SEARCH

SEARCH BY CITATION

Keywords:

  • groundwater models;
  • Lotic invertebrate Index for Flow Evaluation;
  • low flows;
  • macroinvertebrates

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

Despite the pressing environmental, economic and social issues surrounding water abstraction, scientific methods for managing its ecological impacts remain in their infancy. In this paper, we demonstrate statistically significant relationships between in-stream ecological condition using macroinvertebrates and the hydrological effect of groundwater abstraction on surface water flows in streams originating from Permo-Triassic sandstone aquifers in the English midlands. Ecological condition was most strongly correlated to the effect of abstraction on medium-low flows (Q75) compared with effects at other flows, water quality, habitat or seasonal effects. Ecological impacts occurred when the effect of abstraction on Q75 flows exceeded 60%. The same relationships were shown among individual macroinvertebrate taxa, validating the biological responses. The hydroecological model has provided a scientific basis for making local decisions on investigation sites and has helped to focus resources to areas at risk of not meeting Good Ecological Status under the Water Framework Directive because of abstraction.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

For over four decades, issues of water quality have dominated the research and management of riverine ecosystems. Major legislation (e.g. UK Water Acts of 1973 and 1983; US Clean Water Act 1977) has driven substantial investment by the water industry that has resulted in widespread improvement in the water quality of rivers in Europe and North America over recent decades. By contrast, the ecological issues associated with the quantity of water in rivers have received far less attention. This balance is currently being redressed given the increasing use of freshwater resources by humans conflicting with major legislation focusing on the protection of aquatic ecosystems (Acreman & Ferguson 2010; HM Government 2011).

Across Europe, the Water Framework Directive (WFD; 2000/60/EC) has been the primary legislative driver for increasing focus on the ecological impacts of water abstraction in rivers by specifying that ‘hydromorphology’ should underpin good ecological status. Improvements in water quality over recent decades mean that hydromorphological limits on ecological quality are becoming increasingly apparent (Vaughan et al. 2009), driving the need for local investigations to identify abstraction impacts and to inform mitigation measures. Meanwhile, links between ecology and hydromorphology that could lead to the development of the tools needed to provide robust assessments of the impact of water abstraction on aquatic ecology remain elusive. As a result, much current environmental river flow management relies largely on expert opinion with relatively little underpinning science (Acreman et al. 2008; Vaughan et al. 2009; SNIFFER 2012). This is of concern given the potential for unjustified burdens to be placed on water users or for environmental damage to be undetected. There is therefore an urgent need for water scientists to work together with aquatic ecologists to quantify the environmental flow requirements of riverine ecosystems to support good ecological status (Dunbar & Acreman 2001; Acreman & Dunbar 2004; Arthington et al. 2006).

Advances towards a tool for setting ecologically based flow targets in UK rivers have been led by the development of macroinvertebrate indices, such as the Lotic invertebrate index for flow evaluation (LIFE; Extence et al. 1999). LIFE uses expert opinion to weight invertebrate groups according to their preference to higher velocity flows. While LIFE has been correlated with historic hydrological and hydraulic parameters (Extence et al. 1999; Monk et al. 2008; Dunbar et al. 2010a, b) to set river flow targets (Exley 2006) and to characterise the discharge profiles of UK rivers (Monk et al. 2006), no published studies have related LIFE directly to the effects of water abstraction or incorporated it into an impact assessment tool for site investigations.

In this paper, we present a hydroecological tool, based on LIFE, for diagnosing the effects of abstraction on ecological communities in groundwater fed streams and for predicting where abstraction is likely to prevent the attainment of good ecological status under the WFD. We investigated the impacts of groundwater abstraction on surface water flow and macroinvertebrate assemblages in six headwater streams in the English midlands. The purpose of the hydroecological model is to provide a scientific basis for making local decisions on sites under investigation and to focus resources to outstanding areas of uncertainty in subsequent investigations. By using the reference condition approach, a further aim of the hydroecology model was to provide evidence to support the development of a standard classification and assessment method for hydromorphological pressure, consistent with the approach adopted by the WFD.

Background

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

The National Environment Programme (NEP) is a programme of actions and investigations for environmental improvement schemes that ensures that water companies in England and Wales do not cause the failure to meet European Union (EU) Directives, national targets and their statutory environmental obligations. The Environment Agency (UK) produces the NEP after consultation with the water industry and a number of other organisations. The NEP includes investigations into the impact of water abstraction on the environment.

Severn Trent Water Ltd (STWL) is licensed to abstract from 180 groundwater sources, primarily from the Permo-Triassic sandstones in Staffordshire, Shropshire, Worcestershire and Nottinghamshire. Groundwater provides about a third of supplies (about 600 Ml/d) so is a vital component of the Company's operation. During 2005–2010, STWL carried out investigations at 12 catchments under the NEP process. These were predominantly focused around groundwater abstractions from the Permo-Triassic sandstones. In-stream ecological and hydrological surveys were carried out at six of these sites, and these are the focus of this paper.

Site details

Multiple sampling locations were selected within each investigation site (Table 1). Across the six investigation sites, there were 17 paired ecology and hydrology sampling locations. The sampling locations within the six investigation sites were agreed with the Environment Agency (UK) and were selected to represent a range of locations that were both unaffected and potentially affected by groundwater abstraction in each investigation site. The six investigation sites are shown on Fig. 1, and the 17 sampling locations are described in Table 1.

figure

Figure 1. Site locations.

Download figure to PowerPoint

Table 1. Site statistics and flow data
SiteLocationCatchment area (km2)Qmean (Ml/d)Q75 (Ml/d)Q85 (Ml/d)Q95 (Ml/d)
Obs.Est. LTA obs.NaturalisedObs.Est. LTA obs.NaturalisedObs.Est. LTA obs.NaturalisedObs.Est. LTA obs.Naturalised
  1. LTA, long-term average.

Hoo BrookHB228.215.920.218.39.56.36.38.710.510.54.63.73.7
HB322.114.318.114.38.86.06.07.48.78.75.72.02.0
Sher BrookCC58.03.64.36.32.63.75.72.22.64.61.61.63.6
CC67.84.45.36.23.45.06.01.52.53.41.00.00.9
CC76.22.53.04.91.92.94.81.72.03.90.91.02.9
CC87.62.73.36.02.22.75.41.31.94.61.01.13.8
River BlitheRB10.90.30.30.90.10.20.80.10.00.70.00.00.7
RB44.31.61.34.60.61.24.50.50.13.40.20.33.7
RB59.35.14.110.02.33.69.51.91.37.21.41.57.4
Merryhill BrookMH26.83.75.04.60.30.20.20.10.20.20.00.00.0
MH49.62.73.76.60.50.43.20.20.33.10.00.02.8
MH513.23.85.28.90.30.13.80.10.03.70.00.03.7
MH713.63.75.19.20.10.04.20.00.04.10.00.04.1
Lonco BrookLM526.011.59.217.05.15.412.93.63.411.03.23.210.8
WFB33.617.217.222.117.217.211.117.217.29.117.217.28.6
LM735.715.912.723.38.45.312.85.04.311.94.24.311.9
Moddershall BrookMB62.94.43.92.73.84.14.13.44.14.12.63.33.3

As the sites were similar in physical character (unpolluted headwater streams that receive a significant percentage of their flow as baseflow from the underlying Permo-Triassic Sandstone), there was the potential to combine datasets across the sites to provide a region-wide analysis of the hydrogeology, hydrology and ecology of these streams.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

Data sets

Hydrometric data sets

Preliminary desk studies identified that a detailed local understanding of the hydrogeology of each catchment was required in order to quantify abstraction effects accurately. This involved topographical surveys, construction of new observation boreholes, shallow piezometers, and an extensive programme of groundwater level and monthly spot flow gauging over a 3- to 5-year period (see Table 1). This was designed to identify the flow and groundwater level conditions in each catchment under a range of hydrological conditions. This was important both for the impact assessment and to form a baseline against which the effects of any future changes in abstraction could be assessed.

Ecological data sets

Ecological assessment focused on macroinvertebrates that have well-defined associations with river flows (Extence et al. 1999) and are widely used as biological indicators of the integrity of riverine ecosystems (Wright et al. 2000). Macroinvertebrate sampling locations were paired with the spot flow gauging locations and samples were collected at the nearest suitable shallow, gravelly habitat to the flow gauging site (the macroinvertebrate sampling site was chosen a short-distance upstream of the flow gauging site to prevent interference when the flow gauging and macroinvertebrate sampling were conducted at the same time). Sampling followed Environment Agency (UK) kick/sweep sampling protocols and was undertaken on five occasions (summer and autumn 2008, and spring, summer and autumn 2009), generally at the same time as spot flow gauging events. Samples were processed in the laboratory following Environment Agency (UK) protocols to species level (except Sphaeriidae, Oligochaeta, Hydracarina, Chironomidae, Sphaeriidae and other Diptera to genus level).

Data analysis

The approach used to determine the effect of groundwater abstraction on stream flows was relatively simple; relying on Environment Agency (UK) estimates of catchment averaged effective precipitation. However, a similar approach could easily be adapted to use outputs from groundwater models (where these are available).

Conceptual model of streams fed by Permo-Triassic sandstone

The key elements of the conceptual model of the streams being investigated by STWL can be summarised as follows (Shepley & Streetly 2007):

  • Because of the high storage capacity of the Permo-Triassic Sandstone, groundwater levels in the aquifer typically only fluctuate seasonally by a metre or so. This means that baseflow to streams does not usually fluctuate rapidly.
  • However, when exposed to a sequence of dry years, the storage in the Permo-Triassic Sandstone aquifer may become depleted and may take several years to become replenished. Thus, streams (and their associated ecology) are naturally exposed to long sequences of above/below average baseflow. During the period of monitoring, flow conditions varied; flows were lower than average during the first part of the monitoring period (2004–2006) but were above average during the second part (2007–2009), which was also the period during which the ecological surveys were carried out.
  • Most public water supply abstractions from the Permo-Triassic Sandstone operate at fairly constant rate and evidence from both field studies, and groundwater model simulations suggests that their impact develops slowly over a number of years (or even decades) and applies fairly evenly across the range of seasonal flows to which the streams are exposed (i.e. the loss of stream flow in Ml/d is approximately the same at low flows as at high flows). This means that the relative proportion by which flow is reduced is larger at low flows than at high flows.
Determining the effect of abstraction on flows

In order to determine whether flows at each site deviated from expected conditions, it is necessary to make an estimate of flow rates under the current abstraction regime and to compare these with the estimated flows in the absence of abstraction. Although the sites were monitored at monthly intervals for several years, because of the variability of rainfall, there is no guarantee that the flows recorded were ‘typical’ of conditions. Long-term (1990–2007) observed flow statistics were therefore determined by comparison with flow conditions at appropriate continuous stream flow gauging sites (reference sites, i.e. on streams with similar catchments). The approach adopted is summarised concisely as follows:

  1. Estimate long-term (1990–2007) flow statistics (mean, Q75, Q85 and Q95 meaning daily mean flow exceeded 75, 85 and 95% of the time, respectively) for each spot flow gauging site by reference to similar sites at which the Environment Agency (UK) maintains a permanent flow gauge.
  2. Estimate the long-term mean natural outflow at each gauging site (effective precipitation multiplied by catchment area).
  3. Estimate the ‘flow reduction’ (mean natural flow minus estimated long-term mean observed flow). As there were few significant other abstractions or discharges in these catchments, this could largely be equated to the reduction in flow caused by STWL's abstractions (abstraction effect).
  4. Estimate the natural low flow by adding the ‘abstraction effect’ back to the estimated observed long-term low flow data.
Relating the effect of abstraction on flows to ecological impacts

For each sample, the macroinvertebrate assemblage was summarised as a LIFE score, which is an average of abundance-weighted flow groups that indicate the microhabitat preferences of each taxon for higher water velocities and clean gravel/cobble substrata or slow/still water velocities and finer substrata (Extence et al. 1999).

River Invertebrate Prediction and Classification System (RIVPACS) is a software package that is used to assess the biological quality of rivers in the UK (Wright et al. 2000). RIVPACS version III+ (Wright et al. 2000) was used to calculate the expected LIFE scores for each sample under reference conditions without abstraction. The ratio of the observed to expected family LIFE scores (LIFE O/E) was used to quantify impact on the macroinvertebrate community as a result of low flows. This type of analysis is consistent with the reference condition approach adopted by the WFD. Whilst LIFE O/E is not currently used to derive status classifications under the WFD, the Environment Agency (UK) use a lower 10th percentile LIFE O/E of 0.94 as a guideline to help identify if flow is a possible pressure acting on the ecology at a site.

To provide further evidence to show the relationship between macroinverebrate assemblages and abstraction effect on river flows, we examined the taxonomic composition and relative abundance of macroinvertebrate assemblages in relation to abstraction effect and other environmental variables, such as water quality. The multivariate structure of the macroinvertebrate assemblages was summarised into the main axes of variation using principal components analysis (PCA) performed using the CANOCO 4.5 statistics package (Leps & Smilauer 2003). The first and second PCA axis scores that describe the main components of variation in the macroinvertebrate assemblages were extracted and used as ecological indices that were then related to abstraction effect and other environmental variables.

The effect of habitat and seasonal differences might be high at the relatively small spatial and temporal scale of this study. Therefore, the first stage in establishing hydroecological relationships was to measure the significance of abstraction effect on macroinvertebrates in relation to the significance of other environmental variables. Stepwise multiple regression was used to test the relative significance of abstraction effect at different flows (% flow reduction at Qmean, Q75, Q85 and Q95, as explained earlier) in explaining variation in macroinvertebrate indices in relation to the significance of other environmental variables: water quality [biochemical oxygen demand (BOD)], river name (site) and sampling season.

Further ordination analysis was undertaken to directly examine the relative strength of different environmental variables in influencing the taxonomic composition and relative abundance of macroinvertebrates across the study sites in relation to the strength of abstraction effects. Redundancy analysis (RDA) is an extension of PCA that constrains the ordination to a new set of axes that describe the relative strength of significant environmental parameters. The first stage in the analysis is to add environmental parameters in a stepwise fashion to the model in order of their significance in explaining variation in the species data. The environmental parameters that were tested were: wetted width; water depth; pH; conductivity; abstraction effect at Q95, Q85, Q75 and Qmean; BOD; flow type; substratum composition; and macrophyte composition. Flow type, substratum composition and macrophyte composition were each expressed as the first axis of a PCA (PC1) that described the major gradient of variation of each of these parameters across the investigation sites. Each parameter is tested for significance using Monte Carlo permutations (499 random permutations) and then added to the model until no more parameters explain any variation in the species data. Therefore, any redundant or autocorrelated variables were rejected, and the relative strength of abstraction compared with other potentially confounding environmental influences in determining macroinvertebrate distributions was analysed.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

The stepwise multiple regression analysis showed that abstraction effect (flow reduction) at Q75 was the most significant predictor of LIFE O/E (Table 2). BOD, site and season were not related to LIFE O/E, which indicated that the effect of abstraction at Q75 was not confounded by water quality and that all sites and seasons could be combined into one hydroecology model. LIFE O/E was plotted against abstraction effect at Q75 that revealed three categories of abstraction effect and indicative ecological status (Fig. 2):

  • LIFE O/E scores were below 0.94 (indicating possible impacts of low flows on macroinvertebrates) when abstraction effect exceeded 80% of Q75;
  • LIFE O/E scores were largely above 0.945 (indicating no impacts of low flows on macroinvertebrates) when abstraction effect was less than 60% of Q75; and
  • LIFE O/E scores were both above and below 0.945 (indicating no impacts of low flows on macroinvertebrates at some locations and possible impacts at others) when abstraction effect was between 60 and 80% of Q75.
figure

Figure 2. The relationship between Lotic invertebrate Index for Flow Evaluation (LIFE) O/E scores and abstraction effect on medium-low flows (Q75). Horizontal line indicates LIFE O/E 0.94 below which low flows are a possible pressure acting on the ecological community.

Download figure to PowerPoint

Table 2. Results of stepwise multiple regression analysis for LIFE O/E, PCA axis 1 scores and PCA axis 2 scores
PredictorLIFE O/E R2 = 59.5%PCA axis 1 R2 = 64.2%PCA axis 2 R2 = 43.9%
TPTPTP
  1. T indicates the relative strength of each predictor, P indicates the statistical significance of each predictor in the model (significance level ≤ 0.05), and R2 indicates the proportion of variation explained by the whole model.

  2. BOD, biochemical oxygen demand; LIFE, Lotic invertebrate index for flow evaluation; NS, not significant; PCA, principal components analysis.

Qmean abstraction3.66< 0.001−7.97< 0.001−1.13NS
Q75 abstraction−5.77< 0.0014.54< 0.0014.35< 0.001
Q85 abstraction−1.05NS−0.01NS0.84NS
Q95 abstraction2.480.015−0.89NS−2.910.005
BOD1.02NS0.91NS−0.77NS
Site−1.39NS7.47< 0.001−0.89NS
Season−0.38NS−1.75NS−0.55NS

A one-way analysis of variance statistical test showed that mean LIFE O/E scores were significantly different between these abstraction effect categories (F2,84 = 85.27, P ≤ 0.001), indicating that these categories could provide a useful framework for setting abstraction effect thresholds with defined ecological effects (Fig. 3).

figure

Figure 3. Mean Lotic invertebrate index for flow evaluation (LIFE) O/E scores in abstraction effect categories [1 = < 60%; 2 = 60–80%; 3 = > 80% abstraction effect on medium-low flows (Q75)].

Download figure to PowerPoint

The PCA indicated that abstraction effect at Q75 was a strong influence on macroinvertebrate composition and relative abundance, consistent with the relationships described earlier for LIFE O/E (Table 2). The first two PCA axes explained 32.8% of the variation in macroinvertebrate community composition and relative abundance (PCA axis 1 explained 17.5%, and PCA axis 2 explained 15.3% variation). The corresponding Eigenvalues for PCA axis 1 and PCA axis 2 were 0.175 and 0.152, respectively (total variance = 1). These Eigenvalues were above the mean of all of the Eigenvalues (0.015), which indicated that interpretation of PCA axis 1 and axis 2 was meaningful (Leps & Smilauer 2003). PCA axis 1 was strongly correlated to abstraction effect at Q75, and this axis also described highly significant correlations with abstraction effect at mean flows and differences between individual sites (using site name as a predictor). The differences between individual sites were likely to reflect the effect of subtle differences in physical character between the investigation sites, as well as differences in abstraction effect between sites. PCA axis 2, however, was more strongly correlated with abstraction effect at Q75 than any other predictors. This indicated that PCA axis 2 described the impact of abstraction on macroinvertebrates more clearly than PCA axis 1 and was the most suitable index for further analysis.

Figure 4 shows the correlation of PCA axis 2 and abstraction effect at Q75 (P ≤ 0.001; R2 = 30.0%). Examination of the species scores along PCA axis 2 indicated which macroinvertebrate taxa were driving the observed changes in ecological indices along the gradient of abstraction effect and which individual taxa were associated with sites that were significantly impacted by abstraction and those which were not (Table 3). PCA axis 2 was also strongly correlated to LIFE O/E score (P ≤ 0.001; R2 = 62.2%) (Fig. 5). These analyses together provided validation that LIFE O/E described real biological responses to the effects of water abstraction.

figure

Figure 4. Correlation between principal components analysis (PCA) axis 2 scores and the effect of abstraction on river flows at Q75. Closed diamonds = abstraction effect > 80%; open triangles = abstraction effect 60–80%; closed circles = abstraction effect < 60%.

Download figure to PowerPoint

figure

Figure 5. Correlation between family Lotic invertebrate Index for Flow Evaluation (LIFE) O/E score and principal components analysis (PCA) axis 2 scores. Horizontal line indicates LIFE O/E 0.94 below which low flows are a possible pressure acting on the ecological community. Closed diamonds = abstraction effect > 80%; open triangles = abstraction effect 60–80%; closed circles = abstraction effect < 60%.

Download figure to PowerPoint

Table 3. The bottom 20 macroinvertebrate taxa and associated species Lotic invertebrate Index for Flow Evaluation (LIFE) flow groups (Extence et al. 1999) on principal components analysis (PCA) axis 2 that were associated with sites not significantly affected by abstraction (left side) and the top 20 macroinvertebrate taxa that were associated with sites that were affected by abstraction (right side)
TaxaPCA axis 2Flow groupTaxaPCA axis 2Flow group
Elmis aenea−1.96612Radix balthica1.40814
Rhyacophila dorsalis−1.76021Physidae1.2253 
Gammarus pulex−1.71012Asellus aquaticus1.16474
Silo pallipes−1.67371Crangonyx pseudogracilis1.05954
Ancylus fluviatilis−1.56752Haliplus sp.1.0586 
Serratella ignita−1.50962Stagnicola palustris1.03656
Rhyacophila sp.−1.4947 Erpobdellidae1.0022 
Silo sp.−1.4329 Micropterna sequax0.95422
Baetis rhodani−1.32492Dugesia lugubris/polychroa0.9403 
Limnius volckmari−1.31722Tipula (Yamatotipula) sp.0.9384 
Simuliidae−1.1573 Physella acuta group0.8804 
Agapetus sp.−1.1341 Micropterna sp.0.8539 
Hydropsyche sp.−1.1204 Sphaeriidae0.8133 
Sericostoma personatum−1.11042Agabus sp.0.8097 
Hydropsyche pellucidula−1.09442Trocheta subviridis0.79134
Dicranota sp.−1.0712 Oligochaeta0.7898 
Ecdyonurus sp.−1.0361 Hydropsyche angustipennis0.77242
Rhithrogena semicolorata−1.03021Dugesia sp.0.7503 
Alainites muticus−0.972Radix sp.0.7317 
Rhithrogena sp.−0.9637 Haliplus lineatocollis0.71293
Nematoda−0.9351 Haliplus ruficollis group0.6859 
Austropotamobius pallipes−0.92362Lymnaeidae0.6558 
Heptageniidae−0.8686 Coenagrionidae0.6546 
Eloeophila sp.−0.8638 Dugesia tigrina0.65263
Hydropsyche instabilis−0.85642Physella sp.0.642 
Ephemera danica−0.8512Molophilus sp.0.6418 
Wiedemannia sp.−0.8485 Dugesiidae0.6209 
Hydropsyche siltalai−0.82462Gerridae0.6209 
Isoperla grammatica−0.79181Ilybius/Agabus sp.0.6209 
Orectochilus villosus−0.77212Anisus (Disculifer) vortex0.60024

Consistent with the results described earlier, RDA suggested that abstraction effect at Q75 was a strong influence on macroinvertebrates across the study sites (Fig. 6). Conductivity was the most significant parameter that explained macroinvertebrate composition across the investigation sites described along RDA axis 1 (horizontal axis). The effects of abstraction on macroinvertebrates was explained along RDA axis 2 (vertical axis) of the ordination plot that also described a gradient of decreasing wetted width and increasing proportion of sand/silt (substratum PC1), with increasing abstraction impact. The order of individual macroinvertebrate taxa along RDA axis 2 was similar to that described earlier for PCA axis 2, and RDA axis 2 scores were significantly correlated to the species LIFE flow groups of these taxa (P ≤ 0.001, R2 = 23.5%).

figure

Figure 6. Redundancy analysis biplot showing correlations between macroinvertebrate community composition and measured environmental variables, including abstraction effect at different flows. Colours indicate the effect of abstraction on medium-low flows (Q75) at each sampling location (red circles = abstraction effect > 80%; yellow circles = abstraction effect 60–80%; green circles = abstraction effect < 60%).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

In December 2011, the UK Government set out its plans to reform the current system for managing water abstraction by facing up to the challenges of protecting aquatic ecosystems while ensuring the supply of affordable water for all (HM Government 2011). The challenge is not easy given that the first classification of surface water bodies in the UK indicated that about 14% of river water bodies in England and Wales are potentially at risk of failing WFD standards because of water abstraction or flow regulation [http://www.environment-agency.gov.uk/research/planning/33306.aspx (accessed 26/04/2012)]. The burden to be placed on water suppliers to mitigate this scale of environmental impact is potentially high and unjustified unless it is based on good-quality data and sound science (Friberg et al. 2011). The current study is unique in that it demonstrates for the first time significant relationships between standard biotic indices and empirical measures of abstraction effect on river flows.

The results presented in this current study suggested that few ecological impacts were identified at any of the study sites when the abstraction effect at Q75 from abstraction was less than 60%. The analysis of variance showed that significant impacts of abstraction on macroinvertebrates occurred only at locations where the abstraction effect at Q75 exceeded 80%. These locations were characterised by macroinvertebrates that are adapted to still/slow velocity and low dissolved oxygen concentrations [freshwater hoglouse (Asellus spp.), snails (Gastropoda), water beetles (Coleoptera) and leeches (Oligochaeta)] rather than macroinvertebrates that require faster flows and high dissolved oxygen concentrations [mayflies (Ephemeroptera), caddisflies (Trichoptera) and freshwater shrimps (Gammarus spp.)] that characterised the unimpacted reference conditions.

These results are broadly consistent with previous studies that have documented the impacts of abstraction on riverine macroinvertebrates. Studies undertaken over the same period in the chalk stream headwaters of the River Avon in Wiltshire indicated that impacts on macroinvertebrates were only apparent at similar magnitudes of flow impact from groundwater abstraction (APEM 2009). Studies of the impact of abstraction on riverine macroinvertebrates about 20 years ago indicated that impacts of abstraction were most obvious in headwater sites that had substantial dewatering because of groundwater abstraction (Armitage & Petts 1992; Bickerton et al. 1993) or upland sites where flow was drastically reduced by abstractions for power generation (Castella et al. 1995). The results of these studies suggest that current environmental standards for hydrology, which are considered by expert opinion to support good ecological status (Acreman et al. 2008), might be conservative for macroinvertebrates in rivers fed by Permo-Triassic sandstone aquifers.

The LIFE approach to assessing ecological responses to river flow variability has proved useful in assessing the sensitivity of river reaches to abstraction pressure (Extence et al. 1999; Dunbar et al. 2010a, b) and for setting location-specific minimum flow targets (Exley 2006). The strength of the LIFE approach is that it incorporates multi-taxa preferences to current velocity (as opposed to one or just a few species, as for fish) and integrates the influence of antecedent flow conditions (Extence et al. 1999). It is therefore matched to the temporal scales of abstraction effects rather than reflecting just short-term flow variations. We have demonstrated in this study how the response of LIFE to abstraction effect across different sites was mirrored by a response of the individual taxa and their relative abundances to abstraction effect, providing biological validation for the proposed hydroecology model.

At the time of undertaking this study, it was only possible to use RIVPACS at the family level for generating abundance-weighted reference LIFE scores with which to compare with observed LIFE scores. However, the results of the PCA and RDA analysis that examined the individual taxa responses to the same environmental predictors provided additional validation that family LIFE O/E encapsulated real biological response to abstraction impacts and was not confounded by other environmental factors. It also provided useful validation of the accuracy of RIVPACS for predicting site reference values in this study despite ongoing developments of the RIVPACS model to improve its accuracy for assessing hydromorphological impacts at individual sites (Clarke et al. 2003; SNIFFER 2010, 2011).

While the hydroecological relationships observed earlier are strong, the catchments tested are all fairly similar in size and nature. Caution should be used in applying these results in other catchments until similar relationships have been observed more widely. The temporal scale of the ecological data used in this analysis encompasses 2 years that were characterised by above average stream flows (2008–2009), preceded by 1 year of above average flows (2007) and 3 years of lower than average flows (2004–2006). As discussed earlier, hydrological effects of abstraction were normalised for long-term average conditions. Previous studies in lowland Chalk rivers showed that macroinvertebrates recover from multiseason drought within 2 years (Wood & Armitage 2004). In this study, the timing of sampling relative to antecedent low flows suggests that on the Permo-Triassic sandstone, either the chronic effects of groundwater abstraction have impacted the resilience of macroinvertebrate communities to low flow stress or the aquifer systems respond more slowly than the chalk. In either case, these sites should be prioritised for environmental improvement schemes. The relationships established in this study now need to be tested over longer temporal scales incorporating wet and dry years.

The hydroecology model is being developed further as a regional model to incorporate more sites, longer time series, and new developments of groundwater models and RIVPACS. The model is also being used to inform the effectiveness of different water resource options to meet WFD objectives. Finally, while this study was regional in its focus, the results presented here are unique in showing statistical relationships between river ecology and the impacts of groundwater abstraction on surface flows. By using the reference condition approach, this information provides evidence to support the development of a standard classification and assessment method for hydromorphological pressure consistent with the approach adopted by the WFD.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References
  • (1)
    Abstraction effect as a proportion of ‘natural’ Q75 was the most significant predictor of macroinvertebrate LIFE scores across the 17 sampling locations at six investigation sites.
  • (2)
    Few ecological impacts of low flows were identified at locations where the abstraction effect was less than 60% of Q75. Where abstraction effect exceeded 80% of Q75, ecological impacts of low flows were observed.
  • (3)
    The same relationships were shown among individual macroinvertebrate taxa, validating the biological responses.
  • (4)
    The hydroecological model has provided a scientific basis for making local decisions on sites under investigation and has helped to focus resources to outstanding areas of uncertainty in the current abstraction investigations.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References

This work was commissioned by Severn Trent Water Ltd, and the company's support and permission to publish is gratefully acknowledged. We acknowledge the field staff of APEM and Hydro-Logic Ltd for undertaking the ecological sampling and hydrometric gauging, respectively; also APEM's laboratory staff for undertaking the analysis of samples and the identification of macroinvertebrates. We are grateful to Mike Dunbar of the Environment Agency (UK) and the participants of an expert workshop for providing critical comment and guidance in the development of this work, particularly, Dr Chris Extence, Richard Chadd, Anne Taylor, Matilda Beatty, Chris Atkinson, Cecilia Young and Mark Warren. We thank three anonymous reviewers for their constructive comments.

Footnotes
  1. To submit a comment on this article please go to http://mc.manuscriptcentral.com/wej. For further information please see the Author Guidelines at wileyonlinelibrary.com

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Methods
  6. Results
  7. Discussion
  8. Conclusions
  9. Acknowledgements
  10. References
  • Acreman, M. and Dunbar, M.J. (2004) Defining Environmental River Flow Requirements – A Review. Hydrol. & Earth Syst. Sci., 8, 861876.
  • Acreman, M. and Ferguson, J.D. (2010) Environmental Flows and the European Water Framework Directive. Freshw. Biol., 55, 3248.
  • Acreman, M., Dunbar, M., Hannaford, J., Mountford, O., Wood, P., Holmes, N.T.H., Cowx, I.G., Noble, R., Extence, C.A., Aldrick, J., King, J., Black, A. and Crookall, D. (2008) Developing Environmental Standards for Abstractions from UK Rivers to Implement the EU Water Framework Directive. Hydrol. Sci., 53, 11051120.
  • APEM. (2009) River avon SAC low flows ecological investigations. Report to Wessex Water Services Ltd.
  • Armitage, P.D. and Petts, G.E. (1992) Biotic Score and Prediction to Assess the Effects of Water Abstractions on River Macroinvertebrates for Conservation Purposes. Aquat. Conserv.: Marine & Freshw. Ecosyst., 2, 117.
  • Arthington, A.H., Bunn, S.E., Poff, N.L. and Naiman, R.J. (2006) The Challenge of Providing Environmental Flow Rule to Sustain River Ecosystems. Ecol. Appl., 16, 13111318.
  • Bickerton, M., Armitage, P.D., Castella, E. and Petts, G.E. (1993) Assessing the Ecological Effects of Groundwater Abstraction on Chalk Streams: Three Examples from Eastern England. Regul. Rivers Res. & Manag., 8, 121134.
  • Castella, E., Bickerton, M., Armitage, P.D. and Petts, G.E. (1995) The Effects of Water Abstractions on Invertebrate Communities in UK Streams. Hydrobiologia, 308, 167182.
  • Clarke, R.T., Armitage, P.D., Hornby, D., Scarlett, P. and Davy-Bowker, J. (2003) Investigation of the relationships between the LIFE index and RIVPACS. Putting LIFE into RIVPACS. Environment Agency R&D Technical Report W6-044/TR1.
  • Dunbar, M.J. and Acreman, M.C. (2001) Applied Hydroecological Science for the 21st Century. In Acreman, M.C. (ed). Hydroecology: Linking Hydrology and Aquatic Ecology, pp. 117. IAHS Pub. No. 266 118. Centre of Ecology and Hydrology, Wallingford, UK.
  • Dunbar, M.J., Pederson, M.L., Cadman, D., Extence, C.A., Waddingham, J., Chadd, R.P. and Larsen, S.E. (2010a) River Discharge and Local-Scale Physical Habitat Influence Macroinvertebrate LIFE Scores. Freshw. Biol., 55, 226242.
  • Dunbar, M.J., Warren, M., Extence, C.A., Baker, L., Cadman, D., Mould, D.J., Hall, J. and Chadd, R.P. (2010b) Interaction between Macroinvertebrates, Discharge and Physical Habitat in Upland Rivers. Aquat. Conserv.: Marine & Freshw. Ecosyst., 20, 3144.
  • Exley, K.J. (2006) River Itchen Macro-invertebrate Community Relationship to River Flow Changes. Environment Agency Report.
  • Extence, C.A., Balbi, D.M. and Chadd, R.P. (1999) River Flow Indexing Using British Benthic Macroinvertebrates: A Framework for Setting Hydroecological Objectives. Regul. Rivers Res. & Manag., 15, 543574.
  • Friberg, N., Bonada, N., Bradley, D.C., Dunbar, M.J., Edwards, F.K., Grey, J., Hayes, R.B., Hildrew, A.G., Lamouroux, N., Trimmer, M. and Woodward, G. (2011) Biomonitoring of Human Impacts in Natural Systems: The Good, the Bad, and the Ugly. Adv. Ecol. Res., 44, 168.
  • HM Government (2011) Water for Life. Presented to Parliament by the Secretary of State for Environment, Food and Rural Affairs by Command of Her Majesty.
  • Leps, J. and Smilauer, P. (2003) Multivariate Analysis of Ecological Data Using CANOCO. Cambridge University Press, Cambridge, UK.
  • Monk, W.A., Wood, P.J., Hannah, D.M., Wilson, D.A., Extence, C.A. and Chadd, R. (2006) Flow Variability and Macroinvertebrate Community Response within Riverine Systems in England and Wales. Rivers Res. & Appl., 19, 595615.
  • Monk, W.A., Wood, P.J., Hannah, D.M. and Wilson, D.A. (2008) Macroinvertebrate Community Response to Inter-Annual and Regional River Flow Regime Dynamics. River Res. & Appl., 24, 9881001.
  • Shepley, M.G. and Streetly, M. (2007) The Estimation of ‘Natural’ Summer Outflows from the Permo-Triassic Sandstone Aquifer, UK. QJEGH, 40, 213227.
  • SNIFFER. (2010) Enhancements to the river invertebrate classification tool. SNIFFER Research Report WFD 100.
  • SNIFFER. (2011) Further developments of river invertebrate classification tool. SNIFFER Research Report WFD 119.
  • SNIFFER. (2012) Ecological indicators of the severe effects of abstraction and flow regulation; and optimising flows from impoundments. SNIFFER Research Report WFD 21D.
  • Vaughan, I.P., Diamond, M., Gurnell, A.M., Hall, K.A., Jenkins, A., Milner, N.J., Naylor, L.A., Sear, D.A., Woodward, G. and Ormerod, S.J. (2009) Integrating Ecology with Hydromorphology: A Priority for River Science and Management. Aquat. Conserv.: Marine & Freshw. Ecosyst., 19, 113125.
  • Wood, P.J. and Armitage, P.D. (2004) The Response of the Macroinvertebrate Community to Low-Flow Variability and Supra-Seasonal Drought within a Groundwater Dominated Stream. Arch. Hydrobiol., 161, 120.
  • Wright, J.F., Sutcliffe, D.W. and Furse, M.T. (2000) Assessing the Biological Quality of Freshwaters: RIVPACS and Other Techniques. Freshwater Biological Association, Ambleside.