sediment and crop protection products) The study by Holman and Deeks (2007) suggests that point sources (e.g. consented small sewage treatment works) contribute little to nutrient levels in the Upper Thurne broads. Activities should therefore be focussed around reducing diffuse source losses from agriculture. Assuming that the farming community are following CoGAP and Good Agricultural and Environment Condition (GAEC) requirements, it is recommended that:  The nutrient content of manures, where used, are recognised within nutrient budgets to avoid application;  The Catchment Sensitive Farming iniative is rolled out to the Thurne and that P testing is carried out on soils and manures to target applications and to assess whether fertiliser applications can be reduced;  Care is taken to avoid excessive mud on roads during the autumn/winter sugar beet harvesting season, particularly where road runoff is likely to enter drains and watercourses and awareness is raised of the Environment Agency ‘Mud on Road’ campaign;  Voluntary Initiative measures are followed to minimise losses of agrochemical products (pesticide and herbicides) to the environment. This is particularly important with regard to application in under-drained fields and those with wet ditch margins, and to the disposal of sprayer washings, where incentives for installing Biobeds should be sought;  Increased use of extensive buffer strips, headlands and reversion to semi-natural vegetation next to watercourses and minimum tillage on soils which are susceptible to erosion. References Barker, T., Hatton, K., O’Connor, M., Connor L., Bagnell, L and Moss B (2008). Control of ecosystem state in a shallow, brackish lake: implications for the conservation of stonewort communities. Aquatic conservation: marine and freshwater ecosystems, 18(3), 221-240. Broads Authority (1999) Assessment of the proposed cutting of aquatic plants in Hickling Broad : final report by the Assessment team. Broads Authority, Norwich. Broads Authority (2006) Upper Thurne water space. Broads Authority, Norwich.

22 Defra (2001). The future of the Broads National Park. Defra review, 19 Nov 2001. www.broads-society.org.uk/defrareview.html Driscoll, R.J. (1984). Chloride ion concentrations in dyke water in the Thurne catchment area in 1974 and 1983. Unpublished Report for NCC, Norwich, 62pp. ELP and Cranfield University (2005). Feasibility study for solutions to the salinity and ochre issues in the Brograve catchment. Unpubl. Rep for the Broads Internal Drainage Board Evans, L.C. (2004). Factors affecting the distribution and abundance of the submerged macrophyte community in Hickling Broad, Norfolk: is boating activity having an adverse impact. Unpubl. MSc thesis, University College London Gassman P.W., Reyes M.R., Green C.H.. Arnold J.G. (2007). The Soil and Water Assessment Tool: Historical Development, Applications, and Future Research Directions. Transactions of the ASABE, 50(4): 1211-1250. (http://www.card.iastate.edu/environment/items/asabe_swat.pdf) Gurney, R (1904). The fresh and brackish water crustacean of east Norfolk. Trans. Nor. Nor. Nat. Soc., 7, 637-660 Hails, A. J. (ed) (1996). Wetlands, Biodiversity and the Ramsar Convention: the role of the Convention on Wetlands in the Conservation and Wise Use of Biodiversity. Ramsar Convention Bureau, Gland, Switzerland (http://www.ramsar.org/lib/lib_bio_1.htm ) Harding, M., and Smith, K. (2002) Ochre In The Brograve Catchment: Causes and Cures. Happisburgh-Winterton Internal Drainage Board Harris, J. (2000). Hickling Broad Aquatic Plant Cutting and Monitoring Programme- Summary report of results from 1994 to 1999. Unpublished report for the Broads Authority Harris, J. (2004). Hickling Broad Aquatic Plant Cutting and Monitoring Programme- Summary report of results from May 2004 to September 2004. Unpublished report for the Broads Authority Harris, J. (2007). Hickling Broad Aquatic Plant Cutting and Monitoring Programme- Summary report of results from 2000 to 2006. Unpublished report for the Broads Authority Holman, I.P. (1994). Controls on Saline Intrusion in the Crag aquifer of north-east Norfolk. Unpubl. Thesis, University of East Anglia, Norwich. Holman, I.P. and Deeks, L.K. (2007). Upper Thurne modelling for Stage 4of the Review of Consents for the Habitats Directive. Unpublished report for the Environment Agency Holman, I.P. & Hiscock, K.M. (1998). Land drainage and saline intrusion in the coastal marshes of north east Norfolk. Quarterly Journal of Engineering Geology, 31, 47-62. Holman I.P., Rounsevell M.D.A., Shackley S., Harrison P.A., Nicholls R.J., Berry P.M. and Audsley E. (2005). A regional, multi-sectoral and integrated assessment of the impacts of climate and socio-economic change in the UK: I Methodology. Climatic Change, 71, 9-41. Hunter, P.D. (2007). Remote Sensing in Shallow Lake Ecology. Unpublished Ph.D. Thesis, University of Stirling Innes, A.G. (1911). Tidal actions in the Bure and its tributaries. Trans. Norf. Nor. Nat. Soc., 9(2), 244-262. Johnes, P. (1996) Nutrient export modelling - River Bure, Norfolk. Aquatic Environments Research Centre, University of Reading. LaBreche, TMC, Dietrich AM, Gallagher DL and Shepherd N (2002). Copper toxicity to larval Mercenaria mercenaria (hard clam). Environmental Toxicology and Chemistry 21, 760- 766 Lambert, S.J., Thomas K.V and Davy A.J (2006). Assessment of the risk posed by the antifouling booster biocides Irgarol 1051 and Diuron to freshwater macrophytes. Chemosphere 63(5), 734-743 Lambert, S.J. (2007). The environmental range and tolerance limits of British stoneworts (Charophytes). Unpubl. PhD thesis, University of East Anglia, Norwich, UK McHugh M, Wood G, Walling D, Morgan R, Zhang Y, Anthony S and Hutchins M (2002). Prediction of sediment delivery to watercourses from land. Phase II R&D Technical Report No. P2-209, Environment Agency, Bristol. Moss, B. (1977) Conservation problems in the Norfolk Broads and rivers of East Anglia – phytoplankton, boats and the causes of turbidity. Biol. Cons. 12, 95-114. Moss, B (2001). The Broads; the people’s wetland. HarperCollins Publishers, London. Pallis, M. (1911). Salinity in the Norfolk Broads. I. On the cause of the salinity of the Broads of the River Thurne. Geograph. J., 37(3), 284-291.

23 Rose, N.L. and Appleby PG (2005). Sediment accumulation in the Broads. A report to the Broads Authority. Research Report No. 101. Environment Change Research Centre, University College London Sayer, C.D, Hoare DJ, Simpson GL, Henderson ACG, Liptrot ER, Jackson MJ, Appleby PG, Boyle JF, Jones JI and Waldock MJ (2006). TBT causes regime shift in shallow lakes. Environmental Science and Technology 40, 5269-5275 Servera-Martinez, E (2005). Ecological failure criteria for Hickling Board, Norfolk. Unpubl. MSc thesis, Cranfield University Simpson, TB (2007). Understanding the groundwater system of a heavily drained coastal catchment and the implications for salinity management. Unpubl. PhD thesis. Cranfield University Smith, D.C.S. (2003). The population biology of charophytes in the context of shallow lake restoration. Unpubl. PhD thesis, University of East Anglia, Norwich, UK White, SM, Fisher KR and Holman IP (2008). Modelling the dynamics of the Upper Thurne river system and its influence on Hickling Broad. Unpubl report for the Broads Authority. White, S, Deeks L, Apitz SE, Holden A and Freeman M (2005). Desk based study of the sediment inputs to the Broads catchment, with the identification of key inputs and recommendations for further targeted research and management to minimise inputs. Unpubl. Final report: Phases I and II for the Broads Authority Report, Norwich Whitehead, J. 2006. Integrated catchment scale model of a lowland eutrophic lake and river system: Norfolk, UK. Unpubl. PhD thesis. Cranfield University.

24