Lower River Test NEP

Volume 3: Appendices

Southern Water Services

October 2013

Volume 3 Appendices Notice

This document and its contents have been prepared and is intended solely for Southern Water Services’ information and use in relation to the Lower River Test NEP

Atkins Limited assumes no responsibility to any other party in respect of or arising out of or in connection with this document and/or its contents.

Document history Job number: 5099146 Document ref: 5099146 / 079 / DG / 217 Revision Purpose description Originated Checked Reviewed Authorised Date Rev 2.0 Issued for sign-off Atkins HG, AB PS PS 10 10 13

Client signoff

Client Southern Water Services

Project Lower River Test NEP

Document title Lower River Test NEP Investigation Report Volume 3 Appendices- Final

Job no. 5099146

Atkins

Table of Contents

Appendix 1.2.1 NEP Scope Appendix 1.5.1 Testwood PWS Abstraction Appendix 1.6.3 Water Body Summary Sheets Appendix 2.1.1 River Test SSSI Citation Appendix 2.1.2 Lower Test Valley SSSI Citation Appendix 2.3.1 River Corridor Survey 1991, S169-171 Appendix 3.5.1 Frequency of extreme low flows in the Great Test Appendix 3.5.2 Freshwater Flows Appendix 4.1.1 Hydraulic Model Appendix 4.2.1 Wetland Model Appendix 4.3.1 Thermal Model Appendix 5.2.1 Hydraulic Model Outputs Appendix 6.1 Fish Counter data Appendix 6.2 Fish Count Little Test Appendix 6.3 Fish Passage at Testwood Appendix 6.4.1 Salmon Movement Model Manual (Pisces) Appendix 6.4.2 Additional Notes on the Salmon Simulation Model (Pisces) Appendix 7.2 Aquatic Macrophytes Appendix 7.3 Aquatic Macroinvertebrates

Atkins

Appendix 1.2.1 NEP Scope

Atkins

Scope of the Lower Test NEP Investigation

21 June 2011

DG032 Scope of the Lower Test NEP Investigation

Notice

This document and its contents have been prepared and are intended solely for Southern Water Services‟ information and use in relation to The Lower Test NEP Investigation.

Atkins Limited assumes no responsibility to any other party in respect of or arising out of or in connection with this document and/or its contents.

This document has 15 pages including the covers.

Document History

Job number: 5099146 Document ref: DG032 Revision Purpose Description Originated Checked Reviewed Authorised Date For discussion and Rev 1.0 approval with the project Atkins Atkins Atkins Atkins 18/05/11 Steering Group Project Project Rev 2.0 Final Atkins Atkins Steering Steering 21/06/11 Group Group

1

DG032 Scope of the Lower Test NEP Investigation

Table of contents

1. Introduction 4 1.1. Background to the study 4 1.2. Structure of this document 4 2. Scope of NEP Investigation 4 2.1. Study Area 5 2.2. Programme 5 2.3. Steering Group 5 2.4. Key Risks 5 2.5. Communications 5 3. Technical Approach 6 3.1. General 6 3.2. Hydrological and hydraulic assessment 8 3.3. Fish assessment 9 3.4. Ecological assessment 10 3.5. Effect of management 11 3.6. Water temperature 12

2

DG032 Scope of the Lower Test NEP Investigation

3 DG032 Scope of the Lower Test NEP Investigation

1. Introduction

This document gives a background to the Lower Test National Environment Programme (NEP) Investigation, and related projects. This document outlines the scope of work for the Lower Test NEP Investigation to be undertaken by the Atkins project team on behalf of Southern Water as part of its AMP5 water resources work pack. The document has been discussed with, and approved by, the Steering Group that comprises Southern Water, Atkins, Adams Hendry, the Environment Agency and Natural England. Southern Water‟s supply area is divided into three geographically separate supply Areas which are made up of discrete Water Resource Zones (WRZs). The Western Area comprises the Hampshire South and Isle of Wight WRZs which are interconnected, and the discrete Andover WRZ and Kingsclere WRZ. In addition to the Lower Test NEP, the AMP5 programme for the Western Area also includes the Hampshire and Isle of Wight Water Resources Options Appraisal, the Itchen NEP project, and the Isle of Wight HD monitoring investigation. These projects are interrelated.

1.1. Background to the study Southern Water‟s abstraction licence (no. 11/42/18.16/546) for the Testwood Water Supply Works (WSW) on the River Test authorises abstraction at a peak rate of 136 Ml/d up to an annual total of 49,915,080 m3 (equivalent to 366 days at the peak rate). The current capacity of the Testwood works is 105 Ml/d. The focus of this NEP investigation is to assess the potential impacts of different abstraction scenarios upon the hydrological regime downstream of the Testwood intake, and the potential effects upon particular habitats and species of the River Test SSSI and the Lower Test Valley SSSI.

1.2. Structure of this document The following sections of this document comprise the following:  Section 2: Scope of the NEP investigation; and  Section 3: Technical approach.

2. Scope of NEP Investigation

This Draft Scope has been developed taking into account the report by the Environment Agency (March 2010) Testwood Public Water Supply Abstraction Impact Investigation– Statement of Issues and Assessment, Version 2.0, plus subsequent discussions. This Draft Scope also reflects the current and ongoing review of previous studies and available data that has been undertaken during the initial stage of this Southern Water study to evaluate the issues and concerns that are relevant to the Testwood abstraction. This Scope also reflects the outputs to date from discussions with technical experts at the EA and with key landowners and their tenants. Through consultation undertaken to date, it is understood that there are concerns about potential disturbance arising from access to the River channel or banks for surveys or investigations during the fishing season. This position has not changed our Scope, however we will place more focus on using available data and information to undertake the NEP investigation, whilst continuing to engage with these stakeholders to obtain access.

4 DG032 Scope of the Lower Test NEP Investigation

2.1. Study Area The study area for the NEP Investigation will be the reach and associated tributaries of the River Test from the Testwood abstraction intake, downstream to Redbridge, that are directly hydrologically linked to the Testwood abstraction.

2.2. Programme The proposed programme for the NEP can be seen in Table 1. At this stage we propose to issue a Draft Report on the conclusions of the NEP investigations in February 2012. This timeframe reflects the volume of information and data that are available for the study area which can be used in the NEP investigation, and the wider context of the Western Area investigations. There is a need to agree a programme of meetings and review periods with the Steering Group for the Lower Test NEP Investigation to ensure its progression without undue delay.

2.3. Steering Group This first Steering Group meeting will be the basis for an agreement between Atkins, SWS and the statutory bodies on:  Interpretation of the objectives and deliverables of the Lower Test NEP Investigation;  Agreement on what data and analysis is appropriate for this investigation, with particular regard to the situation regarding access to the river; and  Approval of the outlined approach and programme.

2.4. Key Risks The main risks we have identified for the Lower Test NEP Investigation are summarised below. These will need to be regularly reported to and monitored by the Steering Group:  The extent to which existing datasets can be used for the Investigation;  The number and extent of new surveys or data required;  The need for access to undertake surveys and investigations;  The need for collaborative working partnerships between SWS, the statutory agencies, and key stakeholders; and  The ability to simulate the hydrological and hydraulic regimes of the Lower Test using appropriate hydrological and hydraulic models.

2.5. Communications Given the wide ranging interests in the River Test, the NEP scheme and wider activities, we recognised the importance of ensuring communication was established early with the key statutory, non-statutory and other stakeholders with an interest in the scheme. Draft Communications Plans have already been prepared that detail the strategy for stakeholder engagement throughout the NEP study, and also for the wider Western Area schemes. Key stakeholders are landowners and lease holders in the study area and the Test and Itchen Association. A wider stakeholder group would comprise the Salmon & Trout Association, Hampshire

5 DG032 Scope of the Lower Test NEP Investigation

and Isle of Wight Wildlife Trust, WWF, local angling clubs and the Hampshire Biodiversity Information Centre (HBIC) amongst individual interested parties. The interests of Local Planning Authorities will be covered through the wider communications plan for the IoW and Hampshire Options Appraisal Project. Direct face to face engagement with key stakeholders has already commenced and will be repeated at regular stages throughout the work. Direct engagement enables discussion of the existing and licensed abstraction and operational management of the Lower River Test. The results of ongoing consultation will inform the NEP investigation where relevant. Once the initial consultation with all the key stakeholders has been completed, a programme of wider consultation will commence. Work on an AMP 5 Investigation webpage has been started, which will be held on the website of Southern Water. This will assist the dissemination to stakeholders and other interested parties, and will contain factual information on the progress of the Investigation and will be updated as appropriate.

Table 1: The proposed programme for the Lower Test NEP work

c10

Jul11

Oct 11 Oct

Apr 11 Apr Jan 12

Jun11 Feb 12 Mar12

Sep 11 Dec11

Nov11 Aug Aug 11

Topic May11

De Jan 11 Feb 11 Mar11

Development of

Communication Plan *

Liaison with Stakeholders

Data collation Review of previous work and confirm NEP scope * WR Aquator modelling and scenario testing * * Fish Migration assessment * * Hydrological and hydraulic modelling * * * Identification of hydro- ecological requirements * Ecological data collection (macro-invertebrates) Develop hydroecological assessment tools * * Impact Assessment * Identification of target flow regime * Draft NEP Report * Final NEP Report *

Where * indicates an interim report on findings

3. Technical Approach 3.1. General The overall approach to be taken is to ensure that future decisions on the Lower Test with regard to abstraction are made on the basis of a scientific evidence base. Components of the work include: collating and reviewing relevant reports, data and literature; collecting new data where appropriate and possible; updating existing modelling tools such as the Test and Itchen groundwater model; developing new models where appropriate; and providing an

6 DG032 Scope of the Lower Test NEP Investigation

integrated assessment of the potential impact from the Testwood abstraction upon the river and floodplain ecology. The study area is the Lower Test: the reaches downstream of Testwood abstraction point. In order to model the hydrological regime and ecological conditions above and below the abstraction it may be necessary to extend the geographical area for data collection to a suitable distance upstream. The specific technical objectives are to:  quantify the potential flow effects from a range of abstraction scenarios at Testwood;  use hydrological and hydraulic models to translate flow impacts into velocity and water level impacts downstream of the abstraction;  assess the potential effect of the abstraction regime upon fish migration;  assess the potential hydroecological impacts from the different abstraction scenarios based on the key habitats and species using preferred hydroecological requirements;  assess the magnitude of identified impacts from abstraction for the key habitats and species; and  on the basis of these assessments identify the range of river management and abstraction interventions that together would deliver the required outcomes. Consideration will also be given to channel management and how this may affect the flow regime, fish migration, aquatic species and habitats. The above objectives of the study form discrete but interlinked work packages, as presented in Figure 1 below.

Data collation and collection The River Test has been subject to a large number of data collection and other studies covering a range of aspects on different temporal and spatial scales, with many reports and datasets available for the catchment. The recent Environment Agency Lower Test Project has collated datasets and other information relating to the Lower Test area. The documentation from that project has been reviewed to develop an understanding of baseline conditions and consolidate our existing knowledge of the current hydrological and ecological regime in the Lower Test, both upstream and downstream of the Testwood abstraction. As noted above, access to the River channel or banks may not be possible until the end of the 2011 fishing season. While we continue to engage with these stakeholders we will focus on the use of available data and information to progress the NEP investigation.

7 DG032 Scope of the Lower Test NEP Investigation

Figure 1 Flow charts showing flow path of assessment approach for the Lower Test NEP

3.2. Hydrological and hydraulic assessment

3.2.1. Our approach The purpose of this work package is to simulate different abstraction scenarios at Testwood and quantify the potential changes in flow and regime of the Lower Test study area in terms of quantity, timing, magnitude and frequency. The different scenarios will examine the naturalised, historic and fully licensed abstraction scenarios in addition to other possible scenarios that are more relevant to Western Area as a whole.

3.2.2. Work to date We have started to construct an Aquator model of the Hampshire South WRZ to assess how existing and forecast demands can be met from Southern Water‟s groundwater and surface water sources. Outputs from the Test & Itchen groundwater model (used by the EA to produce flow sequences for its CAMS assessments and Habitats Directive Review of Consents) will be used to provide the surface water flows for input to the Aquator model.

3.2.3. Further work River flows from existing scenario runs of the Test & Itchen Groundwater Model will be input to Aquator to explore the effect of different abstraction scenarios on the downstream flow regime. We understand that the Environment Agency proposes to update the length of the hydrological datasets for the Test & Itchen groundwater model. Discussion with EA suggests that this update would not be completed within the timescale for this NEP outlined in Table 1, and so options for accelerating the update programme will be discussed and agreed with the EA.

8 DG032 Scope of the Lower Test NEP Investigation

In order to determine the effect of the change in flow regime from the different abstraction scenarios upon the fish species and aquatic habitats under consideration, we shall construct a hydraulic model using the best available data on channel form, water control structures and topography. As noted above, the need to use existing datasets stems from potential lack of access to the river. We will start with a 1-D hydraulic model that focuses on a reach of the Great Test extending upstream of the Testwood intake to the Nursling Mill and fish farm, extending downstream to Testwood Mill where a series of structures define the tidal limit. We are currently investigating the availability of suitable data with which to build the hydraulic model. Outputs from this model will provide a means to derive robust conclusions regarding the effect of abstraction scenarios upon flow, water levels and velocity. Hydroecological metrics of relevance to the aquatic species and habitats under consideration can then be linked to the hydraulic model outputs to assess impacts. A number of control structures exist in the Lower River Test which split flows into numerous channels and affect velocity and water levels. The hydraulic model will represent these control structures using available information in order to assess the effects of channel management upon the flow regime and aquatic habitats and species. The model can be updated, extended and fully calibrated should further data or access become available.

3.3. Fish assessment

3.3.1. Our approach The Environment Agency (EA) has provided an initial assessment of the aquatic species and habitats of the Lower Test area in their report (March 2010) Testwood Public Water Supply Abstraction Impact Investigation – Statement of Issues and Assessment, Version 2.0. Having reviewed the above report, and following discussion and review of available data and reports with the Environment Agency, we are focusing upon salmon to examine the effects of the Testwood abstraction in the Lower River Test as shown in Table 2. The focus of the assessment will be to determine how fish migration might be affected by different abstraction scenarios for the Testwood WSW. We shall also examine the influence of channel management in order to assess the wider effects upon fish migration in the Lower Test area to include fish passage through the various structures in the Testwood Mill area.

3.3.2. Work to date To date we have collated some of the salmon count data held by the Environment Agency and reviewed the most relevant literature and recently commissioned studies. Whilst it is still considered that a fully calibrated hydraulic model of the Lower Test will be very helpful in understanding the relative impacts of the abstraction, weather events and operation of structures etc, an initial analysis of the salmon data undertaken so far suggests that, in regard to salmon movement past the Testwood abstraction, some firm conclusions are still possible without the results of the model. Some early examination of salmon count data from the Great Test was undertaken in February followed by a very useful meeting with David Solomon (Independent) and Adrian Fewings (EA Fisheries Specialist) in March to discuss some of the issues to be addressed. Further assessment will be undertaken following receipt of additional data from the EA – we understand that release of these

9 DG032 Scope of the Lower Test NEP Investigation

data may be imminent. Once the available data has been received and an initial analysis undertaken, the scope of additional work will be agreed with David Solomon and he will be contracted to undertake that work alongside Atkins with the objective of issuing an initial report in August 2011.

3.4. Ecological assessment

3.4.1. Our approach As with the fisheries assessment, we have reviewed the habitats and species listed in the 2010 Environment Agency report, and have held discussions with Environment Agency staff. We have also reviewed the conservation designations of the area, the outcomes of the Habitats Directive review outcomes and the hydrology of the Lower Test study area. The outputs of this review have formed the scope for the ecological assessment as detailed in Table 2. Each of the species and habitats listed in Table 2 will be subject to a separate hydroecological assessment. We will initially develop maps showing the abstraction impact pathway based on our hydrological understanding of the site. Where available, we will also map the distribution of different habitats and species to assess which of these might be impacted by the Testwood abstraction. For species and/or habitats deemed „at risk‟ we will then determine their preferred hydroecological requirements (water levels, velocity, flows) and assess how different abstraction scenarios Testwood might influence these requirements and how these regimes might be influenced by other factors such as river management interventions.

3.4.2. Work to date

i) Review of previous assessments We have reviewed the condition of the study area using previous studies and reports. The study area comprises the River Test SSSI, Lower Test Valley SSSI (with parts designated as part of the Solent Maritime SAC) and the Solent and Southampton Water SPA and Ramsar site. The Test Valley SSSI is considered by Natural England to be in favourable condition. The River Test SSSI unit which is relevant to this NEP study is Unit 91, which has been assessed by Natural England as being „unfavourable no change‟ due to the reasons of “Inappropriate water levels, Inappropriate weirs dams and other structures, Invasive freshwater species, Siltation, Water abstraction, Water pollution - agriculture/run off and Water pollution - discharge”. We have reviewed the Habitats Directive assessment for these sites which assessed the effect of the Testwood abstraction upon these areas of international importance. The Testwood licence was one of three licences that were considered in both the SPA and SAC Stage 3 and Stage 4 Appropriate Assessment undertaken by the Environment Agency. The Environment Agency affirmed these licenses for Stage 4 stating that that „there is currently no evidence that the current rate of abstraction is having an adverse effect on the designated site[s]. However there is also no evidence to suggest that if the licence were used to its full capacity that this would not result in an adverse effect on the integrity of the designated site. Due to the pressure within this area of the country for additional growth and therefore additional water to meet this demand the increased demand on this particular licence is highly likely. However the Environment Agency does not have sufficient evidence to equate removal of current headroom to a risk of impact from abstraction at rates within the range of current to fully licensed – it is over-precautionary. This option is considered overly precautionary by the Environment Agency and is not considered appropriate.’

10 DG032 Scope of the Lower Test NEP Investigation

The outcome of the Habitats Directive assessment is that the Environment Agency has determined that the Testwood abstraction does not impact on the Solent Maritime SAC and the Solent and Southampton Water SPA site. Therefore, the scope of this NEP will focus on those areas, habitats and species in the SSSI citation for the River Test that are not covered by the HD assessment. In line with guidance received from the Environment Agency it will also be necessary to consider the impacts of abstraction on the Ramsar features of the Lower Test. Similarly, the NEP investigation will consider those Ramsar designated species and habitats that are not included in the citations of the SAC and SPA.

ii) Hydrological walkover survey We have been able to undertake a hydrological walkover survey of the area leased to the Hampshire and Isle of Wight Wildlife Trust (HIoWWT) to better understand the hydrological connectivity of the channel network in the Lower Test study area.

We have used information from this site visit, alongside habitat distribution maps provided by HIoWWT, Ordnance Survey maps and aerial photographs to determine the nature of water level management in the area and the distribution of the main habitat area of relevance to the NEP investigation.

An initial assessment of the existing layout of watercourses, the location and current management of key water control structures and the distribution of different habitat types suggest that the Lower River Test and its floodplain (including the Lower Test Valley SSSI) can be split into a series of distinct hydrological sub-units. Water levels in each of these sub-units are influenced by different sources of fresh water (e.g. Great Test or Little Test) and water control structures. The relative importance of tidal influences within each sub-unit also varies as a result.

The NEP assessment will focus on those sub-units which are in hydrological connectivity with the abstraction impact flow path.

3.4.3. Further work Using the above data, we will be able to verify the condition status of the key species and habitats in the study area.

The preferred hydroecological conditions of the species and habitats under focus are generally well known to us from a number of similar investigations we have undertaken elsewhere in the south of England. We will use these sets of information to identify key hydrological metrics for the impact assessment.

By considering how these conditions might change in response to differing abstraction scenarios and other river management interventions (see Section 3.5) an assessment of the ecological impact can be made.

3.5. Effect of management

This NEP scheme seeks to understand the potential of the Testwood abstraction to affect the flow regime downstream of the intake, and thus affect key aquatic habitats and species.

However, as evidenced by the Natural England assessment of Unit 91 of the River Test SSSI, there are a number of other factors - aside from the Testwood abstraction - that have the potential to affect flow, velocity and levels downstream and hence could affect the key habitats and species. These include:

11 DG032 Scope of the Lower Test NEP Investigation

 The effects of past activities to widen channels for land drainage;  Riparian and in-river weed cutting activities; and  Sluice management.

In assessing the impacts of abstraction on favourable condition we will also need to consider the extent to which these other factors are influencing flows, velocities and levels.

This will establish the need for, and benefits associated with, an integrated river management approach for the lower reaches of the River Test SSSI.

An integrated approach to river management such as this would require collaboration between Southern Water, the Environment Agency and Natural England, tying in with existing projects such as the EA‟s River Restoration Strategy. This approach would also need to engage local landowners and river users.

3.6. Water temperature

The EA has expressed some concern about the potential impact of abstraction on water temperatures downstream of Testwood. This concern relates primarily to potential impacts on salmon, which are known to be particularly sensitive to riverine temperature regimes.

An initial assessment of the potential for temperature impacts should be completed in June 2011 and this will be used as a basis for scoping out additional monitoring and modelling work, if required.

12 DG032 Scope of the Lower Test NEP Investigation

Table 2 Fisheries and ecological assessment for the Lower Test NEP The first four columns comprise information taken from the Environment Agency‟s report (March 2010) Testwood Public Water Supply Abstraction Impact Investigation– Statement of Issues and Assessment, Version 2.0

Monitoring Stated assessment Species req’ments in Comment from the EA Abstraction Impact Report The Scope of the Lower Test NEP Rationale requirements in the Report Report The report states that a methodology to assess the impacts of different abstraction regimes on salmon migration is needed. Examination of salmon behaviour and the understanding of Effect of abstraction regimes on Flow, water how abstraction regimes may impact on the flow regime – total discharge, water levels, hydrological and temperature and We will look at the potential impact of different abstraction scenarios on The outputs of the salmon assessment will guide the consideration of other fish Salmon temperature, velocity profiles etc - are important. Should estimate the effects of potential delays in impact upon migration of salmon temperature and salmon migration species migration migration on salmon populations. Water temperature profile in the lower River Test reaches populations. should also be monitored and the likely impact of increased abstraction on this plus potential climate change scenarios and effect on salmon. Intakes must be adequately screened Although the flow requirements for sea trout differ slightly that detailed The environmental outcomes that are expected are: a flow regime in the lower River Test that for salmon can used for both maintains or improves passage for migrating sea trout, the effective screening of all Flow, water species. However the impact on abstraction intakes to prevent fish being drawn in and trapped at any stage of their life cycle, As for Salmon, however this is dependent upon available datasets on sea Sea trout temperature, sea trout of any changes in water the maintenance of a water temperature profile in the lower River Test which is not raised as trout migration temperature identified in the a result of increased abstraction and is as resilient as possible to climate change. Intakes salmon work should be assessed must be adequately screened separately The report states that "the River Test below the Testwood public water supply intake is not Brown trout n/a n/a characteristic Brown Trout habitat therefore no specific monitoring is required for this species. n/a Not considered required by the Environment Agency Intakes must be adequately screened to prevent fish losses. " Grayling n/a n/a The report states that only need to assess grayling if salmon work shows an impact n/a Not considered required unless salmon assessment shows an impact Eel n/a n/a Report states no specific monitoring is required for this species n/a Not considered required by the Environment Agency An assessment of the effect of different abstraction scenarios on water levels Water level control structures can have a large impact on water levels. Thus Report states that no specific monitoring needed but should assess impact of abstraction on Bullhead n/a Impact on water levels in the reach downstream of the abstraction to the next control structure will be assess the impact of the abstraction alone upon water levels it is important to note water levels downstream of the intake undertaken the effect of control structures The report states that only need to assess if salmon work shows an impact, but does not Not considered required by the Environment Agency unless salmon assessment Lamprey n/a n/a n/a specify what the assessment should be shows an impact Population We will use existing data on water vole distributions and examine effect of The focus of the NEP will be to examine hydrological changes of the abstraction Water vole distribution and Impact on water levels Report requests a survey of distribution, and changes from water levels. different abstractions scenarios upon water level changes in areas noted for and impact on hydroecological requirements. A distribution survey will therefore water level impact this species not be undertaken We will use existing data on otter distributions and examine effect of different Impact on "identified features of abstractions scenarios upon "features of important" where hydrologically Where the "features of importance" cannot be hydrologically linked to abstraction Otter N/a importance to otters" relevant. We will engage in consultation with EA and NE to determine these then theses will not be examined within the NEP (e.g. predation) "features of importance" Report states that need to: collate data on the number and distribution of breeding waders in the Lower Test Marshes over a five year period using data collected by the Hampshire and We will collate existing data on number and distribution of breeding waders. It is not the scope of an NEP investigation to examine effects of water table or Isle of Wight Wildlife Trust, review this data in the light of predicted sea level rise, changes in We will identify if historic, existing and proposed breeding areas are water level changes in relation to non-abstraction related effects such as predicted Breeding n/a Impact on habitat water tables habitat/vegetation distribution and planned land management regimes in the Lower Test hydrologically linked to the effects of abstractions scenarios, and if so sea level rise, changes in habitat/vegetation distribution and planned land waders Marshes and adjacent Manor Farm meadows, assess any likely adverse effects of abstraction determine the impact on water levels and water tables from these scenarios, management regimes in the Lower Test Marshes and adjacent Manor Farm on breeding wader habitat through reductions in ground water levels and flood plain plus the role of water level control structures meadows. inundation within the target areas for breeding wader habitat restoration. Report states that need to: review data on breeding passerines in the Lower Test Marshes We will collate existing data on number and distribution of passerines. We using CBC/BBS results, map locations and habitats of breeding passerines, assess water will identify if historic, existing and proposed breeding areas are hydrologically Passerines n/a Impact on habitat water tables level requirements for habitats supporting breeding passerines, assess effects of abstraction linked to the effects of abstraction scenarios, and if so determine the impact on habitat quality and ensure measures are in place to maintain sufficient flow to support on water levels and water tables from these scenarios, plus the role of water target water levels in terms of both duration and time of year. level control structures Southern n/a n/a Report states no monitoring is required n/a Not considered required by the Environment Agency damselfly Macro- Aquatic Need 2 x samples per season (spring and autumn) at established sites u/s and d/s of the We will undertake aquatic macroinvertebrate surveys from the sites to invertebrates We will assess the historic macroinvertebrate dataset to examine how the dataset macroinvertebrate Impact on flow intake, identified to species level. EA believe existing/historic dataset not considered robust determine the effect of different abstraction scenarios. We will try to generate - main river can be used with appropriate caution. population enough for use due to perceived methodological issues. a robust relationship between macroinvertebrates and flow. channel The focus of the NEP will be to examine hydrological changes of the abstraction "(1) Survey of wetland invertebrate fauna of the Lower Test Valley SSSI and assess this in We will consult EA and NE for the hydroecological requirements for floodplain and impact on hydroecological requirements. A survey will therefore not be Impact on the floodplain wetlands terms of its habitat requirements. (2) Identify the ways in which freshwater flows influence the Wetland invertebrates, and use these to assess the impact of changes in water level undertaken. We consider Requirement (1) to be general research and not linked to Floodplain including flooding period and flow range and wetland habitats of importance to invertebrates within the flood plain including macroinvertebrate and water tables created by different abstraction scenarios in areas abstraction impact assessment. We will undertake the assessment for (2) where invertebrates regime. Flows to the Middle River assessments of flooding period and flow regime. Flows to the Middle River are included in population hydrologically linked to the Testwood abstraction. We will also assess the areas are hydrologically linked to the abstraction impact pathway. Where the are included in this assessment this assessment and integrated into the requirements for fish and in-channel invertebrate impact of water level control structures on preferred conditions. hydroecological requirements for floodplain invertebrates are not available, we will communities." use habitat requirements related to breeding waders. We will consult EA and NE for the hydroecological requirements for white Need to maintain a flow regime in the lower River Test which: does not cause White-clawed White clawed Impact on white clawed crayfish clawed crayfish, and use these to assess the impact of different abstraction N/a Crayfish habitat to be reduced or degraded, provides sufficient flow at all times to maintain The report does not present any hydroecological requirements or preferences. crayfish habitat scenarios in crayfish areas hydrologically linked to the Testwood abstraction. water quality and levels. The main threats are predation pollution and siltation. We will also assess the impact of control structures on preferred conditions. The report states that expected environmental outcomes are the maintenance of a flow The report comments that, “more research is required into the impact of flow on regime in the lower River Test which: Maintains a habitat as characterised by the long-term Discussion with the Environment Agency to date has concluded that no chalk stream macrophytes” and cites that a PhD project looking into these Freshwater n/a Effect on macrophyte habitat health of the macrophyte community, downstream of the abstraction intake; and habitats assessment on macrophyte habitat is required, and that a macroinvertebrate variables on the Itchen will provide the methodology and analysis technique for the macrophytes required to support the rare and diverse flora present in the lower Test River. To deliver these assessment would be representative. Test. However, since the publication of the report, the study has not progressed as outcomes, more research is required into the impact of flow on chalk stream macrophytes. anticipated, and is not considered promising for the NEP investigation. The report states an assessment needs to cover impact from abstraction on surface water We will undertake a hydrological survey of the Lower Test area in order to The report states that “abstraction is only likely to be a key factor… on water levels flooding, groundwater levels and soil moisture in the floodplain. Understanding of how determine the hydrological network and connectivity of the channel, and the if the impact is significant, to the extent that there is insufficient water to manage, Lowland wet Impact on water levels and water abstraction might affect the availability of flows into the network of floodplain ditches. abstraction impact pathway. The location and influence of water level control in terms of flows into ditches and water levels with ditches, and surface water flows grassland n/a tables Understanding of how abstraction could affect water levels within the network of ditches in the structure will also be noted. We will collate information on habitat distribution onto the floodplain for periods in the winter months.” This acknowledges that local habitat floodplain. Assessment of the vegetation communities of the floodplain (including and assess the effect of different abstraction scenarios upon preferred water level control can have a major influence on flows and water levels in such aquatic/wetland brackish and saline and other semi-natural communities). habitat requirements in the potentially affected areas changes. Location of the The impact on the intrusion of The report requests an assessment of the impact of abstraction on the intrusion of seawater Long term monitoring of key species should be undertaken part of the routine mixing zone, or seawater into Southampton Water into Southampton Water and the lower Test. This could be done by monitoring the location of The impact on intertidal habitat can be assessed in terms of the change in Intertidal monitoring activities of the conservation agencies. The supply of organic carbon is key indicator and the lower Test, and on the the mixing zone or the location of key indicator species such as fucoid seaweeds. Also, an flow volume entering the tidal reaches in terms of the change in frequency, mudflats influenced by a large variety of different parameters (as noted in the report), and species , mudflat supply of organic carbon to assessment of increased abstraction on the supply of organic carbon to Southampton Water duration, magnitude and timing. so should be subject to a research exercise funded by the conservation agencies. carbon levels Southampton Water is needed i.e. monitor carbon levels in the mudflats

13 Lower Test NEP Volume 3: Appendices Appendix 1.5.1 Testwood PWS Abstraction

Table of Contents

1. Introduction 1 1.1. Background 1 1.2. Configuration of the Testwood infrastructure 1

2. Monitoring volumes 4 2.1. SWS Approach 4 2.1.1. The Process 4 2.2. Environment Agency operational monitoring 5 2.3. Time series of NALD data 5 2.4. Water taken from the River 5 2.4.1. Calculating the change in storage of Little Testwood Lake 6 2.4.2. Derivation of a time series for the ‘Water taken from the River’ 6

3. Summary 7

Annex A: Data Processing 5

1. Introduction

1.1. Background

Southern Water Services (SWS)’s Testwood water supply works (WSW) abstracts water from the River Test. The Lower Test is currently subject to an NEP investigation, examining the effects of the Testwood abstraction upon the hydrology and ecology of the river. Discussions during a Steering Group meeting of the Lower Test NEP project (23 November 2011) identified a need to ensure that an agreed time series of abstraction at Testwood was being used in the Lower Test NEP investigation.

This Appendix summarises the main flow meters at Testwood, how SWS calculates the volume abstracted from the river, and what constitutes the agreed approach for the derivation and use of a Testwood abstraction time series for the Lower Test NEP Investigation.

1.2. Configuration of the Testwood infrastructure

An indicative schematic showing the Testwood infrastructure is shown in Figure 1. Water is abstracted from the river through twin culverts to a low level pumping station (marked as LLPS on Figure 1). Water is then either pumped directly to the works or is pumped to Little Testwood Lake, a bankside storage reservoir.

Little Testwood Lake came online in 2004 and is used as a balancing storage facility. With a storage capacity of 250 Ml it is used operationally to provide some raw water settlement prior to treatment. As shown in Figure 1, some of the water taken from the River is first stored in Little Testwood Lake before it is treated and passed into the public water supply network. Meters are positioned on the Inflow and Outflow of the Little Testwood Lake.

1 Lower Test NEP Volume 3: Appendices

Generally, it can be stated that in winter, water is pumped from the River to the Lake; in summer water is pumped from the River to the Lake and also from the River direct to the treatment works (for blending with water pumped from the Lake to the works).

Table 1 lists the different flow meters used (as shown in the schematic in Figure 1), their different reference identifiers, and their purpose.

Table 1 Description of the different meters used in the Testwood infrastructure Meter Meter Use Address Description 24 hour PI Tag 15 min PI Tags Ref Testwood Lake Flow meter to TESTLAKE-SWR- TESTLAKE-SWR-A- Abstraction F001 Reservoirs Testwood Lake A-192553 174770 Testwood Lake Flow meter to WSW TESTLAKE-SWR- TESTLAKE-SWR-A- Abstraction F004 Reservoirs from Testwood Lake A-192551 174766 Testwood Surface 185288TESTLAKE- Abstraction Testwood Low Works HA0194 Water Abstraction SWR-A-174866 Testwood Band District Testwood CR0042 Screen TESTWOOD-WSW- Supply Testwood WSW HL to Rownhams R0199 A-185020 Low Lift 1. Clarifiers 1 TESTWOOD-WSW- TESTWOOD-WSW- Supply Testwood WSW HS0206 - 4. A-185288 A-185016 TESTWOOD-WSW- TESTWOOD-WSW- Supply Testwood WSW Low Lift 2. Clarifier 5. HS0207 A-185289 A-184973 Low Lift 3. Clarifiers 6 TESTWOOD-WSW- TESTWOOD-WSW- Supply Testwood WSW HS0208 - 7. A-185290 A-185018 24" Flow to TESTWOOD-WSW- TESTWOOD-WSW- Supply Testwood WSW PX0406 Rownhams (North) A-185292 A-185020 24" Flow to TESTWOOD-WSW- Supply Testwood WSW PX0405 Rownhams (South) A-185020 Esso/IOW High Lift TESTWOOD-WSW- TESTWOOD-WSW- Supply Testwood WSW flow meter. Industrial PX0703 A-185293 A-185021 main. TESTWOOD-WSW- TESTWOOD-WSW- Supply Testwood WSW HL to Waterside PX0301 A-185291 A-185019 NB Different PI tags exists according to the type of data collected (e.g. 15 minute or 24 hour daily averages)

2 Lower Test NEP Volume 3: Appendices

Figure 1 Schematic of the infrastructure at Testwood works NB reference for each flow monitor is given as the flow meter reference, then the PI Tag, for example: HS0206, TESTWOOD-WSW-A-185288.

NB Different PI tags exist according to the type of data collected (e.g. 15 minute or 24 hour daily averages) (see Table 1). The 24 hour daily average PI Tag is shown above, except where noted. The three clarifiers are monitored by flow meters with the tags HS0206, HS0207, and HS0208.

3 Lower Test NEP Volume 3: Appendices 2. Monitoring volumes

2.1. SWS Approach

Water is abstracted from the River Test through twin culverts to a pumping station. Water is then either pumped directly to the works or is pumped to the holding lake (Little Testwood Lake). Water pumped to the works can be a combination of directly extracted water from the river and water from the lake.

Currently Southern Water calculates the volumes submitted to the Environment Agency, for the National Abstraction Licensing Database (NALD), as the sum of flow to the three clarifiers (meters HS0206, HS0207 and HS0208 in Figure 1). The raw water feed to the works can be taken directly from the river; a mixed stream comprising river water and water from Little Testwood Lake; or water from Little Testwood Lake only. The precise mix is a function of operational and water quality reasons. These three meters record the total flow into the works.

The more straightforward approach to calculating the abstraction from the river would be to measure the flow in the twin culverts, however siltation significantly affects the accuracy of the flow measurement. Southern Water is continuing its investigations into how to improve the flow measurement in the culverts in order to directly record the abstraction volume. As a result of this, in January 2012, improved flow monitors have been installed in the culverts from the river in order to directly measure the volume taken from the River.

2.1.1. The Process

Southern Water’s NALD returns are based on the sum of daily flow to the clarifiers:

 Flow meter HS0206, PI tag TESTWOOD-WSW-A-185288 (clarifiers 1 to 4);

 Flow meter HS0207, PI tag TESTWOOD-WSW-A-185289 (clarifier 5); and

 Flow meter HS0208, PI tag TESTWOOD-WSW-A-185290 (clarifier 6 & 7).

Southern Water’s returns process for the Testwood abstraction is as follows:

 Each week (Monday) the Hants Distribution Input (DI) spreadsheet is updated. This spreadsheet is linked to PI tags and provides abstraction, output and transfer data for all SWS’ sites;

 Any anomalous figures are identified and resolved;

 Should the flow figure require changing then a DI alteration process is used which will change the archived PI value to the correct value. This process requires approval from the Water Technical Manager;

 Each month the data from the DI spreadsheet is sent to the respective Area Manager for approval and to be signed off;

 When the approved DI data is received back, staff input the abstraction data into a master spreadsheet that mirrors the EA spreadsheet (i.e. sites are ordered by abstraction licence number, data is in m3/sec, and expressed to three decimal places).

 This process facilitates the submission of the Abstraction Return to the EA: SWS staff download the EA template from the website, copy and paste data from the SWS master spreadsheet into the template and then submit it to the EA.

4 Lower Test NEP Volume 3: Appendices

2.2. Environment Agency operational monitoring

The Environment Agency regards the NALD returns made by Southern Water as the official record of volumes abstracted from the Testwood abstraction.

For operational purposes however, the Environment Agency has continued to monitor a telemetered output from the Testwood Works and have regarded this as the volume abstracted from the river. Upon investigation it is thought that the telemetered data monitored by the Environment Agency does not comprise the total abstracted volume.

The Environment Agency has stated that it ask Southern Water to confirm that the meters within the culverts are now located to accurately measure abstraction from the source of supply and to supply that data to the Agency in the future. It will also ask for a telemetry feed from that meter to replace the signal currently monitored.

2.3. Time series of NALD data

Figure 2 shows the Testwood PWS NALD returns over time. The data shown in Figure 2 exhibits day to day fluctuation in the early period of the dataset which gradually lessens over time. The volume of water is generally within a band of approximately 50 Ml/d to 60 Ml/d; the average over the entire dataset is computed to be 60.20 Ml/d. Figure 2 shows that a minimum value of zero Ml/d is recorded on 1-2 February 2008. There is a note attached to these dates from SWS that states that no abstraction occurred on these days due to high turbidity in the river.

Given that the NALD data are comprised from the volume of water to the clarifiers and into supply, it could be presumed that the volume should not exceed the nominal operating capacity of the work which, currently, is 105 Ml/d. However, there are some instances where Figure 2 shows that values over this threshold have been recorded historically, and there are a number of reasons why this is the case.

The first of these is that the operating capacity of Testwood has not been static over time: there are a number of different treatment streams or units being used, and each has been subject to modification or change in operational capacity over time.

Finally, the use of the works for different end users would also affect the volume passing through the clarifiers and into supply. For example, in the 1990s two treatment streams were operational at Testwood: for industrial and domestic supply. As the industrial stream was not subject to the same quality regulations as that for potable supply, a greater volume could be passed through the treatment units at Testwood as it would be subject to additional treatment elsewhere (e.g. at Broadfields WSW on the IOW and at a large industrial user).

2.4. Water taken from the River

As stated above, the NALD data are determined from totalling the volumes of water pumped to the clarifiers before being distributed into supply, because of difficulties in measuring flow in the culverts.

For the NEP, in order to derive a time series that represents the water taken from the river, rather than the volume treated, it is necessary to account for any changes in storage in Little Testwood Lake. The reason for this is outlined below.

Water is pumped from the River to the Lake for storage. Water is also pumped from the Lake to the clarifiers for treatment and to be put into supply. Therefore on any particular day there will be water pumped to the Lake for storage, and from the Lake for supply. The balance of the water pumped into and out of from the Lake represents the change in storage from the Lake.

The volume of water pumped to the clarifiers (either from the Lake, or directly from the River) is measured by the inlet meters to the clarifiers, and this is included in the NALD data.

5 Lower Test NEP Volume 3: Appendices

What is not included in the NALD data is any change in storage in Testwood Lake, i.e. when more water is pumped from the River to the Lake for storage than is pumped to the clarifiers; and conversely when less water is pumped from the River to the Lake than is sent to the clarifiers.

Accounting for the change in storage of Little Testwood Lake would address this issue: adding the change in storage of the Lake to the NALD data would give the water abstracted from the River on any particular day.

2.4.1. Calculating the change in storage of Little Testwood Lake

There are meters on the Inflow and Outflow of Little Testwood Lake (see Figure 1). The difference in the volumes recorded by these meters represents the change in storage.

Figure 3 shows the data from the meters. The figure shows that the Inflow (as measured by the meter on the ‘inlet’) and the Outflow data (as measured by the meter on the ‘outlet’) closely mirror each other. The figure also, as per the NALD data in Figure 2, shows a day to day fluctuation is evident before November 2006, particularly for the Inflow to the Lake, compared to the Outflow. After November 2008, there is less fluctuation in both the Inflow and Outflow of the Lake.

Figure 3 also shows that there is a period of infilled data in the dataset, between October 2006 to November 2008 (shown by the red and black lines). There is a gap in the inlet and outlet meter datasets over this time period and therefore it has been necessary to infill data so that a continuous dataset can be generated. Annex A details the processing that was undertaken to fill this data gap and generate a continuous dataset.

The change in storage has been determined by subtracting the volume measured by the Lake Inlet meter from the volume recorded by the Lake Outlet meter. Figure 4 shows that the change in storage is both positive and negative. A positive difference indicates that more water is being pumped into the Lake than is being pumped to the clarifier i.e. water is being stored in the Lake. A negative difference indicates that less water is being pumped into the Lake than is being pumped to the clarifier i.e. stored water is being used.

As the volume computed by the sum of the clarifiers includes the water taken from the Lake, the positive and negative change in storage of the Lake can be interpreted as follows:

 A positive change in storage of Testwood Lake = the sum of the clarifiers is less than the volume taken from the river; and

 A negative change in storage of Testwood Lake = the sum of the clarifiers is more than the volume taken from the river.

Table 2 presents summary statistics for the change in storage volumes. Overall the average change in storage is small at -1.56 Ml/d, with a range of +27 Ml/d to -42 Ml/d.

Table 2 Summary statistics for the change in storage at Testwood Lake Positive change in Negative change in Ml/d All data set Storage Storage Daily average -1.56 2.45 -3.36 Daily maximum change 26.75 26.75 0 Daily minimum change -41.89 0 -41.89

2.4.2. Derivation of a time series for the ‘Water taken from the River’

Figure 4 shows the time series that represents Water taken from the River. This time series has been calculated by adding the change in storage of Little Testwood Lake to the NALD data. As the change in storage is both positive and negative, so the time series ‘Water taken from the River’ is sometime higher and lower than the NALD data, respectively.

6 Lower Test NEP Volume 3: Appendices

For comparison purposes, Figure 4 also shows the NALD data, and it can be seen that the two datasets are near identical, however a greater degree of day to day fluctuation is shown by the data series of the ‘Water taken from the River’.

As demonstrated by the summary statistics in Table 2, the average change in storage is small, however there is a difference in volume on a day to day basis.

Therefore it was agreed by the Steering Group that the calculated time series, ‘Water taken from the River’, is used in the NEP investigation assess the effects of Testwood upon the Lower River Test.

3. Summary

The purpose of this Appendix is to present the steps that were undertaken to produce an abstraction time series that best represents the volume of water taken from the River Test at Testwood.

The Steering Group has approved of the method and data recommended to derive the abstraction time series, and therefore it has been used in the Lower Test NEP Investigation.

7 Lower Test NEP Volume 3: Appendices

120

1.30 110 1.20 100 1.10 90 1.00

80 0.90

70 0.80

60 0.70

NALD data Ml/d data NALD 0.60

50 NALD data Ml/d data NALD 0.50 40 0.40 30 0.30 20 0.20

10 0.10

0 0.00

Jul 04 Jul 05 Jul 06 Jul Jul 07 Jul 08 Jul 09 Jul 10 Jul 11 Jul

Jan 08 Jan 09 Jan 10 Jan Jan 04 Jan 05 Jan 06 Jan 07 Jan 11 Jan 12 Jan

Mar Mar 05 Mar 06 Mar Mar 04 Mar 07 Mar 08 Mar 09 Mar 10 Mar 11

Nov04 Nov05 Nov06 Sep08 Sep09 Nov11 Sep04 Sep05 Sep06 Sep07 Nov07 Nov08 Nov09 Sep10 Nov10 Sep11

May 08 May 09 May 10 May May 04 May 05 May 06 May 07 May 11 May

Figure 2 Testwood NALD data

1 Lower Test NEP Volume 3: Appendices

120 1.30 110 1.20 100 1.10 90 1.00 80 0.90

70 0.80

60 0.70

Ml/d m3/s

50 0.60 0.50 40 0.40 30 0.30 20 0.20

10 0.10

0 0.00

Jul11

Jul05 Jul09 Jul04 Jul06 Jul07 Jul08 Jul10

Jan 11 Jan

Jan 04 Jan 07 Jan 08 Jan 12 Jan Jan 05 Jan 06 Jan 09 Jan 10 Jan

Mar 11 Mar

Sep 11 Sep 11 Nov

Mar 04 Mar 08 Mar Mar 05 Mar 06 Mar 07 Mar 09 Mar 10 Mar

Sep 05 Sep 06 Sep 06 Nov 09 Sep 10 Sep 10 Nov Sep 04 Sep 04 Nov 05 Nov 07 Sep 07 Nov 08 Sep 08 Nov 09 Nov

May 11 May

May 04 May 05 May 08 May 09 May May 06 May 07 May 10 May

F001 TSTWD LAKES INLET F004 TSTWD LAKES OUTLET Infilled data - inlet meter Infilled data - outlet meter

Figure 3 Daily Inflow and Outflow to Little Testwood Lake

2 Lower Test NEP Volume 3: Appendices

40 0.46 A positive difference means that the Abstraction Data is lower than the calculated water taken from the River. 0.42 This is because water is being taken from the River and stored in the Lake. 0.38 30 0.34 0.30 0.26 20 0.22 0.18 0.14 10 0.10 0.06

0.02 /s 0 3

-0.02 m Ml/d -0.06 -0.10 -10 -0.14 -0.18 -20 A negative difference -0.22 means that the Abstraction -0.26 Data is higher than the -0.30 calculated water taken from -30 -0.34 the River. This is because -0.38 stored water in the Lake is being used. -0.42

-40 -0.46

Jul11

Jul05 Jul06 Jul07 Jul08 Jul09 Jul10 Jul04

Jan 11 Jan

Jan 05 Jan 06 Jan 07 Jan 08 Jan 09 Jan 10 Jan 12 Jan Jan 04 Jan

Mar 11 Mar

Sep 11 Sep 11 Nov

Mar 04 Mar 05 Mar 06 Mar 07 Mar 08 Mar 09 Mar 10 Mar

Nov 04 Nov 05 Nov 06 Nov 07 Nov Sep 04 Sep 05 Sep 06 Sep 07 Sep 08 Sep 08 Nov 09 Sep 09 Nov 10 Sep 10 Nov

May 11 May

May 04 May 05 May 06 May 07 May 08 May May 09 May 10 May

Figure 4 Daily change in the storage of water in Testwood Lake

3 Lower Test NEP Volume 3: Appendices

120 1.30 110 1.20 100 1.10 90 1.00 80 0.90

70 0.80

Ml/d /s

60 0.70 3 m

50 0.60 0.50 40 0.40 30 0.30 20 0.20

10 0.10

0 0.00

Jul11

Jul04 Jul05 Jul08 Jul09 Jul06 Jul07 Jul10

Jan 11 Jan

Jan 04 Jan 05 Jan 09 Jan 10 Jan Jan 06 Jan 07 Jan 08 Jan 12 Jan

Mar 11 Mar

Nov 11 Nov

Sep 11 Sep

Mar 04 Mar 05 Mar 08 Mar 09 Mar Mar 06 Mar 07 Mar 10 Mar

Sep 04 Sep 04 Nov 08 Sep 08 Nov 09 Sep 09 Nov Sep 05 Sep 05 Nov 06 Sep 06 Nov 07 Sep 07 Nov 10 Sep 10 Nov

May 11 May

May 04 May 05 May 08 May 09 May May 06 May 07 May 10 May

Water taken from the River NALD Return Figure 4 The calculated time series showing Water Taken from the River, with the NALD data

4 Lower Test NEP Volume 3: Appendices

Annex A: Data Processing

As stated in Section 2.4 there is a gap in the data record of the Inlet and Outlet meters between October 2006 and November 2008. It has been necessary to infill this data gap so that a continuous record can be used. This Annex summarises the data processing that has been undertaken to fill the gaps in the dataset of the Little Testwood Lake Inlet and Outlet meters.

The first attempt to infill the data gap was to use information from the Testwood Controllers Log, which record the pumps used to move water around the system. The Controllers log comprises a record of the operational decisions and processes deployed during each shift and it contains information on a range of water quantity and quality aspects. The information needed to infill the data gap is not directly recorded but it was hoped that sufficient information could be gleaned from the logs to make an informed estimate. However it was realised that while relevant information could be gleaned from the logs, the granularity of the information available would mean that only moderate confidence could be given to the estimates of pumped volumes arising.

Therefore an alternative approach was followed, which uses the high quality data available from the Inlet and Outlet meters before and after the data gap to derive a daily estimate of the volumes pumped into and out of Little Testwood Lake.

Each dataset from the Inlet and Outlet meters were processed in turn as follows:

 Each data point was assigned a day-of-year label. Then the period of time relating to the data gaps was removed from the dataset so only the data comprising good quality data remained;

 The data were then ranked according to day of year so all the data points for a particular day were grouped together irrespective of year. For example, all the 1st January days are ranked first then all the 2nd January days and so on;

 An average daily value was demined for each day of year data point. There are 5-6 years of data so the daily averages were compiled for 5-6 data points. Leap years are accounted for in this process, but have fewer data points (2) from which the average was determined; and

 The derived daily average volume for a particular day of year was used to infill data over the missing time period. For example data are missing for 11 Oct 2006, which has the day of year number of 284. Therefore the daily average volume generated for day 284 was used to infill the data gap for this day.

Figure 3 in this document shows the results of infilling the dataset in this way (the infilled data is represented by the red and black lines). Because it is derived from averaged values before and after the period of missing data, the infilled dataset fills the gap with appropriate volumes and bridges the change in pattern from the relatively noisy trend before the data gap, and the steady pattern afterwards.

An alternative option could have been to derive an average volume using the data before the data gap, or afterwards. However, it is evident that the meter data before and after are different in trend, so justification would be required to use the ‘before’ or ‘after’ dataset. By deriving an average from the whole dataset, all available data are being used which gives the derived values higher confidence.

The disadvantage of this approach is that it does not take into account actual operating conditions and what may have actually been taken on day to day basis. However, other than the controller logs (which are not deigned to collect data for this purpose) no other data are available. It is argued that the magnitude of error that might be incurred by the above approach is likely to be small and is not of significance to the end objective.

Therefore it was agreed by the Steering Group that this approach for generating data to infill the data gap should be and is used as part of the dataset to derive the time series of ‘Water taken from the River’.

5 Lower Test NEP Volume 3: Appendices

Appendix 1.6.3 Water Body Summary Sheets

This appendix contains the Environment Agency Summary Sheets for the following WFD Water Bodies:  River Blackwater  Test (Lower)

1

Live Document – subject to change

Water Body Summary Sheet

Water Body Summary Information (Data based on SERBMP Dec 2009)

WATERBODY ID WB NAME CATCHMENT WB TYPE HMWB Gb107042016790 River Blackwater Test & Itchen River No WB COORDINATOR/TEAM AIG LEAD DESK STUDY AUTHOR Dominic Longley Alice Gregory-Morris

Designations Bathing Drinking Shellfish Freshwater Nitrates Urban Waste Wild Birds Habitats and Water Water Water Fish Directive Water Directive Species No Yes No Yes (Sal) Yes – N/A No No No

Elements Other Failing Overall Ecological Confidence WB is Driving Elements Status/Potential less than good Classification (element status) Elements Passing Invertebrates, DO Ammonia (Phys chem and Annex 8), Phosphate, Arsenic, Copper, Iron, Poor Very Certain Fish (poor) None Phenol, Zinc, pH, Temperature Chemistry (see below for list), Quantity and Dynamics of Flow, Morphology

Relevant Monitoring Points Physico- Diatoms Macrophytes Fish Invertebrates Chemical Chemistry 42087 G0003989 G0003989 43099 Hammond’s Farm G0003994 G0003994 Not Monitored Not Monitored 43323 13478 G0004007 G0004007 43353 G0004008 G0004008 43638

Wellow Mill invertebrate site on River Blackwater, Spring 2000

Last saved by DLONGLEY 09/03/2012

Situation

BACKGROUND = This stretch of the River Blackwater runs for over 21km, but does not run through the New Forest National Park. The waterbody begins at West Wellow and runs through Embley Wood and confluences with the River Test at Testwood. This waterbody is similar to others found in the New Forest, it is oligotrophic and acidic, although there is a small carrier that originates from the River Test (Broadlands Estate) and confluences with the Blackwater, above Broadlands and Testwood lakes. This introduces a chalk influence on the bottom end of the Blackwater. This waterbody also includes the tributaries, Red Lodge Stream, Shootash Stream and Canada Stream. A mapping error occurred in the original River Basin Management Plan which included part of the Lower Test waterbody, this is in the process of being corrected and so the Drinking Water designation should be removed.

STATUS = Currently, the bottom waterbody of the River Blackwater is classified as Poor due to the fish failure, but is predicted to improve to Moderate by 2015.

PRESSURES = The river flows through mostly natural wooded habitat, so the watercourse is heavily shaded. There is also a Sewage Treatment Works (STW) called West Wellow. Other pressures thought to impact the waterbody are, alien species (Himalayan balsam), sedimentation from diffuse pollution, pesticides and dangerous substances (but currently all chemical analysis show this waterbody is passing for: Benzo (a) and (k) fluroanthene, Benzo (ghi) perelyene and indeno (123-cd) pyrene, Benzo(a)pyrene, Trichlorobenzenes, para-para DDT, Aldrin, Dieldrin Endrin, Isodrin, Fluoranthene, Hexachlorobenzene, Hexachlorobutadiene and Hexachlorocyclohexane. Also Cadmium, Lead, Mercury and Nickel and their compounds). Sedimentation is due to be addressed through Catchment Sensitive Farming which will begin in Summer 2011 (Action SE0306).

FAILING ELEMENT OVERVIEW = The following is a summary of the current situation for each failing element. This was last updated on 24th Feb 2011.

Fish – Expected to improve by 2015: This waterbody is classified as failing for fish on the basis of one survey conducted at Hammond’s Farm in 2004. However, the catch data and the FCS2 diagnostic data should be disregarded because no stop nets were used, minor species and eels were not recorded and only part of the channel width was surveyed. The species driving the failure were eel, trout, bullhead and salmon. In the 2010/2011 national fish monitoring review two new sites were selected as the permanent fish sampling points for this catchment: Upper Whinwhistle & Broadlands fishery. Both sites were surveyed in September 2011 and bullhead and eels were abundant at both. Trout were abundant at Whinwhistle and present at Broadlands. These data will be used to reclassify the waterbody in the 2012 FCS2 analysis and it is expected that the fish class will improve to “Good”. The class may be suppressed by the expected prevalence for salmon but this is considered erroneous – salmon rarely enter any Test tributary and there is no record of the species in the Blackwater, despite the river supporting abundant migratory and non migratory brown trout.

KEY PARTNERS – Southern Water Ltd

Water Body Action Team (to include external stakeholders where appropriate)

Resource Target Team / Action ID Action Description Estimate Progress Name Date Organisation (Days or £) RBMP Actions Improvements to water company assets at 21 locations in the Test and Itchen Catchment, to deliver benefits against the pressures identified or investigate the need for further investment. The SE0106 2015 Southern Water improvements include: - A scheme to ensure no deterioration in the current classification as a result of increased volumes of discharge from West Wellow STW Establish a 'Regional Better Rivers Programme’ to improve habitat and ecology in a first round of waters. Jo SE0223 2012 EA (SEP) Outcome: Improve ecological status or Simmons potential in 35 water bodies totalling 389 km in length Work with Natural England to target Catchment Sensitive Farming type activities and agri-environment schemes Nigel EA (EM New to ensure adoption of best farming SE0306 2012 Thomas- Forest and practices. Outcome: Reduce diffuse Childs Test) pollution sources from agriculture within water bodies identified as being impacted or at risk. Sub Actions

WB Add on RBMP Action

Sub Actions

New Actions Carry out investigations into the origins, causes of and solutions to WFD fish Dom NA1 2011 Complete EA (A&R) failures where we need to improve Longley certainty Sub Actions Discuss the reasons for the high expected prevalence for salmon with the 1 Feb Dom NA1-1 Complete EA (A&R) FCS2 National Lead – identify a solution 2010 Longley to the problem Conduct new fish population survey at Hammond’s Farm using appropriate Complete methods to ensure sufficient catch 31 Nov Dom NA1-2 (NB new EA (A&R) efficiency for target species, especially 2011 surveys at 2 Longley eels and bullhead – reclassify waterbody alternative sites) on the basis of new data Assess relevance of RBMP measures Dom NA1-3 2011 Complete EA (A&R) below if fish failure confirmed. Longley Redundant RBMP Actions (Of those listed above) Action ID Action Description Reason no longer relevant Carry out investigative riverine and land based field work into the origins, causes SE0199 and solutions to sedimentation. Outcome: This action will be addressed through SE0306 Improve our understanding of problems, in order to take effective action to

Water Body Action Team (to include external stakeholders where appropriate)

Resource Target Team / Action ID Action Description Estimate Progress Name Date Organisation (Days or £) address them. Investigation at Site of Special Scientific This action relates to the Lower Test. Originally attached to the SE0340 Interest perceived to be adversely Blackwater due to mapping error which has been corrected affected by abstraction (SE0340) Tributyltin (TBT) compounds. Investigate the reason for failure: to assess the This action relates to the Lower Test (awaiting confirmation from SE0188 contribution from dredging or disposal Area Planning Team Jenny Stillwell). Originally attached to the activities on EQS compliance as Blackwater due to mapping error which has been corrected appropriate Mitigation Measures (MM) Resource Target Team / MM ID MM Description Estimate Progress Name Date Organisation (FTE or £)

Map of Catchment –

Glossary A&R Analysis and reporting team ASPT Average Score Per Taxa BIOSYS Our main database for storing, manipulating and reporting data from freshwater and marine biological surveys at any taxonomic level BMWP Biological Monitoring Working Party CEO Combined emergency overflow CSF Catchment sensitive farming CSM Customer Self Monitoring (of STPs/WIMS sampling points) CSO Combined sewer overflow D/S Downstream DO Dissolved oxygen EM Environment management team EP Environmental planning team FCS2 Fisheries Classification Scheme version 2 FRB Fisheries recreation and biodiversity team HEVI HydroEcological Validation tool LIFE Lotic Invertebrate index for Flow Evaluation NFPD National Fish Population Database NTAXA Number of taxa P Phosphate RIVPACS River InVertebrate Prediction and Classification System RIVPACS predicts the macro-invertebrate fauna at any site on a river from a small number of environmental parameters derived from maps or measured at the site. SERBMP South East River Basin Management Plan SS Suspended solids STP Sewage treatment plant STW Sewage Treatment works U/S Upstream WB Waterbody WQIP Water Quality Improvement Plan WWTW Waste water treatment works

Live Document – subject to change

Water Body Summary Sheet

Water Body Summary Information (Data based on SERBMP Dec 2009)

WATERBODY ID WB NAME CATCHMENT WB TYPE HMWB GB107042016840 Test (Lower) Test and Itchen River No WB COORDINATOR/TEAM AIG LEAD DESK STUDY AUTHOR TBC Elliot Tinton- Environmental Planning Cath Nelson Emma McSwan

Designations Bathing Drinking Shellfish Freshwater Nitrates Urban Waste Wild Birds Habitats and Water Water Water Fish Directive Water Directive Species No Yes No Yes (Sal) Yes Yes Yes Yes

Elements Other Failing Overall Ecological Confidence WB is Driving Elements Status/Potential less than good Classification (element status) Elements Passing Fish (Good), Invertebrates (High) Phosphate (Both Good), Ammonia (Both High), Dissolved Oxygen (Good), pH, Temperature, 2,4-dichlorophenol, 2,4- dichlorophenoxyacetic Acid, Arsenic, Copper, Cypermethrin, Dimethoate, Iron, Linuron, Mecoprop, Permethrin, Phenol, Toluene, Zinc, 1,2- Macrophytes dichloroethane, Atrazine, Benzene, Benzo (Moderate) (a) and (k) fluoranthene, Benzo (ghi) perelyene and indeno (123-cd) pyrene, Chemistry- Benzo(a)pyrene, Cadmium And Its

Phytobenthos Tributyltin Compounds, Diuron, Fluoranthene, Poor Quite Certain (Poor) Compounds Hexachlorobenzene, (Moderate) Hexachlorobutadiene, Hexachlorocyclohexane, Lead And Its Compounds, Mercury And Its Compounds, Napthalene, Nickel And Its Compounds, Pentachlorophenol, Simazine, Trichlorobenzenes, Trichloromethane, Trifluralin, Aldrin, Dieldrin, Endrin & Isodrin, Carbon, Tetrachloride para - para DDT, Tetrachloroethylene, Trichloroethylene (All High)

Relevant Monitoring Points Physico- Diatoms Macrophytes Fish Invertebrates Chemical Chemistry 13386 (Moorcourt G0003885 Carrier) (River Test at G0003885 (River 43301- Broadlands 43301- Broadlands 13392 (Moorcourt 43301- Broadlands Testwood) Test at Testwood) (Longbridge) (Longbridge) Main) (Longbridge) G0003890 G0003890 (River 30864 (Red River (River Test at Test at Longbridge) Carrier, Broadlands) Longbridge)

Last saved by EMcSwan 09/03/2012 Photographs of catchment

Downstream Main Bridge, Romsey Upstream Longbridge, Broadlands

Situation BACKGROUND = The River Test is one of Southern England’s classic chalk streams. The waterbody runs from Polhampton, and continues through Whitchurch, Stockbridge, Romsey and Southampton. This waterbody begins downstream of Main Bridge in Romsey and ends at the tidal limit of Southampton Water. The river is internationally renowned for its fly fishing. Its calcium-rich, high alkalinity waters support a characteristic flora and fauna, high in productivity of plants, invertebrates and fish. A mapping error has missed the inclusion of the Little Test, this will be corrected. (Action NA1) STATUS = The lower River Test is currently at Poor status, and is not predicted to improve by 2015 as it would be technically infeasible and disproportionately expensive to do so. The Lower Test is a SSSI (Unit 91 in unfavourable condition, no change). After this investigation, we hope this waterbody will improve to Moderate by 2015, based on improvement in the Diatom/Phytobenthos element (Action SE0196-1 and -2). PRESSURES = The lower River Test has been managed for many hundreds of years. The water levels have been altered through the development of the water meadow system since the 17th century and for use by water mills. Today the river is a key provider of the public water in the area. In addition, the river has historic fish farm structures, fisheries and is managed to maintain the in-river fishing interests in particular towards the preservation of the Salmon and Trout populations. Romsey WWTW discharges into the waterbody just above this one. FAILING ELEMENT OVERVIEW = The following is a summary of the current situation for each failing element. This was last updated on 10h December 2010. Phytobenthos: This is most likely impacted from the historical phosphate loading into the river from Romsey WWTW. Phosphate stripping has been installed, and in time we should see the diatom community improve at the one site the classification is based on. This improvement will be monitored by the Environment Agency as part of the WFD Surveillance programme. Also, diatom and chemical samples can be taken from the top of the Test to confirm if phosphate levels, or the classification tool, are of the greater influence upon the diatom failure (Action SE0196-1). The diatom classification tool DARLEQ is being updated to include more high alkalinity sites, this will allow a better, more accurate reflection of the current diatom community status. DARLEQ 2 is due Winter 2011, area staff will re- run the classification to see if the diatom classification has changed to reflect the improved system. (Action SE0196- 3). Macrophytes: Only one macrophyte sample was used for classification. Further macrophyte surveys are needed to increase confidence in the classification and have been programmed into monitoring. This should clarify/confirm the failure this data will be reviewed in SE0196-2. Tributyltin Compounds: We have investigated the TBT failure in this waterbody. The National Evidence Team are aware there are issues with the TBT classification. Current techniques employed by the National Laboratory have minimum detection levels above the TBT compounds EQS. In the Lower Test water body there is uncertainty that TBT causes any ecological impacts (i.e. imposex in dogwhelks) as there is a lack of suitable habitat for dogwelks. However surveys on the impact of TBT on the dogwelk population on the Northern Shore of the Isle of Wight indicate that the population is recovering from the impacts of TBT. No further investigation of TBT is required in this river basin monitoring cycle. Sedimentation: Sedimentation has been identified as a suspected reason for failure and we will be investigating current data (Action SE0196). Catchment sensitive farming assessment is the normal method to locate sources and reduce sediment, however this waterbody is not considered a high priority in comparison with waterbodies further upstream, such as Wallop Brook and River Dever. CSF will address issues in upstream waterbodies and will therefore decrease sediment loading in this waterbody (making SE0306 for this waterbody redundant). Currently, extensive study of suspended solid data from the Lower Test shows that the amount of suspended solids have decreased in the last 20 years.

Situation Drinking Water Protection: There are four actions related to the protection of drinking water (Actions DrWP0032, DrWP0033, DrWP0034 and SE0017). Monitoring for these chemicals will be ongoing throughout the RBMP cycle by water companies, to fulfil WFD Article 7.3 requirements unless this monitoring shows the substance of concern is no longer a risk. Similarly monitoring may identify new risks and pesticides of concern, and these findings will be incorporated into their future monitoring plans. An update on monitoring progress has been requested from the water companies. Water Resources WR WFD Stage 1 is a desktop study to confirm the flow compliance result is correct and ascertain whether the ecological monitoring sites are suitable for assessing abstraction impacts. The ecological status of suitable monitoring sites are noted. Those where flow non-compliance is confirmed and the ecological assessment indicates there is a potential hydroecological problem, progress to WR WFD Stage 2. WR WFD Stage 2 assesses the reasons for the failure and the water resource abstraction pressure upon the failing ecology. Water Resource WFD Stage 1= The flow compliance result in this waterbody has been confirmed as not supporting ‘Good’ status. The ecology is also indicated as failing. This water body will pass to WR WFD investigations Stage2 (Identify cause of failure)

KEY PARTNERS – There are many interested partners on the Lower River Test which we would be keen to work with on many of the actions. These include the Test & Itchen Association, riparian landowners, Southern Water, Hampshire County Council, Southampton City Council

Water Body Action Team (to include external stakeholders where appropriate)

Resource Target Team / Action ID Action Description Estimate Progress Name Date Organisation (FTE or £) RBMP Actions DrWPA0 Further monitoring to confirm risk of 31 Dec In Jenny EA (EP), Water

032 failure of WFD Article 7 2011 Progress Stillwell Companies Natural DrWPA0 Chlortoluron. Re-direct existing CSF 31 Dec In Jenny England, (EA, 033 resource to address issue. 2011 Progress Stillwell EP) Isoproturon. This substance has been EA (EP), banned/shortly to be withdrawn from use Central DrWPA0 31 Dec In Jenny in the UK. No further measures should Government, 034 2011 Progress Stillwell be required although surveillance (Water monitoring will be done to confirm this. Companies) Isoproturon; Atrazine; Aldrin. This EA (EP), substance has been banned/shortly to be Central withdrawn from use in the UK. No further 31 Dec In Jenny SE0017 Government, measures should be required although 2011 Progress Stillwell (Water surveillance monitoring will be done to Companies) confirm this. Tributyltin (TBT) compounds. Investigate the reason for failure: to assess the 31 Dec Emma SE0188 contribution from dredging or disposal Complete EA (A&R) 2012 McSwan activities on EQS compliance as appropriate EA (SEP Ecol) (Riparian owners; Developers; Establish a 'Regional Better Rivers Forestry Programme’ to improve habitat and Commission; ecology in a first round of waters. 31 Dec Natural SE0223 Jo Simmons Outcome: Improve ecological status or 2012 England; Local potential in 35 water bodies totalling 389 Authorities; km in length Environmental NGOs; Countryside Management Service (CMS) Investigation at Site of Special Scientific 31 Dec In Alison Southern SE0340 Interest perceived to be adversely 2012 Progress Matthews Water, EA (EP) affected by abstraction Sub Actions Investigation or liaise with National team SE0188 - (Nicky Cunningham) into suitability of 31 Mar EA (National, Complete Jon Brock 1 using Tributyltin (TBT) compound EQS to 2011 Evidence) measure Tributyltin. SE0188 - Investigation of minimum detectable limit 31 Mar Emma Complete EA (A&R) 2 of TBT Cation with Starcross Laboratory. 2011 McSwan SE0340 - Update of Southern Water Lower Test 31 Dec Alison EA (EP) 1 Project 2011 Matthews Scoping what mitigation measures need to be implemented to deliver Good 31 Dec Lawrence SE0223-1 Ecological Status and identifying lead EA, (SEP Ecol) 2011 Talks functions/ projects/ partners to implement.

WB Add on RBMP Action Carry out a desk study into the origins, causes of and solutions to pollution where we need to improve certainty. 31 Mar In Emma SE0196 EA (A&R) Outcome: Improve our understanding of 2012 Progress McSwan problems, in order to take effective action to address them. Sub Actions Review diatom data and investigate need 31 Mar In Emma SE0196-1 to take further samples from headwaters EA (A&R) 2012 Progress McSwan for comparison to lower reaches Review macrophyte data from 2011 30 Dec Emma SE0196-2 EA (A&R) classification 2012 McSwan Review new classification from DARLEQ 30 Jun Emma SE0196-3 EA (A&R) 2 for diatoms 2012 McSwan New Actions Raise the need to correctly designate Little Test and Lower Test channels as 31 Mar NA1 Completed Sam Orchard EA (A&R) waterbody, rather than Luzborough Lane 2012 Stream and River Blackwater. Romsey WWTW needs adding to the 30 Sep NA2 Liz Kruba EA (SEP WQ) SERBMP Annex D 2011 Sub Actions

Redundant RBMP Actions (Of those listed above) Action ID Action Description Reason no longer relevant Nutrients and Organics. Carry out investigative riverine and land based field This measure has been replaced with SE0196. An initial desk work into the origins, causes of and study of chemistry and ecological data will show if further SE0200 solutions to pollution where we need to chemical or ecological sampling is needed. This action will be improve certainty. Outcome: Improve our re-instated if Action SE0196 shows that the problem has not understanding of problems, in order to been resolved. take effective action to address them. Carry out investigative riverine and land This measure has been replaced with SE0196. An initial desk based field work into the origins, causes study of chemistry and ecological data will show if further and solutions to sedimentation. Outcome: SE0199 chemical or ecological sampling is needed. This action will be Improve our understanding of problems, re-instated if Action SE0196 shows that the problem has not in order to take effective action to been resolved. address them. This sub-action has been replaced with SE0196. An initial desk Work with CSF Officer/EM Team to study of chemistry and ecological data will show if further SE0199 - design a monitoring programme to chemical or ecological sampling is needed. This sub-action will 1 identify origin of sedimentation. be re-instated if Action SE0196 shows that the problem has not been resolved. Work with Natural England to target Catchment Sensitive Farming type This action has been removed because CSF work in upstream activities and agri-environment schemes waterbodies will address this problem more effectively. If the to ensure adoption of best farming SE0306 actions taking place for Catchment Sensitive Farming in practices. Outcome: Reduce diffuse upstream waterbodies does not resolve the issue then this pollution sources from agriculture within action will be reinstated. water bodies identified as being impacted or at risk. Mitigation Measures (MM) Resource Target Team / MM ID MM Description Estimate Progress Name Date Organisation (FTE or £)

Map of Catchment – Test and Itchen - River Test (Lower) Water body no.: GB107042016840

Poor Diatoms Site ID: 43301

Moderate Macrophytes Site ID: 43301

Poor Fish Site ID: 13392

WFD Monitoring Status Elements ) ! Moderate ") Diatoms ! Poor ! Bad !( Fish #* Inverts Moderate Chemistry Symbols Site ID: G0003885 WFD_rivers_050k ") Chemistry # WWTWWTW discharge KEY ELEMENT 2009 GF Macrophytes Phytobenthos XW DO $+ BOD [_ Ammonia %, Phosphate

Date: 13 Nov 2010 SSD Analysis & Reporting © Crown Copyright All rights reserved. Environment Agency 100026380, (2009). 10/09/2009

Glossary A&R Analysis and reporting team ASPT Average Score Per Taxa BIOSYS Our main database for storing, manipulating and reporting data from freshwater and marine biological surveys at any taxonomic level BMWP Biological Monitoring Working Party CEO Combined emergency overflow CSF Catchment sensitive farming CSM Customer Self Monitoring (of STPs/WIMS sampling points) CSO Combined sewer overflow D/S Downstream DO Dissolved oxygen EM Environment management team EP Environmental planning team FCS2 Fisheries Classification Scheme version 2 FRB Fisheries recreation and biodiversity team HEVI HydroEcological Validation tool LIFE Lotic Invertebrate index for Flow Evaluation NFPD National Fish … Database NTAXA Number of taxa P Phosphate RIVPACS River InVertebrate Prediction and Classification System RIVPACS predicts the macro-invertebrate fauna at any site on a river from a small number of environmental parameters derived from maps or measured at the site. SERBMP South East River Basin Management Plan SS Suspended solids STP Sewage treatment plant STW Sewage Treatment works U/S Upstream WB Waterbody WQIP Water Quality Improvement Plan WWTW Waste water treatment works

Appendix 2.1.1 River Test SSSI Citation

Atkins

County: Hampshire Site Name: River Test

Status: Site of Special Scientific Interest (SSSI) notified under Section 28 of the Wildlife and Countryside Act, 1981 (as amended 1985), Section 17 of the Water Resources Act, 1991 and Section 4 of the Water Industry Act, 1991.

Environment Agency Region: Southern Water Company: Southern Water plc

Local Planning Authority: Hampshire County Council, Basingstoke & Dean Borough Council, Test Valley Borough Council, New Forest District Council

National Grid Reference: SU 533498 to Ordnance Survey Sheet 1:50,000: 185, 196 SU 367150 and SU 361145

Length of River SSSI: Approx. 50 km Area: 442.93 (ha)

Date Notified (Under 1981 Act and 1991 Acts: 10 June 1996 Confirmed: 20 February 1997

Other Information: New site; parts of the site are already separately notified as parts of Bere Mill Meadows, Bransbury Common, Chilbolton Common, East Aston Common, Lower Test Valley and Stockbridge Common Marsh SSSIs.

Description and Reasons for Notification:

Key Features and General Character: The River Test is a classic chalk stream. It is one of the most species-rich lowland rivers in England. The bulk of the system drains Cretaceous Chalk, with the lowermost reaches passing over a variety of geological strata. More recent deposits are common in the valley bottom throughout the system. The Test is a longer and larger river than the neighbouring Itchen, showing a greater downstream succession of species in consequence. Nevertheless, the entire river system exhibits a flora characteristic of chalk streams. In the lower reaches, additional species, more commonly found in other river types, add diversity to the assemblage. The Test supports a high diversity of invertebrate species, and is especially rich in aquatic molluscs.

Apart from the very short winterbourne section, substrate and flow regime combine to produce a fairly uniform pattern and gradual succession of habitats down the river. Shallows with gravel bottoms are a major habitat in the upper river while open stretches with deeper water are more common in the lower reaches. The water is naturally base-rich and of great clarity, but like many lowland rivers shows evidence of nutrient enrichment.

The river has been modified over the centuries by the construction of sluice systems and creation of channels for water meadows, water mills and navigation. Many stretches have also been realigned for such purposes and some deepened for land drainage. This has resulted in a multiplicity of water courses, though as a chalk river it would naturally have a braided channel. The Test is world renowned for game fishing, which is largely provided by brown trout, both wild and stocked populations, and to a lesser extent salmon and sea trout. For many years the riverÕs channels, banks and vegetation have received regular management and alterations, to maintain, improve and facilitate the fishing. All these activities have variously contributed to its present character and appearance.

The riverÕs water is abstracted for public and agricultural use from boreholes as well as from its channel, which also receives discharges from sewage treatment works and a paper factory. There are commercial water-cress farms near the headwaters of the Test, and on certain tributaries it is a substantial industry. There are nine fish farms producing trout for human consumption utilising the waters of the Test, the largest are located on the lower half of the river. Also, smaller trout rearing and holding facilities are a part of many of the fisheries. Traditional water meadow management fell into disuse during this century and their unimproved flood pasture swards, together with the swamp and fen vegetation which developed on them, are still present in those meadows which have not been converted for modern intensive grassland or arable production. Areas of riparian vegetation including reed fen and wet woodland are a frequent feature in the upper half of the Test Valley, and at Leckford are particularly extensive and well developed. From Stockbridge Common downstream, until the marshes of the Lower Test Valley Nature Reserve at the riverÕs mouth, such habitats tend to be localised.

Flora: The Test is more species rich than most other lowland rivers, with the most diverse communities being found in the lower reaches where the substrate is more varied. Over 100 species of flowering plant, moss and liverwort have been recorded along its channel and banks. There are distinct successional changes on passing downstream. In the upper river where velocity is fastest and substrate coarsest, lesser water-parsnip Berula erecta and brook water- crowfoot Ranunculus penicillatus var. pseudofluitans predominate in the classic small chalk stream community. Ranunculus is abundant, if not dominant, throughout most of the rest of the river where flow and substrate conditions suit it, together with its main associates, blunt- flowered water-starwort Callitriche obtusangula and fools water cress Apium nodiflorum. This predominance of Ranunculus is one of the objectives of the fisheriesÕ management. In the larger, more silty channels the vegetation is more diverse, with additional chalk stream species such as mareÕs-tail Hippuris vulgaris, opposite-leaved pondweed Groenlandia densa, the nationally-scarce river water-dropwort Oenanthe fluviatilis and horned pondweed Zannichellia palustris. Species typical of slower water are also common: common club-rush Scirpus lacustris, shining pondweed Potamogeton lucens and Canadian pondweed Elodea canadensis. Extensive growths of blanket weed Cladophora have occurred in recent low flow years, an indicator of increased eutrophication. Ivy-leaved duckweed Lemna trisulca, although not abundant, is a species of chalk rivers only.

The stable river margins have several characteristic species, amongst which lesser pond sedge Carex acutiformis, reed canary-grass Phalaris arundinacea and reed sweet-grass Glyceria maxima are abundant throughout the river. Tall perennials such as greater willowherb Epilobium hirsutum, meadowsweet Filipendula ulmaria, purple loosestrife Stachys palustris, yellow loosestrife Lysimachia vulgaris and orange balsam Impatiens capensis (an introduced species) are also widespread. Amongst the more localised and infrequent river bank species are skullcap Scutellaria galericulata and meadow rue Thalictrum flavum. Low growing water edge plants such as watercress Rorippa nasturtium-aquaticum, water mint Mentha aquatica, water forget-me-not Myosotis scorpioides and water speedwell Veronica anagallis-aquatica often carpet the base of the bank and river margins, and the emergent branched bur-reed Sparganium erectum is frequent. Another feature of the Test is the number of plant species which may be regarded as relics of a wetland flora adjacent to the river and now indicate the wetter stretches of river bank. These include water dock Rumex hydrolapathum, greater tussock-sedge Carex paniculata, common reed Phragmites australis, marsh marigold Caltha palustris and bulrush Typha latifolia. The vegetation of drier banks, where the ground rises higher above the river, or has been made up to be so, usually support a greater proportion of coarse grasses and ruderal herbs with false oat grass Arrhenatherum elatius and common nettle Urtica dioica often in abundance. Trees are an important feature along the river, the roots of alder Alnus glutinosa and willows (usually crack willow Salix fragilis and sallow S. cinerea) bind the banks and provide refuges for river animals such as otter.

The site includes former water meadows, fen pasture and rush pasture communities of botanical interest. Soils in the valley derive from alluvium, peat and ÔtufaÕ (calcareous marl). These, combined with the meadowsÕ networks of ridges and drains, result in complex mosaics of dry grassland, rush pasture, fen-meadow, flood pasture and swamp communities. The floristic diversity of these unimproved meadows is high and species-rich communities typical of wet, calcareous, pastures are well represented. Proximity to the river and its carriers maintains the high ground-water levels which are important for the botanical diversity and interest, as are appropriate levels of grazing. Where present, the transitions these wet grasslands can have with the bank vegetation of water courses and other riparian vegetation in the valley are an important component of the habitat diversity.

The fen meadow and flood pasture communities can be considered characteristic of these former water meadows with moist calcareous soils and are the ones of highest botanical interest. The fen-meadow community is typically rich in plant species but is also very variable in its composition and structure, the differences usually due to environmental and management factors such as grazing and mowing. It forms one of the main elements of interest in most of the meadows in the site. Blunt-flowered rush is usually a prominent feature of this type of fen-meadow where it occurs elsewhere in Britain but in the Test Valley site it is largely localised or absent. Some of the communityÕs typical species are often abundant in these meadows. Amongst the most constant species are ÔgeneralÕ grassland ones: creeping bent Agrostis stolonifera, red fescue Festuca rubra, Yorkshire fog Holcus lanatus, common mouse- ear Cerastium fontanum, jointed rush Juncus articulatus, meadow vetchling Lathyrus pratensis and red clover Trifolium pratense. Other prominent associates represent the fen character of the community: water mint Mentha aquatica, fen bedstraw Galium uliginosum, marsh bedstraw G. palustre, greater birdÕs-foot trefoil Lotus uliginosus, marsh horsetail Equisetum palustre, wild angelica Angelica sylvestris, common fleabane Pulicaria dysenterica, ragged robin Lychnis flos-cuculi and meadowsweet Filipendula ulmaria.

The flood pasture community does not often occupy extensive areas. Usually lying in the transition from dry grassland to wet drain or on lower lying ground alongside the river, it is also rich in species with abundant short sedges and localised carpets of mosses (mostly Calliergon cuspidatum). Many of the flood pastureÕs typical species also feature in the fen meadow community and its distinguishing elements include prominent marsh marigold Caltha palustris, water avens Geum rivale, meadow buttercup Ranunculus acris, carnation sedge Carex panicea and brown sedge Carex disticha, together with crested dogÕs tail Cynosurus cristatus, common sorrel Rumex acetosa, ribwort plantain Plantago lanceolata and common spike-rush Eleocharis palustris. The mown fishing paths are often former parts of the adjoining meadows, and can sometimes retain elements the flood pasture community, whilst it has been lost from the agriculturally improved fields.

In the meadowsÕ diverse flora, species of particular note for being strongly associated with unimproved grassland are: adderÕs tongue Ophioglossum vulgatum, betony Stachys officinalis, bogbean Menyanthes trifoliata, common sedge Carex nigra, distant sedge C. distans, devilÕs-bit scabious Succisa pratensis, marsh arrowgrass Triglochin palustre, marsh lousewort Pedicularis palustris, marsh pennywort Hydrocotyle vulgaris, marsh valerian Valeriana dioica, meadow rue Thalictrum flavum, pepper saxifrage Silaum silaus, purple moor-grass Molinia caerulea, southern marsh orchid Dactylorhiza praetermissa and quaking grass Briza media.

The species usually dominating the fen and swamp communities of the drainÕs reed sweet-grass Glyceria maxima, lesser pond sedge Carex acutiformis and reed canary-grass Phalaris arundinacea can be widespread in the meadows and occur in the other communities. The thick tall vegetation of the drains can also include gipsywort Lycopus europaeus, marsh thistle Cirsium palustre, common valerian Valeriana officinalis and meadowsweet, and in places greater tussock-sedge Carex paniculata swamp has developed, especially around springs. Rush pasture, dominated by hard rush Juncus inflexus and jointed rush J. articulatus, also occurs in parts of the meadows and more extensively in those alongside the river at Romsey.

Grazing and mowing maintains the structure and diversity of the unimproved swards, and along certain drains and stretches of river-bank, cattle grazing also creates and maintains particular habitat conditions on which quite specialised invertebrates depend.

Tall riparian vegetation with willows and scrub (including hawthorn Crataegus monogyna, guelder rose Viburnum opulus and elder Sambucus nigra) is a particular feature along many stretches of the river channels. Sometimes separated from the bank by fishing paths it is generally dry, elsewhere it is a continuation of a swampy river margin with beds of common reed and other tall emergents. Though comprising many of the plant species featuring in parts of the grazed meadows and open river banks, these areas provide valuable habitat for wetland birds and invertebrates such as moths, beetles, spiders and the adult phases of mayflies and other ÔfishermanÕs fliesÕ. Wet woodland is also an important component of the river valley, similarly providing particular habitat conditions for its dependent fauna. Alder and willows are usually the main species, together with ash Fraxinus excelsior, oak Quercus robur and occasionally birch Betula pubescens. The woodland is often ÔsecondaryÕ having developed from the more open vegetation, and usually retains many of its plant species. Where the woodland has long been established a more distinct flora can be present. One small area of wet alder woodland at a location which has had a long continuity of wooded cover, supports a very varied flora with quite numerous twayblade Listera ovata and herb Paris Paris quadrifolia, an uncommon and localised species.

Invertebrates

Over 232 invertebrate taxa (species and groups of species) have been recorded from the River Test. The main groups represented are Oligochaete worms, Crustaceans (e.g. the very abundant shrimp Gammarus pulex), Diptera (flies) and Neuroptera (alderflies, lacewings etc.) Ephemeroptera (mayflies) are a major element of the fauna with 21 species from 6 families represented. Reflecting the highly productive nature of the chalk stream environment, maximum numbers of individuals of each of the main groups may reach 4,000/m2. The river is exceptionally rich in aquatic molluscs, with two national rarities (RDB* see note) reported: Valvata macrostoma (RDB2) and Pisidium tenuilineatum (RDB3). The former is a calcicole snail living in still or slow-flowing water amongst abundant submerged vegetation, the latter is a pea mussel also preferring still or slow flowing conditions. Some of the molluscs reported have localised distributions e.g. Gyraulus albus, Valvata piscinalis, Bithynia leachi and Theodoxus fluviatilis. Also worthy of mention amongst the exceptionally rich aquatic fauna are: Niphargus aquilex, an eyeless freshwater shrimp of calcareous springs, Atrichops crassipes a Rhagionid fly (RDB3), the nationally scarce riffle beetles Oulimnius troglodytes, Riolus cupreus and R. subviolaceus, and the alderfly Sialis nigripes which is mainly northern and western in distribution. Many of the nineteen species of mayfly noted are local in distribution. Other species of nationally scarce status reported from the Test include two species of caddisfly Ylodes conspersus and Metalype fragilis, and the mayfly Baetis atrebatinus. Recording of invertebrates has been carried out on the Leckford Abbas Estate section of the valley since the 1940s at the instigation of its owner John Spedan Lewis. A unique record of the invertebrate fauna of the riparian habitats has been built up, and though some records are now dated, it is considered a particularly important site for three groups of fly and the Aculeate family of Hymenoptera (bees and wasps). The records reveal an exceptional species richness which includes numerous nationally rare and scarce species, and demonstrate the quality, as invertebrate habitat, that the riverÕs riparian vegetation can have.

Amongst the nationally rare species recorded from areas of the riverÕs riparian vegetation is Cosmetopus dentimanus (RDB1) a ÒdungÓ fly that has only been recorded twice in Britain, at Leckford and in the Lower Itchen Valley. Others include the Empidid flies Syneches muscarius (pRDB2) and Platypalpus infectus (pRDB3) and Endothenia pullana (pRDB3) a moth whose larvae feed on the roots of marsh woundwort. Two very rare craneflies are reported from shaded river margins: Gonamyia abbreviata (pRDB3) and Arctoconopa melanpodia (pRDB2). The Southern damselfly Coenagrion mercuriale (RDB3) and DesmoulinÕs whorl snail Vertigo moulinsiana (RDB3), both species considered of European Interest and listed in Annex II of the ÔHabitats and SpeciesÕ Directive, have also been recorded in the site. There are many records of nationally scarce invertebrates from the valley, and examples of those species dependent on the wetland vegetation include the reed dagger Simyra albovenosa a moth whose larvae feed on common reed, the scarlet tiger Callimorpha dominula a moth whose larvae feed on comfrey, Tetanocera phyllophora a snail-killing fly and Chrysolina menthastri a bright green beetle which lives on water mint.

Birds The Test and its adjoining vegetation provides valuable habitat for wetland birds. The diverse range of characteristic riverine species breeding in the site includes almost ubiquitous kingfisher Alcedo atthis, grey wagtail Motacilla cinerea and little grebe Tachybaptus ruficollis. In the dense vegetation along its margins coot Fulica atra and moorhen Gallinula chloropus are frequent and tufted duck Aythya fuligula, pochard A. ferina and mute swan Cygnus olor also nest. Sedge warbler Acrocephalus schoenobaenus and reed warbler A. scirpaceus can be numerous in the tall vegetation with scattered scrub along the water courses. In this same habitat, the formerly rare CettiÕs warbler Cettia cettia is becoming widely established whilst the grasshopper warbler Locustella naevia has become quite scarce. Water rail Rallus aquaticus, though seldom seen, also breed in the dense wetland vegetation. Numbers of wet-grassland breeding birds of the Test Valley such as snipe Gallinago gallinago, redshank Tringa totanus and lapwing Vanellus vanellus, have all undergone declines in recent years, and few breed in the site. Passage species using the riverÕs margins include common sandpiper Actitis hypoleucos and green sandpiper Tringa ochropus. Kingfisher and grey heron Ardea cinerea are the riverÕs resident and most commonly seen fish-eating birds, although cormorant Phalacrocorax carbo now increasingly also range along the river and bittern Botaurus stellaris uses the river margins and reed beds on passage and in winter.

Fish The Test has developed a very important recreational game fishery. Almost the entire river is managed to maintain and facilitate fishing for trout (brown and rainbow), with fishing for sea trout, salmon and coarse fish also taking place along the lower reaches. Pike and other coarse fish are regularly removed along most of the river, but still maintain a presence. In its range of species, the fish fauna of the Test is typical of lowland chalk-rivers, though the community has been modified by introductions of rainbow trout, grayling and hatchery-reared brown trout, and the removal of other species. In the uppermost reaches of the Test system native populations of brown trout Salmo trutta are believed to persist, and strong populations of bullhead Cottius gobbo and brook lamprey Lampetra planeri are notable elements of the natural fish fauna. The riverÕs runs of salmon Salmo salar fluctuate markedly. Numbers have shown a steady decline since the 1960s which has increased sharply since the late 1980s, giving cause for much concern and efforts to remedy the situation. The increased silt loads in the river and their deposition on the river bed gravels, adversely affecting salmonid spawning, is one of the causes implicated in this decline. The remedial measures undertaken have entailed river channel habitat management and the release of large numbers of young hatchery-reared fish.

Mammals Otters Lutra lutra have been reported from certain parts of the site, but the Test no longer has an established population. Water voles Arvicola terrestris are common in places, but their numbers are thought to have declined as has been noted elsewhere in Britain.

Notes: * Red Data Book (RDB) identifies the status of BritainÕs rarest invertebrate species: RDB1 = Endangered; RDB2 = Vulnerable; RDB3 = Rare; (pRDB = proposed status).

Appendix 2.1.2 Lower Test Valley SSSI Citation

Atkins

File ref:

County: Hampshire Site Name: Lower Test Valley SSSI

Local Planning Authority: Hampshire County Council, New Forest District Council, Southampton City Council, Test Valley Borough Council

National Grid Reference: SU 360153

Ordnance Survey Sheet 1:50,000: 185/196 1:25,000: SU 31

Area: 138.7 (ha) 342.7 (ac)

Date Notified (Under 1949 Act): 1971 Date of Last Revision: –

Date Notified (Under 1981 Act): 2.12.1986 Date of Last Revision: –

Other Information:

110 ha is leased as a nature reserve by the Hampshire and Isle of Wight Naturalists’ Trust.

Reasons for Notification:

The site comprises the upper estuary of the River Test and exhibits a gradation from salt through brackish to freshwater conditions. It consists of one of the most extensive reed Phragmites beds on the south coast with flanking unimproved meadowland intersected by numerous tidal creeks: and flooded on high water spring tides.

The brackish grassland in the south of the site supports a varied flora with several species characteristic of salt marsh habitat, for example, sea arrow-grass Triglochin maritima, sea aster Aster tripolium and sea milkwort Glaux maritima. The rare bulbous foxtail Alopecurus bulbosus, a species whose distribution is rapidly contracting, occurs here, together with other uncommon species such as brookweed Samolus valerandi, spike-rush Eleocharis uniglumis, and the hybrid saltmarsh grass Puccinellia 3 Krusemaniana.

Above the termination of tidal influence are extensive unimproved neutral meadows containing a colourful and species-rich flora dominated by grasses such as Yorkshire fog Holcus lanatus, sweet vernal-grass Anthoxanthum odoratum, ryegrass Lolium perenne and meadow fescue Festuca pratensis, with abundant sedges Carex species and rushes Juncus species. Several plants now rather uncommonly found owing to modern intensive agricultural methods are common here, including ragged robin Lychnis flos-cuculi, water avens Geum rivale, lesser valerian Valeriana dioica, green-winged orchid Orchis morio, flowering rush Butomus umbellatus, water whorl-grass Catabrose aquatica and large bitter-cress Cardamine amara. Groups of willows Salix species are widespread along drains and creeks and support notable populations of the nationally rare green-flowered helleborine Epipactis Phyllanthes var Vectensis in an atypical habitat, together with numbers of common helleborine Epipactis helleborine. Over 450 species of flowering plants have been recorded for the site as a whole.

The site is also important for wetland breeding birds and as a wader and duck feeding and roosting ground. The reed beds support large breeding populations of reed warblers Acrocephalus scripaceus and sedge warblers A. schoenobaenus. They also function as an autumn roost site for swallows and martins, and a pre- migratory feeding site for various passerine birds, notably reed and sedge warblers. Status: Site of Special Scientific Interest (SSSI) notified under Section 28 of the Wildlife and Countryside Act, 1981

Appendix 2.3.1 River Corridor Survey, 1991, Sections 169-171

Atkins

Lower Test NEP Volume 3: Appendices

Appendix 3.5.1 Frequency of extreme low flows in the Great Test

Introduction As the NEP investigation of the Testwood abstraction has progressed, the focus has increasingly been placed on the potential impacts of abstracting at the full licensed quantity in periods of extreme low flow. To assess the risks associated with the licence, it is therefore important to understand the frequency and duration of such low-flow events.

Frequency and Duration of Extreme low flows In order to set the context for the NEP investigation and assess periods of extremely low flows, attempts have been made to look further back in time beyond the start of observed data from the Broadlands GS in 1957. Due to the lack of hydrometric data from the River Test before 1957, output from the Test and Itchen Groundwater Model has been used, as the simulated flow time series extends from 1920. This assessment has been undertaken in order to understand the frequency and duration of extreme low flow events, and also the potential impacts of full licensed abstraction during such events.

A long term flow time series has been derived based on two datasets:  Modelled flows at Broadlands GS (1920 to 1957) using flow outputs from the EA’s Test & Itchen Groundwater Model; and  Measured flows at Broadlands GS (1958 to 2011)

The flows at Broadlands were then transformed into estimated flows at the MRF location on the Great Test using a very simple methodology based on the difference in flows between the two locations on 11th September 2006 (the day of lowest flow at the MRF point in 2006 using the agreed river flow time series). This difference was about 163 Ml/d (Broadlands 362 Ml/d and calculated MRF 199 Ml/d) and the rate of abstraction at Testwood was 64 Ml/d. The outputs are shown in Figure 1.

The difference of 163 Ml/d was then applied to both records at Broadlands GS to provide a long-term estimate of flows under the abstraction and flow management conditions operating in the river in early September 2006. The data were then adjusted on the assumption that 136 Ml/d rather than 64 Ml/d was abstracted at the Testwood PWS works. The outputs are shown in Figure 2.

It is relevant to note that at this time, the split in flow between the Great and Little Test was not in accordance of the Coleridge Award (see Section 3.7.5) of 66% and 33% respectively, but was approximately 56% and 44%. This means that flows in the Great Test were lower than they should have been under rules of this historic river management agreement.

Given the complexity of the flow management issues downstream of Broadlands, and the essentially “high level” nature of the issue being considered in this note, there would appear to be little benefit of using a more complex approach than that described above.

Points to note Some key points from Figure 1 and 2 include the following observations:  This high level assessment has shown that the hydrological conditions of 2006 provides a useful benchmark when looking at long-term flow record because not only were flows very low (only 3 or 4 years in the last 87 would flows have been lower than 2006), but key data exist (salmon movement, water temperature etc) for 2006 and it is relatively fresh in recent memory.  The area between the green and orange dashed lines on Figures 1 and 2 marks the range of flows below those recorded in 2006 (2.2 m3/s) as low as the MRF (1.05 m3/s). There are no historic observations of salmon movement (although there are some fish catch data) or water temperature data in this flow zone at the MRF point. This zone therefore defines the range of flows in which an evaluation of risk will need to be undertaken in the absence of monitoring data. The gap can be filled in part by the use of the models developed for the NEP investigation but some judgement will still be required.  A rough estimate based on the data in Figure 1 suggests that with abstraction at historic levels, flows such as those in 2006 would occur once in every 15 to 20 years. With continuous abstraction at 136 Ml/d, Figure 2 shows that the recent “2006 flows” could occur every 4-5 years.  Although this approach is highly simplified, it is possible that under the conditions of 1921 and 1976, with fully licensed abstraction, and with the Coleridge Award flow split operating at 56% and 44%, flows would be close to Lower Test NEP Volume 3: Appendices

or below the MRF. This emphasises the importance of understanding the flow and temperature regime under these conditions, and whether the MRF is set at an acceptable level.

It is important to emphasise that the purpose of this analysis to provide some clear reference points for the NEP assessment, not to assess whether the frequency and extent of these low flow events is acceptable or not. In this context,Potential flows however, at the MRF point the based results on historic are abstraction considered (64 Ml/d) (Estimates to constitute based on 2006 a data reasonable at Broadlands GS and worst the MRF case point) given that, in reality, the abstraction would be highly unlikely to operate continuously at 136 Ml/d in the manner shown. 8

7 600

6 500

5 400

4

300

3

Flow atFlow MRF Point(m³/s) Flow atFlow MRF Point(Ml/d) 200 2

100 1

0 0

1920 1922 1932 1934 1944 1946 1956 1958 1960 1968 1970 1972 1982 1984 1994 1996 2006 2008 1924 1926 1928 1930 1936 1938 1940 1942 1948 1950 1952 1954 1962 1964 1966 1974 1976 1978 1980 1986 1988 1990 1992 1998 2000 2002 2004 2010 2012 MRF Lowest flow at the MRF location in 2006, historic abstraction (64 Ml/d) Using T&I Model Data at Broadlands GS (1920-1957) Broadlands GS Data (1958 to 2011)

Figure 1 Potential flows at the MRF location based on historic abstraction, 1920 to 2011 Based on thePotential abstraction flows at the MRF and point flow based onmanagement continual fully licensed conditions abstraction (136 Ml/d)operating (Estimates basedin the on 2006 river data at in Broadlands early GS September and the MRF point) 2006

8

7 600

6 500

5 400

4

300

3 Flow at MRF Point (m³/s) Point MRF at Flow

200 (Ml/d) Point MRF at Flow 2

100 1

0 0

1920 1926 1930 1934 1940 1944 1954 1958 1968 1972 1978 1982 1986 1992 1996 2000 2006 2010 1922 1924 1928 1932 1936 1938 1942 1946 1948 1950 1952 1956 1960 1962 1964 1966 1970 1974 1976 1980 1984 1988 1990 1994 1998 2002 2004 2008 2012 MRF Lowest flow at the MRF location in 2006, historic abstraction (64 Ml/d) Using T&I Model Data at Broadlands GS (1920-1957) Broadlands GS Data (1958 to 2011) Figure 2 Potential flows at the MRF location based on fully licensed abstraction, 1920 to 2011 Based on the abstraction and flow management conditions operating in the river in early September 2006 Lower Test NEP Volume 3: Appendices

Appendix 3.5.2 Freshwater Flows

Table of Contents

1. Introduction 2 1.1. Background 2 1.2. Previous studies 2 1.3. Objective 2

2. Flow gauging locations 2 2.1. Relevant flow gauging datasets 2 2.2. Site A: Main Test downstream of Longbridge GS and Upstream of the Little and Great Test flow split 4 2.2.1. Broadlands GS 4 2.2.2. Test Back Carrier GS 4 2.2.3. Calculation of flow at Site A 4 2.3. Site B: Great Test downstream of the Little and Great Test flow split and Upstream of the Nursling Fish Farm offtake 5 2.3.1. Calculation of flow at Site B 5 2.4. Site C: Great Test downstream of the Nursling Fish Farm Carrier and upstream of the Testwood Abstraction 5 2.4.1. Location of returning water 6 2.4.2. Abstraction Volume 6 2.4.3. Calculation of flow at Site C 7 2.5. Site D: Great Test downstream of the Testwood Abstraction and upstream of the Blackwater confluence 7 2.5.1. Calculation of flow at Site D 7 2.6. Site E: Great Test just downstream of the Blackwater confluence 7 2.6.1. Testwood GS 7 2.6.2. Ower GS 7 2.6.3. Broadlands Fish Farm Carrier 8 2.6.4. Discussion Point 2 10 2.6.5. Generation of flow at Site E 10 2.7. Site F: Great Test at the MRF location 11 2.7.1. Discussion Point 3 11 2.7.2. Generation of flow at Site F 11 2.8. Site G: Great Test downstream of the MRF and just upstream of Testwood Pool 11 2.8.1. Calculation of flow at Site G 12 2.9. Site H: The Little Test downstream of the Little and Great Test flow split 12 2.9.1. Calculation of flow at Site H 12

3. Summary 12 3.1. Summary for Sites A to H 12 3.2. Consideration of timescale of different time series 12 3.3. Conclusion 12

1 Lower Test NEP Volume 3: Appendices

1. Introduction

1.1. Background

Southern Water Services (SWS) operate a public water supply (PWS) surface water abstraction on the Lower River Test at Testwood. The Lower Test is currently subject to an NEP investigation, examining the effects of the Testwood PWS abstraction upon the hydrology and ecology of the river. There are numerous locations where river flow is measured within the study area. Different flow gauging techniques are used and flow time series are available for different time periods. During a Steering Group meeting of the Lower Test NEP project, the Environment Agency raised concerns about the flow values at some of the gauging stations used to compile the flow time series used to date for the NEP investigation. It was therefore decided that a Lower Test flow time series should be compiled by Atkins for discussion between SWS and the Environment Agency, in order to agree the flow time series used in the investigation. This Appendix summarises the key characteristics of the different gauges on the Lower Test and the selection of gauges from which the approved flow time series is generated.

1.2. Previous studies

This Appendix builds upon previous studies on the hydrology of the area notably:  The Hydrology and hydrogeology chapter of the Environment Agency (2009) Lower Test Project Phase 1 Baseline Data Report; and

 Environment Agency (2009) Lower Test Project Phase 1 Flow Diversion Scoping Report.

1.3. Objective

The objective of this exercise is to produce a flow time series for key points along the River Test of interest to the NEP study. The key locations of interest for the Lower Test NEP study, for which a flow time series is needed, are shown in Table 1. It is not the scope of this Appendix to detail all the available flow gauging datasets available, and repeat the work and discussions already documented: the reader is referred to the previous studies for a thorough understanding of the hydrometric network and data for the Lower River Test. This Appendix builds upon these reports so that only the justification of the chosen gauges is presented.

Table 1 The key locations of interest for the Lower Test NEP study No. Site

A Main Test downstream of Longbridge GS and upstream of the Little and Great Test flow split

B Great Test downstream of the Little and Great Test flow split and upstream of the Nursling Fish Farm offtake

C Great Test downstream of the Nursling Fish Farm Offtake and upstream of the Testwood Abstraction

D Great Test downstream of the Testwood Abstraction and upstream of the Blackwater confluence

E Great Test just downstream of the Blackwater confluence F Great Test at the MRF location G Great Test downstream of the MRF and upstream of Testwood Pool H The Little Test downstream of the Little and Great Test flow split

2. Flow gauging locations

2.1. Relevant flow gauging datasets

Figure 1 shows the flow gauging locations in the Lower Test, and Table 2 gives a summary of the hydrometric network operated by the Environment Agency.

2 Lower Test NEP Volume 3: Appendices

© Crown Copyright. All rights reserved. Environment Agency, 10002638, 2011 1 To (! © CEH. Some features of this map Ower GS are based on digital spatial data licensed from the Centre for

(!2

r e i r r

a r

C e

i

3 r

(! m r r a

a C

F

k

h c s

i a R F B iv e s t r

s d t

B e n s

T

a e

l l a

d T c

a r k o e w r v a i t B er R Site A 4 6 7 (! (! (! (!8 5 (! (!9

Site B (!(!11 Site H t Nursling Fish Farm Offtake es t T Litt rea le T G est R iv e r B la c Site C k w 12 Site D a (! te r (!13 Site E & F Site G Scale 1:20,600 0 0.25 0.5 1 Kilometers 1:20,600 ± Lower_Test_Gauging_Stations (! 7, M27 TV2 rivers_50k (! 1, Broadlands (! 8, M27 TV3 WLMP_Lower (! 2, Longbridge (! 9, M27 Nursling Lower_Test_SSSI (! 3, Broadlands Fish Flow 10, Test Back Carrier Site of Interest: (! Site "x" Ref Atkins NEP Tech Note (! 4, M27 Blackwater (! 11, Conagar Bridge (! 5, M27 Test (! 12, Testwood (! 6, M27 TV1 (! 13, Testwood Bridge (ADCP & Spot Flow)

Figure 1 The Lower Test Hydrometric network operated by the Environment Agency Produced by the Environment Agency and based on from Figure 7 in Environment Agency (2009) Lower Test Project Phase 1 Baseline Data Report. NB The above figure shows that the channel monitored by Gauge no 9 flows into the Main Test just upstream of the boundary marked by the “wlmp_units_lower” polygon, i.e. upstream of the Little and Great Test Flow Split. Atkins believes that this is incorrect, and that the channel flows toward Gauge no 10. Also note that the label of Gauge 11 obscures Gauge 10 Test Back Carrier.

3 Lower Test NEP Volume 3: Appendices

Table 2 Summary of hydrometric network Reference Monitoring Site River NGR Site Type Data record on Figure 1 Timsbury Bridge Not shown Test SU 35164 23404 Ultra Sonic 13/02/2004 to current Romsey (Tadburn Not shown Tadburn Lake SU 36233 21212 Open-Channel 03/11/1977 to current Lake) Ower * Not shown Blackwater SU 36233 21212 Weir 01/10/1976 to current Broadlands * 1 Test SU 35406 18879 Chart Recorder 01/10/1957 to current Longbridge 2 Test SU 35491 17791 ElectroMag 01/10/1981 to current Broadlands Fish Spot flow 3 Broadlands Fish Carrier SU 35092 17288 06/01/1998 to current Farm gaugings only M27 Blackwater 4 Blackwater SU 34639 16380 Ultra Sonic 03/02/2004 to current M27 Main Test 5 Test SU 35458 16344 Ultra Sonic 02/02/2004 to current M27 TV1 * 6 Broadlands Fish Carrier SU 34816 16388 Ultra Sonic 03/02/2004 to current Nursling Fish Farm Not shown Nursling Off-Take SU 35168 15761 Gauging Site 1983 to 1997 Test Back Carrier * 10 Test Back Carrier SU 35535 15904 Chart Recorder 10/01/1986 to current Conagar Bridge * 11 Little Test SU 35512 15898 ElectroMag 01/01/1982 to current Testwood EM * 12 Great Test SU 35396 15268 ElectroMag 11/05/1987 to current Testwood Bridge 13 Great Test+ Blackwater SU 35610 15174 ADCP 02/02/2004 to current Source: Table 4 in Environment Agency (2009) Lower Test Project Phase 1 Baseline Data Report Where * denotes a gauging site used to derive a flow time series as detailed in this Appendix

2.2. Site A: Main Test downstream of Longbridge GS and Upstream of the Little and Great Test flow split It is proposed that the flow data from the Broadlands Gauging Station (Gauge no 1 on Figure 1) is used to represent flow in the River Test at this location, plus that from the Test Back Carrier.

2.2.1. Broadlands GS The Broadlands GS records stage; flow is estimated by the Environment Agency using the “RIVTEST” program. Flow data are available from 1957. While the Longbridge GS (Gauge no 2) and the M27 Main Test gauge (Gauge no 5) are installed further downstream, in the Baseline Data Report, the Environment Agency states that, “...the most useful data to use to represent the flow going into the Lower Test catchment is that of the Broadlands GS.” This is because over time, the data from the Longbridge GS has become unreliable, and the M27 Main Test site has suffered instrument problems.

2.2.2. Test Back Carrier GS The Test Back Carrier Gauging station is located on the channel called the “Test Back Carrier” which exits the Main Test between Broadlands & Longbridge, and also receives some inputs from Longbridge Lake. The carrier confluences with the Little Test just downstream of Conagar Bridge GS. The Baseline Data report states that discharge from the Test Back Carrier is generally low, except during storm periods. The current channel at its splits with the Main Test is not maintained and is over-grown and silted up since about 2004. Therefore, in the summer the carrier is likely to get water only from the Longbridge Lake, if at all (it has been known to dry up) and in winter it is fed by the Lake and also from Test floodwaters when levels are elevated in the Test. This trend can be seen when looking Figure 2 which shows flows in the Test Back Carrier, which is very flashy and strongly follows the pattern shown by the Broadlands GS, indicating the link between the two channels.

2.2.3. Calculation of flow at Site A Therefore, in order to account for the water exiting the Main Test and flowing down the Carrier, the discharge recorded at Test Back Carrier GS should be subtracted from the Broadlands GS. (In order to account for the water joining the Little Test, the discharge recorded at Test Back Carrier GS should be added to that of Conagar Bridge gauging station to give the flow of the Little Test downstream of its confluence, see section 2.9).

Flow at Site A = Flow at Broadlands GS - Flow at Test Back Carrier

4 Lower Test NEP Volume 3: Appendices

300 3000

250 2500

200 2000

150 1500

100 1000

Broadlands GS flow Ml/d flow GS Broadlands Test Back Carrier Flow Ml/d Flow Carrier Back Test

50 500

0 0

Jan 88 Jan 91 Jan 94 Jan 95 Jan 97 Jan 98 Jan 00 Jan 01 Jan 04 Jan 07 Jan 10 Jan Jan 86 Jan 87 Jan 89 Jan 90 Jan 92 Jan 93 Jan 96 Jan 99 Jan 02 Jan 03 Jan 05 Jan 06 Jan 08 Jan 09 Jan 11 Jan

Test Back Carrier Broadlands GS

Figure 2 The flow pattern of the Test Back Carrier and Broadlands gauging stations

2.3. Site B: Great Test downstream of the Little and Great Test flow split and Upstream of the Nursling Fish Farm offtake It is proposed that the flow at this location in the river can be represented by subtracting the flow measured by Conagar Bridge on the Little Test arm of the river, Gauge no 10 on Figure 1, from the flow computed from Site A. The description of Conagar Bridge gauges from the Environment Agency’s Baseline Data Report is that it has “a good and long flow record from 1982” and that the gauge “calibrates well against its spot flow gaugings”.

2.3.1. Calculation of flow at Site B The flow at Site B can be calculated as follows: Flow at Site B = Flow at Site A - Flow at Conagar Bridge

Flow at Site B = (Flow at Broadlands GS - Flow at Test Back Carrier) - Flow at Conagar Bridge

2.4. Site C: Great Test downstream of the Nursling Fish Farm Carrier and upstream of the Testwood Abstraction The Nursling Fish Farm Offtake, or Carrier, diverts water from the Great Test approximately 1 km above Testwood (GS and abstraction location) into the Fish Farm complex. The abstraction licence has a maximum daily licence of 45.5 Ml/d (since 1990), but with no requirement to measure discharges. There are two issues to resolve for the Nursling Fish Farm abstraction:  The location of the returning water; and  The volume of water abstracted from the Main Test over time. These issues are considered in turn.

5 Lower Test NEP Volume 3: Appendices

2.4.1. Location of returning water This fish farm carrier diverts water from the Great Test upstream of Testwood (GS and abstraction) into the Fish Farm. The returning water from Nursling Fish Farm can re-enter the Great Test in several different places: either directly downstream of the Fish Farm back into the River Test upstream of the Testwood abstraction, or via a more indirect route entering the River Test downstream of the Blackwater confluence. Given the size of the abstraction licence, the volume of water diverted could potentially significantly affect downstream river flows. The Environment Agency believes that, in the past, the outflow from Nursling Fish Farm re-entered the main river channel just downstream of Testwood Bridge and was far more significant than it is now. It is believed that in recent years the location of the returning water has changed, although exactly when this occurred is unknown. Given this uncertainty, it has been agreed that the Nursling Fish Farm abstraction will be considered as wholly returning to the Great Test downstream of the Blackwater confluence. This will represent a precautionary position with regard to river volume upstream of the Testwood Abstraction for the NEP investigation.

2.4.2. Abstraction Volume The Nursling Fish Farm Abstraction volume is of particular importance due to its inclusion in the calculation of the Minimum Residual Flow (see Section 2.7). As stated in the Baseline Data Report, there is a gauging station located at the inlet to Nursling Fish Farm GS (SU 35168 15761) which recorded flows from 1983. Spot flow gaugings also exist for this time period. While it is acknowledged that it is important to have good data at this site, the gauged and spot flows records do not calibrate well against each other, and no data have been recorded since 1991. Notwithstanding the above, the gauged flow data from 1983 to 1991 indicate flows in the range of 50–100 Ml/d. The Environment Agency undertook spot flow gauging from September 2004 which showed “nearly immeasurable flow” was being discharged from Nursling Fish Farm, however the abstraction returns from the licence holder did show that some water was being diverted under the conditions of the abstraction licence. Abstraction return data from 2001 to 2008 suggest an average flow of 11 Ml/d and from 2009 suggest 5Ml/d. The reduction in water diverted under the terms of the abstraction licence reflects the change in use as the site is no longer managed as a fully operational fish farm. The uncertainty over the abstracted volume over time, and the location where the abstracted water is returned to the channel, leads to uncertainty in the flow volume upstream of the intake for the Testwood abstraction. Monthly abstraction returns are available for use, however, the method of how the abstraction returns are calculated by the license holder, and their accuracy are unknown. Given this uncertainty, other studies have modelled the Nursling Fish Farm abstraction volume as a fixed quantity. The decision of what volume to use to present Nursling Fish Farm Abstraction is explored below in Discussion Point 1.

2.4.2.1. Discussion Point 1 To support the NEP investigations, the Environment Agency has contacted the license holder, who has agreed that its monthly abstraction data can be used in the NEP investigations. Therefore the following options are available for the Lower Test NEP:  Option 1 - apply a constant value of 11 Ml/d to represent the Nursling Fish Farm Abstraction, following the same approach used in the Test and Itchen Groundwater Model, and hence CAMS;  Option 2 - apply a value of 45 Ml/d to represent the Nursling Fish Farm Abstraction before the year 2000 and 0 Ml/d thereafter;  Option 3 - apply a value of 45 Ml/d to represent the Nursling Fish Farm Abstraction before the year 2000 and 11 Ml/d thereafter which uses the best known previous estimates of the abstracted volume;  Option 4 - apply a value of 45 Ml/d to represent the Nursling Fish Farm Abstraction before the year 2000 and 11 Ml/d between 2000-2009, and 5 Ml/d thereafter which uses the current best known estimates of the abstracted volume; and  Option 5 – apply the abstraction return values submitted by the license holder. In conclusion, it has been decided to use the monthly abstraction returns data for the NEP investigations. However, the period of time over which the data are available is only January 2001 to March 2011. Therefore in order to use the returns data, but extend the time series, Options 4 and 5 will be adopted, i.e. use the monthly data when available, but use the recommended constant values before and after this dataset begins and ends, in order to extend the period over which the time series is available.

6 Lower Test NEP Volume 3: Appendices

2.4.3. Calculation of flow at Site C In order to present a precautionary position with regard to river volume upstream of the Testwood Abstraction for the NEP the Nursling abstraction will be considered as returning to the Great Test downstream of the Blackwater confluence.

Therefore, the flow for Site C can be derived by subtracting the volume diverted under the Nursling Fish Farm abstraction licence from the previous step i.e.

Flow at Site C = Flow at Site B - volume diverted at Nursling Fish Farm

Flow at Site C = (Flow at Broadlands – Flow at Test Back Carrier – Flow at Conagar Bridge) – volume diverted at Nursling Fish Farm

2.5. Site D: Great Test downstream of the Testwood Abstraction and upstream of the Blackwater confluence It is proposed that the flow data from the Testwood Gauging Station (Gauge no12 on Figure 1) is used to represent flow in the River Test at this location. The Baseline Data Report comments that the Testwood GS flow data does not display any indication of major problems at the site, and that there is consistency between the spot flow gaugings and the gauged flow across the full flow range.

2.5.1. Calculation of flow at Site D The flow at Site D can be calculated as follows: Flow at Site D = Flow at Testwood Gauging Station

2.6. Site E: Great Test just downstream of the Blackwater confluence There is no gauging station which measured the flow of the Great Test downstream of the River Blackwater. To estimate flow at this location a combination of flow data sets must be used. The flow at the location is made up of the flow at Site D, plus the flow of the Blackwater. The flow of the Blackwater at this location includes flow from the Broadlands Fish Farm Carrier. It is proposed that the flow data from a number of sources is used to represent flow in the River Test at this location as follows: Testwood GS, Ower GS and the Broadlands Fish Farm Carrier. Each of these sources is discussed below.

2.6.1. Testwood GS As stated above in 2.1.4, the data from Testwood GS is considered to be reliable and accurate so its data will be used.

2.6.2. Ower GS The Blackwater confluences with the Great Test, downstream of Testwood GS. There are two gauging stations with available flow data on the Blackwater: Ower GS and M27 Blackwater GS. The Ower GS is located upstream of the M27 Blackwater before the confluence of the Cadnam River with the Blackwater (it is not shown on the Figure 1 and Table 1 as it is outside the Lower Test area). The M27 Blackwater GS sits on the Blackwater at its entrance to the Lower Test Area (Gauge no 4 on Figure 1). The Environment Agency’s Baseline Data Report considers that data from the Ower GS has a reasonably good quality, with a long flow record (data are available from 1976). The M27 Blackwater GS was installed in 2003 to measure flood flows beneath the M27. It was hoped that the station would provide a full range of flow but the site physically cannot represent mid and low flows and only captures peak flows to any reasonable degree. Therefore, the Ower GS is the only reliable source of flow data for the Blackwater. In order to ensure the Ower GS is representative of the Blackwater at the location of the M27 Blackwater site, the flow recorded at the GS is factorised to account for accretion along its length.

7 Lower Test NEP Volume 3: Appendices

The factoring approach involves determining the discharge per km2 for Ower GS flow at the location of the M27 gauge by accounting for the increase in catchment area from 104 km2 to 146 km2 respectively. The EA Flow diversion scoping report shows that when factorised to account for the increase in catchment area, the Ower GS flow matches well to spot flow gaugings carried out at the M27 Blackwater location. There is some over-estimation of peak flows using the process, but as it is the low flows that are of most concern to this investigation then the Environment Agency recommend that this approach is followed. The Flow Diversion Scoping Report also states that there is very little gain in flow between the M27 gauge and the Blackwater confluence apart from that from the Broadlands Fish Farm Carrier (FFC). Therefore, flow for the Broadlands FFC must be added to the Ower GS factorised flows to estimate flows at the confluence of the Blackwater and the Great Test.

2.6.3. Broadlands Fish Farm Carrier The volume derived for the Broadlands FFC is of particular importance due to its inclusion in the calculation of the Minimum Residual Flow (see Section 2.7). The Broadlands Fish Farm Carrier (FFC) splits from the main Test just upstream of Broadlands GS and diverts water to the River Blackwater downstream of the M27 Blackwater GS (see Figure 1). The carrier was initially created to feed the Broadlands water meadow system, but more recently part of the Broadlands Fishery has been used for angling. Currently there is no functioning operational control over the quantity of water that is diverted down the FFC. The Environment Agency spot flows on the Broadlands FFC at approximately 1.5 km away from Broadlands GS since 1998; there is no long term continuous flow record. Continuous data are available from the M27 TV1 gauge which is also located on the FFC, approximately 0.9 km downstream of the FFC spot flow site. The Baseline Data Report advises against using the M27 TV1 in place of the Broadlands FFC spot flow measurements as the former is not considered viable due to instrument problems. However, further work since the Flow Diversion Option Report (and CAMS modelling) by the Environment Agency on the Broadlands FFC has corrected the M27 TV1 flow data and the Agency state that the data from the M27 TV1 can be used in the Lower Test NEP.

2.6.3.1. Atkins’ approach to generating a continuous dataset The Test and Itchen Groundwater Model represents the Broadlands FFC as a constant diversion of 40 Ml/d. Flows are however known to be more varied: spot flow data from the Broadlands FFC over 1998 to 2008 indicate average flows of 72 Ml/d and a Q95 of 46 Ml/d. Therefore it was considered that a more robust representation of the Broadlands FFC was required for the Lower Test NEP. In order to more robustly represent the Broadlands FFC for the Lower Test NEP, two relationships were derived between spot flow measurements along the Broadlands FFC and at Timsbury GS to try to determine a more robust representation of the Broadlands FFC. Timsbury GS was chosen as it is positioned upstream of the Broadlands FFC and so is unaffected by the offtake (unlike Broadlands GS). A linear relationship was determined based on comparing flows at the Broadlands FFC and Timsbury GS (referred to as the absolute flow method) while a curvilinear relationship was determined based on comparing the FFC flows as a percentage of Timsbury flow vs. flows at Timsbury (referred to as the percentage split method), see Figure 3. On average, the Broadlands FFC was found to take about 9.4% of flow from the main Test at Timsbury. Model runs were carried out to determine the fit of both relationships. Two runs were done based on the percentage split method, one with a cut off set at 5.4% representing the lower limit for which the relationship remained valid. Below 5.4% (equivalent to Timsbury flows greater than 1,800 Ml/d approx.), Broadlands FFC flows were found to decrease with increasing flows in the main channel of the River Test. The cut off was therefore implemented to ensure that Broadlands FFC flows consistently increased with increasing Timsbury flows. Flow hydrographs of modelled and observed flows along the Broadlands FFC are shown in Figure 4. It is clear that a constant flow of 40 Ml/d is an underestimation especially during high flow events. High flows were better simulated in Runs 16 and 17 although both runs consistently overestimated flows throughout much of the model period. It was decided that Run 16 (i.e. percentage split method with 5.4% cut off) provided the best representation of Broadlands FFC flows.

8 Lower Test NEP Volume 3: Appendices Correlation based on absolute flow method Correlation based on percentage split method 180 20 y = 0.054x + 36.558 y = -5.376ln(x) + 45.658 160 18 R² = 0.5601 R² = 0.5844 140 16 120 14 100 12 10 80

BFFC (Ml/d) BFFC 8 60 6 5.4% cut off 40 4 20 2 0 (%) BFFC at flow Proportion ofTimsbury 0 0 500 1000 1500 2000 0 500 1000 1500 2000 Timsbury (Ml/d) Timsbury (Ml/d)

Figure 3 Relationship between Broadlands Fish Farm Carrier and Timsbury GS flows based on the percentage split approach

250

200

150 Flow (Ml/d) 100

50

0 1995 1996 1996 1997 1998 1999 2000 2001 2002

Run 14 Run 15 Run 16 Run 17 BFFC actual Figure 4 Broadlands Fish Farm Carrier modelled vs. observed flows Key: Run 14 (constant diversion of 40 Ml/d); Run 15 (based on the percentage split method); Run 16 (based on the percentage split method with cut off set at 5.4%); and Run 17 (based on the absolute flow method)

2.6.3.2. Environment Agency’s approach to generating a continuous dataset As stated above, further work has been undertaken by the Environment Agency on deriving a continuous flow time series for the Broadlands FFC by the Environment Agency since the Baseline Data Report. Bethan Davies (pers comm.) states that the EA hydrometric archive WISKI was used to produce a synthesised flow time series for M27 TV1 by manipulating the Broadlands GS and calibrating and fitting resulting flows to the gaugings at M27 TV1, the Broadlands FFC site and the limited M27 TV1 daily mean flow. The flow was manipulated using the WISKI archive internal software functions of vertical stretching (expansion factors) and vertical shifting. The processes and values used are recorded in the data quality control audit trail within WISKI. The resulting time series can be seen in Figure 5 and the Environment Agency believe that it shows a good match between the synthesised series and the M27 TV1 actual spot gauged and daily mean flows. Therefore, it is reasonable to hindcast the series to the date of the first gaugings on the carrier at the BFF spot flow site using those gaugings to calibrate results.

9 Lower Test NEP Volume 3: Appendices

6 40

5 30

4 20

3 10

2 0 Broadlands GS Flow (m3/s) Flow Broadlands GS

1 -10 M27 TV1, TV1, (m3/s) FlowBroadlandsCarrier Fish FarmM27

0 -20

01 Jan 11 Jan 01

01 Jan 96 Jan 01 99 Jan 01 03 Jan 01 06 Jan 01 07 Jan 01 10 Jan 01 01 Jan 97 Jan 01 98 Jan 01 00 Jan 01 01 Jan 01 02 Jan 01 04 Jan 01 05 Jan 01 08 Jan 01 09 Jan 01 12 Jan 01

Synthesised flow for Broadlands FFC at M27 TV1 M27 TV1 Gaugings BFF Gaugings M27 TV1 Broadlands GS Figure 5 Hindcast Synthesised Flow time series for the Broadlands Fish Farm Carrier compared to Broadlands GS flow, M27 TV1 gauged flow, and spot flows from BFF spot flow site and M27 TV1

2.6.4. Discussion Point 2 Given the derivation of a continuous time series for the Broadlands FFC from the TV1 gauge, there needs to be a review over the data quality with this gauge, given that he Baseline Data Report (p28) stated the, “M27 TV1 site (Plate 2) is also on the Broadlands FF…… M27 TV1 could be used to replace the Broadlands FF … However…the site data is still not a viable source of flow data as… the equipment is still not calibrated and has intermittent problems.” Since the production of the Baseline Data Report, the Environment Agency has stated (Bethan Davies, pers. comm.) that while the Report showed the M27 TV1 data in high resolution as having a “spiky” variation in 15-min flow the variation in 15-min data is due to instrument calibration and it being fine tuned. However, when calculated into a daily mean time series the variation is smoothed out, so this time series has been adjusted to calibrate with the Agency’s monthly spot gaugings. The results show that the daily mean time series matches well with the gaugings and any deviation is well within tolerance margins. The spot gauging continues monthly in order to keep the calibration going. Therefore, the data are of good quality and usable. Therefore, the Environment Agency state the synthesised flow time series for the Broadlands Fish Farm Carrier should be used to represent the flow of this channel. Data set length The original synthesised dataset started in January 1998 as this is the period for which spot flow data are available. Upon request, the Environment Agency has hindcast the time series to start in January 1996 in order that the start of the time series matches the period for which fish count data are available for the study area. However, the Environment Agency believes that it would not be robust to hindcast the time series earlier than this date. This is because of the lack of spot flow readings against which to check the synthesized flows. Therefore the period of data available for the Broadlands Fish Flow Carrier is from January 1996.

2.6.5. Generation of flow at Site E The generation of a flow time series for the Site E is therefore: Flow at Site E = Flow at Site D + catchment factorised Ower GS flow + synthesised Flow of Broadlands FFC

Flow at Site E = (Flow at Testwood GS) + catchment factorised Ower GS flow factored up to the M27 Blackwater site + synthesised flow of Broadlands FFC

10 Lower Test NEP Volume 3: Appendices

2.7. Site F: Great Test at the MRF location The Testwood PWS abstraction has a Minimum Residual Flow (MRF) condition, located where the Great Test enters the Lower Test Valley SSSI: approximately 600 m downstream of Testwood Bridge, and approximately 500 m downstream of the confluence of the Blackwater. The computation of the MRF is stated on the licence as follows: “The authority must cease or reduce the rate of abstraction hereby authorised so as to not cause either the flow in the river Test d/s of the Testwood Pumping Station intake as measured at SU 359 150 or the aggregate of; the flow in the River Blackwater as measured at the Ower GS at SU 328 175, the flow in the river Test d/s of the Testwood Pumping Station intake at SU 354 153 and the flow in the Nursling Fish Farm carrier below Nursling Mill as measured at SU 351 158 to fall below 20 mgd (91,000 m3 per day1)”; The intention of this condition is to protect flows at the specified grid reference. As there is no gauging station at that point, the licence stated how flows could be calculated by adding together the flows at three gauging stations. Since the licence was issued, it has become apparent that the summation of those flows does not equal the flow at the specified grid reference as there is additional water in the river which is not measured by any of the three gauging stations listed. Therefore in enforcing the licence condition, the Environment Agency would always seek to take spot flow measurements at the specified grid reference in order to ensure a true measure of the flow at that point.

The Environment Agency has installed a simple flow measurement gauge at Testwood Bridge. However there is only a flow record at this point from 2007 and so an alternative means of calculating flow at this location is needed to assess historic flow patterns.

2.7.1. Discussion Point 3 Flows calculated at the MRF point are dependent upon:  The values chosen for the Nursling Fish Farm abstraction;  The location of where the Nursling Fish Farm abstraction is returned; and  The value chosen for the Broadlands FFC. Therefore due to the potential different approaches to derive the volume or time series for the Nursling and Broadlands Fish Farm carrier and offtake respectively, the value of the flow that is tested against the MRF licence condition could vary. Therefore in order to standardise the value of residual flow downstream of the intake for the Testwood abstraction which is compared against the MRF, it is recommended that the flow is computed as outlined in this report for Site E, plus the agreed Nursling Fish Farm abstracted volumes (see Section 2.4).

2.7.2. Generation of flow at Site F Following the above discussion, the flow at Site F is calculated as follows:

Flow at Site F = Flow at Site E + Nursling Fish Farm abstraction

Flow at Site F = (Flow at Testwood GS + catchment factorised Ower GS flow factored up to the M27 Blackwater site + flow of Broadlands FFC) + Nursling Fish Farm abstraction

2.8. Site G: Great Test downstream of the MRF and just upstream of Testwood Pool There is an offtake between the Blackwater confluence and Testwood Pool, termed the Middle Test. However, the channel is not free flowing, as there is control structure which prevents water movement. There is also a breach of water around the structure. Information about the structure has been collected as part of the NEP investigation in order for it to be represented within the hydraulic model that has been constructed. The hydraulic model will use the information about the structure to determine what flows would be able to pass through the structure. Therefore flows at Site G will be determined using the hydraulic model, and subtracted from the flow time series derived for Location F.

1 NB: 91,000 m3 per day is equivalent to 1.05 m3/s or 88.83 Ml/d

11 Lower Test NEP Volume 3: Appendices

2.8.1. Calculation of flow at Site G The flow at Site G can be calculated as follows: Flow at Site G = Flow at Site F - offtake to the Middle Test

Flow at Site G = (Flow at Testwood GS + catchment factorised Ower GS flow factored up to the M27 Blackwater site + flow of Broadlands FFC + Nursling Fish Farm abstraction) - offtake to the Middle Test

2.9. Site H: The Little Test downstream of the Little and Great Test flow split It is proposed that the flow data from Conagar Bridge Gauging Station (Gauge no 11 on Figure 1) and from the Test Back Carrier gauge (Gauge no 10 on Figure 1) is used to represent flow in the River Test at this location. See Section 2.3 for details of Conagar Bridge, and Section 2.2.2 for the Test Back Carrier.

2.9.1. Calculation of flow at Site H The flow at Site H can be calculated as follows: Flow at Site H = Flow at Conagar Bridge GS + Test Back Carrier GS

3. Summary

3.1. Summary for Sites A to H The preceding sections have explained the different flow time series available and the optimal approach to generating different time series for key points along the lower River Test. Table 3 presents a summary of the justification made for each location in deriving a flow time series. Figure 6a shows the flows calculated for all locations; figures 6b to 6e show the flows for a pair or small group of sites for ease of comparison. As to be expected, the charts show that Sites C and D have the lowest flows in the Great Test after the abstractions in that particular reach of the river. The confluence of the Backwater and return of the Nursling Fish Farm abstraction to the main channel raises flows so that by Site F, flows are at their highest for the section downstream of the Great and Little Test split.

3.2. Consideration of timescale of different time series Table 4 presents the different datasets used for the derivation of approved flow for the NEP investigation and their length of data. The NEP investigation needs to have data from all gauges in order to ensure a complete flow trim series for the locations along the reach under focus. Table 4 shows that the earliest date when information is available from all gauges or datasets is January 1996. Therefore the period which will be used for the NEP investigation will range from January 1996 to December 2011. This is an appropriate time period for use as it is lengthy (15 years) and covers a range of flow scenarios e.g. wet winter (e.g. December 2000), dry winter (e.g. December 2005), wet summer (e.g.2007) and dry summer (e.g. 2008).

3.3. Conclusion To conclude, the purpose of this Appendix has been to present the steps that were undertaken to produce an abstraction time series that best represents the volume of water taken from the River Test at Testwood.

The Steering Group has approved of the method and data recommended to derive the flows time series, and therefore it has been used in the Lower Test NEP Investigation.

12 Lower Test NEP Volume 3: Appendices

Table 3 Key Locations and flow gauging datasets Flow gauging dataset or No. Site calculation required Notes (Gauge No) Main Test downstream of Broadlands GS (No1) minus

A Longbridge GS and upstream of the Test Back Carrier (No10) the Little and Great Test flow split Great Test downstream of the Broadlands GS (No1) minus Little and Great Test flow split and

B the Test Back Carrier (No10) upstream of the Nursling Fish minus Conagar Bridge (No11) Farm offtake Conclusion of Discussion Point 1: The volume of the Nursling Fish Farm abstraction is Broadlands GS (No1) minus to be taken from monthly abstraction returns. Great Test downstream of the the Test Back Carrier (No10) Constant values will be used before Jan 2001 Nursling Fish Farm Offtake and

C minus Conagar Bridge (No11) (45 Ml/d) and after March 2011 (5 Ml/d) when upstream of the Testwood minus the volume abstracted monthly data are not available (Option 4 & 5). It is Abstraction at Nursling Fish Farm assumed that the abstracted water is returned to the Great Test downstream of the Blackwater confluence. Great Test downstream of the Testwood Abstraction and

D Testwood GS (No. 12) upstream of the Blackwater confluence Testwood GS (No 12) plus Conclusion of Discussion Point 2: Great Test just downstream of the factorised flow for Ower GS

E The EA’s flow time series of the Broadlands FFC is Blackwater confluence plus the volume Broadlands used Fish Farm Carrier (No.6) Conclusion of Discussion Point 3: The current method to compute the flow downstream As for Location E plus Nursling F Great Test at the MRF location of the intake for the Testwood abstraction and Fish Farm against which the MRF is compared should be used using the agreed values as outlined in this report As for Location F minus a flow Great Test downstream of the The flows moving down the Middle Test will be time series derived for the G MRF and upstream of Testwood determined using the hydraulic model prepared for offtake channel (Middle Test) if Pool the NEP and so is not presented here. possible Conagar Bridge GS (No. 11) The Little Test downstream of the H plus the Test Back Carrier Little and Great Test flow split (No10)

Table 4 Key Locations and flow gauging datasets No. Flow gauging datasets required Data set length Latest start date Notes  Broadlands GS  Jan 1957 to current

A Jan 1986  Test Back Carrier  Jan 1986 to current  As for Location A  Jan 1986 to current

B Jan 1982  Conagar Bridge  Jan 1982 to current  As for Location B Jan 1986 if assume  Jan 1986 C  Nursling Fish Farm fixed Nursling values  Jan 2001 to March 2011 Abstraction before 2001

D  Testwood GS  May 1987 to current May 1987  As for Location D  May 1987 The Broadlands Fish Farm

E  Factorised flow for Ower GS  Jan 1976 to current Jan 1996 Carrier has the latest start  Broadlands Fish Farm Carrier  Jan 1996 to current date for all datasets  As for Location E  Jan 1996 to current F  Nursling Fish Farm Jan 1996  Jan 2001 to March 2011 Abstraction  Jan 1996 to current  As for Location F  Derived from hydraulic G  flow time series derived from Jan 1996 model so determined by the hydraulic model other datasets  Conagar Bridge GS  Jan 1982 to current H Jan 1986  Test Back Carrier  Jan 1986 to current

13 Lower Test NEP Volume 3: Appendices

3500 Figure 6c Flows for Sites C and D 3250 3500 3000 3250 2750 3000 2500 2750 2250 2500 2000 2250 1750 2000 1500 Flow (Ml/d) 1750 1250 1500 1000 Flow (Ml/d) 1250 750 1000 500 750 250 500 0 250

0

Jan 98Jan 99Jan 02Jan 05Jan 06Jan 09Jan 10Jan Jan 97Jan 00Jan 01Jan 03Jan 04Jan 07Jan 08Jan 11Jan Jan 96Jan

Site A Main Test d/s of Longbridge GS and upstream of the Little and Great Test flow split

Jan 98Jan 99Jan 02Jan 05Jan 06Jan 09Jan 10Jan Jan 97Jan 00Jan 01Jan 03Jan 04Jan 07Jan 08Jan 11Jan Figure 6a Flows for each locationSite B Great discussed Test d/s ofin the this Little Appendix and Great Test flow split and upstream of the Nursling Fish Farm offtake 96Jan Site C Great Test d/s of the Nursling Fish Farm Offtake and upstream of the Testwood Abstraction Site D Great Test d/s of the Testwood Abstraction and upstream of the Blackwater confluence Site D Great Test d/s of the Testwood Abstraction and upstream of the Blackwater confluence 3500 Site E Great Test just d/s of the Blackwater confluence Figure 6d Flows for Sites D and E Site F Great Test at MRF location 3250 Site H The Little Test d/s of the Little and Great Test flow split 3500 3000 Site E Great Test just d/s of the Blackwater confluence 3250 2750 3000 2500 2750 2250 2500 2000 2250 1750 2000 1500 Flow (Ml/d) 1750 1250 1500 1000 Flow (Ml/d) 1250 750 1000 500 750 250 500 0 3500 3250 250 3000

2750 0

Jan 98Jan 99Jan 02Jan 05Jan 06Jan 09Jan 10Jan Jan 97Jan 00Jan 01Jan 03Jan 04Jan 07Jan 08Jan 11Jan Jan 96Jan 2500 2250 2000 1750

Site A Main Test d/s of Longbridge GS and upstream of the Little and Great Test flow split 1500

Jan 98Jan 99Jan 02Jan 05Jan 06Jan 09Jan 10Jan Jan 97Jan 00Jan 01Jan 03Jan 04Jan 07Jan 08Jan 11Jan 1250 96Jan Figure 6b Flows for Sites A, B and H 1000

Flow (Ml/d) 750 500 Site B Great Test d/s of the Little and Great Test flow split and upstream of the Nursling Fish Farm offtake 250 0 3500 Figure 6e Flows for Sites E and F

3250 Site H The Little Test d/s of the Little and Great Test flow split Site E Great Test just d/s of the Blackwater confluence Site F Great Test at MRF location

Jan 96Jan 97Jan 00Jan 01Jan 05Jan 06Jan 09Jan 10Jan Jan 99Jan 02Jan 03Jan 04Jan 07Jan 08Jan 11Jan 98Jan 3000 2750 Legend entries for all figures: 2500 2250 Site A Main Test d/s of Longbridge GS and upstream of the Little and Great Test flow split 2000 1750 Site B Great Test d/s of the Little and Great Test flow split and upstream of the Nursling Fish Farm offtake 1500 Flow (Ml/d) Site C Great Test d/s of the Nursling Fish Farm Offtake and upstream of the Testwood Abstraction 1250 Site D Great Test d/s of the Testwood Abstraction and upstream of the Blackwater confluence 1000 750 Site E Great Test just d/s of the Blackwater confluence 500 Site F Great Test at MRF location 250

0 Site H The Little Test d/s of the Little and Great Test flow split

Jan 98Jan 99Jan 02Jan 05Jan 06Jan 09Jan 10Jan Jan 97Jan 00Jan 01Jan 03Jan 04Jan 07Jan 08Jan 11Jan Jan 96Jan Site C Great Test d/s of the Nursling Fish Farm Offtake and upstream of the Testwood Abstraction 14

Site D Great Test d/s of the Testwood Abstraction and upstream of the Blackwater confluence Lower Test NEP Volume 3: Appendices

Appendix 4.1.1 Calibration of the Hydraulic Model

1 Introduction

A hydraulic model of the Lower River Test has been constructed in order to assess the effects of an abstraction upon river levels within channel under low flow conditions. The model is of a short reach of the river in the downstream end of the catchment. The downstream extent of the model is the Testwood Pool; a complex series of structures control the outflow from the River Test into the Testwood Pool. A location plan showing the extent of the model is given in Figure 1.1.

This Technical Note describes the hydraulic model and the processes and decisions taken during its construction.

Figure 1.1 - River Test Location Plan and Model Extent Contains Ordnance Survey data © Crown copyright and database right 2012

1

Lower Test NEP Volume 3: Appendices

2 Model Construction

The hydraulic model is a one-dimensional (1D) model constructed using Infoworks RS v11.5.6. The model was constructed using Infoworks as it allows for easy adaptation of the model and inclusion of 2D elements if required. This was considered important at the start of the construction phase as it was unknown whether 2D representation of the wetlands and other channels such as the Little Test would be required in order to achieve the objectives of the modelling; however, as the modelling progressed it was found to not be required.

2.1 Study Data The model of the Lower Test has been constructed twice, originally the model was constructed using old survey data from the 1940s to create the River Channel cross sections. While the model was constructed and successfully tested, there were issues associated with using such old data for the cross sections. Near the upstream end of the Testwood Pool, the survey data was found to have been inconsistent with drawings of the structures that control outflow of the River Test into the Testwood Pool, with the bed level shown being higher than the sill level of the structures so manual modification of the sections was required. In October 2011, the Environment Agency collected additional cross section data using a boat equipped with Acoustic Doppler Current Profiler (ADCP) equipment. The ADCP returns water depths at locations across the river channel and when used in conjunction with survey of river water level, it can be used to create survey river channel cross sections. This survey data was used to create an improved version of the Lower River Test model. The two versions of the survey data showed considerable variation in river bed level between each other; this is not surprising given the significant changes in channel morphology likely to have taken place in the 60-70 years between the two sets of data. Light Detection and Ranging (LiDAR) data, obtained from the Environment Agency, was also used for construction of the hydraulic model in order to extend cross sections into the floodplain.

2.2 River Sections The final version of the hydraulic model was constructed using the ADCP river section data previously described. The ADCP data collects an overly dense cross section for use in a hydraulic model, so the sections were individually altered within excel to reduce the number of data points in each cross section, whilst retaining an accurate representation of channel geometry. These altered cross sections were then imported into Infoworks RS and the 1D channel network was constructed from the data. Despite being predominantly a low flow model, the sections required extending into the floodplain as the maximum elevation on the ADCP surveyed sections was equivalent to the water level at the time of the survey. Before the model was extended into the floodplain, it was unstable during periods of higher flows; however, this was rectified by extending the sections; the sections were extended into the floodplain using LiDAR data.

2.3 Structures A significant number of hydraulic structures control the interaction between flows in the River Test and the Testwood Pool, a plan of these structures is shown in Figure 3.1. The representation of the structures has been based upon previous surveys of the structures and recent inspection of them. Following recent inspection of the structures, the overflow sluice has been kept shut for all scenarios, as the inspection revealed that the structure had either failed, or boards had been placed within it to prevent flow through it. The majority of structures that control flow between the River Test and the Testwood Pool have been included within the model as a penstock, so that variable operation of them can be included for the various scenarios that were to be tested. The Testwood Flow gauge has been included as a spill unit as the ADCP survey of it indicated that it was an irregular weir. In addition a series of sharp crested weirs have been included to represent the point where the river overflows into the Middle Test. A further weir unit has been included to control the flow between the Testwood Pool and the downstream tidal reach.

2.4 Model Roughness Model roughness is accounted for in Infoworks RS using values of Manning’s n. Values of 0.03 have been used in channel and 0.05 for the floodplain. These values are consistent with those regularly used hydraulic

2

Lower Test NEP Volume 3: Appendices modelling of rivers of the nature of the River Test in this location and were based upon observations of the river while on a site visit. The model has been sensitivity tested for changes in manning’s n values by both increasing and decreasing the values of Manning’s n by 20%. At the downstream end of the river adjacent to Testwood Pool, there was zero change in river levels. At the upstream end, the total difference in river levels between the +20% and - 20% model runs was approximately 40mm. These results indicate that the model is not sensitive to changes in channel roughness. This is unsurprising given that this reach of the river is impounded; meaning that the settings of the gates have a more significant impact on river levels that the roughness coefficients chosen.

2.5 Model Run Parameters and Performance The hydraulic model can be run with an adaptive time-step and runs quickly. It is possible to simulate 1 years worth of river flows in approximately two and a half minutes, making long term simulations which are required for low-flow modelling to be performed.

3 Model Calibration / Verification

3.1 Calibration Aims Initial hydraulic model runs indicated that for some periods of time, the fit between the modelled and observed level series wasn’t good for the whole of the historic model run. Where there was a difference between the observed and the modelled level, it was unknown whether this was due to gate operation, the input flow series or the downstream boundary tidal data that was used. A number of model runs were carried out in order to calibrate / verify the hydraulic model using data for 2011. The calibration process involved changing a number of features in order to assess their impact on the performance of the model. The parameters altered were:  Structure Operation  The upstream inflow boundary series  The downstream tidal boundary

3.2 Calibration Setup

1.1.1. Gate settings A detailed calibration of the hydraulic model is not possible, as this would require detailed time-series operation of the various manually operated hydraulic structures that control the flow of water between the Lower Test and the Testwood Pool. Given the lack of structure operation data, calibration of the model against long term recorded data for 2011 has been carried out using 3 different structure operation scenarios for the structures in the Testwood Pool. The three scenarios tested are fully closed, fully open and partially open. The exact state of each of the structures for the three structure operation scenarios is described in Table 3.1.

1.1.2. Inflow Series The inflow boundary used originally was an approved flow series provided by the Environment Agency, this was provided as a daily flow series for the period of interest. After the initial model runs, a difference in the timing of the modelled and observed peak water levels was observed. As it was unclear whether this was due the downstream boundary, the gate settings or the inflow boundary, calibration scenarios were run using observed flow series from the Testwood flow gauge located at the upstream end of the model; this allowed an assessment of the impact of using a daily input flow series on the model results.

1.1.3. Tidal Boundary Initially, the use of observed tidal data was investigated and rejected due to the amount of data gaps in the long term record. As a result, long term time-series were generated using the simplified admiralty method and the harmonic constants for Redbridge given in the Admiralty Tide tables. (United Kingdom Hydrographic Office , 2010). Comparison of the synthetic tide series with that from the Eling tide gauge suggested that at times there was a significant difference between the two, possibly due to atmospheric conditions that are not accounted for

3

Lower Test NEP Volume 3: Appendices within the simplified admiralty method. Sensitivity testing was conducted on the tidal boundary conditions and is reported on in section 1.1.5.

3.3 Calibration Results

1.1.4. Impact of changing flow series Initial assessment of the model results had indicated that some of the difference between modelled and observed water levels could be due to the approved inflow series being a daily rather than a sub-daily flow series. To test this, the model was run for the calibration period using the gauged inflow series from the Testwood Gauge instead of the historic approved flow series. Figure 3.2 and Figure 3.3 shows the effect of using measured 15 minute flow data derived from either 15 minute measured flows (blue line) vs. the use daily data divided to get 15 minute data bits (red line) as the upstream boundary. The effect of using daily data is to incur a slight difference in the timing of peaks but this is not considered to be significant for the NEP investigation. It was concluded that the fit achieved using daily data was sufficiently accurate for the purposes of the NEP investigation. This was due to the close similarity of the two datasets (15 minute measured data and disaggregated daily data).

Table 3.1 - Structure position for each of the gate opening scenarios

Structure Fully Open Partially Open Closed Weed Sluice Open 1.0m Open 0.3m Fully closed Closed – Inspection has revealed that Overflow Sluice the structure appears to have been Closed Fully closed modified or failed Fully closed - Structure rarely operated Eel Rack Sluice Fully closed Fully closed and only for Eel fishing Triple Sluices Open 0.72m Open 0.3m Fully closed Closed. Fish pass within Closed – However, fish pass and small gate is open smaller Fully closed except for fish Flood Sluices penstock within gate set to fully open penstock within gate is pass cut into gate closed

4

Lower Test NEP Volume 3: Appendices

Figure 3.1 - Diagram of structure location between the River Test and Testwood pool

Source: provided by the Environment Agency

5

Lower Test NEP Volume 3: Appendices

AP1 - Testwood GS 3.10

3.00 Measured (Testwood GS)

2.90 Historic partially open (Eling tidal boundary + 15 min flow data) 2.80 Historic partially open (Eling tide boundary + mean 15 minute flows from daily) 2.70

2.60

2.50

Water level (m OD) level(m Water 2.40

2.30

2.20

2.10

2.00

01 Dec 11 Dec 01 11 Dec 02 11 Dec 06 11 Dec 07 11 Dec 08 11 Dec 12 11 Dec 13 11 Dec 17 11 Dec 18 11 Dec 19 11 Dec 23 11 Dec 24 11 Dec 28 11 Dec 29 03 Dec 11 Dec 03 11 Dec 04 11 Dec 05 11 Dec 09 11 Dec 10 11 Dec 11 11 Dec 14 11 Dec 15 11 Dec 16 11 Dec 20 11 Dec 21 11 Dec 22 11 Dec 25 11 Dec 26 11 Dec 27 11 Dec 30 11 Dec 31

04 Nov 11 Nov 04 11 Nov 05 11 Nov 06 11 Nov 10 11 Nov 11 11 Nov 15 11 Nov 16 11 Nov 20 11 Nov 21 11 Nov 22 11 Nov 26 11 Nov 27 08 Nov 11 Nov 08 11 Nov 09 11 Nov 12 11 Nov 13 11 Nov 14 11 Nov 17 11 Nov 18 11 Nov 19 11 Nov 23 11 Nov 24 11 Nov 25 11 Nov 28 11 Nov 29 11 Nov 30 07 Nov 11 Nov 07 Figure 3.2 - Chart showing a comparison of measured water level at Assessment Point 1 with simulated water level using the approved historic inflow and the gauged inflow from the Testwood Gauge

(See figure 4.1.1.1 in the main report for locations of assessment points)

AP5 - Testwood Mill water levels 2.90

2.80 Measured

2.70 Historic partially open (Eling tidal boundary + 15 min flow data)

2.60 Historic Partially open (Eling tide boundary + mean 15 minute flows from daily) 2.50

2.40

2.30

2.20

2.10 Water level (m OD) level(m Water 2.00

1.90

1.80

1.70

1.60

01 Dec 11 Dec 01 11 Dec 02 11 Dec 06 11 Dec 07 11 Dec 08 11 Dec 12 11 Dec 13 11 Dec 17 11 Dec 18 11 Dec 19 11 Dec 23 11 Dec 24 11 Dec 28 11 Dec 29 03 Dec 11 Dec 03 11 Dec 04 11 Dec 05 11 Dec 09 11 Dec 10 11 Dec 11 11 Dec 14 11 Dec 15 11 Dec 16 11 Dec 20 11 Dec 21 11 Dec 22 11 Dec 25 11 Dec 26 11 Dec 27 11 Dec 30 11 Dec 31

04 Nov 11 Nov 04 11 Nov 05 11 Nov 06 11 Nov 10 11 Nov 11 11 Nov 15 11 Nov 16 11 Nov 20 11 Nov 21 11 Nov 22 11 Nov 26 11 Nov 27 08 Nov 11 Nov 08 11 Nov 09 11 Nov 12 11 Nov 13 11 Nov 14 11 Nov 17 11 Nov 18 11 Nov 19 11 Nov 23 11 Nov 24 11 Nov 25 11 Nov 28 11 Nov 29 11 Nov 30 07 Nov 11 Nov 07 Figure 3.3 - Chart showing a comparison of measured water level at Assessment Point 5 with simulated water level using the approved historic inflow and the gauged inflow from the Testwood Gauge

1.1.5. Impact of tidal data The impact of the varying tidal data sets on the model was assessed by running the model with various different tidal data sets, these were:

6

Lower Test NEP Volume 3: Appendices

 Tidal series generated using the simplified admiralty method  Data from Eling Mill  A tide series created from a correlation of the generated data set with the Eling Mill data set  The Eling Mill data set in filled with data from the correlated series While it would have been preferable to use observed data for the entire model simulation period, there are significant gaps in the data record for Eling Mill. This lead to the initial decision to use a tide series generated using the simplified admiralty method. Comparison of the two tide series noted some significant differences in peak water levels due to the lack of atmospheric influence in the generated tide series. The model was run for the calibration period using the four different tide series listed above. Figure 3.4 shows modelled water levels from assessment point 5 using the various different tidal boundary data sets. It shows a snapshot of a week in November, but the calibration was run for a whole year (2011). For the correlated data, an attempt was made to scale the synthetic tidal data as it was not reproducing all the peaks shown in the Eling Mill tide measurements. The conclusion has been to compile a tidal data series for use in the hydraulic model that comprises the Eling Mill tidal data where it is available, and to use the data generated from the correlation to infill any gaps in the Eling Mill dataset.

AP5 - Testwood Mill water levels 2.60 Measured

Historic partially open (Eling tidal boundary) 2.50 Correlated tidal data

2.40

2.30 Water level (m OD) Water 2.20

2.10

2.00

01 Dec Dec 1101

24 Nov 1124 Nov 1125 Nov 1126 Nov 1130 28 Nov 1128 Nov 1129 27 Nov 1127 Figure 3.4 - Chart to show water levels at Assessment Point 5 using the various different tidal inflow boundaries

3.4 Final Model Configuration The hydraulic model has been constructed using historic information regarding structure dimensions, channel topography with the addition of data collected on site where possible. Access limitation has meant that it has not been possible to collect all information that could possibly be used to fully represent the Lower Test system. However, it is considered that the model includes the best available data and provides a realistic simulation of the complex hydraulics of the Lower Test river system. To provide some context to the scale of changes seen by varying the inflow boundaries and the downstream tidal boundary, a comparison of water levels at assessment point 2 with common inflow and tidal boundaries is shown in Figure 3.5 for the fully closed, fully open and partially open structure scenarios. It shows that these clearly have a far greater influence on water levels within the modelled river reach than any combination of changing the inflow boundaries or the downstream tidal boundary will have on model results. It is generally considered that the partially open scenario represent a reasonable representation of how the

7

Lower Test NEP Volume 3: Appendices structures are operated for the majority of time; however, it is also known that some of the gates are operated on a relatively ad-hoc basis for which there are no records in order to regulate water levels in the river. The results of the calibration period show that given the lack of structure operation data, the model is representing the hydraulics of the river reach reasonably accurately. As the NEP investigation is interested in model results for all possible operation scenarios of the structures, then even without a detailed record of the operation of the structures, the model is of sufficient accuracy for the purposes of the NEP investigation. The final model configuration was to use the daily approved flow series provided by the Environment Agency, along with the in-filled Eling tide data. The model has been run for a 16 year period from 1996-2011 for the fully open, partially open and fully closed scenarios.

AP2 2.40 1.43 2.30 1.33 2.20 1.23 2.10 1.13 2.00 1.03 1.90 0.93 1.80 0.83 1.70 0.73

1.60 0.63 Water depthWater (m)

Water Level(mAOD)Water 1.50 0.53 1.40 0.43 1.30 0.33

1.20 0.23

01 Jul Jul 11 01

01 Jan Jan 11 01

01 Jun 11 Jun 01

01 Oct 11 Oct 01

01 Apr Apr 11 01

01 Feb 11 Feb 01

01 Sep 11 Sep 01

01 Dec 11 Dec 01

01 Aug Aug 11 01

01 Nov 11 Nov 01

01 Mar11 01 01 May 11 May 01

Simulated Fully Open Simulated Closed Simulated Partially Open Dummy for second axis

Figure 3.5 - Water levels for the fully open, partially open and fully closed scenarios using the common inflow and tidal boundary

Bibliography

United Kingdom Hydrographic Office . (2010). Admiralty Tide Tables 2011 Volume 1 - United Kingdom and Ireland. United Kingdom Hydrographic Office.

8

Lower Test NEP Volume 3: Appendices

Appendix 4.2.1 The Wetland Model

1 Introduction

Water table fluctuations within the Lower Test Valley SSSI have been modelled using an adaptation of a published water table model (Armstrong, 1993; Armstrong and Rose, 1998; and Swetnam et al., 1998). The model was chosen for its simplicity and limited data demands.

The use of the wetland model has important advantages:  Includes an explicit representation of the properties of the stratigraphy underlying the Lower Test Valley SSSI;  Represents the influence of rainfall, evapotranspiration, ditch water levels and tidal variations on an hourly basis;  Provides a way of linking abstraction at the Testwood Works to impacts on the ecological integrity of the site; and  Quantifies the wetland water balance. However, it also has important and recognised limitations, in particular regarding the simplification in the relationship of the wetland water table to groundwater heads in the underlying sandy gravels. There are few data describing water levels in the gravels underlying the Lower Test Valley. The model is also 1- dimensional and does not take account for 3-dimensional flow. 1.1 The Model The soil water table model is 1-dimensional and provides a point estimate of water table elevation in a field bound by watercourses, simplifying the modelling of water tables in wetland areas by avoiding application of the Richards equation of unsaturated flow, which is notoriously difficult to solve and imposes excessive demands upon computing resources (Armstrong and Rose, 1998).

Figure 1 shows the conceptual model for the hydrological functioning of a field within the Lower Test Valley SSSI.

Within the water table model, the water table elevation (MLt) can be estimated as a function of the water table elevation in the preceding time-step (MLt -1) by

MLt = MLt -1 + (P – ET - Qd) (eqn G1) f where

2 2 Qd = 4K (ML – DL ) (eqn G2) L where P is precipitation (m), ET is evapotranspiration (m), Qd is discharge to or from watercourses (m), f is specific yield, DL is the level of the ditches on day t (m), K is hydraulic conductivity (m/day) and L is the distance between the location where the water table elevation is to be estimated (in m) and the closest watercourse.

The model assumes that there is no soil moisture deficit. This assumption is commonly used in groundwater models for riparian and wetland areas where the water table is close to the surface.

1

Lower Test NEP Volume 3: Appendices

1.2 Model period The water table model has been set up to provide water table level predictions on an hourly basis for the year of 2011. The hourly time-step is required to enable the inclusion of the tidally-induced water level variations observed in the Lower Test Valley Marshes (see Section 3.6).

No water table level data have been historically collected across the Lower Test Valley SSSI therefore no data was available for model calibration. However, a detailed network of surface water level recorders has been established on the River Test and the Lower Test Valley SSSI (See Section 5) and has been used as part of model development where relevant.

E P

Silty Sandy CLAY (0.2 – 0.7m) Soft brown PEAT (0.5 – 1.0m) Silty Sandy CLAY (0.4 – 0.5m) Silty SAND (0 – 0.5m) -Qd +Qd Sandy GRAVEL (3 – 4m)

V

M L DL L t-1 DL

Figure 1 Conceptual model of the Lower Test Valley SSSI (soil profile based on Figure 3.2.3.)

1.3 Hydrological processes modelled 1.3.1 Precipitation (P)

The contributions of precipitation to the water table were estimated in line with the methodology set out by (Armstrong 1993) where:

P = (p /1000) (eqn G3) and p is precipitation (mm). Precipitation records were obtained from the Romsey raingauge some 7.5km from the study area. A tipping bucket raingauge (TBR) is located close by at the Testwood Works. However, the TBR record is unreliable and rainfall is underestimated as the bucket ‘sticks’ (A Roberts, Environment Agency, Pers. Comm.). No lag response times were incorporated within the model and precipitation falling on day t generated runoff on day t. Rainfall that exceeds the infiltration capacity of local soils, one of the required model parameters, does not contribute to the soil water table model; when hourly rainfall exceeds the infiltration capacity, it is assumed that rain in excess of the infiltration capacity is conveyed into the ditches as runoff. The infiltration capacity of the soils was set at 0.5mm/hr.

1.3.2 Evapotranspiration (ET)

The rate of ET within the wetland model was adjusted based on water table depth in line with the calculation of water-table evapotranspiration within MODFLOW. The following parameters are included in the wetland model:  the evaporation surface elevation; and

2

Lower Test NEP Volume 3: Appendices

 the extinction depth. The evaporation surface elevation defines the elevation that the water table needs to fall to before the rate of ET drops below the maximum rate. This was taken as 0.25m in line with the approximate depth of top soil in the Lower Test Valley SSSI (Figure 3.2.3). Once the water table falls below this level, the rate of ET declines asymptotically until the water table reaches the second parameter, the extinction depth, at which point ET ceases. For the Lower Test Valley Marshes, the extinction depth was taken as 0.9m in line with approaches used in MODFLOW for riparian areas. This also corresponded closely with the depth of the alluvium in borehole logs closest to the area of interest (Section 3.2). Once the water table level falls below the extinction depth, no water can be extracted from the soil water table and no evaporation can occur. An hourly time-series of ET has been calculated using monthly MOSES data by assuming a constant rate on evapotranspiration on each hour in day in the month. A series of coefficients have been applied to MOSES data to account for the differences between evapotranspiration in wetlands and the reference conditions estimated by the MOSES methodology. The evaporation coefficients proposed by the Environment Agency for use in lowland wetlands are shown in Table 1.

Table 1. Evaporation coefficients used in the wetland model.

Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

ET kc 0.90 0.90 0.90 1.05 1.05 1.20 1.20 1.20 1.20 0.70 0.70 0.90

1.3.3 Interaction with watercourses (Qd)

In the Lower Test Marshes, there is a layer of alluvium 1.0m thick consisting of clays, silty clays, sandy clays and soft brown peats (see Section 3.2). Suitable values of K for use in the calculation of Qd were identified as part of the literature review. A value of 0.01md-1 was applied as describing the conductivity of the silty sandy clays and peats based on data included in standard hydrogeological text books. Distances applied within the model were 25, 50 and 125m to assess water table variations in the Lower test Valley SSSI from areas close to watercourses to field centres.

1.3.4 Exchange with the underlying aquifer (V)

A component to account for exchange with the underlying aquifer (V) has been included in the soil water table model so that Equation G1 becomes:

MLt = MLt -1 + (P - ET - V - Qd) (eqn G4) f where P is precipitation (m), ET is evapotranspiration (m), Qd is discharge/recharge to or from watercourses and f is specific yield (dimensionless).

In the case of the Lower Test Valley, the value of V accounts for the exchange between the alluvium and the underlying gravels. Borehole logs analysed for the Lower Test Valley SSSI indicate that the gravel extends beneath the whole site (Section 3.2) and is exposed in the tidal channel bed near Redbridge where high flows limit the deposition of tidal mud. The gravels themselves are isolated from the underlying Chalk aquifer by Tertiary deposits (Section 3.2). The exchange of groundwater between the alluvium and the gravels (V) was estimated based on Darcy’s Law:

V=KvI (eqn G5)

3

Lower Test NEP Volume 3: Appendices

Where Kv is the hydraulic conductivity of the alluvial deposits (the main constraint to vertical groundwater flow between the gravels and the alluvium, in m/day), I is the hydraulic gradient between the gravels and the alluvium at each model time-step (in m) and d is the depth is the thickness of the alluvium (in m).

At each model time-step the value of I is calculated by

MLt - GLt (eqn G6)

where MLt is the predicted water table level on day t and GLt is the head in the underlying gravels (both in m OD).

As previously discussed, in the Lower Test Valley, the gravels are isolated from the underlying Chalk aquifer by the Tertiaries. In the absence of gravel water level data, a synthetic time-series of gravel head (GLt) under the Lower Test Marshes was generated. It was assumed that water levels in the underlying gravels were mainatined at a constant at a level of 2.5m OD (equivalent to the field surface) throughout the year. This approach simulates the continuous down-valley flow of water through the gravels that is typical in Chalk catchments. For the assessment of the impacts of abstraction it has the added advantage of maintaining the value of V at a constant level to isolate the effects of abstraction from other processes. The gravel water level in the model can also be changed however to assess the sensitivity of the site to gravel water levels, if required.

1.4 Model set up Research during the model build focussed on the identification of suitable values of Kh and Kv. The vertical hydraulic conductivity (Kv) controls the vertical interactions between the wetland and the underlying Gravels. The horizontal hydraulic conductivity (Kh) controls interactions between the field water table and ditches. No field-based estimates for these parameters were available. Values were therefore necessarily taken from the literature and previous studies on soils similar to those found in the Lower Test Valley. The final parameter data set for the Lower Test Valley model is given in Table 2 below. A value of 0.01 m/day was used for both the vertical (Kv) and horizontal hydraulic conductivity (Kh). A value of 0.2 has been used for the specific yield (f) in line with the optimum value determined by Armstrong (1993) and Acreman et al. (2007) for alluvial soils. The depth of the deposits is taken as 1.00m in line with results of soil coring on the Lower Test Marshes SSSI (see Section 3.2). The infiltration capacity was taken as 0.05 mm/hr based on previous modelling studies in coastal wet grassland areas. The extinction depth corresponded with the mean depth of the alluvium in borehole logs in the central part of the Lower Test Valley SSSI. The evapotranspiration surface elevation was taken as 0.25m, the mean depth of top soil in the central part of the Lower Test Valley SSSI. The extinction depth was 0.9m based on the application of MODFLOW to riparian areas.

Table 2 Final calibration parameter data set for the Lower Test Valley SSSI water table model Value Source Kh Alluvium hydraulic conductivity 0.01 m/day Literature Sands, silts and unconsolidated Chalk Kv 0.01 m/day Literature hydraulic conductivity 0.2 Armstrong (1993); S Peat specific yield (dimensionless) Acreman et al. (2007) Ki Infiltration rate 0.05 mm/hr - ET/K surface depth 0.25 m MODFLOW - Extinction depth 0.9 m MODFLOW

4

Lower Test NEP Volume 3: Appendices

Appendix 4.3.1 The Atkins’ Aquatic Heat Model

1. Introduction 2 1.1. Surface Heat Transfer and Shading 2 1.2. Description of Foliage 3 1.3. Meteorological Data 3

2. River Model 3 2.1. Basis of Model 3 2.2. Channel Shapes 4

3. Surface Heat Transfer 5 3.1. Introduction 5 3.2. Detailed Mathematical Description 5 3.3. Convective Heat Exchange 7 3.4. Vaporisation 7 3.5. Short Wave Radiation 8 3.5.1. Short Wave Radiation at the Earths Surface 8 3.5.2. Earth Sun Distance 8 3.5.3. Solar Declination, Sunrise and Sunset Times 8 3.5.4. Extraterrestrial Radiation 9 3.5.5. Radiation Scattering and Absorption 10 3.5.6. Dewpoint Temperature 11 3.5.7. Cloudiness 12 3.5.8. Albedo 12 3.5.9. Beer’s Law 13 3.6. Long Wave Radiation 13 3.6.1. Atmospheric Long Wave Radiation 13 3.6.2. Long Wave Back Radiation 13 3.6.3. Vegetation Re-radiation Component 14 3.6.4. Vegetation Re-radiation Stream Surface View Factor 14 3.7. Stream Bed Heat Balance 15 3.7.1. Friction Generated Stream Heat 15 3.8. Solution Technique 15 3.8.1. Fractional Step Method 15 3.8.2. Time and Space Step 16

4. Application of the Model to the Lower River Test 17 4.1. Hydrodynamics & Transport 17 4.2. Cross Section & Processed Parameters 17 4.3. Foliage 17 4.4. Flows & Abstractions 18 4.5. Meteorological Data 18 4.6. Downstream Boundary 19 4.7. Upstream Boundary 20 4.8. Groundwater & Soil Interaction 21

5. Calibration and Verification Results 21 5.1. Calibration 21 5.2. Verification 22

6. Conclusions 23

1

Lower Test NEP Volume 3: Appendices

1. Introduction

The Atkins Aquatic Heat Model [AQHM] was developed to simulate the heat balance in slowly moving water bodies such as lakes and docks as well as streams. It includes the effects of shading and a full description of the surface heat transfer which, in broad terms, is a function of air temperature, wind speed and direction, relative humidity and cloud cover. The distribution and shape of foliage is modelled with appropriate view factors. The mathematical background is described in the sections below. The model focuses on the thermal balance and heat transport in the river as a function of riparian shading. Both trees and shrubs as means of providing shade can be simulated. In particular the sensitivity of the downstream temperature gradient has been investigated as a function of:  Stream flow  Shading type – shrubs or trees  Stream orientation  Variation in wind speed  Variation in air temperature

1.1. Surface Heat Transfer and Shading

In the model the effect of foliage or trees on the short and long wave radiation is modelled in some detail as accurate representation of the surface heat transfer is critical. It is independent of the water body model type and only requires a local water surface temperature from the water body model. The temperature distribution is not independent of the water body type but it is assumed that the water is fully mixed over the channel cross section so the vertical and lateral horizontal temperature is spatially invariant and only the longitudinal temperature is not. A brief description is given here to clarify the effect of foliage or trees. The presence of foliage modifies all the terms except conduction at the surface, generation of heat within the channel due to friction and interaction with the ground. Short wave directional solar radiation is intercepted by the foliage and branches then re-radiated at long wavelengths in a diffuse manner. At the water surface therefore the short wave radiation component is reduced, and partly replaced by diffuse long waver radiation. The amount of reduction is a function of the foliage density which is seasonal and varies with the shape of the tree or shrub. Foliage or trees also modify the wind speed at the water surface due to shielding. The foliage will reduce the direct solar radiation by shielding, but heat is re-radiated at longer wavelengths and will be absorbed by the stream. Shading therefore does not reduce heating as much as might be supposed due to cutting off sunlight. Re-radiation is essentially a function of the foliage distribution and its temperature, and the latter assumed to be equal to the air temperature. At night the amount of effective back radiation from the stream surface will therefore be reduced as the shrubs or trees will continue to provide diffuse radiation heating to the stream, although of course that will reduce as the air temperature drops. Nevertheless, the effect will be to increase the minimum night time temperature of the stream. The maximum daytime temperature will be reduced. The amount of diffuse long wave radiation transmitted to the stream surface is a function of the shape of the tree or shrub and the distribution of foliage and is characterised by the view factor. In the model the seasonal changes in foliage, associated changes in the effect on wind speed reduction and the effect on solar shading and re-radiation are linked so that only one description of the seasonal changes to the foliage is required.

2

Lower Test NEP Volume 3: Appendices

1.2. Description of Foliage

In all the simulations shading and shielding from the wind is derived from either shrubs or trees, which are defined geometrically in the model. The important distinction is that the shrubs provide some shielding from the effects of wind right down to ground level, whereas the trees are specified in such as way that the wind is allowed to pass more or less unhindered beneath the lowest branches.

1.3. Meteorological Data

The surface heat transfer sub model requires as input the following meteorological descriptors

 Wind speed and direction  Relative humidity  Cloud cover  Air temperature

Note that solar radiation is the major source of heat at the water surface and an important part of the surface heat transfer sub model is directed to evaluating this. It is a function of a number of variables, including

 Latitude  Cloud cover  Dust content of the air  Relative humidity  Altitude

In particular the time of day and season dictate the angle at which the sun strikes the ground and the column of air through which the radiation must travel and is therefore reduced. If the water surface temperature is also know the heat transfer rate can then be found. As the model relies on detailed time stepping simulations hourly measurements for these five parameters have been used. The most influential feature is the variation of air temperature which can be seen to vary as we would expect, dropping to a minimum in the early hours of the day and rising to a maximum in the early afternoon. This typical trend is modified by cloud cover, and relative humidity to some extent, which are also related to rainfall. Wind speed also shows a slight daily variation reaching a peak during the middle of the day and falling to a minimum at night. Relative humidity is highest during the night. These all have some effect on the total surface heat transfer at the water surface and hence the temperature.

2. River Model

2.1. Basis of Model

The river is treated as a fully mixed one dimensional flow vessel. The flow is assumed to be locally uniform and steady. The assumption of local uniformity ensures that Mannings equation can be used to quantify the stage discharge. It is also assumed that the flow is uniform over the entire length of the channel for which the thermal profile is required. For the low flow conditions that are being considered this a reasonable assumption. They will vary only very slowly in time and space. The assumption of complete mixing is reasonable since the channels are relatively shallow and wide. The only temperature gradient that we need consider is therefore longitudinal. We further assume that diffusion term is small and is therefore neglected. This is reasonable since the temperature gradients are small. 3

Lower Test NEP Volume 3: Appendices

In the model the shading from either low foliage such as shrubs, or trees, is disposed on either bank. The heat transfer distribution will therefore not be quite uniform across the width of the channel, but in line with the assumption of fully mixed flow. The total surface heat balance across the river surface is thus modified according to the distribution of shading. The heat at a cross section is then advected downstream. The 1D equation for the heat balance is as follows. The fundamental equation that is solved by the 1D model is given by the following Partial Differential Equation (PDE). This is equation is one dimensional diffusion or heat equation:-

   2  u  D  s Equation 2 t x L x 2

At the upstream end of the channel

(0,t)  (t) HW Equation 4

Where at the water surface:-

Q  f SHTM (T,t) Equation 5 t z0

f SHTM (T,t) is the temperature change due to surface heat transfer t is the time [s] T is the temperature at time t [°C] -1 -1 cp is the specific heat of water [4186 Jkg °C ] -3 ρ0 is the density of water [kgm ]

2.2. Channel Shapes

Two channels have been considered but are all treated in the same way. In the model the flow is considered to be steady and uniform. As the cross section shape is defined all the cross section properties are derived directly from this geometry. These include  Cross section  Width  Wetted perimeter  Conveyance The stage discharge curve is derived using Manning’s uniform flow equation as follows:

1 A3/ 2 Q(z)  S n R Equation 4 h

4

Lower Test NEP Volume 3: Appendices

3. Surface Heat Transfer

3.1. Introduction

Accurate representation of the surface heat transfer is critical. It is influenced by many factors including time of day and date, latitude, wind speed, air temperature and relative humidity. It is independent of the water body model type and only requires a local water surface temperature from the water body model. Figure 1 illustrates the processes that combine to create a net transfer of heat through the water surface at a given point in time. A detailed mathematical description is given here. The model includes:  Short wave radiation  Long wave incoming radiation modified by foliage  Long wave back radiation from stream surface  Convective heat transfer at the surface  Conduction at the surface  Conduction at the stream bed  Heat generation from frictional losses in the flow The model also accounts for the modification of short wave radiation due to shading from buildings directly from the building geometry and solar azimuth and bearing. The effect of buildings on net long wave radiation is also included although this is a calibrated effect. Buildings also affect the amount of short wave radiation that occurs in the shadows, due to reflection. In reality this is complex to compute and in the LTHM model it the effect is calibrated.

Variation in net surface heat transfer due to the following factors are all included:  Wind speed  Wind direction  Relative humidity  Cloud cover  Dustiness of the air  Ground reflectance  Air temperature  Foliage shading

3.2. Detailed Mathematical Description The heat balance is given by:-

q  qio  qss  q p  qc  qs  qsr  qsu  ql  qlr  qlu  qg  qsed  qv Equation 14

Where:-

Δq Change in heat capacity of the water body qio heat transfer from inflow and outflow qss heat transfer from sources and sinks qp heat transfer from precipitation qv loss of energy due to evaporation qc convective heat transfer qs short wave radiation qUsrU reflected short wave radiation qsu short wave radiation emitted from the water body ql long wave radiation above the sea surface qlr reflected long wave radiation qlu emittance of long wave energy from the water body 5

Lower Test NEP Volume 3: Appendices qg heat exchange between ground and water body qsed heat exchange between sediment and water body

The last two terms are generally left out of models of water bodies as the effects are generally relatively small. As will be seen in the following sections many of the terms can be lumped together into a single term. In general four terms are therefore considered:-

 Heat flux due to convection  Latent heat flux [due to vaporisation]  Net short wave radiation  Net long wave radiation

The source term at the surface for a non 3D model is therefore:-

qv  qc  qsr,net  qlr,net H  Equation 15 0C p

Where:-

H net source term [°Cms-1]

0 density of water Cp specific heat of water [4186 Jkg-1°C-1]

Cloud Cover

Short Wave Radiation Directional

Back Radiation Long Wave Radiation

Wind Speed modified by trees

Convective and conductive exchange

Frictional heat generation & ground heat transfer

Figure 1 Physical Processes contributing to Surface Heat Transfer

6

Lower Test NEP Volume 3: Appendices

3.3. Convective Heat Exchange

Assuming a turbulent boundary layer then:-

For air temperature greater than or equal to the water temperature

qc  airCairCcW10Twater Tair  Equation 16

For air temperature less than the water temperature

qc  airCwCcW10Twater Tair  Equation 17

Where:-

-3 air density of air kgm -1 -1 Cair specific heat of air 1007 Jkg °C -1 -1 Cw specific heat of water 4186 Jkg °C -1 Q10m wind speed 10m above surface ms Tw absolute temperature of water surface °K Tair absolute temperature of air °K -3 Cc sensible transfer coefficient 1.41x10

3.4. Vaporisation Heat exchange due to vaporisation is based on Daltons Law as follows:-

qv  LCe a1  b1W2m Qwater  Qair  Equation 18

Where:-

L latent heat of vaporisation 2.5x106 Jkg-1 Ce moisture coefficient 1.32x103 -1 W2m wind speed 2m above the water surface ms Qwater vapour density close to the surface Qair vapour density in the atmosphere a1 and b1 user specified coefficients [default values are a1=0.295 and b1=0.371]

Vapour densities can be related to vapour pressure which in turn can be related to relative humidity and near surface temperatures as follows:-

   1 1    1 1  exp K   R.exp K      Tk Twater    Tk Tair  qv  Pv a1  b1W2m    Equation 19  Twater Tair     

7

Lower Test NEP Volume 3: Appendices

Where:-

-3 Pv latent constant 4370 J°Km K a constant temperature 5418°K Tk a constant temperature 273.15°K

3.5. Short Wave Radiation

3.5.1. Short Wave Radiation at the Earths Surface This exists mostly in the range 4000 to 9000 Å. The intensity is a function of sun to earth distance, radiation striking the upper atmosphere, cloud cover, humidity and the angle of incidence of the radiation. It varies daily and seasonally and is also a function of the dampening effect of the earth’s atmosphere and reflection from the water surface.

The solar radiation that strikes the surface of the earth is given by

qsr,nett  H 0at 1 RS Ca Equation 20

Where qsr,nett Short wave radiation intensity H0 Short wave radiation reaching the earths outer atmosphere at Attenuation coefficient due to scattering and absorption Rs Albedo or reflection coefficient Ca Fraction of solar radiation not absorbed by clouds [ie transmitted]

3.5.2. Earth Sun Distance

The ratio between the mean distance r0, and the actual distance r, is given by an equation which is a function of the Julian day of the year dn, as follows:-

2 d 1   n Equation 21 365

and

 r0  E0     1.000110  0.034221cos 0.001280sin 0.000719cos2 0.000077sin2  r  Equation 22

3.5.3. Solar Declination, Sunrise and Sunset Times The daily rotation of the earth contributes to a 24 hours variation in solar radiation. In addition there are seasonal changes, which vary as the angle of declination. The equation for the declination angle is:-

8

Lower Test NEP Volume 3: Appendices

  0.006918  0.399912cos 0.07257sin  0.006758cos2 0.000907sin2 Equation 23  0.002697 cos3 0.00148sin3

From geometry the sunrise angle, wsr is given by:-

1 sr  cos  tan tan  Equation 24

From which the sunrise and sunset times can be obtained as

 sr  TR  121  Equation 25   

And

 sr  TS  121  Equation 26   

Where TR and TS are the hours of sunrise and sunset.

3.5.4. Extraterrestrial Radiation The flux of short wave radiation that strikes the earth’s outer atmosphere is

 12  H 0  H sc E0sin sin  cos cos sinhe  sinhb  Equation 27   

-2 -1 Where Hsc is a solar constant with value 1390Jm s

Where

H0 Solar radiation flux that strikes earths atmosphere E0 The ratio between the mean and actual earth sun distance  correction factor for exposure to diurnal flux  Latitude  Declination he Hour angle at end of the time period over which H0 is computed hb Hour angle at start of time period over which H0 is computed

The hour angles are calculated from

9

Lower Test NEP Volume 3: Appendices

   hb  hr 1 ts  a 12  2 b Equation 28 12 

   he  hr  ts  a 12  2 b Equation 29 12 

Where hr Hour of the day

a = +1 for hr  12 a = -1 for hr  12 b varies with the magnitude of the term in square brackets [B]for both hb and he as b = -1 if [B] > 2 b = +1 if [B]  0 b = 0 for 0  [B]  2

ts is the fraction of the 15° increment by which the local meridian is west of the standard meridian for the time zone and is given by

E t  a L  L  Equation 30 s 15 sm lm

 is a step function such that

 = 1 for TR  t  TS and 0 otherwise

3.5.5. Radiation Scattering and Absorption Radiation entering the atmosphere is reduced by scattering and absorption. The fraction of radiation reaching the water surface after reduction by scattering and absorption is given by

a2  0.51 a1  cd  at  Equation 31 1 0.5Rg 1 a1  cd 

Where

Cd Dust coefficient ranging from 0.0 to 0.13 with a typical value of 0.06

Rg Reflectivity of the ground, which varies with the type of ground cover.

The mean atmospheric transmission coefficient after scattering and absorption is

a2  exp 0.4650.134Pwc 0.179  0.421exp 0.721am am Equation 32

The mean atmospheric coefficient is

10

Lower Test NEP Volume 3: Appendices

a1  exp 0.4650.134Pwc 0.129  0.171exp 0.88am am Equation 33

am Dimensionless optical mass Pwc mean daily precipitable atmospheric water content given by

Pwc  0.85exp0.11 0.0614Td  Equation 34

Where

Td is the dewpoint temperature in °C.

The optical air mass can be obtained from the elevation of the site and the solar altitude as follows

5.256 2.25694x105 z  am  Equation 35 sin  0.15  3.8551.253

Where z altitude of site  solar altitude in radians

The extraterrestrial intensity, q0, and the hour angle wi are given by:-

 24  24   q 0 q scE0 SinSin  sin cos cos cosi  Equation 36      

  Et 4  i  12  tsummer   Ls  Le TL  Equation 37 12  60 60 

The motion of the earth is not quite uniform so a correction is made to account for this which is referred to as the “equation of time”. This is the purpose of the term Et which is given by:-

Et  229.180.000075  0.001868cos  0.32077sin   0.014615cos 2  0.04089sin 2

Equation 38

Note that Ls is the standard longitude for the time zone and Le is the local longitude. For example in UK Le would be 0, corresponding to Greenwich Meridian.

3.5.6. Dewpoint Temperature The dewpoint temperature can be found from

log En  a Td  C Equation 39 b  a  log En 

The actual vapour pressure is given by

11

Lower Test NEP Volume 3: Appendices

En  Es Rh Equation 40

The saturated vapour pressure is given by

  bT  E  expa   a  s      c Ta 

Where

Td Dewpoint temperature a Constant log(611.2) b Constant 17.62 c Constant 243.12 En Actual vapour pressure Es Saturated vapour pressure Ta Air temperature Rh Relative humidity [as a fraction]

3.5.7. Cloudiness For short wave radiation the clouds are a sink rather than a source as for long wave radiation. The fraction of radiation passing through the clouds is given by

2 Ca 1 0.65Cl Equation 41

Where Cl the fraction of sky covered by clouds.

3.5.8. Albedo Solar radiation which strikes the earth’s surface is partly reflected, and is referred to as the albedo. It is dependent on the angle at which the solar radiation strikes the surface of the each and the amount and type of cloud cover.

b Rs  a Equation 42

Where Rs albedo a,b constants which depend on cloud cover and cloud type a solar altitude in angular °

Table 1 Values of Coefficients a and b For Different Cloud Cover Fractions

Cloud Description Cloud Cover Fraction a b Overcast C> 0.9 0.33 -0.45 Broken 0.5

12

Lower Test NEP Volume 3: Appendices

3.5.9. Beer’s Law Beer’s law describes the attenuation of light intensity in the near surface region of a body of water as follows:-

d Id  1  I 0e Equation 43

Where

I(d) Light intensity at a depth d below the water surface I0 Light intensity just below the water surface b Infrared near surface absorption factor  A light extinction coefficient

Values for b and l are 0.2 to 0.6 and 0.5 to 1.4m-1 respectively with defaults of 0.3 and 1.

For a 3D model the above expression is applicable, however, for a depth or area integrated model this expression must be integrated over the depth and is therefore not specifically required in such models. It only effects the vertical distribution of heat, but this will have an effect on the actual surface temperature gradient.

3.6. Long Wave Radiation

3.6.1. Atmospheric Long Wave Radiation Long wave radiation occurs between 9000 Å to 25000Å which is the infra-red range and is emitted from the atmosphere and the sea surface. The difference between solar radiation at the top of the atmosphere and at the water surface is the radiation absorbed by the clouds and atmosphere. The atmospheric long wave radiation is often the greatest source of heat at the water surface on cloud days. The magnitude of long wave radiation varies directly with moisture but is also affected to a lesser degree by ozone, carbon dioxide and other materials such a sulphur dioxide.

Atmospheric radiation is computed using Stefan Boltzmann’s law modified for the emissivity of the air as follows.

4 qlr   a sbTair Equation 44

Where qlr Long wave radiation  Emissivity -8 -2 -4 sb Stefan Boltzmann constant [5.67x10 Wm °K ]

The emissivity is given by

2  a  0 1 0.17Cl Ta  Equation 45

Where -5 0 Constant having a value 0.937x10 Cl Fraction of sky covered with clouds Ta Absolute temperature 2m above the surface

3.6.2. Long Wave Back Radiation The back radiation from the water surface represents a loss of heat which is most significant at night when short wave radiation is absent. The equation for back radiation is It is similar to that for atmospheric long wave radiation as follow.

4 qlr,b   a sbTs Equation 46

Where Ts Water surface absolute temperature 13

Lower Test NEP Volume 3: Appendices

3.6.3. Vegetation Re-radiation Component Vegetation either riparation or tree like, intercepts all forms for radation and then re-emits radiation charcteristic of that foliage type. The characteristation is based on the emmissitivity, but the temperature can be assumed to be the same as the ambient air tempepature. The radiation emmited by the vegetation is: 4 q    T Equation 46 ve ve sb a

Where Ta Ambient air absolute temperature °C qve Re-radiation from vegetation Wm-2 σ Stefan boltzman constant WJm-2s-1°K4 The total radiation in the shade is modified by the degree of cover and porosity of the radiating vegetation.

3.6.4. Vegetation Re-radiation Stream Surface View Factor The amount of vegetation re-radiation reaching the surface of the stream is a function of the form of the foliage and its position relative to the point on the stream surface that is being considered. In the present model the foliage view factor is represented by a continuous rectangular wall displaced vertically and laterally from the stream surface. The amount of radiation that reaches the water surface is also factored by the ratio of foliage to open space whcih is referred to as the coverage. This form of the view factor will exagerate the back radiation from foliage to stream surface slightly. A better form is to consider each individual bush or tree as being represented by the frustrum of a cone. However, the view factor is less simple. Using Stokes theorem it is possible to transform view factor double surface integrals into contour integrals. However, a numeircal integration for the veiw factor need only be done once and is not burdensome for the resolution required in this applicaiton. Even this may exagerrate the back radiation as the effective re-raditating surface is not smooth but is rather somewhat specular, although als a function of leaf type and orientation distrbution. For simplicity in the present model the the rectangular form is adopted and is illustrated in Figure 2. However, it is recognised that this would be an area for improvement.

Y2

Y1 Ground surface Water surface

X1

X2

Figure 2 Geometry of Foliage View Factor

14

Lower Test NEP Volume 3: Appendices

x2  y2   x2  y2   x2  y2   x2  y2  F  1 2 2 1 2 2 1 1 FR 2 x  x Equation 39  1 2 

3.7. Stream Bed Heat Balance Conduction occurs at the streambed and the process is well understood although the necessary data is not always easily obtained. The heat conduction at the stream bed is a function of the difference between the temperature at the stream bed and at an equilibrium ground temperature which will be some distance below the stream bed. The soil conductivity and thermal capacity will also influence this local heat balance. The heat flux through the stream bed is given by:

qd  Kg Tg Tw  zg Equation 46

qd heat flux through stream bed

Kg thermal conductivity of streambed soil or rock [1.65 Jm-1s-1°C-1 for water saturated sands and gravels[Pluhowski 1970]

Tg stream bed equilibrium temperature

Tw Stream temperature at the soil water interface

Δzg depth from the stream bed water interface at equilibrium temperature

3.7.1. Friction Generated Stream Heat Energy is lost from the flow but is transformed into heat energy which alters the water temperature. The effect is quite small but for steep streams in colder conditions can be noticeable.

q f  Cef QS f Ac Equation 46

qf heat generated through friction

Cef Coefficient of energy conversion [ρg]

Sf friction slope

Ac stream flow area

3.8. Solution Technique

3.8.1. Fractional Step Method The flow is uniform in space and steady in time. Furthermore the temperature gradients are very small and do not involve the quite steep thermal fronts that for example are associated with outfalls. It is possible to generate an analytical solution for the 1D heat transport equation, but the first and second differentials for the surface heat transfer are required. Neither of these terms is continuous about zero due to the discontinuity in the convective heat exchange at the water surface which is directionally dependent. In addition, an analytical solution, at least if it is continuous

15

Lower Test NEP Volume 3: Appendices throughout the reach of interest necessarily assumes a constant disposition of shading. A numerical solution is not computationally burdensome, and since it allows for greater flexibility, is the method chosen here.

The river reach is discretised in elements of length ∆x so that there are n=Lr/∆x elements in the computational domain. The temperature at the centre of an element at time t will be derived from a point upstream a distance x such that

Q x  t A Equation 49

Since this point will not necessarily fall exactly at an upstream node interpolation is required to obtain the correct temperature at the upstream point from which the advection over the time ∆t has occurred. The temperature chance in each element is given by A q  q  q  q   s v c sr,nett lr,nett Equation 15 Ve 0CP

Where:-

H net source term [°Cms-1]

0 density of water Cp specific heat of water [4186 Jkg-1°C-1]

3.8.2. Time and Space Step The choice of time and space step is perhaps somewhat arbitrary as this very simple scheme is unconditionally stable. Numerical experiments have shown that, at least for the flow and temperature regimes considered in this report, space steps of 10m and time steps of 60 seconds ensure that the model is almost independent of the discretisation.

16

Lower Test NEP Volume 3: Appendices

4. Application of the Model to the Lower River Test

4.1. Hydrodynamics & Transport The reach in question is relatively short and we have used a steady state approach so that the discharge is constant through the reach of interest. Velocities and water surface elevation will in space and time. The solution to the steady state flow equation is accomplished by integration from the downstream boundary where the stage discharge relationship is defined by the configuration of the structures. Inter cross section distance in the computations is between 10m and 20m. Friction is defined by Manning’s number. The river geometry is defined conventionally by a set of cross sections lying along a centerline which are fully georeferenced. The computational cross sections, from which all the processed parameters for the model are derived, are interpolated onto an accurate river centerline in order to be correctly placed relative to the foliage and oriented relative to the sun and wind. The river banks are therefore also georeferenced. The temperature gradients are very small and a simply fractional step method is used to transport temperature downstream.

4.2. Cross Section & Processed Parameters Cross sections for the model were derived from a recent EA survey. In the computations the cross sections are interpolated to ensure that the required spatial resolution is achieved and processed parameters as a function of elevation are derived. These include hydraulic radius, cross section area, river width and others that are required for the hydraulic model. The vertical resolution used is 0.1m. These so called processed parameters are pre computed by the modeling system before simulations commence. The upper levels are set to be well above what is expected during the simulations. Should the level be exceeded during simulations, the processed parameters are extrapolated, and glass walling with a vertical wall at the geometric edge of the raw cross section data is avoided as this give a false hydrodynamic extension.

4.3. Foliage Foliage provides shielding from direct solar radiation. Foliage itself will radiate at long wave at approximately air temperature. This is very slightly lower when transpiration occurs or when the leaves are wet. The foliage type can be either shrubs or trees. The foliage density varies with season so the shielding and net radiation varies seasonally. Shading of short wave solar radiation varies with the sun angle and the position of the somewhat diffuse but direct shadow will move over the river water surface. The effect on the water surface is based on the geometric intersection of the shadow and river surface. Foliage also shields against the wind which is a function of foliage density, base height and diameter at the base of the tree and near the crown. Data describing the position and approximate size and density of trees for the Test model has been derived from Google Earth. There is very little precise data on the relative shading of short wave radiation throughout the year. In order to measure this images of individual trees are required at sufficiently short intervals to provide data for analysis. Starting with an approximate description of the shading effect this has been adjusted as part of the calibration process. The model allows for large numbers of individual lines of trees or shrubs which grow sufficiently close to the river or stream to influence the solar radiation on the water surface. Although in the modeling system each individual line of trees can have the coverage individually specified, in the Test model insufficient detailed information on each line of trees meant that a global seasonal coverage was used. The final calibrated values for each month are listed in Table 2. For all trees the back radiation emissivity was set to 0.9526.

17

Lower Test NEP Volume 3: Appendices

Table 2 Seasonal Coverage for Trees

Coverage Month %age of Max Area January 10 February 10 March 15 April 30 May 70 June 85 July 100 August 100 September 100 October 70 November 40 December 10

A total of 26 groups of trees line the river banks in the model. Their positions were defined along a set of line segments within each group of trees. Within each group it was assumed that the trees were similar and their shape defined by the parameters in Table 3. The data in the table is only for one group of trees, for the full data set the calibration data in the “*.fol” files. These files define the complete foliage configuration.

Table 3 Tree Shape Definition

Parameter Value [m] Base Elevation 6.6 Height of crown 20.0 Top Diameter 8.00 Base Diameter 3.0

4.4. Flows & Abstractions Daily average naturalised, historic and flows with fully licensed abstractions were provided for the years 1996 to 2011. The model was calibrated for the year 2008 with detailed meteorological data obtained from MeteoArchive. Daily average flows are interpolated linearly to provide flows at intermediate time steps which are sufficiently well resolved for the model. In the calibration only known historic abstractions were included.

4.5. Meteorological Data The surface heat exchange model, in particular the incoming solar radiation, relies on good quality meteorological data and must include:  Wind speed and direction  Air temperature  Relative humidity  Cloud cover

Full details of the physics included in the model are given Appendix A. The data is shown in Figure 3 and Figure 4. River flow is included for convenience.

18

Lower Test NEP Volume 3: Appendices

Meteorlogical Data for Calibration & Verification

30 1.0

1] - 25 0.8

20 0.6

15

C] & Wind Speed [ms Speed Wind & C] °

0.4 Cover Cloud 10

0.2

5

Air rempterature [ [ rempterature Air

Apr

May

Mar

Aug

Sep

Jul

Oct

Dec

Nov Jan

0 Jan 0.0

Feb Feb

Jun

-

- -

-

2008

-

-

-

- -

-

-

2008 2009

-

- 2008

2008

2008

-

2008

2009 2008

2008

2008

2008

2008 2008

Air Temp Wind Speed Cloud Cover

Figure 3 Air Temperature Wind Speed & Cloud Cover

Meteorlogical Data for Calibration & Verification

30 100

90 25 80

70 20

60 1

- 15 50

s 3

40 Relative Humidity % RelativeHumidity Flow m Flow 10 30

20 5 10

0 0

Aug

Apr

Sep

Jun

Nov

Dec

Oct

May

Mar

Feb Feb

Jan Jul Jan

-

- -

-

2008

-

- -

-

-

-

-

2008 2009

2008

-

-

2008

2008 2009

2008

2008

2008

-

2008

2008

2008 2008

River Flow Relative Humidity %

Figure 4 River Flow and Relative Humidity %age

4.6. Downstream Boundary The downstream boundary is derived from the Info works model results although in the present model this represents the combined behavior of the entire Testwood Pool arrangement. The Q(η) relationship is shown in Figure 5. The upper limit is well above the maximum flow in the model calibration.

19

Lower Test NEP Volume 3: Appendices

Figure 5 Discharge Elevation Relationship

4.7. Upstream Boundary The model upstream flows are in reality derived from groundwater and streams. In the model cross section data only allowed the model to extend less than 3km upstream from Testwood Pool. If thermal profile models are to simulate the downstream temperatures well, the upstream feed waters must be as close as possible to being in correct thermal equilibrium when they reach the upstream part of the section of interest. Naturally, as measured temperature data was available at Testwood Pool there is no particular reason why a good calibration could not be achieved. The upstream boundary was therefore represented as a feeder lake. The characteristics of this lake are adjusted to provide the appropriate feed water into the river reach. This approach emulates the upper streams well. The feed lake was given a water depth of 1.5m m, which is the major influence on thermal inertia and therefore the diurnal temperature fluctuation amplitude. The area is only important when related to the shading parameters and was 20,000m2. A total area maximum area of shading for the lake was 10,000 m. The seasonal variation of shading was adjusted to minimize the difference between the simulated and measured temperatures at Testwood Pool. The effective maximum shading is listed in Table 4.

Table4 Effective maximum Shading on Virtual Feeder Lake

Month Effective Maximum Shading Jan 0.00 Feb 0.00 Mar 0.00 Apr 0.10 May 0.30 Jun 0.50 Jul 0.60 Aug 0.90 Sep 0.80 Oct 0.20 Nov 0.00 Dec 0.00 The variation shown reflects the natural variation of shading by trees and shrubs that will in reality occur on the upper reaches of the feeder streams.

20

Lower Test NEP Volume 3: Appendices

4.8. Groundwater & Soil Interaction Some proportion of the feed water is assumed to be derived from the groundwater flow and that has an associated temperature. The model assumes that the ground temperature varies sinusoidally throughout year but with the minimum and maximum temperature lagging the air temperature. In the model a proportion of the upstream flow is assumed to be derived from the groundwater and is assumed to be at the ground temperature at that point in time. An upper limit on the total groundwater flow is also assumed. The soil equilibrium temperature variations in the calibrated model are as follows:  Groundwater mean temperature : 10.5 °C  Ground water amplitude : 3.0 °C  Time of maximum : 1st July  Time of minimum : 15th January  Fraction of flow derived from g/w : 25%  Upper limit derived from g/w : 15 m3s-1

These are derived entirely empirically based on adjustment to give an optimal calibration for the summer simulations.

5. Calibration and Verification Results

5.1. Calibration The principle aim of the calibration is to ensure that the model predicts the temperatures at Testwood Pool. Numerous simulations were undertaken and parameters were adjusted. Only the final calibration results are presented here along with a single verification simulation. The model was calibrated for the year 2008 principally due to the availability of very good quality meteorological data.

Comparisons were made of the temperatures at the Testwood Pool (at the very downstream end of the model) without any abstraction. Noting that the temperature changes are a function of both depth and velocity there is a non linear relationship between flow reduction due to abstraction for example, and temperature change.

Calibation Results Run 07

20 6.0 19 18 5.0 17 16 4.0 15 14 3.0 13

12 2.0

C C

11 ° ° 10 1.0 9 8 0.0 7 6 -1.0 5 4 -2.0

3

Apr

May

Mar Aug

Sep

Jul

Oct Dec

2 Nov -3.0

Jan Jan

Feb Feb

Jun

-

- -

-

2008

-

-

-

- -

-

-

2008 2009

-

-

2008

2008

2008

-

2008

2008 2009

2008

2008

2008

2008 2008

Testwood Simulated Tempetature Testwood Error

Figure 6 Calibration Result Broadlands Flow (2008)

21

Lower Test NEP Volume 3: Appendices

Figure 7 Figure 6 shows the simulated and measured water temperatures at Testwood Pool. The time series is filtered to reduce the high frequency variations for clarity.

The maximum –ve deviation is -5.376°C (model under predicts), the maximum +ve 2.07°C (model over predicts) with a standard deviation of 1.18°C and a mean error of -0.34°C. Note that the model under predicts the temperatures in two periods in November and December. It is not clear from the data why this discrepancy has occurred as the remainder of the year matches well. For the month April to the end of September, during which period higher temperature are more of a concern, the maximum –ve temperature deviation is -2.46°C, the maximum +ve 1.9°C with a standard deviation of 0.643°C and a mean error of +0.154°C. Figure 7 shows a detailed section of measured and simulated temperatures at Testwood Pool from the model calibration. The long term trend temperature has been removed so that a comparison with the diurnal fluctuations can be made. The temperature phasing is very good and peak temperatures match well often with errors of only 0.1°C.

Calibation Results Run 07 Mid Year

1.0

0.8

0.6

0.4

0.2

C 0.0 Model ° Testwood -0.2

-0.4

-0.6

-0.8

03 17

-1.0 10

20 27 13

- - -

- - -

Aug Aug Aug

Jul Jul Jul

- - -

2008 2008 2008

- - -

2008 2008 2008

Figure 7 Detailed Time Series Segment from Summer Months

5.2. Verification The verification run serves only to check that the model still performs accurately for different scenarios. For this run all parameters that define the model behaviour must be kept identical to those for the model calibration run. Only the upstream flow and abstraction has been altered, with naturalised flows upstream and historic abstraction. The model is clearly performing well. The error in the summer month is again quite small and that largest errors occur in November and December. Note however that the shapes of the simulated and measured temperatures are very similar. It is not clear why this discrepancy occurs particularly given the good performance of the model during the summer months. It is possible that the temperature sensor suffered a drift form mid October or that an additional warmer flow into the river has been unaccounted for by the model. Note that the air temperature correlates closely with the water temperature and that in this case the water temperature during the period of greatest error in the model actually is closer to the measured air temperature than the measured water temperature.

22

Lower Test NEP Volume 3: Appendices

Verification Run 1

20 6.0 19 18 5.0 17 16 4.0 15 14 3.0 13 12 2.0

C 11

° C

10 1.0 ° 9

8 0.0 Error 7 6 -1.0 5 4 -2.0

3

Apr

May

Mar

Sep Aug

Jul

Oct Dec

2 Nov -3.0

Jan Jan

Feb Feb

Jun

-

- -

-

2008

-

-

-

- -

-

-

2008 2009

-

- 2008

2008

2008

-

2008

2009 2008

2008

2008

2008

2008 2008

Testwood Verification Error

Figure 8 Verification Result Naturalised Flow with Historic Abstraction (2008)

6. Conclusions

A very detailed model of the Lower River Test has been developed. The model has been calibrated using only one measured point for temperature but the resulting performance shows good accuracy for diurnal temperature variation particularly in the summer months. Although we have made a major assumption about the groundwater influence, the model appears to simulate the thermal evolution well. In particular the effect of abstractions on the temperatures at Testwood Pool should be modeled to a good degree of accuracy. The assumptions regarding groundwater temperature and influence appear to hold well until flows start to increase in the autumn following significant rainfall events, increased runoff and groundwater recharge. More data relating to upstream water temperatures would be required to improve the calibration of the model in these periods, although for the purposes of the evaluation of temperature effects in the summer months and early autumn in the Lower Test NEP investigation, the model is considered to be fit for purpose.

23

Lower Test NEP Volume 3: Appendices

Appendix 5.2.1 Hydraulic Model Outputs

1 Introduction

This Appendix comprises the output of the hydraulic model for Assessment Points 1 to 5 (see Figure (4.1.1 in the Main Report) in the form of tables and charts.

2 Flow discharges

Table 1 Flow discharges at different exceedance values for the different scenarios (Ml/d) for Assessment Point 1

Closed Partially open Fully open Q- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 1675 1759 1811 1675 1759 1811 1675 1759 1811 10 1295 1375 1432 1295 1375 1432 1295 1375 1432 50 496 572 632 496 572 632 496 572 632 80 230 305 367 230 305 367 230 305 367 90 158 233 295 158 233 295 158 233 295 95 126 190 262 126 190 262 126 190 262 98 91 161 228 91 161 228 91 161 228 40 3,250

35 3,000

2,750 30 2,500

2,250 25

2,000

/s) 3 20 1,750

1,500

Flow (m Flow Flow (Ml/d) Flow 15 1,250

1,000 10 750

500 5 250

0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 1 Flow duration curve for Assessment Point 1 1

Lower Test NEP Volume 3: Appendices

Table 2 Flow discharges at different exceedance values for the different scenarios (Ml/d) for Assessment Point 2

Closed Partially open Fully open Q- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 2256 2338 2393 2256 2338 2393 2256 2338 2393 10 1702 1786 1838 1701 1786 1838 1701 1786 1838 50 644 720 780 644 720 780 644 720 780 80 341 415 478 341 415 478 342 416 478 90 260 334 397 261 334 397 261 334 397 95 228 294 364 228 294 364 228 294 364 98 189 257 325 188 258 325 188 257 324

40 3,250

35 3,000

2,750 30 2,500

2,250 25

2,000

/s) 3 20 1,750

1,500

Flow (m Flow Flow (Ml/d) Flow 15 1,250

1,000 10 750

500 5 250

0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 2 Flow duration curve for Assessment Point 2

2

Lower Test NEP Volume 3: Appendices

Table 3 Flow discharges at different exceedance values for the different scenarios (Ml/d) for Assessment Point 3

Closed Partially open Fully open Q- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 2256 2339 2392 2256 2339 2392 2256 2339 2393 10 1702 1786 1838 1702 1787 1838 1699 1786 1838 50 644 721 780 644 721 780 644 721 780 80 341 415 478 342 416 478 343 417 478 90 260 334 397 262 334 397 262 334 397 95 228 294 364 229 294 364 227 293 364 98 188 257 325 188 258 324 186 257 323 40 3,250

35 3,000

2,750 30 2,500

2,250 25

2,000

/s) 3 20 1,750

1,500

Flow (m Flow Flow (Ml/d) Flow 15 1,250

1,000 10 750

500 5 250

0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 3 Flow duration curve for Assessment Point 3

3

Lower Test NEP Volume 3: Appendices

Table 4 Flow discharges at different exceedance values for the different scenarios (Ml/d) for Assessment Point 4

Closed Partially open Fully open Q- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 2102 2175 2223 2120 2194 2241 2161 2234 2282 10 1610 1685 1732 1628 1703 1749 1664 1742 1789 50 644 720 779 644 721 780 644 721 780 80 341 415 478 342 416 478 343 417 479 90 260 334 397 262 334 397 262 335 398 95 228 294 364 229 294 364 227 293 364 9 188 257 325 188 258 324 185 256 323 8

40 3,250

35 3,000

2,750 30 2,500

2,250 25

2,000

/s) 3 20 1,750

1,500

Flow (m Flow Flow (Ml/d) Flow 15 1,250

1,000 10 750

500 5 250

0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 5 Flow duration curve for Assessment Point 4

4

Lower Test NEP Volume 3: Appendices

Table 5 Flow discharges at different exceedance values for the different scenarios (Ml/d) for Assessment Point 5

Closed Partially open Fully open Q- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 1196 1244 1276 1269 1317 1348 1372 1418 1450 10 872 921 951 950 998 1027 1077 1111 1138 50 323 361 390 424 460 489 553 607 647 80 176 212 243 293 323 348 282 352 410 90 137 172 203 238 289 316 206 273 333 95 121 153 187 205 266 302 171 236 299 98 102 135 168 166 231 282 132 197 261

40

2,000 35 1,750 30 1,500 25

1,250

/s) 3 20

1,000

Flow (m Flow Flow (Ml/d) Flow 15 750

10 500

5 250

0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 5 Flow duration curve for Assessment Point 5

5

Lower Test NEP Volume 3: Appendices

3 Flow velocities

Table 6 Flow velocities at different exceedance values for the different scenarios (m/s) for Assessment Point 1

Closed Partially open Fully open V- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 0.76 0.77 0.77 0.82 0.83 0.83 1.03 1.04 1.04 10 0.74 0.75 0.76 0.81 0.82 0.82 1.01 1.02 1.03 50 0.57 0.61 0.64 0.64 0.69 0.71 0.73 0.78 0.82 80 0.38 0.46 0.52 0.45 0.54 0.60 0.45 0.55 0.62 90 0.30 0.40 0.46 0.34 0.45 0.54 0.34 0.45 0.54 95 0.25 0.34 0.42 0.28 0.39 0.49 0.28 0.39 0.50 98 0.20 0.30 0.39 0.21 0.34 0.45 0.21 0.34 0.45

1.2

1.0

0.8

0.6 Velocity (m/s) Velocity

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 6 Velocity duration curve for Assessment Point 1

6

Lower Test NEP Volume 3: Appendices

Table 7 Flow velocities at different exceedance values for the different scenarios (m/s) for Assessment Point 2

Closed Partially open Fully open V- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 0.59 0.59 0.60 0.62 0.62 0.62 0.69 0.69 0.69 10 0.57 0.58 0.58 0.60 0.61 0.61 0.68 0.68 0.68 50 0.36 0.39 0.40 0.42 0.43 0.45 0.56 0.58 0.59 80 0.25 0.28 0.31 0.34 0.35 0.36 0.52 0.53 0.53 90 0.21 0.24 0.27 0.32 0.32 0.34 0.51 0.52 0.53 95 0.19 0.22 0.26 0.32 0.32 0.33 0.48 0.50 0.51 98 0.16 0.20 0.24 0.30 0.31 0.32 0.43 0.47 0.49

1.2

1.0

0.8

0.6 Velocity (m/s) Velocity

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 7 Velocity duration curve for Assessment Point 2

7

Lower Test NEP Volume 3: Appendices

Table 8 Flow velocities at different exceedance values for the different scenarios (m/s) for Assessment Point 3

Closed Partially open Fully open V- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 0.44 0.44 0.45 0.45 0.46 0.46 0.52 0.52 0.52 10 0.39 0.40 0.40 0.41 0.42 0.42 0.52 0.52 0.52 50 0.27 0.28 0.29 0.31 0.33 0.34 0.41 0.44 0.46 80 0.18 0.21 0.23 0.23 0.24 0.26 0.33 0.35 0.37 90 0.15 0.18 0.20 0.21 0.22 0.24 0.30 0.33 0.35 95 0.13 0.16 0.19 0.21 0.21 0.23 0.28 0.31 0.33 98 0.11 0.14 0.17 0.20 0.21 0.21 0.25 0.29 0.32

1.2

1.0

0.8

0.6 Velocity (m/s) Velocity

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 8 Velocity duration curve for Assessment Point 3

8

Lower Test NEP Volume 3: Appendices

Table 9 Flow velocities at different exceedance values for the different scenarios (m/s) for Assessment Point 4

Closed Partially open Fully open V- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 0.36 0.37 0.37 0.38 0.39 0.39 0.47 0.47 0.47 10 0.33 0.33 0.33 0.35 0.36 0.36 0.47 0.47 0.47 50 0.24 0.25 0.26 0.27 0.29 0.30 0.35 0.37 0.39 80 0.15 0.18 0.20 0.18 0.21 0.23 0.25 0.27 0.30 90 0.12 0.15 0.17 0.17 0.18 0.20 0.21 0.24 0.27 95 0.11 0.14 0.16 0.16 0.17 0.19 0.19 0.23 0.25 98 0.09 0.12 0.15 0.15 0.17 0.18 0.17 0.21 0.24

1.2

1.0

0.8

0.6 Velocity (m/s) Velocity

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 9 Velocity duration curve for Assessment Point 4

9

Lower Test NEP Volume 3: Appendices

Table 10 Flow velocities at different exceedance values for the different scenarios (m/s) for Assessment Point 5

Closed Partially open Fully open V- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 0.24 0.25 0.25 0.27 0.27 0.28 0.34 0.35 0.35 10 0.20 0.20 0.21 0.22 0.23 0.23 0.33 0.34 0.34 50 0.11 0.12 0.12 0.17 0.17 0.17 0.30 0.31 0.31 80 0.08 0.09 0.09 0.16 0.16 0.16 0.23 0.25 0.27 90 0.06 0.08 0.08 0.16 0.16 0.16 0.19 0.22 0.25 95 0.06 0.07 0.08 0.16 0.16 0.16 0.17 0.21 0.23 98 0.05 0.06 0.07 0.15 0.15 0.16 0.15 0.19 0.22

1.2

1.0

0.8

0.6 Velocity (m/s) Velocity

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 10 Velocity duration curve for Assessment Point 5

10

Lower Test NEP Volume 3: Appendices

4 Water levels

Table 11 Water levels at different exceedance values for the different scenarios (m) for Assessment Point 1

Closed Partially open Fully open WL- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 3.16 3.18 3.20 3.13 3.16 3.18 3.08 3.11 3.12 10 2.99 3.02 3.03 2.96 2.99 3.00 2.88 2.92 2.94 50 2.50 2.55 2.59 2.41 2.47 2.51 2.31 2.35 2.38 80 2.27 2.34 2.38 2.18 2.22 2.27 2.18 2.22 2.25 90 2.20 2.27 2.32 2.13 2.18 2.21 2.13 2.18 2.21 95 2.17 2.23 2.29 2.11 2.15 2.20 2.11 2.15 2.19 98 2.12 2.20 2.26 2.08 2.13 2.17 2.08 2.13 2.17

4.0

3.5

3.0

2.5

2.0 Stage (m) Stage 1.5

1.0

0.5

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 11 Water level duration curve for Assessment Point 1

11

Lower Test NEP Volume 3: Appendices

Table 12 Water levels at different exceedance values for the different scenarios (m) for Assessment Point 2

Closed Partially open Fully open WL- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 3.07 3.10 3.11 3.04 3.07 3.08 2.98 3.00 3.02 10 2.91 2.94 2.96 2.88 2.90 2.92 2.79 2.82 2.84 50 2.47 2.51 2.55 2.36 2.41 2.45 2.11 2.17 2.22 80 2.25 2.31 2.35 2.06 2.16 2.22 1.79 1.88 1.95 90 2.18 2.25 2.30 1.87 2.04 2.13 1.68 1.77 1.85 95 2.15 2.21 2.27 1.75 1.97 2.09 1.63 1.72 1.80 98 2.11 2.18 2.24 1.61 1.84 2.03 1.57 1.67 1.76

4.0

3.5

3.0

2.5

2.0 Stage (m) Stage

1.5

1.0

0.5

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 12 Water level duration curve for Assessment Point 2

12

Lower Test NEP Volume 3: Appendices

Table 13 Water levels at different exceedance values for the different scenarios (m) for Assessment Point 3

Closed Partially open Fully open WL- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 2.97 2.99 3.00 2.94 2.96 2.97 2.85 2.87 2.88 10 2.84 2.86 2.87 2.79 2.82 2.83 2.68 2.71 2.73 50 2.44 2.48 2.51 2.32 2.37 2.41 2.03 2.09 2.13 80 2.24 2.29 2.33 2.03 2.13 2.19 1.70 1.80 1.87 90 2.17 2.23 2.28 1.84 2.02 2.10 1.59 1.69 1.77 95 2.14 2.20 2.25 1.70 1.95 2.06 1.53 1.63 1.72 98 2.11 2.17 2.22 1.53 1.81 2.00 1.46 1.57 1.67

4.0

3.5

3.0

2.5

2.0 Stage (m) Stage

1.5

1.0

0.5

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 13 Water level duration curve for Assessment Point 3

13

Lower Test NEP Volume 3: Appendices

Table 14 Water levels at different exceedance values for the different scenarios (m) for Assessment Point 4

Closed Partially open Fully open WL- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 2.95 2.97 2.98 2.91 2.93 2.94 2.82 2.84 2.85 10 2.82 2.84 2.85 2.77 2.80 2.81 2.65 2.68 2.70 50 2.43 2.47 2.50 2.31 2.36 2.40 2.01 2.07 2.10 80 2.23 2.29 2.33 2.03 2.12 2.18 1.69 1.78 1.85 90 2.17 2.23 2.27 1.83 2.01 2.10 1.57 1.67 1.75 95 2.14 2.20 2.25 1.70 1.94 2.05 1.51 1.62 1.71 98 2.10 2.17 2.22 1.52 1.81 2.00 1.45 1.56 1.65

4.0

3.5

3.0

2.5

2.0 Stage (m) Stage

1.5

1.0

0.5

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 14Water level duration curve for Assessment Point 4

14

Lower Test NEP Volume 3: Appendices

Table 15 Water levels at different exceedance values for the different scenarios (m) for Assessment Point 5

Closed Partially open Fully open WL- Fully Fully Fully Percentile Historic Naturalised Historic Naturalised Historic Naturalised licensed licensed licensed 5 2.94 2.96 2.97 2.90 2.92 2.93 2.80 2.82 2.83 10 2.81 2.83 2.84 2.76 2.78 2.80 2.64 2.67 2.68 50 2.43 2.47 2.50 2.31 2.36 2.39 2.00 2.05 2.09 80 2.23 2.29 2.33 2.02 2.12 2.18 1.68 1.77 1.84 90 2.17 2.23 2.27 1.83 2.01 2.09 1.57 1.66 1.74 95 2.14 2.20 2.25 1.69 1.94 2.05 1.51 1.61 1.70 98 2.10 2.17 2.22 1.51 1.80 1.99 1.45 1.55 1.64

4.0

3.5

3.0

2.5

2.0 Stage (m) Stage

1.5

1.0

0.5

0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Exceedance (%)

Fully licenced, closed Historic, closed Naturalised, closed Fully licenced, partially open Historic, partially open Naturalised, partially open Fully licenced, fully open Historic, fully open Naturalised, fully open

Figure 15 Water level duration curve for Assessment Point 5

15

Lower Test NEP Volume 3: Appendices

Appendix 6.1 Fish counter data

1. Introduction

Various forms of fisheries data has been provided by the Environment Agency for use in the Testwood NEP. This includes records of net catches from Mudeford, time series of salmon counts from the River Test fish counters, rod catch data and fisheries statistics including estimated returning stock, spawning escapement and emigrating smolt estimates. This data forms the basis of the fish assessment for the Testwood NEP, and much of this data has not been presented in the level of detail considered within the report, particularly to such a wide audience. In the interests of being open and transparent it is important to understand and communicate how the data is captured and derived, what some of the limitations of the data might be, and how significant they are to the overall conclusions of the project. The purpose of this technical note is to provide this information as an appendix accompanying the NEP report.

2. Mudeford Net Catch Data

There is a licensed commercial salmon seine net fishery in operation at Mudeford at the entrance of Christchurch Harbour, known as the “Mudeford catch”. During the netting season, June and July each year, salmon that are caught are weighed and measured, and these measurements are then used to indicate sea age of each fish. This data is reported to the Environment Agency. It is important to note that the fish passing through Mudeford are not all destined to enter Southampton Water and the River Test, many will end up migrating up the south west chalk rivers such as the Avon and Frome, although some may migrate up the Rivers Test and Itchen. It is generally accepted that the Mudeford catch data is a good indication of the timing of arrival of salmon from sea. For the purposes of this NEP investigation, Mudeford catch data has been supplied by the Environment Agency for the period of 2000 to 2010 and this has been used to indicate timing of salmon arrival from sea rather than an indication of absolute numbers entering Southampton Water and the River Test.

3. Summary Salmon Data

Salmon populations on the River Test have been monitored in one form or another for decades. Historically this was done primarily by collating rod catch information from all the fisheries along the River, and these records are available as far back as 1928. In addition to this, other population measures and estimates have been made including: estimates of returning stock; spawning escapement; and emigrating smolt estimates. The EA have provided this “summary salmon data” for the period of 1990 to 2008. Further data of this type is available in the publication “Annual Assessment of Salmon Stocks and Fisheries in England and Wales 2010” (CEFAS and Environment Agency, 2010). Summary data of this type includes:

 Estimated Returning Stock – The Returning Stock Estimate (RSE) intends to give an estimate of the stock before taking account of losses during the return migration. Historically the fisheries would have removed a significant number of salmon prior to spawning, although this is now reduced to practically nil since adoption of the catch and return policy.

1

Lower Test NEP Volume 3: Appendices

Prior to catch and return, RSE was defined as the sum of the validated salmon count and the reported catch below the counter. For the years since 100% catch and return has been in operation the RSE is calculated based on fish counter data alone.

However, it is important to note that the sum of up count data used to indicate migration activity on any given day will often differ from the RSE for a number of reasons including that the RSE includes a correction for downtime of the fish counter (see Table 1) and includes an element of down fish counts. The frequency of up fish events is thus a measure of migration activity, and the RSE is intended as a best estimate of returning stock magnitude.

 Spawning Escapement - The best estimate of the number of spawners after the salmon fishery has completed. Used to estimate the egg deposition  Rod Catch – data obtained from rod catch returns of individual fishermen to the Environment Agency as part of rod licensing requirements. This is not the same as the fishery records of catch that individual fisheries hold, and there can be differences between the two if returns or catches are not reported correctly either to the fishery or to the EA.  Catch & Release - The number of salmon reported by the salmon fisheries that are returned  Egg deposition – The estimated number of eggs deposited based on the estimated number of spawners, sex ratio and fecundity  Conservation Limit Compliance - The proportion of eggs deposited compared to the conservation limit (eggs)  Emigrating Smolt Estimate – derived from mark-recapture procedures for the purpose of population estimation. Estimated emigrating salmon smolt population based on mark-recapture procedures, estimates are from sites between near the Testwood intake and Romsey depending on the year

4. Historic Rod Catch Data

Rod catch records have generally been kept by the individual fisheries as far back as 1928. As part of this investigation rod catch data held by Broadlands, Nursling and Testwood fisheries has been transcribed from records into digital format and used in this assessment. This was undertaken by David Solomon with permission from the relevant fisheries.

5. Fish Counter Data

In addition to the high level information discussed above, data from the fish counters has also been supplied by the EA in the form of upwards migration counts used in the fish assessment and the following section provides background to this. N.B this data is upwards movement only and does not take account of downwards movements. This is considered by the EA to give a good indication of migratory activity but should not be considered ‘net upwards count’.

5.1. Background There are two fish counters on the River Test, one on the Great Test at Nursling Mill and the other on the Little Test at the bifurcation of the Great Test / Little Test just upstream of Conagar Bridge (the Coleridge Award structure). Figure 1 shows the exact locations.

Both fish counters are the resistivity type which work by electrical methods. The general principle is to mount a series of electrodes across the channel and pass an electric current through the water between the electrodes. Water, as a good conductor, allows the current to be transmitted and then detected by the electrodes. When something passes between the electrodes it disrupts the electrical signal thereby giving a change in the resistivity data pattern. Analysis of these data patterns can then detect if the signal disturbance is typical of a passing salmon or something else such as debris, ice etc. Video stills from a camera mounted over the counter channel are also taken and this is used for verification purposes to check that the electrical signature is correct and detecting a fish passing.

2

Lower Test NEP Volume 3: Appendices

Figure 1: Location of the Fish Counters on the Great Test and the Little Test Contains Ordnance Survey data © Crown copyright and database right 2012

Figure 2: The Great Test salmon counter at Nursling Mill

3

Lower Test NEP Volume 3: Appendices

The direction of travel of the fish can be detected from the time step change in signal between the three electrodes and thus the direction of travel (upstream or downstream) can be detected and movement back and forth across the counter can be detected.

These counters have been in place for approximately 25 years are considered by specialists in the field as one of the best examples of fish counters in the country. They are however not perfect and because this assessment relies heavily on these datasets an assessment of the limitations of the datasets needs to be understood.

5.2. Fish ‘leakage’ around the counter

Figure 2 shows the fish counter at Nursling (Great Test). The counter is located on a weir structure across the main channel. It is generally accepted that this weir provides adequate fish passage under all flow conditions. Although it is physically possible for salmon to leap the weir alongside the counter (and hence bypass the counter) it is generally thought that they are far more attracted to the white water coming through the fish pass and therefore particularly with salmon it is thought that the vast majority of fish moving up and down the river are counted at this counter.

The Little Test fish counter is located in an enclosed fish pass that runs in parallel to the main flow control structure (the Coleridge Award structure). The Coleridge award consists of two bottom-opening penstock gates situated in parallel, and at the time of visit on 21st July 2011 one penstock was fully open and the other was fully shut. Such a configuration would provide attractive flows to salmon moving up the Little Test and it is likely that this attraction would be stronger than that encountered through the fish counter. As such there is the potential risk that salmon bypass the counter in favour of the stronger attractive flow through the penstock – a situation that could lead to a significant undercount at the fish counter in the parallel channel.

To reduce this risk, an electric screen as been applied to the open penstock gate. This screen operates in a similar way to electro fishing, except at a lower level with the intention to deter fish rather than incapacitate them.

Flat plates are mounted either side of the penstock, with the electrode on the upstream side of the penstock acting as an anode. As fish approach the penstock they encounter a progressively increasing electric field which impairs their swimming as they get closer to the hatch. This is supposed to discourage them from passing through the penstock, despite the more attractive flow signature, and encourage them through the fish counter channel where they can be included in the count.

It is important to note however that the level of confidence in the effectiveness and consistency of operation of the fish screen is generally regarded as low. When the electric screen breaks down it is impossible to detect without being on site and measuring the signal between the two plates.

There is therefore a reasonable risk that salmon are able to bypass the Little Test counter and evade being included in the count, which would lead to under count.

5.3. Data processing to derive salmon counts

It is important to note that the fish counters are not ‘salmon counters’, they detect changes in resistivity signal across the electrodes which can be caused by various types of fish or by debris or ice. The data signals received from the fish counters therefore need careful processing and verification to extract salmon counts with a high degree of confidence. The EA undertake this data processing and the methodology is not already documented, however the following information has been provided by the EA which outlines the process followed and highlights areas of potential inaccuracies.

1) The wave forms and still images that the counters record ( the “events”) are compiled into a monthly database (N.B. an “event does not necessarily constitute a salmon movement, it is merely defined as a record made by the counter and so could also be due to weed, general ‘noise’ or could be a combination of all three) 2) Bespoke software, developed by the EA local fisheries representative, then interrogates the database, stepping through the “events” and verifying the counts against wave form to ensure the 4

Lower Test NEP Volume 3: Appendices

event constitutes a genuine salmon movement. This process is verified using the video still images where necessary although sometimes the images are obscured by turbidity / spiders / weed etc and hence can only used as a backup to wave form data. This produces a list of events with an operator affirmed identity although where a fish cannot be seen but the waveform indicates it has fish characteristics it is identified as such. 3) The software verification process produces a list giving all the signal parameters from an individual “event” 4) The EA then undertakes a manual checking and filtering process to reconcile the raw data against the outputs from the software (which is considered to be insufficiently accurate alone) and to remove “non-salmon” events (this process is done by eye). 5) Prior observations of fish and their wave form size given an indication of a threshold to separate fish such as salmon and sea trout as well as down migrating eels. 6) Events that are evaluated as noise or weed etc. are indicated as such. 7) These data are compiled into groups by month from May to December. 8) Where downtime has occurred (see Table 1) or data lost an attempt to account for the downtime with a pro-rata scaling of the collected data for that month. In many cases, best guesses are made when equipment has been faulty or when a confused signal has been received in the data. As such the data is not considered 100% accurate in terms of absolute numbers – however it does give a reasonable indication of the net upstream movement of fish.

Table 1 Fish Counter Summary: counts and down times 1996 - 2010

Total Great Test Salmon Up Salmon Up % Counter % Counter Year Salmon Up Count Count Downtime Downtime Count May-Aug Sept-Dec May-Aug Sept-Dec

1996 368 126 (34%) 242 (66%) 0 0

1997 264 70 (27%) 194 (73%) 0 6% (7d)

1998 705 164 (23%) 541 (77%) 0 0

1999 656 204 (31%) 452 (69%) 0 0

2000 579 109 (19%) 470 (81%) 0 0

2001 No data No data No data No data No data

2002 943 161 (17%) 782 (83%) 0 0

2003 406 106 (26%) 300 (74%) 0 0

2004 577 66 (11%) 511 (89%) 0 5% (6d)

2005 727 107 (15%) 620 (85%) 0 0

2006 634 28 (4%) 606 (96%) 0 0

2007 517 110 (21%) 407 (79%) 0 0

2008 402 116 (29%) 286 (71%) 70% (85d) 12% (15d)

2009 202 120 (59%) 82 (41%) 12% (15d) 40% (49d)

2010 591 164 (28%) 427 (72%) 25% (30d) 2% (3d)

5

Lower Test NEP Volume 3: Appendices

Appendix 6.2 Fish Count Charts: Little Test

Charts showing yearly information over 1997 to 2007 for the Little Test are presented below.

1997: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

1998: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

1999: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

1 Lower Test NEP Volume 3: Appendices

2000: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon L Test Flow (m3/s)

2001: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec

L Test Salmon Count L Test Flow (m3/s)

2002: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

2 Lower Test NEP Volume 3: Appendices

2003: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

2004 : Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

2005: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

3 Lower Test NEP Volume 3: Appendices

2006: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec L Test Salmon Count L Test Flow (m3/s)

2007: Little Test 4 80

70

3 60

50

2 40 Flow

30 Count Salmon

1 20

10

0 0 01May 01Jun 01Jul 01Aug 01Sep 01Oct 01Nov 01Dec

L Test Salmon Count L Test Flow (m3/s)

4 Lower Test NEP Volume 3: Appendices

Appendix 6.3 Solomon 2005 Report to the Environment Agency

Fish passage at Testwood and on the Little River Test.

Dr David Solomon, Fisheries Consultant. February 2005.

Declaration of interest.

I have fished for salmon at both Testwood and Nursling on a number of occasions in the past and hope to do so in the future.

Introduction

In September 2004 I was asked by Mr Andy Thomas of the Environment Agency to provide an independent view of the fish passage situation on the two main branches of the Lower Test; at Testwood on the main river, and throughout the Little River Test. The proportion of the whole river catch taken by the two fisheries on these two branches has been increasing in recent years. Concern has been expressed by salmon angling interests further upstream that some aspect of the in-river structures on the two branches, or the way that they were managed, may be causing an increased hindrance to the free migration of the fish.

Qualifications and experience

I am a fishery scientist with extensive experience of chalkstream salmon. I have been involved in studying salmon migration on the Itchen, Test, Avon, Frome and Piddle, mainly by radio tracking; I have personally radio-tagged and tracked more than 1000 salmon on the rivers of Southern Britain. I have a good knowledge of the salmon fisheries of the Test having undertaken a major study on behalf of the proprietors of the Testwood, Nursling and Broadlands fisheries which resulted in the production of a report entitled “River Test Salmon Fisheries – an analysis of their history and factors affecting their performance”. In recent years I have conducted surveys of obstructions and fish passage facilities on the Avon, Piddle and Frome for the Environment Agency.

I have known the sites of concern to this report for almost 20 years and have visited them on many occasions, including three times in 2004. On September 21st I visited the river specifically to inspect the various structures, in the company of Mr Graham Purbrick (Proprietor of both the Testwood and Nursling fisheries) and Mr Martin Donovan (Manager of the Nursling fishery). I stress that this was a visual inspection only, and no measurements or detailed surveys were undertaken.

Migratory behaviour of summer salmon in chalk streams

The tendency for salmon returning in the summer to remain for some months in the lowermost reaches has been observed on all the Southern chalk streams. The 1 Lower Test NEP Volume 3: Appendices tendency is greatest in summers with low freshwater flows, and associated high water- temperatures. In turn, the tendency is reduced in years with high summer flows. It occurs in the absence of structures at or close to the tidal limit; for example, radio tracking and other studies have shown that large numbers of fish build up around the tidal limits on the Frome and Piddle, where the nearest in-river structures (both easily-passed gauging weirs) are several kilometres upstream. Although there are structures on the Avon close to the tidal limit (Knapp Mill and the Great Weir on the Royalty Fishery), detailed radio tracking of several hundred individual fish shows that the large numbers that build up downstream of there, especially in low-flow summers, have not ascended as far as the structures before they decide to stop in the area. Very few fish move upstream beyond this lowermost fishery at residual flows at the tidal limit below 8m3/sec; again, however, it is stressed that this is a behavioural disinclination to migrate upstream under such conditions and not due to an inability to do so due to obstructions.

The situation is quite different in the spring, with few of the fish entering the rivers before the end of May remaining for significant periods in the lowermost reaches.

The situation on the Test

There is no doubt that a large proportion of the run of salmon into the Test spends much of the summer around the tidal limit of the river at Testwood, and that the fishery there makes large catches of these fish. Similarly, many salmon also spend many weeks in the Little River, and significant numbers are caught by anglers, although here they are rather more evenly distributed.

Testwood Mill and Pool

The current situation at Testwood Mill represents a classic example of the tendency for summer fish to remain in the lowermost reaches for a number of reasons.

First, grilse currently dominate the run on the Test, and these fish typically enter the river between late June and early August, well into the period of flow recession. Only 17.5% of the fish caught in 2004 at Testwood were over 8.5 lb, the locally applied division between grilse and two-sea-winter fish. At the time of the peak catches at Broadlands between about 1955 and 1975, when catches there averaged in excess of 300 fish, the run was dominated by two-sea-winter fish entering the river in April through June, with a fair numbers of three-sea-winter fish entering between January and April.

Second, the last two years (2003 and 2004), when perhaps the perception of this issue have been strongest, experienced summer flows significantly below average.

Third, the abstraction for Public Water Supply at Testwood (upstream of Testwood Mill, but downstream of the split with the Little River), is likely to be exaggerating the low flow effect on the lowermost reaches of the main river. The Testwood abstraction started in September 1967 and the original licence was for 90Ml/d (a little over 1m3/second); this was increased to 136 Ml/d (1.57 m3/second) in 1982. However, the abstracted volume has generally been well within the licensed volume.

Fourth, Testwood Pool represents a very good environment for salmon to hold safely and in comfort, with deep water, steep banks, breakwaters, rocks, gravel banks, weed 2 Lower Test NEP Volume 3: Appendices beds, cover represented by white water, and a wide range of flow velocities in a limited area. Similarly, the reach immediately upstream of Testwood Pool, with deep water and numerous breakwaters, represents an attractive environment for fish inclined to take up residence; it is worthy of note that at least 50 of the Testwood catch of about 320 fish in 2004 were taken in the few hundred yards above Testwood Pool

Fifth, Testwood Pool represents good angling conditions, with the fish being easily located and amenable to a range of specialised methods that have been developed for such situations. It is also likely that the recent trend in catch and release has boosted overall catches. If we assume that being caught once does not alter the likelihood of future capture, catch and release will boost the overall catch of a fishery with a high exploitation rate much more than one with a low exploitation rate. For example, if a fishery exerted a 50% catch rate, a quarter of the stock (half of the fish released) would be caught a second time, an eigth three times, and so on, virtually doubling the overall catch figure. However, if the catch rate was only 10%, then 1% would be caught a second time, 0.1% a third time, and so on, boosting the overall catch by only a little over 11%.

Sixth, the stock of salmon in the river has undoubtedly been lower in recent years than it was at the peak of catches forty years ago. It is likely that there is a tendency for the river to “fill up from the bottom” during summer, with limited runs of fish finding good holding conditions in the lowermost reaches. As runs increase there will be increased competition for the best lies and a greater tendency for part of the stock to move upstream.

It is my view that the structures at Testwood Mill do not represent a significant obstacle to upstream migration of fish inclined to go. At Testwood Mill the main hatches represent an easy migration route, especially with the downstream level at half tide or above (Figure 1). The head difference is reduced to just a few tens of centimetres, and the three hatches are generally well open; on the occasions of the three site visits in 2004 the gates were lifted by about 50 cm. At 15oC, a typical 60 cm grilse can maintain a burst speed of about 4.7 m/sec for up to a minute (Beach 1984); such a velocity would only be exceeded through the main hatches when the head difference was in excess of 1.25 m. The considerable volume passing through the main hatches, with the flow passing across the width of the pool, represents an ideal lead to enable fish to find this route.

Concern has been expressed that fish could become “trapped” in the area of the pool beyond the main hatches, and possibly attracted to the flow passing beneath the flood control gates yet be unable to pass. There is a fish pass from this area of the pool, and a third flow through a culvert from the river upstream of the main hatches. In my view any fish in this part of the pool that decided it wished to migrate upstream would have no difficulty in locating the main hatches, irrespective of the passability of the other three potential routes upstream. The distance from the flood control sluices to the main hatches is only of the order of 70 m, and the underwater sound of the flow from the main hatches would be readily detected by, and attractive to, fish deciding to migrate. It is my experience that salmon that are having difficulty ascending an obstruction are generally apparent, constantly jumping in the white water and swimming rapidly but vainly against the flow. Although salmon and sea trout are often seen jumping, swirling and generally “sloshing about” in the pool, there is no build-up of frustrated fish close to the hatches. 3 Lower Test NEP Volume 3: Appendices

Figure 1. The main hatches at Testwood, with the downstream level somewhat below high tide.

Although the hatches at Testwood are in my view quite passable by salmon and grilse, there is no doubt that Testwood Pool does, for the reasons discussed above, act as a “honeypot” for fish that are inclined to stop in the general area. If the mill and pool were not there it is likely that the fish would spread out more evenly over the lowermost kilometre or two of the river; and this is what happens in the Little River.

The Little River Test (Nursling Fishery)

Although there are two hatch pools on the Little River (the “Drawing Room Pool” and Lock Hatches) these do not generally hold a disproportionate number of salmon and are not particularly productive fishing spots; the fish, and the catches, are well distributed over a kilometre or two of river. On the occasion of my visit in September 2004 numbers of salmon were visible in Cottage Hole, close to the tidal limit, and behind the various breakwaters upstream of there. There are no obstructions of any kind in this lower part of the Little River – the old hatch structure at Cottage Hole is derelict (see Figure 2). Once again the fish remain here because there are good holding areas where the fish feel safe and comfortable, and, in low flow years at least, fish arriving in summer are not inclined to progress further until later in the year. This is an entirely natural and appropriate behaviour pattern.

4 Lower Test NEP Volume 3: Appendices

Figure 2. The derelict Cottage Hole hatches on the Little River at Nursling.

Conclusions

Although large concentrations of salmon occur around the tidal limit on the both the main river and Little River in summer, with consequent concentration of angling catches in these lowermost reaches, I consider this to be the result of natural behaviour and not due to passage problems. The effect appears to have been particularly pronounced in recent years probably due to the strong preponderance of grilse in the returning stock, lower overall stocks, low-flow summers, and abstraction upstream of Testwood Mill. Further, it is likely that the move to virtually total catch and release has maintained overall catches in the fisheries in the lowermost reaches to a greater extent than those further upstream.

5

Handbook for Pisces River Test Salmon Simulation Model Peter Henderson & Richard Seaby

April 2012

Pisces Conservation Ltd IRC House The Square Pennington Lymington Hampshire SO41 8GN UK pisces@pisces‐conservation.com www.irchouse.demon.co.uk www.pisces‐conservation.com

Phone: 44 (0) 1590 674000 Fax 44 (0) 1590 675599 .

Contents Introduction ...... 1 The study area ...... 2 A summary of the model structure ...... 3 The arrival model ...... 3 The movement into the above counter reach ...... 9 Random generation of movement ...... 9 Probability of movement in relation to flow ...... 9 Temperature probability ...... 11 Further field observations supporting the use of temperature as a variable determining movement12 The division between the Greater Test and the Little Test ...... 13 The movement into the upper Test ...... 14 Additional attempts at finding relationships...... 15 Showing uncertainty ...... 16 Model validation ...... 18 Response to flow ...... 18 Typical seasonal pattern ...... 19 Information on the program and how to use the screens ...... 22 Population settings ...... 22 Below to Above Counter ...... 22 Above Counter to Upper ...... 24 Flow data ...... 24 Temperature data ...... 25 Results grids ...... 25 Chart ...... 26 Waiting time charts ...... 27 Multiple results ...... 27 Multiple charts ...... 28 Appendix 1 ...... 30 Fish arrival procedure ...... 30 Procedure for movement above the counter in relation to flow ...... 31 Procedure for taking temperature into account for movement above the counter ...... 33

Page | 1

Introduction

This report describes the salmon simulation model for the movement of salmon into the River Test, Hampshire. The program is an individual fish model - essentially, 1000 individuals divided into 1-, 2- and 3-winter fish are modelled as they progress upriver.

The program was developed using Delphi, which is based on Pascal for Windows. The program is designed to work with Windows 7, but will probably work with earlier Windows versions.

The arrival flow and temperature data sets used by the program were prepared using Excel and are read into the program as CSV files. These files must be present in the same folder as the executable code for the program to work.

To start the program, locate the executable file named SalmonModelReduced.exe and double click on the icon. The introductory screen will open and await your input.

The program is run using the 3 buttons along the top edge. They must each be pressed in turn.

Load - loads the arrival, flow and temperature datasets.

Run & Report - Runs the a single simulation and creates the salmon and their position and creates the graphs and datasets that are saved to the disk.

Multiple run - Runs the simulation the number of times stated in No. Runs. The user can change the timing of the arrival model by using Arrival Shift. To shift the arrival dates type in the number separated by commas that you wish to move the arrival model by. In this example the simulation will be run 3 times, for three different timings of the arrival model; default timing (0), 7 days before (-7), and 7 days later (7).

If you want to look at the variability within a single model, simply only have one number in the Arrival shift box

The Progress bar shows the progression of each of the runs, with the different arrival timings.

Help - Runs the help system.

The output of the model can be opened in Excel and graphical output can be exported into a large number of formats.

Page | 2

The study area

Map of the study region showing the position of the fish counter and other features is shown below, together with a sketch map.

Figure 1 A map of ther lower River Test showing the location of the fish counter

Figure 2 A sketch map of the Lower River Test system showing locations mentioned in the text.

Page | 3

A summary of the model structure The salmon movement model is an individual based model. Essentially a fixed number of fish (usually set at 1000) arrive over time into the estuary and then the movement upstream of each is simulated. Because the probability of movement of each fish on any particular day is determined in part by random simulation of a probability of movement no two simulations are identical.

The basic time unit used for the simulation is 1 day.

The habitat is envisaged as divided into a series of spatial components. The individual fish move between these components according to defined rules. Random number generators are used to determine whether over each time step a particular fish will make the transition between the different reaches of the river. The general structure of the model is like a Markov Chain, however the transition probabilities are not fixed. They can vary with flow and water temperature. They also vary with the season and potentially with the age (size) of the fish. However at present the age of the fish is not implemented explicitly as a separate variable. It is taken into account implicitly with the variation in the probability of movement over the seasons because fish which spend more than 1 winter at sea tend to arrive earlier in the year.

The movement chain is as follows:

The arrival model Movement from the estuary across the counter to the river  Movement into the Upper Test River

Because the movement of each fish is modelled individually the final output is the history of each fish plus average statistics.

Using a time step of 1 day the model covers 365 days from January 1st to December 31st. There is no account taken for leap years. The model does not force every salmon to enter the river, although in practice the vast majority, if not all, will do so. Should a salmon not pass the counter – it is considered to wait from the day of arrival to December 31st for calculation of average waiting times.

For each time step the movement of each fish present is considered in turn. The basic protocol is to calculate for the flow and temperature of the river on a particular day a probability that the fish will move. A random number generator is then used to determine if the fish actually moves. This approach results in every run of the model being unique. However, it is possible to make the program repeat the same series of random numbers by using the same random number generation seed.

The arrival model The data from the commercial fishermen at Mudeford, Christchurch Harbour entrance, can give insight into the returning behaviour of salmon. Since 1999 the fishermen are only licensed to catch during June and July. The Mudeford capture data show a distinct bimodality, the peaks are around weeks 22 and 30. In comparison, the main peak at the River Test counter is around week 42 . This initially suggests that the Mudeford fishermen are not catching fish that will contribute to the main late summer-autumn run. However, salmon radio-tagged on the Itchen before the end of July often remained at the head of tide until the autumn. All of the late-running fish could be accounted for by arrival coincident with the observations at Mudeford. It therefore seems most likely that the relationship between these time series is that shown in Figure 1 below. This explanation is also consistent with the known distribution of single and multi-winter sea fish between the early and late runs. Note that in the figure, the second peak in the captures at Mudeford is dominated by single winter sea fish, as is the late summer-autumn run on the River Test.

Page | 4

Figure 3 The temporal pattern of captures at Mudeford, Dorset and the total counts from the River Test.

The model therefore uses a seasonal arrival distribution based on the pattern of capture of 1-, 2- and 3-winter fish observed at Mudeford. The Mudeford seasonal catch data for different sea-age fish is presented in Table 1 (supplied by A. Fewings – Mudeford nets.xls).

Page | 5

Table 1 The pattern of seasonal capture observed at Mudeford during commercial salmon netting. Data supplied by Adrian Fewings.

Sea age

Week of year 1 2 3 Total 19 11 22 72 9 23 2401759 24 1542681 25 4481163 26 13 34 13 60 27 33 33 7 73 28 82 35 8 125 29 110 49 2 161 30 134 72 1 207 31 74 30 1 105 32 22

A plot of these data is shown below. Note the essentially bimodal form of the overall arrival pattern.

Figure 4 A plot of the seasonal pattern of captures at Mudeford between 2000 and 2009. Data supplied by Dr Adrian Fewings.

The arrival model is implemented using a random number generator and the data supplied to the program in the file Arrival Model.csv. Note that this file can be opened and edited using Excel or similar spreadsheet. The actual contents of this file are tabulated in Table 2 below. This table gives tabulated

Page | 6 probabilities that any fish waiting to enter the estuary on a particular day will actually do so. Note that there are different probabilities for 1, 2 and 3 winter fish. The proportions of each age group in the population are assumed by the program in the Population settings tab and can be altered prior to a simulation run.

The program initially starts with 1000 salmon which are divided between the 3 age groups. Then for each day the probability of entering the estuary below the counter is looked up in the table. So, for example, on day 154 the probabilities are 0.000568, 0.030057 and 0.060995 or 1 2 and 3 winter fish. Now if there are 100 2-winter fish in the sea on this day then using a random number generator on average about 3 of them should arrive below the counter (100 x 0.03 = 3). Remember this is chosen by random number so it may be a slightly lower or higher number. Now once these 3 (say) have arrived the number at sea waiting to arrive is reduced to 97 and the simulation proceeds on to another day. Note that before day 126 all the values are zero – no fish can arrive and after day 237 all the values are 1 which means that all the available fish have arrived irrespective of age or random variation. See Appendix 1 for a listing of the main procedure.

Table 2 The data held in the file Arrival Model.csv which is used to simulate salmon arrival.

Day 1 winter 2 winter 3 winter 1 0 0 0 2 to 124 Deleted to save space 125 0 0 0 126 0 0 0.001605 127 0 0 0.00321 128 0 0 0.004815 129 0 0 0.006421 130 0 0 0.008026 131 0 0 0.009631 132 0 0 0.011236 133 0 0 0.011236 134 0 0 0.011236 135 0 0 0.011236 136 0 0 0.011236 137 0 0 0.011236 138 0 0 0.011236 139 0 0 0.011236 140 0 0 0.011236 141 0 0 0.011236 142 0 0 0.011236 143 0 0 0.011236 144 0 0 0.011236 145 0 0 0.011236 146 0 0 0.011236 147 0 0.002364 0.014446 148 0 0.004728 0.017657 149 0 0.007092 0.020867 150 0 0.009456 0.024077 151 0 0.01182 0.027287

Page | 7

152 0 0.014184 0.030498 153 0 0.016548 0.033708 154 0.000568 0.030057 0.060995 155 0.001136 0.043566 0.088283 156 0.001704 0.057075 0.11557 157 0.002272 0.070584 0.142857 158 0.00284 0.084093 0.170144 159 0.003408 0.097602 0.197432 160 0.003976 0.111111 0.224719 161 0.00426 0.129348 0.266453 162 0.004544 0.147585 0.308186 163 0.004828 0.165822 0.34992 164 0.005112 0.184059 0.391653 165 0.005396 0.202297 0.433387 166 0.00568 0.220534 0.47512 167 0.005964 0.238771 0.516854 168 0.0071 0.254981 0.53451 169 0.008236 0.271192 0.552167 170 0.009372 0.287403 0.569823 171 0.010508 0.303614 0.58748 172 0.011644 0.319824 0.605136 173 0.01278 0.336035 0.622793 174 0.013917 0.352246 0.640449 175 0.017609 0.363728 0.661316 176 0.021301 0.375211 0.682183 177 0.024993 0.386694 0.70305 178 0.028685 0.398176 0.723917 179 0.032377 0.409659 0.744783 180 0.036069 0.421142 0.76565 181 0.039761 0.432624 0.786517 182 0.049134 0.443769 0.797753 183 0.058506 0.454914 0.808989 184 0.067878 0.466059 0.820225 185 0.077251 0.477204 0.831461 186 0.086623 0.488349 0.842697 187 0.095995 0.499493 0.853933 188 0.105368 0.510638 0.865169 189 0.128657 0.522459 0.87801 190 0.151945 0.534279 0.890851 191 0.175234 0.546099 0.903692 192 0.198523 0.55792 0.916533 193 0.221812 0.56974 0.929374 194 0.245101 0.58156 0.942215 195 0.26839 0.593381 0.955056

Page | 8

196 0.299631 0.609929 0.958266 197 0.330872 0.626478 0.961477 198 0.362113 0.643026 0.964687 199 0.393354 0.659574 0.967897 200 0.424595 0.676123 0.971108 201 0.455836 0.692671 0.974318 202 0.487078 0.70922 0.977528 203 0.525135 0.733536 0.979133 204 0.563192 0.757852 0.980738 205 0.60125 0.782168 0.982343 206 0.639307 0.806484 0.983949 207 0.677364 0.8308 0.985554 208 0.715422 0.855117 0.987159 209 0.753479 0.879433 0.988764 210 0.774496 0.889564 0.990369 211 0.795513 0.899696 0.991974 212 0.816529 0.909828 0.993579 213 0.837546 0.919959 0.995185 214 0.858563 0.930091 0.99679 215 0.87958 0.940223 0.998395 216 0.900596 0.950355 1 217 0.909969 0.95542 1 218 0.919341 0.960486 1 219 0.928713 0.965552 1 220 0.938086 0.970618 1 221 0.947458 0.975684 1 222 0.95683 0.98075 1 223 0.966203 0.985816 1 224 0.969895 0.987504 1 225 0.973587 0.989193 1 226 0.977279 0.990881 1 227 0.980971 0.99257 1 228 0.984663 0.994259 1 229 0.988356 0.995947 1 230 0.992048 0.997636 1 231 0.993184 0.997974 1 232 0.99432 0.998311 1 233 0.995456 0.998649 1 234 0.996592 0.998987 1 235 0.997728 0.999325 1 236 0.998864 0.999662 1 237 1 1 1 238 ‐ 364 Deleted to save space 365 1 1 1

Page | 9

The movement into the above counter reach We use the data from the fish counter on the Greater Test to model this transition from the Greater and Little Tests to the above counter reach.

Two variables are used to determine the likelihood that an individual fish will move on any one day, (1) Flow and (2) water temperature.

The actual probability of movement of any one day is calculated as

Actual probability = probability for the flow x probability for the temperature.

As with the seasonal arrival model from the sea a random number generator is used in conjunction with this probability to determine if an individual fish waiting below the fish counter will actually move. See Appendix 1 for listings of the main procedure.

Random generation of movement To determine if a fish moves, a simple random number generator is used. To illustrate the method used, consider a situation where the flow and temperature predict that the probability of movement on a particular day is 0.1 (10%) and there are 100 fish waiting below the counter to enter the reach above the counter.

For each fish in turn a random number between 1 and 100 is generated. If this number is less than 11 the fish is moved in to the reach above the counter, otherwise it remains below the fish counter. Given 100 fish waiting to move upriver on average 10 should move, however because movement is determined by a random generator, this can vary to some degree.

The age of the fish is not taken into account when calculating these probabilities.

Probability of movement in relation to flow There is a clear tendency for movement to be modulated by flow and change in flow. Further there is a change in behaviour over the season.

The provability of movement in relation to flow is modelled differently for the two periods as described below.

Spring and summer up to August 31st For this period the probability is determined solely by the magnitude of the flow. The following relationship was used:

3.0251. Probability = 0.00003 x flow

This relationship was derived empirically from the data by regression as shown in Figure 3.

Page | 10

Figure 5 The average probability of movement across the counter prior to September calculated from empirical data

Consideration was given as to the role of a change in flow. For this period there is not a clear relationship between a change in flow and fish movement. Many fish have been observed to move when the flow is declining. The low number of observations over this period does not allow a more sophisticated model to be developed or tested.

Autumn from 1st September It has previously been noted that upstream movements tend to occur on a positive change in flow. The model therefore uses a different equation relating the probability of upstream movement depending on whether the between-day change in flow is positive or negative. The two curves used in the model and the data from which they are derived are presented below.

For positive changes in flow between days

2.5556 Probability of movement = 0.0002 x Flow .

For negative changes in flow between days

2.2901 Probability of movement = 0.0003 x Flow .

These equations have been derived empirically and fitted by regression (see Figure 4 ).

The maximum probability is capped at MaxNeg and MaxPos of 0.1. In other words on any one day the maximum probability of movement through the fish counter is 10%. This value is derived from an analysis of movement during high flow periods later in the year when it is believed all the fish have entered the tidal river and are available to move upstream. It is clear that their actual time of movement varies greatly between individuals and is never greater than about 10% of the fish available to move.

Page | 11

Figure 6 The probability of movement across the fish counter on a rising and falling river flow for the autumn period after the 1st September. The data are derived from empirical observations. Probabilities were calculated using the arrival model generated by the Mudeford landings to predict the number of fish waiting below the counter to enter the River Test.

During initial model development it was assumed that the magnitude of the change in flow would be a key variable. In fact fish are often observed to move on extremely small positive changes. Increased willingness to move is therefore best described as a two part trigger - (1) positive change in flow and (2) absolute magnitude of flow.

Temperature probability Salmon movement is known to be affected by temperature, and a temperature effect is apparent in the River Test data. Figure 5 below shows the relationship between the numbers of salmon counted moving upriver and the water temperature.

Page | 12

Figure 7 The relationship between the number of fish counted as moving upstream and the river water temperature.

Between water temperatures of 12 and 16 C, the upper bound of the count declines from about 100 to 10, and above 20C it is 5 or less.

To reflect the unwillingness of salmon to move at higher temperatures, a simple model was used, as follows.

For temperatures below 12C probability of movement = 1.0 (it is determined by flow only).

For temperatures above 12C and below 16.4C probability of movement = -0.225 x river temperature + 3.7.

For temperatures above 16.4C probability of movement = 0.

Further field observations supporting the use of temperature as a variable determining movement

The first line of evidence derives from the study of Clarke et al (1994)1 .They reported the probability of telemetered salmon ever entering freshwater, in relation to the water temperature after release. They tagged 260 salmon over three years and tracked the progress of the fish as they entered or left the estuary. These data were reported in temperature bands of two degrees Celsius with no confidence limits on the observations. To allow prediction of salmon probability of entry at temperature increments of 0.5°C, WRc (Water Research Council) were asked by the Environment Agency to employ general linear modelling techniques, with the constraint that at temperatures less than or equal to 12°C, the probability of river entry was a constant.

1 Clarke, D. R. K. Evens, D. M., Ellery, D. S. and Purvis, W. K. 1994. Migration of Atlantic salmon (Salmo salar L.) in the River Tywi estuary during 1988, 1989 and 1990. RT/WQ/RCEU/94/7 Nation Rivers Authority Welsh Region

Page | 13

The results are presented graphically in Figure 6 below. Note that the probability of entry into freshwater drops rapidly above 12 C.

The resultant model was used to construct a lookup table, which allowed the provision of probability estimates with 90 percentile confidence limits. The mean values used for the model were considered an underestimate of probability of salmon entry into the Cleddau. The most likely real relationship is expected to lie between the mean relationship and unity; the upper 90%-ile relationship may therefore be considered a fair working estimate.

Figure 8 The probability of river entry in relation to water temperature as presented by Clark et al (1994).

General Linear Model of probability of river entry data.

The second line of evidence derives from the study of Solomon and Sambrook (2004)2. They found that the failure to enter southern English rivers was related to water quality. For the Hampshire Avon and the Exe in Devon, they studied the upstream migration of adult salmon by tagging salmon at the entrance of an estuary or harbour and observing their runs up river. Solomon & Sambrook noted that:“On the Avon, increased river temperature was associated with decreasing tendency to enter within 10 days of tagging, and virtually all fish entering the river on days when the 09.00-hour river temperature was in excess of 17°C did so in the late evening and early morning, thus avoiding the time of day of highest water temperatures.” and: "Fewer fish enter the river promptly at high river temperatures and more at low temperatures,...."

This study makes it clear that upstream movement is delayed as water temperature increases.

The division between the Greater Test and the Little Test This division was considered for inclusion in the model, but following discussion has been dropped.

In case this topic should arise again the original idea is described below.

2 Solomon D. J. and Sambrook H. T., 2004, Effects of hot dry summers on the loss of Atlantic salmon, Salmo salar, from estuaries in South West England Fisheries Management and Ecology, 11, 353–363

Page | 14

Fish may move upstream using either the Greater or the Little Test channels. These two channels are therefore modelled as separate spatial components. The division in fish between these two components is modelled using a flow model based on Fewings who published the following graph.

The equation used was as follows:

Proportion of Salmon using greater test = a*(b-exp( -c * (flow in river - little test flow)))

The movement into the upper Test Adrian Fewings has presented data comparing the catch of salmon in the upper and lower reaches of the river Test. These data have been used to analyse the transference probability between these two components. Figure 7 below shows these data.

Page | 15

Figure 9 The relationship between the total number of salmon within the River Test and the number which enter the Upper reaches of the river. Data from Adrian Fewings.

Note that below 270 fish no salmon enter the upper section. This seems logical - the salmon fill the lower section of the river first and only when this holds 270 fish do they start to move on into the upper section.

Above 270, about 40% of the fish are inclined to travel to the upper section. The model deals with this by randomly allocating to 40% of the individual fish the tendency to travel up to the top reach. However, these fish will only undertake this movement if they are not amongst the first 270.

Additional attempts at finding relationships.

In an effort to extend and improve upon the salmon model, we looked at the possibility of partitioning into one or two month segments the probability of movement through the counter. The idea was to have different transition probabilities with respect to flow, temperature etc at different months of the year. There are certainly hints that this may be the case. However, we failed in this attempt because we do not have the data required to estimate the probabilities of movement in the early part of the season. In the later part of the season, the main autumn run, we can either estimate the proportion of the fish that have arrived from the sea or, later in the run, assume that all the fish have arrived from the sea and are waiting somewhere to move into the river (or have already made their move). Given the total size of the run, plus the observed number that have already made their move, we can therefore calculate the probability of movement for an observed set of conditions as the number of fish moving through the counter divided by the estimated total number waiting.

Early in the season it is clear that only a fraction of the fish that will be observed for the year have arrived. We have the seasonal arrival pattern using the Mudiford data, but, this cannot give a reliable estimate of the number waiting early in the season. We found that it was not possible to

Page | 16 calculate sensible transition probabilities across the counter for the early months alone. This was a clear sign that estimated number waiting to move up river was not correctly estimated. Further, we are at a considerable disadvantage early in the year because of the lack of observations for fish crossing the counter. The data during the early stages, as we have observed in the past, has the characteristics of a Poisson process. It is essentially indistinguishable from a random model. We are not suggesting that the fish are behaving randomly, rather we are observing the random nature of their arrival together with their variable behaviour and the small numbers moving into the river which, taken together, create a situation which we cannot distinguish from random.

We did try a number of different divisions of time to try to generate a model with varying transition probabilities for different time periods, but reluctantly came to the conclusion that it was not possible. In fact, we concluded that there was no defensible evidence that the behaviour of the fish was different over the season and that therefore it was best to use the same transition probability model relating movement to flow throughout the annual run. The seasonal change in behaviour can be viewed as taken into account with a temperature rule.

Showing uncertainty No single run of the model shows the range of possible movements for salmon. This is in part because the movement of individual fish is modelled using a randomization procedure. However, the real issue is that we cannot be certain about the behaviour of individual fish, and so it is better if we present the model output as a range of possibilities.

Two variables come into play when determining the range of possible outcomes. First, the individual variation between fish in deciding when to travel past the fish counter. Second, the between-year variation in the arrival model. It is certain that climate and possibly other unknown factors determine when the fish arrive in the estuary ready to migrate upstream.

By undertaking repeated runs of the model and allowing the arrival model to be shifted backwards and forwards in time, a range of possible outcomes can be generated, which illustrate the envelope within which the predictions reside.

For example the illustration below shows the model using the 2002 data, under the assumption that the arrival model could be between 7 days earlier and 7 days later than that assumed by the model. This chart was generated from 50 runs. It shows the clear pattern of peaks in predicted movement, but also shows the range in possible outcomes. For example between days 260 and 280 counts on any day might range between 0 and about 12.

To give an illustration of how these results relate to the actual observations, the second graph below shows the actual results for 2002 and the predicted result from a single run of the model with the standard arrival time model assumed. Note that both the models do capture the main features such as the sudden increase in movement after day 289. However, the possible range of outcomes towards the end of the year can only be captured from the simulated range of outcomes shown in the first chart.

At the bottom a graph with 1000 replications for 2002 is shown. These can take some time to calculate and to plot.

Page | 17

Page | 18

Model validation At present no systematic model validation procedure has been undertaken. The parameters used in the model have been derived from the field observations collected over the years 1996 to 2007. It is therefore hoped that data for 2008-2011 will be shortly available to test against the model predictions. However, a number of checks have been undertaken to ensure the model gives reasonable predictions. The questions we have addressed are as follows: 1. Does the model produce the typical pattern of response to changes in flow? 2. Does the model produce the typical seasonal pattern of passage across the fish counter? 3. Does the model respond to extreme low flow/ high flow and high temperatures as experience has shown to occur?

Response to flow

As shown in the plot below, the movement of fish across the counter does show the typical response to flow expected, with large peaks occurring when flow increases.

Page | 19

Typical seasonal pattern

As shown in the plot below the model does show the expected seasonal pattern. In this case (for the year 2002) movement increased in the autumn, just as was actually observed.

As a second example the graph below is a model run and the observed counts in 1996. Note that both show an increase in movement after day 265. The seasonality of the predictions is better shown from a large number of model runs (100), as shown in the following graph.

Page | 20

Page | 21

Page | 22

Information on the program and how to use the screens

Information on the tabbed screens is given below.

Population settings This screen sets the population parameters for the returning population.

For simplicity the total number of fish moving up river is defaulted to 1000, and is set in the No. Fish Arriving box. There is no fish mortality in the model so these fish are distributed between the various localities.

The proportion of single-, two- and three-winter fish in the population is set using the Prop Age 1, 2 and 3+ boxes. It must add up to 1. You cannot add a number into age 3+, it is calculated as 1 - Age 1 - Age 2. It will show in red if the number cannot add up to 1.

Prop. fish to upper gives the proportion of the fish in the above the counter reach that can potentially enter the upper reach. The default is determined by historical data.

Below to Above Counter The movement into reach above the counter is based on an analysis of movement across the fish counter.

The relationship is in two parts - the early season relationship and the late season relationship.

There are 2 models offered, one which starts the later relationship on the 1st September, the second model switches between models on the 1st August.

To chose which model you want to run click the checkbox by Use Model

The effect of temperature can be taken into account or not by checking the box by Use Temperature relationship

Page | 23

The parameters for the early year relationship are given in the first Below to Above Counter boxes

The default day of the change between the two in model 1 is the 1st September (day 244). This can be altered in the second row on the screen in the Day of change from early relationship to late box.

In the early part of the year there is less data and no difference was found with the change in flow (delta flow).

For the later part of the year two equations are used to generate the probability of an individual fish moving in relation to flow. One is the probability given an increase in flow over a 24-hr period. The second is the probability given a decrease in flow. See The movement into the above counter reach

The parameters for the equations are given under Below to Above Counter - Pos Delta and Neg Delta

The flow probability is further controlled over the entire year by the temperature relationship given in Below to Above Counter - Temperature a and b

Page | 24 the default is set to start reducing migration at 12°C and stop it by 16.4°C. This can be changed manually, or by using the calculate slope option. Here just enter the lower and upper temperature and press calculate, this will fill the relationship boxes with the new values.

Above Counter to Upper This part of the model assumes that a certain number of fish fill the are above the counter reach before any further arrivals may enter the upper reach. The default number is 270 fish.

The proportion that may enter the upper reach of those that are potentially available to colonise (eg number above 270) is given by Prop. move after fill.

See movement to upper Test for data supporting this division.

Flow data This grid allows the user to choose the flow data to be used. This file can be edited in Excel.

The default column (column 1) is the average flow data for the river over the last 10 years.

Page | 25

Temperature data This grid allows the user to choose the temperature data to be used. This file can be edited in Excel.

The default column (column 1) is the average temperature data for the river over the last 10 years.

Results grids This screen allows you to examine the data sets that are output by the program as CSV files. It therefore allows a look at files that can be opened in Excel.

Results.csv - Is shown in the upper grid. It holds the division of fish between year classes and the average time the fish of each age waited below the fish counter before entering the reach above the counter. The next 3 columns give number of days waiting for each individual fish. The last four columns show for each month, the number of fish arriving, waiting and moving past the counter.

Salmon Position.csv - This holds for each of the 1000 salmon its position within the system on each day of the year. It also gives the winter age of the fish.

Summary.csv - This data set summarizes how many fish are in each part of the system on each day.

All these files can be opened in Excel, and are saved in the folder containing the salmon model program. They over-write on each run, so if you want to save a version of the data you must remove the files before you re-run the program.

Page | 26

Chart This screen presents a graphical output of the model run, showing where fish are on each day of the year.

Use the tick boxes on the right to select the variables for plotting.

The lower checkboxes plot flow and temperature variable on the same chart using the right axis.

The toolbar buttons (below) allow the user to print, save, and alter almost every aspect of the chart. Edit - This button will offer a wide range of options to change the style of your graph. It is also used to export or copy your graph to file, and even to email it using the 'Send' button. For more information on chart editing use the TeeChart help system available from the Help button on the chart edit box. Print - Use this button to print the graph Copy - Use this option to copy the graph to the clipboard. Save - Save the file in a variety of different formats.

Page | 27

Waiting time charts This screen graphically summarizes the waiting time distribution of the fish below the counter.

Use the check boxes to select fish age.

Use the bin size to select a suitable bin size (in days) for the histogram. To recreate the graph press the Replot button.

Multiple results This tab shows the results of multiple runs of the simulations.

Running multiple simulations

1. Set the number of runs for each simulation in No. Runs 2. Set the shift in the arrival model. 0 = no movement -1, the fish arrive a day earlier, 1 the fish arrive a day later. In this example 0, -7, 7 will run the model without movement and 7 day either side of that date. Number must be entered as 1, 2, 3 with commas between the numbers.

To investigate the variation of a single simulation by running it 100 times with the arrival model in it original position (0) the setting would look as follows.

Page | 28

The results are presented as rows for each simulation. The first cell shows the age group, the first cell of each row shows the arrival time offset. The cells show how many fish of each age passed the counter each day.

Multiple charts The results of a multiple run are plotted showing all the points and a line showing the average number per day.

Only the all age data is plotted, data for each age separately can be obtained from the Multiple Results tab

Page | 29

Page | 30

Appendix 1

The major procedures of the program are listed below.

Fish arrival procedure procedure TForm1.FishArrive(ArrivalShift: Integer); var FishNo, j, K: Integer; Test, ArrivalVal: Single; begin FSalmonList.DayofFirstArrival := 366;

for FishNo := 0 to FSalmonList.count - 1 do begin Test := Random(); for j := Day1 + ArrivalShift to Day365 + ArrivalShift do begin case FSalmonList.Salmon[FishNo].YearClass of 1: ArrivalVal := Y1Arrival[j - ArrivalShift]; 2: ArrivalVal := Y2Arrival[j - ArrivalShift]; 3: ArrivalVal := Y3Arrival[j - ArrivalShift]; end; if Test > ArrivalVal then begin FSalmonList.Salmon[FishNo].ArrivalDay := j; end else Begin for K := j - 1 to Day365 do FSalmonList.Salmon[FishNo].FishPosistion[K] := InBelowCounter; end; end; FSalmonList.DayofFirstArrival := min(FSalmonList.Salmon[FishNo].ArrivalDay, FSalmonList.DayofFirstArrival); end; end;

Page | 31

Procedure for movement above the counter in relation to flow procedure TForm1.MoveAboveCounter; var FishNo, j, K: Integer; ProbMove: D1_Sng_array; Test. TemperatureProp, TempA, TempB:Single; BelowToAboveEarly_a,BelowToAboveEarly_b, Est_To_Low_Neg_A, Est_To_Low_Neg_B, Est_To_Low_Pos_A, Est_To_Low_Pos_b, M2BelowToAboveEarly_a,M2BelowToAboveEarly_b, M2Est_To_Low_Neg_A, M2Est_To_Low_Neg_B, M2Est_To_Low_Pos_A, M2Est_To_Low_Pos_B:Single; begin Setlength(ProbMove, 366); // calculated the probablity arrays for each day

// get all the screen inputs for later use - saves cycles TempA:= AE_Est_Low_Temp_A.FloatValue;// -0.225 TempB:= AE_Est_Low_Temp_B.FloatValue;//3.7;

//Model 1 BelowToAboveEarly_a:=AE_BelowToAboveEarly_a.FloatValue; BelowToAboveEarly_b:=AE_BelowToAboveEarly_b.FloatValue; Est_To_Low_Neg_A:=AE_Est_To_Low_Neg_A.FloatValue; Est_To_Low_Neg_B:=AE_Est_To_Low_Neg_B.FloatValue; Est_To_Low_Pos_A:=AE_Est_To_Low_Pos_A.FloatValue; Est_To_Low_Pos_B:= AE_Est_To_Low_Pos_b.FloatValue;

//Model 2 M2BelowToAboveEarly_a:=M2_AE_BelowToAboveEarly_a.FloatValue; M2BelowToAboveEarly_b:=M2_AE_BelowToAboveEarly_b.FloatValue; M2Est_To_Low_Neg_A:=M2_AE_Est_To_Low_Neg_A.FloatValue; M2Est_To_Low_Neg_B:=M2_AE_Est_To_Low_Neg_B.FloatValue; M2Est_To_Low_Pos_A:=M2_AE_Est_To_Low_Pos_A.FloatValue; M2Est_To_Low_Pos_B:= M2_AE_Est_To_Low_Pos_b.FloatValue;

// get Prob of movement for the day for j := Day1 + 1 to Day365 do begin

If GB_UseTemperature.CheckBox.Checked then DailyTemperatureEffect(TempB, j, TempA, TemperatureProp) else TemperatureProp:=1;

if AdvGroupBox2.CheckBox.Checked then ProbForModel1(Est_To_Low_Pos_b, Est_To_Low_Neg_B, Est_To_Low_Neg_A, j, ProbMove, BelowToAboveEarly_b, BelowToAboveEarly_a, Est_To_Low_Pos_A) else ProbForModel1(M2Est_To_Low_Pos_B, M2Est_To_Low_Neg_B, M2Est_To_Low_Neg_A, j, ProbMove, M2BelowToAboveEarly_b, M2BelowToAboveEarly_a, M2Est_To_Low_Pos_A);

ProbMove[j] := ProbMove[j] * TemperatureProp; end;

// randomly choose a no. and check prob for FishNo := 0 to FSalmonList.count - 1 do begin

Page | 32

for j := Day1 to Day365 do begin Test := Random(); // a new chance each day if (FSalmonList.Salmon[FishNo].FishPosistion[j] = InBelowCounter) then begin if Test < ProbMove[j] then begin FSalmonList.Salmon[FishNo].MoveIntoAboveCounter := j; for K := j to Day365 do FSalmonList.Salmon[FishNo].FishPosistion[K] := InAboveCounter; // this stops it triggering again end; end; end; end; end;

Page | 33

Procedure for taking temperature into account for movement above the counter procedure TForm1.DailyTemperatureEffect(TempB: Single; j: Integer; TempA: Single; var TemperatureProp: Single); begin TemperatureProp := TempA * Temperature[j] + TempB; TemperatureProp := Max(TemperatureProp, 0); TemperatureProp := Min(TemperatureProp, 1); end; procedure TForm1.ProbForModel1(Est_To_Low_Pos_b: Single; Est_To_Low_Neg_B: Single; Est_To_Low_Neg_A: Single; j: Integer; var ProbMove: D1_Sng_array; BelowToAboveEarly_b: Single; BelowToAboveEarly_a: Single; Est_To_Low_Pos_A: Single); const MinPos = 0; MaxPos = 0.1; MinNeg = 0; MaxNeg = 0.1; begin // from here if AE_ChangeDay.IntValue >= j then begin ProbMove[j] := BelowToAboveEarly_a * Power(FlowData[j], BelowToAboveEarly_b); if ProbMove[j] < MinPos then ProbMove[j] := MinPos; end else begin // Flow relationship if DeltaFlow[j] < 0 then begin // neg flow ProbMove[j] := Est_To_Low_Neg_A * Power(FlowData[j], Est_To_Low_Neg_B); if ProbMove[j] < MinPos then ProbMove[j] := MinPos; if ProbMove[j] > MaxPos then ProbMove[j] := MaxPos; end else begin // pos flow ProbMove[j] := Est_To_Low_Pos_A * Power(FlowData[j], Est_To_Low_Pos_B); if ProbMove[j] < MinNeg then ProbMove[j] := MinNeg; if ProbMove[j] > MaxNeg then ProbMove[j] := MaxNeg; end; end; end;

Page | 34

procedure TForm1.MoveIntoUpper; var I, j, K, FishNo, TempPos, StartToMove: Integer; Test: Single; ProbMoveNew: D1_Sng_array; NoFishAboveCounter: D1_Sng_array; begin StartToMove := 1; Setlength(ProbMoveNew, 366); Setlength(NoFishAboveCounter, 366);

for FishNo := 0 to FSalmonList.count - 1 do begin if FSalmonList.Salmon[FishNo].MoveIntoAboveCounter<> 0 then begin TempPos := FSalmonList.Salmon[FishNo].MoveIntoAboveCounter; NoFishAboveCounter[TempPos] := NoFishAboveCounter[TempPos] + 1; end; end;

for I := Day1 to Day365 do begin NoFishAboveCounter[I] := NoFishAboveCounter[I - 1] + NoFishAboveCounter[I]; if NoFishAboveCounter[I] > AE_Upper_FillNo.IntValue then begin StartToMove := I; Break; end;

end;

for FishNo := 0 to FSalmonList.count - 1 do begin for j := Day1 + 1 to Day365 do begin // if Above Counter filled if FSalmonList.Salmon[FishNo].MoveIntoAboveCounter > StartToMove then begin if FSalmonList.Salmon[FishNo].IsAnUpperfish then begin Test := Random(); // a new chance each day if (FSalmonList.Salmon[FishNo].FishPosistion[j] = InAboveCounter) then begin if Test < AE_Upper_ProbAfterFill.FloatValue then begin FSalmonList.Salmon[FishNo].MoveIntoUpper := j; for K := j to Day365 do FSalmonList.Salmon[FishNo].FishPosistion[K] := InUpper; // this stops it triggering again end; end; end; end; end; end; end;

Page | 35

procedure TForm1.RunModel(ArrivalShift: Integer); begin FishArrive(ArrivalShift); // arrive in below counter MoveAboveCounter; // dependant on flow MoveIntoUpper; end;

Lower Test NEP Volume 3: Appendices

Appendix 6.4.2 Additional Notes on the Salmon Simulation Model (Pisces) Additional information from Pisces: P. A. Henderson & R. M. H. Seaby Friday, 09 November 2012

1. Introduction 2 1.1. Data used for the analysis 2 1.2. Simulating uncertainty 2 1.3. Arrival model 2 1.4. Temperature 5 1.4.1. Further field observations supporting the use of temperature as a variable determining movement 6 1.5. Flow 8 1.5.1. Pisces method 8 1.5.2. Spring and summer up to August 31st 8 1.5.3. Autumn from 1st September 9 1.5.4. The Fewings probability of movement function 10 1.5.5. Rainfall as a predictor 10 1.6. The willingness to move 11 1.7. Goodness of fit measures 12

2. How to run the model 13

1

Lower Test NEP Volume 3: Appendices

1. Introduction At a meeting at Pisces on 12 June 2012 it was decided to implement a series of additions to the Pisces salmon model. The major additions it was decided to add to the program are introduced below, with descriptions of the other significant parts of the model for completeness 1. Dr Adrian Fewings has derived a function for the change in the probability of movement of salmon with respect to flow, which differs from that derived by Dr Peter Henderson of Pisces and previously applied in the model. Dr Fewings’ idea is for a more complex function which does not have a simple linear relationship between flow and the probability of movement. It is suggested that Dr Fewings’ model period be applied over August and up to about 15th September. 2. Dr David Solomon presented evidence that the desire or impetus to move changes through time. In his experience, once a salmon stops moving it tends to settle down and the impetus to move may not increase until after 15th September or possibly October, when it starts to rise again. It was therefore decided to add to the model a function which expresses this behaviour. 3. The Atkins team have considered the relationship between rainfall and salmon movement. They note that rainfall may be a powerful predictor of salmon movement and requested that this possibility be explored and, if appropriate, a model with rainfall as an independent variable initiating movement be included as an option in the Pisces salmon model. 4. Dr David Solomon initiated discussion of how to assess the goodness of fit of the model to the actual data. It was agreed this was an area that needed some work and Pisces indicated that they had a number of approaches that can be implemented. 5. It was requested that further clarifications of how equations were derived was presented. These notes need to be read in conjunction with earlier documentation.

1.1. Data used for the analysis The salmon count, flow, temperature and rainfall data for years between 1996 and 2007 were used for the statistical analysis. Only years which had complete or almost complete data runs of salmon counts were used. Data for 2003 was therefore excluded. 1.2. Simulating uncertainty The salmon movement model is an individual based model. Essentially a fixed number of fish (usually set at 1000) arrive over time into the estuary and then the movement upstream of each is simulated. Because the probability of movement of each fish on any particular day is determined in part by random simulation of a probability of movement no two simulations are identical. The basic time unit used for the simulation is 1 day. The habitat is envisaged as divided into a series of spatial components. The individual fish move between these components according to defined rules. Random number generators are used to determine whether over each time step a particular fish will make the transition between the different reaches of the river. For each day of the year, and for each section of the system, the probability of movement of a single salmon moving is calculated. This probability is based on the various models described below. Using a random number generated for each fish on each day, the behaviour of the fish is decided. For example, assume the probability of movement section 1 and 2 is calculated to be 0.5. Three fish are in section 1 and these fish have randomly been assigned test probabilities of 0.2, 0.7 and 0.6. These random assignments are generated by using a random number generator to generate a real number between 0 and 1. On this day, fish 2 and 3 will move, while fish 1 will stay in area 1. Since this Test probability is randomly generated, each time the simulation is run; it is possible for different numbers of fish to move, even with the same probability of movement. 1.3. Arrival model The data from the commercial fishermen at Mudeford, Christchurch Harbour entrance, can give insight into the returning behaviour of salmon. Since 1999 the fishermen are only licensed to catch during June and July. The Mudeford capture data show a distinct bimodality, the peaks are around weeks 22 and 30. In comparison, the main peak at the River Test counter is around week 42 . This initially suggests that the Mudeford fishermen are not catching fish that will contribute to the main late summer-autumn run. However, salmon radio-tagged on the Itchen before the end of July often remained at the head of tide until the autumn. All of the late-running fish could be accounted for by arrival coincident with the observations at Mudeford. It therefore seems most likely that the relationship between these time series is that shown below. This explanation is also consistent with the known distribution of single and multi- winter sea fish between the early and late runs. Note that in the figure, the second peak in the captures at Mudeford is dominated by single winter sea fish, as is the late summer-autumn run on the River Test.

2

Lower Test NEP Volume 3: Appendices

The temporal pattern of captures at Mudeford, Dorset and the total counts from the River Test

The model therefore uses a basic arrival model based on the pattern of capture of 1-, 2- and 3-winter fish observed at Mudeford. The Mudeford seasonal catch data for different sea-age fish follows (supplied by A. Fewings – Mudeford nets.xls)

3

Lower Test NEP Volume 3: Appendices

Sea age Week of year 1 2 3 Total 19 1 1 22 7 2 9 23 2 40 17 59 24 1 54 26 81 25 4 48 11 63 26 13 34 13 60 27 33 33 7 73 28 82 35 8 125 29 110 49 2 161 30 134 72 1 207 31 74 30 1 105 32 2 2

A plot of these data is shown below. Note the essentially bimodal form of the overall arrival pattern.

In this simulation the time step is a single day. A polynomial was used to fit each of these time series and generate the probability of movement (from sea to below the counter) for each day – see the table below.

4

Lower Test NEP Volume 3: Appendices

Day 1 SW salmon 2 SW salmon 3 SW salmon 124 0 0 0 125 0 0 0 126 0 0 0.001605136 127 0 0 0.003210273 128 0 0 0.004815409 129 0 0 0.006420546 130 0 0 0.008025682 131 0 0 0.009630819 132 0 0 0.011235955 133 0 0 0.011235955 134 0 0 0.011235955 135 0 0 0.011235955 136 0 0 0.011235955 137 0 0 0.011235955 138 0 0 0.011235955 139 0 0 0.011235955 140 0 0 0.011235955 141 0 0 0.011235955 142 0 0 0.011235955 143 0 0 0.011235955 144 0 0 0.011235955 145 0 0 0.011235955 146 0 0 0.011235955 147 0 0.002364066 0.014446228 148 0 0.004728132 0.017656501 149 0 0.007092199 0.020866774 150 0 0.009456265 0.024077047 151 0 0.011820331 0.027287319 152 0 0.014184397 0.030497592 153 0 0.016548463 0.033707865 154 0.00056802 0.030057413 0.060995185 155 0.001136041 0.043566363 0.088282504 156 0.001704061 0.057075312 0.115569823 157 0.002272082 0.070584262 0.142857143

1.4. Temperature Salmon movement is known to be affected by temperature, and a temperature effect is apparent in the River Test data. The graph below shows the relationship between the number of salmon counted moving upriver and the water temperature. This graph has been generated using data for all years for which a full record of count data were available.

5

Lower Test NEP Volume 3: Appendices

The focus of the analysis is on the upper boundary of this scatter plot as this is taken to represent the upper boundary of counts that can occur at any temperature. Whether this number of counts actually occurs will depend on other factors such as the number of waiting fish, river flow and possibly rainfall. We seek to identify how this upper boundary declines with temperature. Between water temperatures of 12 and 16 °C, the upper bound of the count declines from about 100 to 10, and above 20°C it is 5 or less. To reflect the unwillingness of salmon to move at higher temperatures, a simple model was used, as follows. For temperatures below 12°C probability of movement = 1.0 (it is determined by flow only). For temperatures above 12°C and below 16.4°C probability of movement = -0.225 x river temperature + 3.7. The linear model used over this temperature range was estimated by drawing a line through the upper boundary counts observed between 12 and 16.4 °C and calculate the slope of the line. It is important to note that the counts in the plot above are presented on a log scale. There is therefore an approximately order of magnitude drop in the upper bound count used to calculate the change in probability of movement relating to temperature between 12 and 16 °C . For temperatures above 16.4°C probability of movement = 0. In practice temperatures above 16.4 °C are infrequently observed and very few fish are counted moving at such elevated temperatures. It makes little difference to the simulated outcome if the probability of movement at temperatures above this threshold are given a zero or very small probability of movement. The values of these parameters can be changed in the model. The probability of movement related to water temperature is combined with other factors to produce the probability of movement for each day.

1.4.1. Further field observations supporting the use of temperature as a variable determining movement The first line of evidence derives from the study of Clarke et al (1994) Clarke, D. R. K. Evens, D. M., Ellery, D. S. and Purvis, W. K. 1994. Migration of Atlantic salmon (Salmo salar L.) in the River Tywi estuary during 1988, 1989 and 1990. RT/WQ/RCEU/94/7 Nation Rivers Authority Welsh Region They reported the probability of telemetered salmon ever entering freshwater, in relation to the water temperature after release. They tagged 260 salmon over three years and tracked the progress of the fish as they entered or left the estuary. These data were reported in temperature bands of two degrees Celsius with no confidence limits on the observations. To allow prediction of salmon probability of entry at temperature increments of 0.5°C, WRc (Water Research Council) were asked by the Environment Agency to employ general linear modelling techniques, with the constraint that at temperatures less than or equal to 12°C, the probability of river entry was a constant. 6

Lower Test NEP Volume 3: Appendices

The results are presented graphically below. Note that the probability of entry into freshwater drops rapidly above 12 °C. The resultant model was used to construct a lookup table, which allowed the provision of probability estimates with 90 percentile confidence limits. The mean values used for the model were considered an underestimate of probability of salmon entry into the Cleddau. The most likely real relationship is expected to lie between the mean relationship and unity; the upper 90%-ile relationship may therefore be considered a fair working estimate.

General Linear Model of probability of river entry data

The second line of evidence derives from the study of Solomon and Sambrook (2004) Solomon D. J. and Sambrook H. T., 2004, Effects of hot dry summers on the loss of Atlantic salmon, Salmo salar, from estuaries in South West England Fisheries Management and Ecology, 11, 353–363 Solomon and Sambrook (2004) found that the failure to enter southern English rivers was related to water quality. For the Hampshire Avon and the Exe in Devon, they studied the upstream migration of adult salmon by tagging salmon at the entrance of an estuary or harbour and observing their runs up river.

Solomon & Sambrook noted that: “On the Avon, increased river temperature was associated with decreasing tendency to enter within 10 days of tagging, and virtually all fish entering the river on days when the 09.00-hour river temperature was in excess of 17°C did so in the late evening and early morning, thus avoiding the time of day of highest water temperatures.” and: "Fewer fish enter the river promptly at high river temperatures and more at low temperatures,...." This study makes it clear that upstream movement is delayed as water temperature increases.

7

Lower Test NEP Volume 3: Appendices

1.5. Flow 1.5.1. Pisces method There is a clear tendency for movement to be modulated by flow and change in flow. Further there is a change in behaviour over the season. Graphical examination of all the annual salmon count sequences shows a standard seasonal pattern. The count data clearly divides into two components a spring- summer run which occurs between 1st May and 1st September and a late summer-autumn run which occurs from 1st September to 31st December. The spring-summer run is known to comprise the greatest proportion of fish that spend more than one winter at sea. The division into the two waves of migration is seen clearly in the figure below.

The probability of movement in relation to flow is modelled differently for the two periods as described below. The years 1996 to 2007 were analysed separately for movement before and after August 31st. The actual point division between the two runs is arbitrary; the 1st of September is used here because no salmon has been counted moving upstream on this day between 1996 and 2007 and is therefore the minimum point between the two distributions. Environment Agency staff has normally used 30th September as the start of the autumn run as large autumn peaks in salmon count have rarely been observed prior to this date.

1.5.2. Spring and summer up to August 31st For this period the probability is determined solely by the magnitude of the flow. The data for the years 1996 to 2007 excluding 2006 were used. For each year the number of fish waiting on each day was estimated using the total size of the run and the known pattern of arrival. The proportion of the waiting fish which moved on each day was then calculated. These proportions were assumed to represent the probability of movement. The data for each day was classified into flow classes 4-4.5, 4.5-5 m3/s etc. The probabilities in each flow class were then averaged to give the data presented in the figure below. Examination of the plot indicated that an exponential curve would be a suitable model for the flow relationship. The following relationship was used: Probability = 0.00003 x flow3.0251.

8

Lower Test NEP Volume 3: Appendices

This relationship was derived empirically from the data by regression as shown in the graph below.

Consideration was given as to the role of a change in flow. For this period there is not clear relationship between a change in flow and fish movement. Many fish have been observed to move when the flow is declining. The low number of observations over this period do not allow a more sophisticated model to be developed or tested. 1.5.3. Autumn from 1st September During the autumn period the probability of movement was calculated in similar fashion to the spring/early summer period above. However, a further complexity was introduced because it was noted that fish behaved differently on a rising compared with falling water level. It has previously been noted that upstream movements tend to occur on a positive change in flow. The model therefore uses a different equation relating the probability of upstream movement depending on whether the between-day change in flow is positive or negative. Examination of the plot indicated that an exponential curve would again be the most suitable model for the flow relationship. The two curves used in the model and the data from which they are derived are presented below. For positive changes in flow between days Probability of movement = 0.0002 x Flow2.5556 For negative changes in flow between days Probability of movement = 0.0003 x Flow2.2901 These equations have been derived empirically and fitted by regression. The maximum probability is capped at MaxNeg and MaxPos of 0.1. In other words on any one day the maximum probability of movement through the fish counter is 10%. This value is derived from an analysis of movement during high flow periods later in the year when it is believed all the fish have entered the tidal river and are available to move upstream. It is clear that their actual time of movement varies greatly between individuals and is never greater than about 10% of the fish available to move. For all higher flow levels the probability is capped at 0.1 or 10% per day.

9

Lower Test NEP Volume 3: Appendices

During initial model development it was assumed that the magnitude of the change in flow would be a key variable. In fact fish are often observed to move on extremely small positive changes. Increased willingness to move is therefore best described as a two part trigger - (1) positive change in flow and (2) absolute magnitude of flow.

1.5.4. The Fewings probability of movement function This was supplied by Dr Fewings, and implemented by approximating the function using a 5th order polynomial. Of the form Probability = constant+ Power1 * X + Power2 * X2 + Power3 * X3 + Power4 * X4 + Power5 * X5 The model only predicts movement in a narrow range of flows, and over the late summer period. The model to be used when this non-linear relationship is not employed can be chosen. The values for the equation are shown below.

1.5.5. Rainfall as a predictor As a first stage a General Linear Model was used to identify if rainfall was a significant predictor of the tendency to move. The analysis was undertaken on data for years with effectively complete records for rainfall and salmon counts; these were 1996, 1997, 1998, 1999, 2000, 2002, 2003, 2005, 2006, 2007. It was found that rainfall with a lag of 1 day was indeed a significant predictor of salmon count. However, there was the need to bring in a second variable, which could be either flow or seasonality, to produce a useful predictive model. The results from the best model identified were as follows: N = 1463 Missing Observations = 1

10

Lower Test NEP Volume 3: Appendices

R = 0.436 Rsqr = 0.190 Adj Rsqr = 0.189

Standard Error of Estimate = 0.099

Coefficient Std. Error t P VIF Constant -0.0173 0.00407 -4.257 <0.001 Flow 0.00608 0.000424 14.335 <0.001 1.262 lag1rain 0.00231 0.000595 3.884 <0.001 1.262

Analysis of Variance: DF SS MS F P Regression 2 3.348 1.674 171.246 <0.001 Residual 1460 14.273 0.00978 Total 1462 17.621 0.0121

Column SSIncr SSMarg Flow 3.201 2.009 lag1rain 0.147 0.147

The dependent variable Probability can be predicted from a linear combination of the independent variables: P Flow <0.001 lag1rain <0.001

All independent variables appear to contribute to predicting Probability (P < 0.05).

Normality Test (Shapiro-Wilk) Failed (P = <0.001)

Constant Variance Test: Failed (P = <0.001)

Power of performed test with alpha = 0.050: 1.000

Having decided upon a model that included both flow and rainfall with a single day lag, a multiple regression analysis was undertaken to identify the best equation, which was found to be:

Probability = -0.0173 + (0.00608 * Flow) + (0.00231 * lag day-1rain)

Implemented in this form the model would not work, as we need to make sure that the probability remains positive or zero. With zero rain, P = -0.0173 + 0.00608 * Flow, so zero probability occurs at about 2.845 m3/s flow. That salmon will not move when the flow is this low seems very reasonable.

1.6. The willingness to move Dr Solomon supplied the willingness to move parameters. 11

Lower Test NEP Volume 3: Appendices

Below is his note: Input to Pisces model with respect to “willingness to migrate”. David Solomon. As discussed, fish appear to have a relatively high willingness to migrate when they first arrive in the river, which falls to a very low level after about 2 weeks. They then remain fairly torpid until a combination of re-aroused willingness and stimulating conditions combine to move them in the autumn. I think a few iterations may be necessary to arrive at the correct balance of value and timing of this “willingness” input to the model. In the first instance I suggest the following:- On the day after arrival in the river the willingness to migrate is at 0.5. It then falls in a straight line down to 0.005 on the fifteenth day after arrival. It remains at this level until September 15. From September 16 it starts to climb in a straight line, reaching a value of 1 on November 30. The idea of setting values of 1 and below is that this is the probability of migrating and passing the counter when river conditions (absolute flow, temperature, rainfall, change in water flow etc) are all optimal.

June 14 2012.

This relationship can be applied from the Below to Above Counter tab.

1.7. Goodness of fit measures To help decide which models give the best fit, 3 measures of goodness of fit are offered by the program:  Nash Sutcliffe Efficiency (NSE)  Percent Bias (PBias)  Ratio Root Mean Square (RSR) The Nash–Sutcliffe model efficiency coefficient was originally used to assess the predictive power of hydrological models. It is defined as:

t where Qo is observed discharge, and Qm is modelled discharge; Qo is observed discharge at time t. Nash–Sutcliffe efficiencies can range from −∞ to 1. An efficiency of 1 (E = 1) corresponds to a perfect match of modelled discharge to the observed data. An efficiency of 0 (E = 0) indicates that the model predictions are as accurate as the mean of the observed data, whereas an efficiency of less than zero (E < 0) occurs when the observed mean is a better predictor than the model or, in other words, when the residual variance (described by the numerator in the expression above), is larger than the data variance (described by the denominator). Essentially, the closer the model efficiency is to 1, the more accurate the model is.

12

Lower Test NEP Volume 3: Appendices

Percent bias (PBIAS) measures the average tendency of the simulated values to be larger or smaller than their observed ones. The optimal value of PBIAS is 0.0, with low-magnitude values indicating accurate model simulation. Positive values indicate overestimation bias, whereas negative values indicate model underestimation bias. The Ratio Root Mean Square is calculated by

obs sim 2 obs Mean 2 RRMS = √ ∑ ( Yi - Yi ) ) / ( √ ∑ ( Yi - Yi ) )

Where obs Yi = Observed data sim Yi = Predicted or simulated data Mean Yi = the mean of the observed data

The RRMS ranges from large negative to large positive values, and a value <= 0.7 indicates a good fit.

In addition, a plot of observed against predicted can be generated with a 1:1 line showing perfect correspondence.

2. How to run the model The movement to Above Counter is based on an analysis of movement across the fish counter. There are 4 possible relationships which relate to flow and to additional factors. There are several possible combinations that can be run. First the user must choose a model, then decide whether to use the temperature and/or willingness to move functions. So, a model might entail: Model1, with Temperature taken into account or Flow & Rain, with Willingness to move used

13

Lower Test NEP Volume 3: Appendices

Appendix 7.2 Aquatic macrophytes

Table of Contents

Appendix 7.2.1 Macrophyte Survey Species List 1

Appendix 7.2.2 Mean Flow Rank calculations 3

Appendix 7.2.1 Macrophyte Survey Species List

Species details Reach Code Scientific name Common name USTW_01 USTW_02 Distribution notes for survey reach Alisma plantago aquatica Common Water-plantain Not recorded Apium nodiflorum Fool's Water-cress C2 Occasional Berula erecta*** Lesser Water-parsnip C2 C4 Throughout occurring more abundantly as submerged form Butomus umbellatus Flowering Rush Not recorded Callitriche agg. Starwort C1 C3 Throughout although more abundant in downstream section Carex species (acutiformis***/riparia) Sedges * Rare bankside Cladophora agg. Filamentous Green Algae C1 C1 Occurring within macrophyte stands Elodea canadensis Canadian Water-weed C1 Rare Enteromorpha Tubeweed C1 C1 Rare Epilobium hirsutum Great Willowherb * Rare bankside species Eupatorium cannabinum Hemp-agrimony * * Rare bankside species Filipendula ulmaria Meadowsweet Not recorded Fontinalis antipyretica*** Moss C1 C3 Throughout although more abundant in downstream section Glyceria maxima Reed Sweet-grass C1 C1 Rare marginal species Hippuris vulgaris Mare's Tail Not recorded Impatiens capensis Orange Balsam Not recorded Impatiens glandulifera Indian Balsam * * Common bankside species Iris pseudacorus Yellow Iris Not recorded Juncus effusus Soft Rush Not recorded Lemna minor Common Duckweed C1 C1 Rare and sporadically distributed Lemna trisulca Ivy-leaved Duckweed C3 C2 Commonly occurring within macrophyte beds Lycopus europaeus Gipsywort Not recorded Lythrum salicaria Purple Loosestriffe * Rare marginal species Mentha aquatica Water-mint C1 Rare marginal species Mimulus guttatus Monkey Flower Not recorded Myosotis scorpioides Water Forget-me-not C2 C2 Common bankside species Myriophyllum spicatum Spike Water-milfoil Not recorded Nuphar lutea Yellow Water-lilly C1 Rare and restricted to deeper pools in submerged form only Oenanthe crocata Hemlock Water-dropwort Not recorded Phalaris arundinacea Reed Canary-grass C1 Rare marginal species Phragmites australis Common Reed C1 Rare marginal species Polygonum amphibium Amphibious Bistort Not recorded Polygonum hydropiper Water-pepper Not recorded Potamogeton crispus Curled Pondweed C3 Throughout downstream section Potamogeton lucens Shining Pondweed Not recorded Potamogeton pectinatus Fennel Pondweed Not recorded Potamogeton perfoliatus Perfoliate Pondweed C2 C4 Widespread and often abundant Pulicaria dysenterica Common Fleabane * Rare bankside species Ranunculus penicillatus pseudofluitans*** Brook water-crowfoot C6 C7 Widespread and often abundant Rorippa nasturtium-aquaticum Water-cress C3 C3 Widespread and often abundant Rumex hydrolapathum Water-dock C1 Rare marginal species Sagittaria sagittifolia Arrowhead Not recorded Schoenoplectus lacustris Common Club-rush C2 Rare and sporadically distributed Scrophularia aquatica Water Figwort Not recorded Solanum dulcamara Bittersweet * C1* Common bankside / rare marginal Sparganium emersum Unbranched Bur-reed C2 C2 Rare and local Sparganium erectum Branched Bur-reed C2 C2 Rare and sporadically distributed Stachys palustris Marsh Woundwort Not recorded Symphytum officinale Common Comfrey * * Common bankside species Typha latifolia Greater Reedmace C2 Rare marginal species Veronica anagallis-aquatica/cat*** Water speedwell hybrid Not recorded Veronica beccabunga Brooklime Not recorded Zanichellia palustris Horned Pondweed Not recorded Broad classification Submerged fine leaved C6 C7 Submerged broadleaved C2 C5 Submerged liner leaved C2 C3 Floating leaved rooted Emergent broadleaved herbs C3 C4 Emergent reeds/rushes/sedges C2 C3 Free floating C1 C1 Open water (%) 50 35

1 Lower Test NEP Volume 3: Appendices

Appendix 7.2.1 Macrophyte Survey Species List (cont.)

Species details Reach Code Scientific name Common name DSTW_01 DSTW-02 DSTW-03 DSTW-04 DSTW-05 Distribution notes for survey reach Alisma plantago aquatica Common Water-plantain C1 C1 C1 C1 Rare and sporadically distributed Apium nodiflorum Fool's Water-cress C1 C1 C1 Rare marginal emergent Berula erecta*** Lesser Water-parsnip C1 C1 C2 C1 C1 Throughout occurring sporadically in both emergent and submerged form Butomus umbellatus Flowering Rush C1 C1 Rare marginal emergent Callitriche agg. Starwort C1 C1 C1 Rare and local Carex species (acutiformis***/riparia) Sedges * C1* C1 C2 C1* Occasional within marginal stands, commonly occurring bankside species Cladophora agg. Filamentous Green Algae C1 Rare submerged species Elodea canadensis Canadian Water-weed C1 C1 Rare submerged species Enteromorpha Tubeweed Not recorded Epilobium hirsutum Great Willowherb * * * * * Common bankside species Eupatorium cannabinum Hemp-agrimony * * Rare bankside species Filipendula ulmaria Meadowsweet * Rare bankside species Fontinalis antipyretica*** Moss C1 Rare Glyceria maxima Reed Sweet-grass C1 C3 C2 C2 C1 Common occurring within established marginal reed fringes Hippuris vulgaris Mare's Tail C1 Rare and local Impatiens capensis Orange Balsam * * Rare invasive occurring on banksides only Impatiens glandulifera Indian Balsam C1* C1* C1* C1* Invasive occurring throughout and abundant on bank tops Iris pseudacorus Yellow Iris C1 C1 C1 C1 Occasional stands throughout Juncus effusus Soft Rush C1 Rare Lemna minor Common Duckweed C1 C1 C1 Rare and local Lemna trisulca Ivy-leaved Duckweed C1 C1 Rare and local Lycopus europaeus Gipsywort C1 Rare marginal Lythrum salicaria Purple Loosestriffe C1 C1 Rare marginal emergent Mentha aquatica Water-mint C1 C1 Rare marginal Mimulus guttatus Monkey Flower C1 Invasive, rare marginal Myosotis scorpioides Water Forget-me-not C1 C1 C1 Rare marginal Myriophyllum spicatum Spike Water-milfoil C2 C2 Rare and local Nuphar lutea Yellow Water-lilly C3 C2 C3 C2 C2 Occasional throughout, occurring more commonly as submerged form Oenanthe crocata Hemlock Water-dropwort C1 C1 C1 C1 Rare and sporadically distributed Phalaris arundinacea Reed Canary-grass C1 C1* C1* C3* C2 Commonly occurring within established marginal reed fringes Phragmites australis Common Reed C1 C1 Rare within established marginal reed fringes Polygonum amphibium Amphibious Bistort C1 C1 Rare marginal Polygonum hydropiper Water-pepper C1 Rare marginal Potamogeton crispus Curled Pondweed Not recorded Potamogeton lucens Shining Pondweed C3 C3 C4 C6 Widespread and often abundant Potamogeton pectinatus Fennel Pondweed C6 C7 Abundant in lower 500m of survey reach Potamogeton perfoliatus Perfoliate Pondweed C4 C2 C2 Commonly occurring in upper survey reaches Pulicaria dysenterica Common Fleabane * Rare bankside species Ranunculus penicillatus pseudofluitans*** Brook water-crowfoot C2 C2 C2 C1 Occasional in upper reaches Rorippa nasturtium-aquaticum Water-cress C2 C2 C2 C1 C1 Widespread marginal but never dominant Rumex hydrolapathum Water-dock C1 C1 C1 C1 C1 Rare but sporadically distributed Sagittaria sagittifolia Arrowhead C1 C2 C3 C1 C1 Commomly occuring in submerged form, emergent form confined to marginal areas Schoenoplectus lacustris Common Club-rush C4 C5 C6 C5 Widespread and abundant throughout Scrophularia aquatica Water Figwort C1 C1 C1 Rare marginal Solanum dulcamara Bittersweet * * * * Widespread bankside species but never dominant Sparganium emersum Unbranched Bur-reed C6 C6 C3 C7 C5 Widespread and abundant Sparganium erectum Branched Bur-reed C2 C1 C3 C5 C3 Common throughout occurring as established marginal reed fringes Stachys palustris Marsh Woundwort * * * Rare bankside species Symphytum officinale Common Comfrey * Rare bankside species Typha latifolia Greater Reedmace C2 C1 C1 Rare marginal emergent Veronica anagallis-aquatica/cat*** Water speedwell hybrid C1 C1 C1 C1 Commonly occurring within established marginal reed fringes Veronica beccabunga Brooklime C1 C1 Rare marginal Zanichellia palustris Horned Pondweed C2 Rare and local Broad classification Submerged fine leaved C2 C2 C3 C6 C7 Submerged broadleaved C4 C4 C4 C5 C6 Submerged liner leaved C6 C6 C5 C7 C6 Floating leaved rooted C1 C1 C1 C2 Emergent broadleaved herbs C3 C3 C3 C3 C2 Emergent reeds/rushes/sedges C2 C3 C4 C5 C4 Free floating C1 C1 C1 Open water (%) 40 60 80 30 20 Notes: *** against species name indicates typical chalk stream plant, * against percentage cover code indicates recorded on bankside / banktop as well as in-channel, * on its own indicated recorded on bankside / banktop only. 2 Lower Test NEP Volume 3: Appendices

Appendix 7.2.2 Mean Flow Rank calculations

Species Cover Values (SCV) Macrophyte Flow Rank USTW-01 USTW-02 Scientific name (MFR) Alisma plantago aquatica 1 Carex species (acutiformis/riparia) 1 Cladophora agg. 1 1 1 Enterommorpha 1 1 1 Glyceria maxima 1 1 1 Iris pseudacorus 1 Lemna minor 1 1 1 Rumex hydrolapathum 1 1 Sparganium erectum 1 2 2 Typha latifolia 1 2 Apium nodiflorum 2 2 Elodea canadensis 2 1 Lemna trisulca 2 3 2 Myriophyllum spicatum 2 Nuphar lutea 2 1 Potamogeton crispus 2 3 Potamogeton lucens 2 Potamogeton pectinatus 2 Potamogeton perfoliatus 2 2 4 Rorippa nasturtium-aquaticum 2 3 3 Sagittaria sagittifolia 2 Schoenoplectus lacustris 2 2 Sparganium emersum 2 1 2 Veronica anagallis-aquatica/cat 2 Zanichellia palustris 2 Berula erecta 3 2 4 Butomus umbellatus 3 Callitriche agg 3 1 3 Fontinalis antipyretica 3 1 3 Hippuris vulgaris 3 Oenanthe crocata 3 Ranunculus penicillatus pseudofluitans 4 6 7

Cover Value Score (∑MFRi*SCVi) 62 102 Species Cover Value (∑ of individual SCV) 28 42 Flow Score (∑CVS/∑SCV0) 2.21 2.43

3 Lower Test NEP Volume 3: Appendices

Appendix 7.2.2 Mean Flow Rank calculations (cont.)

Species Cover Values (SCV) Macrophyte Flow Rank DSTW-01 DSTW-02 DSTW-03 DSTW-04 DSTW-05 Scientific name (MFR) Alisma plantago aquatica 1 1 1 1 1 Carex species (acutiformis/riparia) 1 1 1 2 1 Glyceria maxima 1 1 3 2 2 1 Iris pseudacorus 1 1 1 1 1 Lemna minor 1 1 1 1 Rumex hydrolapathum 1 1 1 1 1 1 Sparganium erectum 1 2 1 3 5 3 Typha latifolia 1 2 1 1 Apium nodiflorum 2 1 1 1 Elodea canadensis 2 1 1 Lemna trisulca 2 1 Myriophyllum spicatum 2 2 2 Nuphar lutea 2 3 2 3 2 2 Potamogeton lucens 2 3 3 4 6 Potamogeton pectinatus 2 6 7 Potamogeton perfoliatus 2 4 2 2 Rorippa nasturtium-aquaticum 2 2 2 2 1 1 Sagittaria sagittifolia 2 1 2 3 1 1 Schoenoplectus lacustris 2 4 5 6 5 Sparganium emersum 2 6 6 3 7 5 Veronica anagallis-aquatica/cat 2 1 1 1 1 Zanichellia palustris 2 2 Berula erecta 3 1 1 2 1 1 Butomus umbellatus 3 1 1 Callitriche agg 3 1 1 1 Fontinalis antipyretica 3 1 Hippuris vulgaris 3 1 Oenanthe crocata 3 1 1 1 1 Ranunculus penicillatus pseudofluitans 4 2 2 2 1

Cover Value Score (∑MFRi*SCVi) 61 70 93 89 70 Species Cover Value (∑ of individual SCV) 31 36 47 48 38 Flow Score (∑CVS/∑SCV0) 1.97 1.94 1.98 1.85 1.84

4 Lower Test NEP Volume 3: Appendices

Appendix 7.3 Aquatic Macroinvertebrates

Table of Contents

Appendix 7.3.1 Statistical assessment 2

Appendix 7.3.2 MDS Outputs 9

Appendix 7.3.3 ANOSIM Outputs 10

Appendix 7.3.4 SIMPER Outputs 15

Appendix 7.3.5 Macroinvertebrate biotic indices 21

1

Lower Test NEP Volume 3: Appendices

Appendix 7.3.1 Statistical assessment

Introduction The need for the statistical assessment arises from concerns relating to a change in sampling procedure in 2010, from one involving proportional habitat sampling across the whole river cross section (‘Old method’, 2002 to 2008), to one in which proportional habitat sampling is confined to a defined area of the channel (‘New method’, 2010 to 2011). This change was implemented by the Environment Agency to minimise the potential for sampling errors caused by surveyors not sampling across the entire cross-section of the river, and thus not sampling certain microhabitats within the cross-section affecting the diversity and abundance of the collected sample. However, as a result of this methodological change theoretically there is potential for certain community traits to be either, under or over represented, by one or other, of the sampling procedures, by limiting the area in which the sampling is to be undertaken. For example, if the new method were to exclude a certain habitat type from being sampled, or result in proportionally more time being spent sampling a particular habitat (when compared to the old method) then one may question the validity of using the entire sampling record in certain community based analysis. The two River Test aquatic macroinvertebrate sampling sites under investigation are:  Upstream of Testwood Abstraction (NGR SU 35250 15350, EA site code 90401); and  Downstream of Testwood Abstraction (NRG SU 35350 15300, EA site code 90402). Figure 2.4.1 in the Figures Volume shows the location of these sites, and Table 1 below provides a summary of the sampling sites and sampling method adopted (‘Old Method’ or ‘New Method’) over the period of record from 2002 to 2011. In total eighteen samples have been collected using the ‘old’ method: nine from each of the two sampling sites. Ten samples have been collected using the new method of sampling: seven from the Downstream site and three from the Upstream site. All samples have been collected in either the spring season (March–May) or autumn season (September–November) with the exception of the sample collected in June from the Downstream site.

Table 1 River Test sampling site details, sampling months and method

R. Test Upstream of Testwood R. Test Downstream of Testwood

Abstraction Abstraction

(NGR SU 35250 15350) (NRG, SU 35350 15300) Year Old Method New Method Old Method New Method

2002 May; Sept. May; Sept.

2003 May; Oct. May; Oct.

2004 May; Oct. May; Oct.

2005 May

2006

2007 Nov. Nov.

2008 April April x 2

2009

2010 May, Nov. X2 April; May; Nov. x 2

2011 April, June, Nov

Total No. 9 3 9 7 of Samples

2

Lower Test NEP Volume 3: Appendices

The aim of the statistical assessment is to determine whether or not the entire dataset, collected before and after the methodological change, can be used for the assessment of flow related stress through use of the community derived Lotic-Invertebrate Flow Evaluation (LIFE) sample scores (after Extence et al. 1999). The LIFE index has been developed to be sensitive to subtle changes in the proportion and abundance of taxa based on their flow requirements and, as such, should be considered susceptible to alterations in sampling method although no detailed investigation of how changes in sampling method affect the metric have been undertaken. The objectives of the statistical assessment are therefore:  To collate and standardise the historic macroinvertebrate data from the Upstream of Testwood and the Downstream of Testwood abstraction EA sampling site;  To investigate temporal and spatial trends in species community structure;  To ascertain if the community and species data evidence the presence of sampling bias as a result of the switch in sampling method; and  If no strong sampling bias is evidenced then to undertake temporal and spatial analysis of biotic scores to investigate abstraction effects on macroinvertebrate communities. The statistical investigation has been undertaken using the statistical package Primer Ver. 6 (Clarke and Gorley, 2001) commonly used in multivariate analysis of biological and abiotic data. In this instance Multi- Dimensional Scaling (MDS), Analysis of Similarities (ANOSIM) and Similarity of Percentages (SIMPER) has been used to detect trends in similarity, to determine factor significance and to investigate macroinvertebrate species responsible for differences between factors. In this instance the word factor relates to a user defined set of sample groups i.e. samples grouped by location, season, year or sampling method. All the statistical analysis discussed below has been undertaken on standardised sample taxa lists to account for difference in identification level with square-root transformed abundance data.

Multi Dimensional Scaling (MDS) The multi-dimensional scaling (MDS) ordination plot provided as Figure 1 below, has been produced to allow visualisation of trends in sample similarity based on the aquatic macroinvertebrate communities of the individual samples. The 2-dimensional display can be used to identify the potential gradients that may be of importance in shaping the macroinvertebrate communities recorded (see Appendix 7.3.2 for axis scores). For each site the code prefix indicates sampling location (US = Upstream of Testwood Abstraction; DS = Downstream of Testwood Abstraction), followed by the month (mmm) of sampling and then year (yy). If more than one sample has been collected in the same month of any single year then a suffix of either A or B is shown. All of the Upstream of Testwood abstraction macroinvertebrate sampling sites are labelled in black text, the Downstream sites in blue text. In addition, green (Old method) and red (New method) indicate the sampling method adopted.

3

Lower Test NEP Volume 3: Appendices

1.5 DSOCT03

1 USNOV10A DSNOV11 USOCT03 DSOCT04 USNOV07 DSSPT02 USOCT04 0.5 DSNOV10B DSNOV07 USNOV10B USSPT02

DSNOV10A 0 -1.5 -1 -0.5 0.5 USMAY04 1 1.5 USMAY05 DSAPR11 DSAPR10 DSAPR08 DSMAY04 DSMAY03 -0.5 USAPR08 USMAY03 USMAY02

USMAY10

-1 DSMAY10 DSMAY02 DSJUN11

-1.5 Old sampling method

New sampling method

-2 Figure 1 MDS sample ordination showing new sampling and seasonal influence on sample cluster

Interpretation of sample MDS ordination The MDS “all sample” ordination has resulted in a separation of the new method samples from the old method samples to along the x axis, with the new method sites plotting with lower axis scores. Furthermore, there appears to be a strong “sampling year” component to the ordination with earlier samples tending to plot with higher x-axis scores, this is particularly evident for the samples collected downstream of the abstraction sites. The new method grouping could, therefore, be a reflection in year-on-year change in community structure and not simply a result of a change in sampling method, or a combination of the two. In addition a very strong seasonal influence is evident with all of the autumn samples plotting with greater y axis scores. This seasonal response is likely to be related to subtle turnover in species due to life-cycle characteristics which would account for the outlying nature of the only sample collected in the summer from the Downstream site in June 2011 (DSJUN11). The Upstream and Downstream sites are distributed throughout the ordination indicating a number of shared community traits between the samples and site locations. This is to be expected due to the proximity of the sample sites within the same river system (less than 100m apart). The lack of separation of Upstream from Downstream sites indicates that macroinvertebrate community structures are less strongly influence by location than they are by the aforementioned gradients. The fact that the new method sites all plot with higher x-axis scores shows there to be some level of difference in community structure. This could be a result of sampling bias through adoption of the new method, although it is also possible that the ordination may represent a general year-on-year change in community structure. The result of the ordination therefore warrants further investigation of the individual species, and general species traits, which are driving the output in order to fully explore the ecological reasons underpinning the results. This has been undertaken in section below (ANOSIM). In summary, the MDS sample ordination may indicate that differences between samples may in part be related to changes in sampling method. The sample ordination can equally be described by temporal gradients such as sampling season and sampling year although further investigation of which species / species traits may be causing the gradient between the new and old sampling sites is warranted.

4

Lower Test NEP Volume 3: Appendices

Analysis of similarities (ANOSIM) A pre-requisite to interpreting community differences between sites is a demonstration that there are statistically significant differences to interpret. ANOSIM tests for statistically significant differences between samples and is analogous to the standard univariate ANOVA tests. The test is based on a simple non- parametric permutation applied to the similarity matrix (in this instance a Bray-Curtis similarity matrix). A test statistic, R, is computed reflecting the observed differences between sets of groups, contrasted with differences among replicate within the sets of samples. This statistic is then recomputed under permutations of the sample labels and a significance level is calculated by referring to the observed value of R to its permutation distribution. The R value scales from +1 to −1. +1 indicates that all the most similar samples are within the same groups. R = 0 occurs if the high and low similarities are perfectly mixed and bear no relationship to the group. A value of −1 indicates that the most similar samples are all outside of the groups. To test for significance the ranked similarity within and between groups is compared with the similarity that would be generated by random chance. Essentially the samples are randomly assigned to groups 1000 times and R calculated for each permutation. The observed value of R is then compared against the random distribution to determine if it is significantly different from that which could occur at random. ANSOIM has been undertaken to assess the level of similarity between factors (or groupings) within the sample data set. These are:  Old method versus New method;  Spring samples versus Autumn samples;  Upstream sites versus Downstream sites; and  Year groups. The ANOSIM results are presented in Table 2 with Primer outputs files provided as Appendix 7.3.3.

Table 2 Analysis of similarities (ANOSIM) test outputs

Similarity test R statistic Significance

Old method versus New method 0.491 0.1*

Upstream sites versus Downstream sites 0.088 9.2 Spring samples versus Autumn samples 0.428 0.1*

Year groups Global R 0.427 0.1* Pairwise 02–03, 04–05 0.054 28.1 Pairwise 02–03, 06–07 0.358 4.4*

Pairwise 02–03, 08–09 0.56 2.2

Pairwise 02–03, 10–11 0.64 0.1* Pairwise 04–05, 06–07 0.455 4.8*

Pairwise 04–05, 08–09 0.855 4.8*

Pairwise 04–05, 10–11 0.58 0.1*

Pairwise 06–07, 08–09 1.0 33.3

Pairwise 06–07, 10–11 0.18 27.7

Pairwise 08–09, 10–11 0.172 18.2 *Note: Significant result is indicated by value of <5.

ANOSIM analysis has confirmed the trends observed from the MDS plot, with both the New method versus Old method sample grouping (R 0.491, significance 0.1), the Autumn versus Spring sample grouping (R 0.428, significance 0.1) and the year grouping (R 0.427, significance 0.1) being assessed as having 5

Lower Test NEP Volume 3: Appendices significant group similarity. In contrast, analysis of the Upstream and Downstream samples indicates no significant similarity within the sample grouping. This indicates that the sampling method, season and year of sampling have a greater influence on the macroinvertebrate community structure of the samples than the site location. It is important to note that although significant differences between groups have been identified, R values of <0.5 for method, season and year groups indicate a relatively high similarity within the entire data set. The significant result obtained for the method ANOSIM may be concomitant with the fact that all the New method samples have been collected in a relatively short time frame in relation to the rest of the sample record i.e. because all of the New samples were collected in 2010 and 2011, they are by default more similar to each other than they are to the other samples. In other words the significance is as likely to be a result of the fact that there is a naturally occurring difference between samples collected in later years (as a result of natural species turnover), as it is to do with the method change. The influence of sampling year is further evidenced by the similarities identified for the year groups and significant pairwise tests for a range of year groupings. In summary, ANOSIM has identified a significant difference between the Old and New method groups but this is potentially a result of the effect of sampling period and the year-on-year species turnover evidenced in the data set, as it is to do with the sampling method change. Having demonstrated that there are statistically significant differences to interpret between the sample method groupings, the next step is to undertake similarity analysis to ascertain if the change in sampling method is reflected within the taxa that best describe differences between the groups.

Similarity of Percentages (SIMPER) This section reports on SIMPER analysis conducted on the Old and New sampling method grouping. SIMPER is an exploratory program used in conjunction with ANOSIM to identify which taxa primarily account for the observed differences between groups. By looking at the overall percentage contribution each taxon makes to the average dissimilarity between two groups the programme lists taxa in decreasing order of their importance in discriminating the two sets of samples, in this instance Old and New method samples. SIMPER outputs are provided as Appendix 7.3.4 and for completeness include the outputs for analysis conducted on the season and year groupings, although these are not discussed here. Table 3 schedules the taxa that contribute cumulatively, 50% of the dissimilarity between samples grouped by sampling method. The taxa are ordered in decreasing individual contribution to the dissimilarity between the groups. For each taxon the average abundance (square root transformed) is provided together with the macroinvertebrate flow score to indicate the flow preference of the taxa (after Extence et al. 1999).

Interpretation of SIMPER output The change in sampling method was undertaken to reduce the potential for sampling errors to be caused by surveyors not sampling across the entire cross-section of the river, and thus not sampling certain microhabitats. Theoretically certain macroinvertebrate community traits may be either, under or over represented, by one or other, of the sampling procedures, by limiting the area in which the sampling is undertaken following the switch in 2010 to the New method. For example, if the New method were to exclude a certain habitat type from being sampled, or result in proportionally more time being spent sampling a particular habitat (when compared to the Old method) then one may question the validity of using the entire sampling record in certain community based analysis. It is hypothesised here that by not sampling the deeper faster sections as a result of the change to the New sampling method that species associated with such environments may be underrepresented in the New method samples and those associated with slacker marginal habitats overrepresented: review of taxa flow preferences has been undertaken to assess this. There are two flow scores that represent the selected taxa (Extence et al. 1999):  Flow Score 2 (FS2): Taxa primarily associated with moderate to fast flows  Flow Score 4 (FS4): Taxa associated with flowing (usually slow) and standing waters. Review of the taxa, that best discriminate between the methods, identifies ten taxa as FS2, three as FS4 and two as having no flow preference. Eight of the ten taxa primarily associated with moderate to fast flows (FS2) show a greater abundance in the Old method samples indicating, on average, a greater abundance of these taxa in the New method samples. This lends strength to the hypothesis that greater sampling of faster habitats occurred with the Old method.

6

Lower Test NEP Volume 3: Appendices

Table 3 Analysis of similarities test outputs

Average Average Cumulative Flow Contribution Taxa abundance abundance contribution score* % (Old method) (New method) %

Brachycentrus subnubilus 2 2.47 15.09 6.94 6.94

Baetis sp. 2 13.02 4.7 4.94 11.88

Chironomidae n/a 12.48 11.94 3.93 15.81

Lepidostomatidae 2 7.94 5.34 3.51 19.32

Oligochaeta n/a 10.1 14.6 3.48 22.8

Caenis pusilla 2 6.96 3.44 3.37 26.17

Cheumatopsyche lepida 2 6.93 2.09 3.21 29.37

Limnius sp 2 8.86 3.4 3.17 32.54

Serratella ignita 2 4.56 3.69 2.85 35.39

Gammarus pulex 2 9.13 11.73 2.77 38.17

Planaria 4 6.69 2.07 2.68 40.84

Oulimnius 2 5.4 3.29 2.31 43.15

Simuliidae 2 5.5 2.31 2.05 45.2

Mystacides sp. 4 4.63 3.38 1.97 47.17

Sphariidae 4 7.65 4.85 1.96 49.13 Note: Shaded cell indicates group in which higher average abundance occurs. Abundance values are square root transformed. *Flow Score obtained after Extence et al. 1999, where Flow 1 is associated with rapid flow, and Flow score 6 associated with drying tor drought affected sites.

However, the species that best discriminates between the samples (identified as contributing nearly 7% of the dissimilarity between the Old and New method grouping) is the cased caddisfly Brachycentrus subnubilus. Notably, and in contrast to the majority of the FS2 species, B. subnubilus occurs at a much higher average abundance in the New method sample grouping. This species has a very strong association with submerged macrophyte abundance (in particular Ranunculus sp.) and is susceptible to population changes as a result of weed cutting activities i.e. weed removal can significantly reduce densities at a local scale (Gunn, 1985 and Dawson et al., 1991). It is, therefore, possible that biotic interactions such as these have acted to create the disparity between Old and New sample groups, rather than the direct change in sampling methodology. It is, however, noted that detailed historical information on macrophyte abundance and weed cutting practises would have be needed to ascertain the importance of such species associations. Only three of the taxa identified as best describing the dissimilarity between Old and New methods are associated with slow flowing habitats. In both cases these were found in greater average abundance in the Old method sites. That is to say that there is limited evidence to suggest that the change to the New method has resulted in a over-representation of species associated with slower habitats and that the differences are better described by subtle changes in the numbers of a few species associated with moderate to fast flows. Interestingly, none of the taxa associated with rapid flow (Flow Score 1) which were present within the macroinvertebrate record were identified as accounting individually for more than 1% of the dissimilarity. If the change to the New method were responsible for noticeable differences in species traits then one might have expected that the taxa associated with more rapid flow would account for a greater difference between the groups. In summary, SIMPER analysis has identified that differences between the Old and New method sampling group is described by a higher abundance of a few taxa associated with moderate to fast flow although the greatest dissimilarity is described by a species strongly associated with macrophyte cover.

7

Lower Test NEP Volume 3: Appendices

There is no evidence to suggest that the change to the New method has resulted in sampling bias creating disproportionate representation of species associated with either rapid or slow flowing conditions.

Suitability of the data for use in flow investigations ANOSIM analysis has shown there to be a statistically significant difference between samples when grouped as either being taken with the Old or New sampling method. In addition, significant seasonal and sampling year influence on the macroinvertebrate communities is also evident, and this is identified as being related to subtle differences in species composition throughout the historical record. Notably, no significant differences have been identified between the Upstream and Downstream samples indicating that the macroinvertebrate communities are less strongly influenced by location than they are by sampling method, season or year of sampling. SIMPER analysis, conducted to see if the change in sampling method is evident in the species flow traits, has identified that the difference between the macroinvertebrate communities sampled using either the Old or New method is best described by a higher abundance of a few taxa associated with moderate to fast flow in the Old method samples. Interestingly, the one species most strongly associated with the New method samples, Brachycentrus subnubilus, is also associated with moderate to fast flows and is known to be susceptible to changes in macrophyte cover / weed management thus alluding to the potential influence of management practises on the differences observed. Furthermore, there is no evidence to suggest that the change to the New method has resulted in disproportionate representation of species associated with either rapid or slow flowing conditions as hypothesised. In summary, there is no conclusive evidence to support the theory that the change in sampling method has accounted for significant differences identified and it is, therefore, concluded that whole data record can be use in flow investigations. In all instances it is advisable that any analysis undertaken combining Old and New method samples makes reference to the change in sampling method. Should the Environment Agency wish to further investigate the effects of the change in the sampling method, then it is recommended replicate sampling be adopted to provide a direct comparison of methods without the influence of temporal gradients.

8

Lower Test NEP Volume 3: Appendices

Appendix 7.3.2 MDS Outputs

Best 2-d configuration (Stress: 0.17) Sample 1 2 % DSMAY02 0.51 -1.05 8.3 DSSEPT02 1.02 0.73 3.7 DSMAY03 0.71 -0.46 2.2 DSOCT03 0.32 1.29 3.7 DSMAY04 0.58 -0.36 5.8 DSOCT04 -0.19 0.73 5.8 DSNOV07 -0.32 0.47 2.2 DSAPR08 -0.50 -0.24 3.9 DSAPR10 -1.05 -0.32 4.7 DSMAY10 -0.50 -1.03 2.6 DSNOV10A -0.57 0.22 3.4 DSNOV10B -0.66 0.56 2.7 DSAPR11 -0.84 -0.13 2.8 DSJUN11 -1.10 -1.42 8.7 DSNOV11 -0.79 0.72 3.7 USMAY02 1.06 -0.74 1.8 USSPT02 1.00 0.23 3.5 USMAY03 0.65 -0.51 1.6 USOCT03 0.49 0.61 1.7 USMAY04 0.75 -0.22 2.6 USOCT04 0.73 0.67 2.4 USMAY05 0.99 -0.10 3.3 USNOV07 0.20 0.61 3.8 USAPR08 -0.37 -0.43 7.8 USMAY10 0.06 -0.91 3.1 USNOV10A -1.11 0.63 2.5 USNOV10B -1.06 0.48 1.9

9

Lower Test NEP Volume 3: Appendices

Appendix 7.3.3 ANOSIM Outputs

Method (Old / New)

ANOSIM Analysis of Similarities

One-Way Analysis

Resemblance worksheet Name: Resem1 Data type: Similarity Selection: All

Factor Values Factor: Method old new

Factor Groups Sample Method DSMAY02 old DSSEPT02 old DSMAY03 old DSOCT03 old DSMAY04 old DSOCT04 old DSNOV07 old DSAPR08 old USMAY02 old USSPT02 old USMAY03 old USOCT03 old USMAY04 old USOCT04 old USMAY05 old USNOV07 old USAPR08 old DSAPR10 new DSMAY10 new DSNOV10A new DSNOV10B new DSAPR11 new DSJUN11 new DSNOV11 new USMAY10 new USNOV10A new USNOV10B new

Global Test Sample statistic (Global R): 0.491 Significance level of sample statistic: 0.1% Number of permutations: 999 (Random sample from 8436285) Number of permuted statistics greater than or equal to Global R: 0

10

Lower Test NEP Volume 3: Appendices

Season (Spring / Autumn)

ANOSIM Analysis of Similarities

One-Way Analysis

Resemblance worksheet Name: Resem1 Data type: Similarity Selection: All

Factor Values Factor: Season spring autumn

Factor Groups Sample Season DSMAY02 spring DSMAY03 spring DSMAY04 spring DSAPR08 spring DSAPR10 spring DSMAY10 spring DSAPR11 spring DSJUN11 spring USMAY02 spring USMAY03 spring USMAY04 spring USMAY05 spring USAPR08 spring USMAY10 spring DSSEPT02 autumn DSOCT03 autumn DSOCT04 autumn DSNOV07 autumn DSNOV10A autumn DSNOV10B autumn DSNOV11 autumn USSPT02 autumn USOCT03 autumn USOCT04 autumn USNOV07 autumn USNOV10A autumn USNOV10B autumn

Global Test Sample statistic (Global R): 0.428 Significance level of sample statistic: 0.1% Number of permutations: 999 (Random sample from 20058300) Number of permuted statistics greater than or equal to Global R: 0

11

Lower Test NEP Volume 3: Appendices

Location (Upstream / Downstream) ANOSIM Analysis of Similarities

One-Way Analysis

Resemblance worksheet Name: Resem1 Data type: Similarity Selection: All

Factor Values Factor: Location ds us

Factor Groups Sample Location DSMAY02 ds DSSEPT02 ds DSMAY03 ds DSOCT03 ds DSMAY04 ds DSOCT04 ds DSNOV07 ds DSAPR08 ds DSAPR10 ds DSMAY10 ds DSNOV10A ds DSNOV10B ds DSAPR11 ds DSJUN11 ds DSNOV11 ds USMAY02 us USSPT02 us USMAY03 us USOCT03 us USMAY04 us USOCT04 us USMAY05 us USNOV07 us USAPR08 us USMAY10 us USNOV10A us USNOV10B us

Global Test Sample statistic (Global R): 0.088 Significance level of sample statistic: 9.2% Number of permutations: 999 (Random sample from 17383860) Number of permuted statistics greater than or equal to Global R: 91

12

Lower Test NEP Volume 3: Appendices

Location (Year groups) ANOSIM Analysis of Similarities

One-Way Analysis

Resemblance worksheet Name: Resem1 Data type: Similarity Selection: All

Factor Values Factor: Year group 02-03 04-05 06-07 08-09 10-11

Factor Groups Sample Year group DSMAY02 02-03 DSSEPT02 02-03 DSMAY03 02-03 DSOCT03 02-03 USMAY02 02-03 USSPT02 02-03 USMAY03 02-03 USOCT03 02-03 DSMAY04 04-05 DSOCT04 04-05 USMAY04 04-05 USOCT04 04-05 USMAY05 04-05 DSNOV07 06-07 USNOV07 06-07 DSAPR08 08-09 USAPR08 08-09 DSAPR10 10-11 DSMAY10 10-11 DSNOV10A 10-11 DSNOV10B 10-11 DSAPR11 10-11 DSJUN11 10-11 DSNOV11 10-11 USMAY10 10-11 USNOV10A 10-11 USNOV10B 10-11

Global Test Sample statistic (Global R): 0.427 Significance level of sample statistic: 0.1% Number of permutations: 999 (Random sample from a large number) Number of permuted statistics greater than or equal to Global R: 0

13

Lower Test NEP Volume 3: Appendices

Pairwise Tests

Significance Possible Actual Groups R Statistic Number >=Observed Level % Permutations Permutations 02-03, 04-05 0.054 28.1 1287 999 280 02-03, 06-07 0.358 4.4 45 45 2 02-03, 08-09 0.56 2.2 45 45 1 02-03, 10-11 0.64 0.1 43758 999 0 04-05, 06-07 0.455 4.8 21 21 1 04-05, 08-09 0.855 4.8 21 21 1 04-05, 10-11 0.58 0.1 3003 999 0 06-07, 08-09 1 33.3 3 3 1 06-07, 10-11 0.18 22.7 66 66 15 08-09, 10-11 0.172 18.2 66 66 12

14

Lower Test NEP Volume 3: Appendices

Appendix 7.3.4 SIMPER Outputs

Method (Old / New) SIMPER Similarity Percentages - species contributions

One-Way Analysis

Data worksheet Name: Data1 Data type: Abundance Sample selection: All Variable selection: All

Parameters Resemblance: S17 Bray Curtis similarity Cut off for low contributions: 90.00%

Factor Groups Sample Method DSMAY02 old DSSEPT02 old DSMAY03 old DSOCT03 old DSMAY04 old DSOCT04 old DSNOV07 old DSAPR08 old USMAY02 old USSPT02 old USMAY03 old USOCT03 old USMAY04 old USOCT04 old USMAY05 old USNOV07 old USAPR08 old DSAPR10 new DSMAY10 new DSNOV10A new DSNOV10B new DSAPR11 new DSJUN11 new DSNOV11 new USMAY10 new USNOV10A new USNOV10B new

Group old Average similarity: 57.81 Av.Abund Av.Sim Sim/SD Contrib% Cum.%

Baetis sp. 13.02 4.85 2.65 8.39 8.39 Chironomidae 12.48 4.31 1.94 7.45 15.84 Oligochaeta 10.10 3.95 4.17 6.84 22.68 Gammarus pulex 9.13 3.35 2.77 5.80 28.48 Sphariidae 7.65 2.89 3.16 4.99 33.47 Elmis sp. 7.96 2.82 3.70 4.88 38.35 Limnius sp 8.86 2.80 2.01 4.84 43.19 Lepidostomatidae 7.94 2.38 1.31 4.11 47.30 Planaria 6.69 2.06 1.75 3.56 50.87 Cheumatopsyche lepida 6.93 1.72 0.97 2.97 53.84 Simuliidae 5.50 1.66 1.72 2.88 56.72 Ephemera sp 4.32 1.58 2.31 2.73 59.45 Oulimnius 5.40 1.45 1.24 2.51 61.97 Athripsodes sp. 4.36 1.28 1.63 2.22 64.19 Caenis pusilla 6.96 1.22 0.68 2.11 66.30 Mystacides sp. 4.63 1.20 1.04 2.08 68.38 Serratella ignita 4.56 1.10 0.95 1.91 70.29 Alainites muticus 3.79 0.92 1.24 1.59 71.87 Hydroptila sp. 3.24 0.90 1.68 1.56 73.43 Hydropsyche siltalai 3.27 0.89 1.12 1.55 74.98 15

Lower Test NEP Volume 3: Appendices

Av.Abund Av.Sim Sim/SD Contrib% Cum.%

Aphelocheirus aestivalis 2.49 0.89 2.06 1.54 76.52 Brachycentrus subnubilus 2.47 0.85 1.47 1.47 77.99 Asellus aquaticus 2.78 0.80 1.31 1.38 79.37 Oecetis sp 2.88 0.79 1.31 1.37 80.74 Sericostoma personatum 2.20 0.68 1.40 1.17 81.91 Heptagenia sp. 2.01 0.66 1.33 1.14 83.06 Polycentropidae 2.45 0.65 0.92 1.12 84.18 Bithyniidae 1.84 0.63 1.91 1.08 85.26 Rhyacophilidae 2.17 0.60 1.31 1.03 86.30 Crangonyx pseudogracilis 2.20 0.59 0.85 1.02 87.32 Ithytrichia sp. 2.27 0.56 0.90 0.97 88.28 Erpobdellidae 1.45 0.50 1.47 0.86 89.14 Caenis luctuosa 1.79 0.48 0.81 0.83 89.98 Psychomyia pusilla 1.65 0.39 1.00 0.68 90.66

Group new Average similarity: 58.55

Species Av.Abund Av.Sim Sim/SD Contrib% Cum.% Oligochaeta 14.60 6.87 4.22 11.74 11.74 Brachycentrus subnubilus 15.09 5.72 2.58 9.77 21.51 Gammarus pulex 11.73 5.68 5.30 9.71 31.22 Chironomidae 11.94 5.23 2.68 8.93 40.15 Elmis sp. 5.67 2.98 5.90 5.09 45.23 Sphariidae 4.85 2.21 2.49 3.78 49.01 Athripsodes sp. 4.18 2.05 4.01 3.50 52.51 Ephemera sp 5.10 2.00 1.26 3.41 55.93 Baetis sp. 4.70 1.67 2.07 2.85 58.78 Sericostoma personatum 3.42 1.64 3.75 2.80 61.58 Limnius sp 3.40 1.52 1.80 2.60 64.18 Lepidostomatidae 5.34 1.49 1.24 2.54 66.71 Mystacides sp. 3.38 1.21 1.27 2.07 68.78 Caenis pusilla 3.44 1.19 1.11 2.04 70.82 Polycentropidae 2.86 1.17 1.57 2.01 72.83 Oulimnius 3.29 1.06 0.80 1.81 74.64 Asellus aquaticus 2.63 1.06 1.17 1.80 76.44 Simuliidae 2.31 0.95 2.47 1.62 78.06 Planaria 2.07 0.94 3.20 1.60 79.66 Ithytrichia sp. 2.93 0.92 0.86 1.56 81.22 Potamophylax sp 2.08 0.74 1.49 1.26 82.49 Heptagenia sp. 2.20 0.74 1.25 1.26 83.75 Oecetis sp 1.67 0.59 1.07 1.00 84.75 Aphelocheirus aestivalis 1.39 0.57 1.72 0.98 85.73 Chaetopteryx villosa 2.67 0.56 0.69 0.96 86.69 Cheumatopsyche lepida 2.09 0.52 0.78 0.90 87.59 Serratella ignita 3.69 0.49 0.43 0.84 88.43 Crangonyx pseudogracilis 1.98 0.46 0.56 0.79 89.21 Hydropsyche pellucidula 1.40 0.45 0.86 0.77 89.99 Bithyniidae 1.40 0.44 0.91 0.75 90.74

Groups old & new Average dissimilarity = 48.67 Species Group old Group new Av.Diss Diss/SD Contrib% Cum.% Av.Abund Av.Abund Brachycentrus subnubilus 2.47 15.09 3.38 1.29 6.94 6.94 Baetis sp. 13.02 4.70 2.40 1.72 4.94 11.88 Chironomidae 12.48 11.94 1.91 1.25 3.93 15.81 Lepidostomatidae 7.94 5.34 1.71 1.37 3.51 19.32 Oligochaeta 10.10 14.60 1.69 1.14 3.48 22.80 Caenis pusilla 6.96 3.44 1.64 1.08 3.37 26.17 Cheumatopsyche lepida 6.93 2.09 1.56 1.35 3.21 29.37 Limnius sp 8.86 3.40 1.54 1.30 3.17 32.54 Serratella ignita 4.56 3.69 1.39 1.21 2.85 35.39 Gammarus pulex 9.13 11.73 1.35 1.22 2.77 38.17 Planaria 6.69 2.07 1.30 1.37 2.68 40.84 Oulimnius 5.40 3.29 1.12 1.29 2.31 43.15 Simuliidae 5.50 2.31 1.00 1.09 2.05 45.20 Mystacides sp. 4.63 3.38 0.96 1.28 1.97 47.17 Sphariidae 7.65 4.85 0.95 1.40 1.96 49.13 Alainites muticus 3.79 0.87 0.89 1.05 1.82 50.95 Elmis sp. 7.96 5.67 0.84 1.12 1.73 52.67 Ephemera sp 4.32 5.10 0.83 1.46 1.71 54.38 Hydropsyche siltalai 3.27 0.77 0.76 1.35 1.55 55.93

16

Lower Test NEP Volume 3: Appendices

Species Group old Group new Av.Diss Diss/SD Contrib% Cum.% Av.Abund Av.Abund Athripsodes sp. 4.36 4.18 0.72 1.39 1.48 57.41 Hydroptila sp. 3.24 1.46 0.72 1.15 1.48 58.89 Chaetopteryx villosa 0.38 2.67 0.72 0.98 1.48 60.37 Ithytrichia sp. 2.27 2.93 0.68 1.36 1.40 61.77 Crangonyx pseudogracilis 2.20 1.98 0.61 1.35 1.24 63.01 Polycentropidae 2.45 2.86 0.58 1.42 1.19 64.20 Asellus aquaticus 2.78 2.63 0.58 1.26 1.19 65.38 Oecetis sp 2.88 1.67 0.55 1.03 1.12 66.51 Potamophylax sp 0.20 2.08 0.54 1.34 1.12 67.62 Caenis luctuosa 1.79 1.64 0.51 1.27 1.05 68.67 Labiobaetis atrebatinus 1.75 0.43 0.51 0.73 1.04 69.71 Isoperla grammatica 1.81 0.97 0.49 1.08 1.00 70.71 Sericostoma personatum 2.20 3.42 0.48 1.41 0.98 71.69 Limnephilidae 0.30 1.70 0.47 0.99 0.96 72.65 Rhyacophilidae 2.17 0.78 0.45 1.19 0.93 73.59 Heptagenia sp. 2.01 2.20 0.44 1.27 0.91 74.49 Leuctra geniculata 1.50 0.83 0.43 1.03 0.89 75.38 Psychomyia pusilla 1.65 0.42 0.42 1.26 0.85 76.24 Ancylus fluviatilis 1.60 0.74 0.42 1.10 0.85 77.09 Hydropsyche pellucidula 1.57 1.40 0.41 1.15 0.85 77.94 Tipulidae 1.56 0.00 0.41 0.60 0.84 78.78 Centroptilum luteolum 1.50 0.40 0.40 0.78 0.83 79.60 Ostracoda 1.57 1.10 0.40 1.22 0.83 80.43 Mystacides azurea 0.00 1.38 0.40 0.49 0.81 81.24 Aphelocheirus aestivalis 2.49 1.39 0.38 1.40 0.78 82.02 Bithyniidae 1.84 1.40 0.33 1.24 0.68 82.70 Gyrinidae 1.13 1.21 0.32 1.27 0.66 83.36 Hydropsyche contubernalis 0.93 0.36 0.32 0.66 0.65 84.01 Erpobdellidae 1.45 0.91 0.30 1.46 0.63 84.63 Riolus subviolaceus 1.04 1.08 0.30 1.15 0.62 85.26 Agapetus sp. 0.93 0.72 0.29 1.11 0.59 85.84 Goera pilosa 1.18 0.52 0.28 1.39 0.58 86.42 Calopteryx splendens 0.82 0.89 0.28 1.16 0.57 86.99 Ceratopogonidae 1.25 0.58 0.27 1.08 0.55 87.54 Planorbidae 0.57 0.87 0.23 1.11 0.48 88.02 Procloeon sp 0.81 0.10 0.23 0.67 0.47 88.50 Dicranota 0.56 0.50 0.23 0.76 0.47 88.97 Limnephilus lunatus 0.47 0.71 0.22 0.99 0.46 89.43 Helobdella stagnalis 1.06 0.92 0.21 1.08 0.43 89.86 Anisus vortex 0.62 0.41 0.21 0.95 0.42 90.28

17

Lower Test NEP Volume 3: Appendices

Location (Downstream / Upstream) SIMPER Similarity Percentages - species contributions

One-Way Analysis

Data worksheet Name: Data1 Data type: Abundance Sample selection: All Variable selection: All

Parameters Resemblance: S17 Bray Curtis similarity Cut off for low contributions: 90.00%

Factor Groups Sample Location DSMAY02 ds DSSEPT02 ds DSMAY03 ds DSOCT03 ds DSMAY04 ds DSOCT04 ds DSNOV07 ds DSAPR08 ds DSAPR10 ds DSMAY10 ds DSNOV10A ds DSNOV10B ds DSAPR11 ds DSJUN11 ds DSNOV11 ds USMAY02 us USSPT02 us USMAY03 us USOCT03 us USMAY04 us USOCT04 us USMAY05 us USNOV07 us USAPR08 us USMAY10 us USNOV10A us USNOV10B us

Group ds Average similarity: 54.60 Species Av.Abund Av.Sim Sim/SD Contrib% Cum.% Chironomidae 13.61 5.76 2.49 10.55 10.55 Oligochaeta 11.95 5.18 2.93 9.48 20.03 Gammarus pulex 8.89 3.90 2.63 7.14 27.17 Baetis sp. 8.06 3.08 1.79 5.64 32.81 Elmis sp. 5.69 2.73 5.23 4.99 37.80 Sphariidae 6.23 2.68 2.61 4.92 42.71 Brachycentrus subnubilus 8.85 2.29 1.06 4.20 46.91 Ephemera sp 4.78 1.86 1.42 3.41 50.32 Limnius sp 4.18 1.64 1.70 3.00 53.33 Lepidostomatidae 5.76 1.58 1.03 2.89 56.21 Athripsodes sp. 3.76 1.56 2.37 2.86 59.07 Simuliidae 3.43 1.33 2.25 2.43 61.50 Mystacides sp. 3.53 1.21 1.34 2.22 63.72 Oulimnius 4.08 1.19 0.98 2.19 65.90 Sericostoma personatum 2.74 1.13 1.80 2.07 67.97 Planaria 3.06 1.06 1.49 1.94 69.91 Polycentropidae 2.78 1.04 1.42 1.91 71.82 Caenis pusilla 4.04 1.03 0.82 1.89 73.71 Cheumatopsyche lepida 4.17 1.01 0.91 1.84 75.55 Asellus aquaticus 2.85 0.98 1.17 1.79 77.34 Serratella ignita 3.66 0.82 0.72 1.50 78.84 Aphelocheirus aestivalis 1.91 0.79 1.87 1.45 80.29 Crangonyx pseudogracilis 2.43 0.73 0.85 1.33 81.62 18

Lower Test NEP Volume 3: Appendices

Species Av.Abund Av.Sim Sim/SD Contrib% Cum.% Heptagenia sp. 1.94 0.62 0.99 1.14 82.76 Ithytrichia sp. 2.35 0.62 0.75 1.13 83.89 Hydroptila sp. 2.36 0.60 1.03 1.09 84.98 Oecetis sp 2.02 0.57 0.87 1.05 86.02 Caenis luctuosa 1.88 0.53 0.76 0.97 86.99 Ostracoda 1.69 0.46 0.79 0.84 87.84 Bithyniidae 1.29 0.42 0.95 0.77 88.61 Glossiphonia complanata 0.94 0.38 1.27 0.70 89.31 Alainites muticus 1.82 0.38 0.62 0.69 90.00

Group us Average similarity: 56.76

Species Av.Abund Av.Sim Sim/SD Contrib% Cum.% Gammarus pulex 11.60 4.47 3.33 7.87 7.87 Oligochaeta 11.54 4.41 3.63 7.77 15.64 Baetis sp. 12.30 3.59 1.51 6.33 21.96 Chironomidae 10.62 3.43 2.06 6.04 28.00 Limnius sp 10.16 3.16 2.14 5.57 33.57 Elmis sp. 8.89 3.14 4.90 5.54 39.11 Lepidostomatidae 8.50 2.55 1.65 4.49 43.59 Sphariidae 7.10 2.41 2.90 4.25 47.84 Planaria 7.39 2.14 1.68 3.76 51.60 Ephemera sp 4.39 1.55 2.18 2.73 54.33 Athripsodes sp. 4.97 1.50 1.75 2.64 56.97 Oulimnius 5.29 1.41 1.20 2.48 59.45 Cheumatopsyche lepida 6.35 1.34 0.81 2.37 61.82 Simuliidae 5.42 1.34 1.23 2.36 64.18 Caenis pusilla 7.68 1.29 0.79 2.28 66.46 Mystacides sp. 4.96 1.21 0.91 2.14 68.60 Brachycentrus subnubilus 5.00 1.07 1.02 1.88 70.48 Serratella ignita 4.96 0.91 0.71 1.59 72.08 Hydropsyche siltalai 3.36 0.90 1.12 1.59 73.67 Oecetis sp 2.95 0.88 2.76 1.54 75.21 Rhyacophilidae 2.73 0.83 2.08 1.46 76.67 Sericostoma personatum 2.55 0.79 1.47 1.39 78.06 Heptagenia sp. 2.26 0.78 2.10 1.37 79.43 Asellus aquaticus 2.56 0.76 1.40 1.34 80.76 Bithyniidae 2.15 0.76 2.98 1.33 82.10 Ithytrichia sp. 2.71 0.74 1.04 1.31 83.41 Alainites muticus 3.82 0.72 1.01 1.27 84.68 Aphelocheirus aestivalis 2.30 0.67 1.68 1.19 85.87 Hydroptila sp. 2.86 0.65 1.14 1.15 87.02 Isoperla grammatica 2.42 0.61 1.15 1.08 88.10 Polycentropidae 2.37 0.57 0.84 1.01 89.11 Psychomyia pusilla 2.02 0.48 0.97 0.85 89.97 Hydropsyche pellucidula 1.74 0.47 1.02 0.83 90.80

Groups ds & us Average dissimilarity = 45.87 Species Group ds Group us Av.Diss Diss/SD Contrib% Cum.% Av.Abund Av.Abund Brachycentrus subnubilus 8.85 5.00 2.08 0.86 4.54 4.54 Baetis sp. 8.06 12.30 2.01 1.53 4.37 8.91 Chironomidae 13.61 10.62 1.86 1.25 4.07 12.98 Caenis pusilla 4.04 7.68 1.77 1.08 3.86 16.83 Limnius sp 4.18 10.16 1.68 1.35 3.65 20.49 Lepidostomatidae 5.76 8.50 1.63 1.36 3.56 24.04 Cheumatopsyche lepida 4.17 6.35 1.51 1.46 3.29 27.33 Oligochaeta 11.95 11.54 1.37 1.12 2.99 30.32 Gammarus pulex 8.89 11.60 1.32 1.18 2.87 33.19 Planaria 3.06 7.39 1.32 1.28 2.87 36.06 Serratella ignita 3.66 4.96 1.29 1.16 2.82 38.87 Oulimnius 4.08 5.29 1.12 1.33 2.44 41.31 Mystacides sp. 3.53 4.96 1.02 1.27 2.22 43.54 Simuliidae 3.43 5.42 1.01 1.08 2.20 45.74 Elmis sp. 5.69 8.89 0.97 1.20 2.11 47.85 Sphariidae 6.23 7.10 0.86 1.35 1.88 49.72 Alainites muticus 1.82 3.82 0.86 0.99 1.87 51.60 Athripsodes sp. 3.76 4.97 0.76 1.24 1.65 53.24 Hydropsyche siltalai 1.53 3.36 0.74 1.44 1.61 54.85 Ephemera sp 4.78 4.39 0.73 1.38 1.58 56.43 Hydroptila sp. 2.36 2.86 0.66 1.14 1.44 57.87 Ithytrichia sp. 2.35 2.71 0.63 1.34 1.38 59.24

19

Lower Test NEP Volume 3: Appendices

Species Group ds Group us Av.Diss Diss/SD Contrib% Cum.% Av.Abund Av.Abund Polycentropidae 2.78 2.37 0.57 1.36 1.25 60.49 Isoperla grammatica 0.76 2.42 0.57 1.28 1.24 61.73 Crangonyx pseudogracilis 2.43 1.72 0.57 1.28 1.23 62.96 Oecetis sp 2.02 2.95 0.56 1.02 1.22 64.18 Asellus aquaticus 2.85 2.56 0.55 1.21 1.20 65.38 Rhyacophilidae 0.79 2.73 0.52 1.32 1.14 66.52 Chaetopteryx villosa 1.68 0.67 0.52 0.76 1.14 67.65 Labiobaetis atrebatinus 1.61 0.82 0.50 0.78 1.08 68.73 Tipulidae 0.32 1.82 0.49 0.67 1.06 69.80 Leuctra geniculata 0.77 1.86 0.48 1.10 1.05 70.84 Caenis luctuosa 1.88 1.56 0.48 1.25 1.04 71.88 Psychomyia pusilla 0.54 2.02 0.45 1.33 0.99 72.87 Sericostoma personatum 2.74 2.55 0.43 1.40 0.93 73.79 Hydropsyche pellucidula 1.32 1.74 0.42 1.25 0.92 74.71 Ancylus fluviatilis 1.06 1.56 0.42 1.13 0.91 75.62 Heptagenia sp. 1.94 2.26 0.40 1.29 0.88 76.50 Ostracoda 1.69 1.03 0.40 1.24 0.87 77.37 Centroptilum luteolum 1.43 0.67 0.39 0.82 0.85 78.21 Mystacides azurea 0.00 1.15 0.36 0.44 0.79 79.00 Potamophylax sp 0.96 0.82 0.36 0.86 0.79 79.79 Limnephilidae 0.78 0.86 0.36 0.79 0.78 80.56 Hydropsyche contubernalis 0.60 0.87 0.35 0.69 0.75 81.32 Aphelocheirus aestivalis 1.91 2.30 0.35 1.27 0.75 82.07 Bithyniidae 1.29 2.15 0.33 1.19 0.72 82.79 Gyrinidae 0.93 1.45 0.33 1.42 0.71 83.50 Agapetus sp. 0.50 1.30 0.32 1.15 0.71 84.21 Riolus subviolaceus 1.07 1.05 0.30 1.06 0.65 84.86 Erpobdellidae 1.12 1.41 0.27 1.34 0.58 85.44 Calopteryx splendens 1.07 0.57 0.27 1.14 0.58 86.02 Goera pilosa 0.75 1.18 0.26 1.26 0.57 86.59 Ceratopogonidae 0.92 1.10 0.26 1.05 0.56 87.16 Dicranota 0.38 0.73 0.24 0.74 0.52 87.68 Glossiphonia complanata 0.94 0.35 0.23 1.45 0.49 88.17 Procloeon sp 0.79 0.25 0.22 0.71 0.49 88.66 Planorbidae 0.76 0.58 0.22 1.05 0.47 89.13 Helobdella stagnalis 0.90 1.15 0.21 1.10 0.46 89.59 Limnephilus lunatus 0.72 0.35 0.21 0.96 0.46 90.05

20

Lower Test NEP Volume 3: Appendices

Appendix 7.3.5 Macroinvertebrate biotic indices

Table 1 Biological metric scores for the Upstream and Downstream of abstraction site

FAMILY SPECIES Site Sample BMWP ASPT No Of Taxa CCI LIFE LIFE location date (observed) (observed) 22-May-02 201 6.09 33 29.43 7.23 7.62 20-Sep-02 251 6.28 40 32.86 7.21 7.73 21-May-03 240 6.15 39 27.78 7.23 7.45 01-Oct-03 207 5.91 35 22.77 7.03 7.36 19-May-04 185 6.17 30 16.45 6.93 7.25 22-Oct-04 166 6.15 27 16.33 7.04 7.39 Testwood 06-Nov-07 210 6.00 35 22.08 7.11 7.33 Downstream 15-Apr-08 196 5.94 33 15.91 6.97 7.38 abstraction 13-Apr-10 212 6.06 35 24.63 7.06 7.55 28-May-10 186 6.41 29 18.97 7.34 7.76 04-Nov-10 201 6.48 31 25.71 7.39 7.89 29-Nov-10 225 6.25 36 25.53 7.20 7.68 07-Apr-11 187 6.23 30 17.50 7.20 7.62 02-Jun-11 167 6.42 26 21.00 7.46 7.92 02-Nov-11 169 6.04 28 17.00 7.14 7.77 22-May-02 219 6.64 33 29.77 7.40 8.00 20-Sep-02 235 6.35 37 28.26 7.29 7.74 21-May-03 246 6.47 38 26.05 7.50 8.02 01-Oct-03 255 6.22 41 33.65 7.10 7.63 19-May-04 237 6.58 36 26.67 7.57 7.70 Testwood 22-Oct-04 192 6.19 31 17.83 7.27 7.67 Upstream abstraction 06-May-05 228 6.51 35 27.50 7.61 7.92 06-Nov-07 178 5.93 30 16.33 7.22 7.50 15-Apr-08 198 6.19 32 16.70 7.38 7.51 28-May-10 219 6.44 34 25.61 7.45 7.98 04-Nov-10 206 6.24 33 20.00 6.84 7.31 29-Nov-10 203 6.15 33 22.94 7.06 7.72 Unpaired t- test (p 0.036 0.023 0.091 0.175 0.0345 0.047 value) Downstream 207 6.16 34 23 7.14 7.53 Mean Upstream 218 6.33 34 24 7.31 7.73 Mean

21 Lower Test NEP Volume 3: Appendices

Table 2 LIFE (F) Observed to Expected (O:E) ratios

Sample FAMILY LIFE Family O:E Family Site location date (Observed) (Expected)* 22-May-02 7.23 6.975 1.037 20-Sep-02 7.21 6.95 1.037 21-May-03 7.23 6.975 1.037 01-Oct-03 7.03 6.95 1.012 19-May-04 6.93 6.975 0.994 22-Oct-04 7.04 6.95 1.013 06-Nov-07 7.11 6.95 1.023 Downstream Abstraction 15-Apr-08 6.97 6.975 0.999 13-Apr-10 7.06 6.975 1.012 28-May-10 7.34 6.975 1.052 04-Nov-10 7.39 6.95 1.063 29-Nov-10 7.20 6.95 1.036 07-Apr-11 7.20 6.975 1.032 02-Jun-11 7.46 6.9 1.081 02-Nov-11 7.14 6.95 1.027 22-May-02 7.40 6.975 1.061 20-Sep-02 7.29 6.95 1.049 21-May-03 7.50 6.975 1.075 01-Oct-03 7.10 6.95 1.022 19-May-04 7.57 6.975 1.085 22-Oct-04 7.27 6.95 1.046 Upstream Abstraction 06-May-05 7.61 6.975 1.091 06-Nov-07 7.22 6.95 1.039 15-Apr-08 7.38 6.975 1.058 28-May-10 7.45 6.975 1.068 04-Nov-10 6.84 6.95 0.985 29-Nov-10 7.06 6.95 1.016 Note: Expected scores calculate from RIVPACS analysis.

22

Ben Piper Atkins Oasis Business Park Eynsham Oxford OX29 4AH

[email protected] 07703 647128

© Atkins Ltd except where stated otherwise.

The Atkins logo, ‘Carbon Critical Design’ and the strapline ‘Plan Design Enable’ are trademarks of Atkins Ltd.