https://orcid.org/0009-0006-0676-2307
Figures
Abstract
To mitigate the morphological and ecological impacts of direct (e.g. dredging) and indirect (e.g. damaged river function) sediment loss, gravel augmentation is commonly practiced in river systems globally. Despite this, the effectiveness of this practice remains poorly understood, especially in less often considered systems such as chalk streams which present uncommon conditions (e.g. low stream power, stable flow) and may respond to interventions in ways that differ from systems more commonly studied. This study quantified immediate (0–1 years) and short-term (1–2 years) physical and ecological responses to gravel augmentation at two English chalk stream restoration sites: Home Stream (HS; River Test) and East Lodge (EL; River Itchen). We quantified habitat (depth, velocity, substrate composition), cover of different macrophytes, and macroinvertebrate (before-after-control-impact) abundance and community structure. Restoration reduced depth and increased gravel cover in both sites and decreased the cover of filamentous green algae in HS. Macroinvertebrate communities became more dominated by silt-intolerant taxa, while abundance [HS only] and taxon richness increased 1–2 years post-restoration. Whilst the responses found were generally positive in light of the restoration goals, the effects varied across sites, post-restoration time periods and ecological groups, emphasising the need for the more holistic monitoring of restoration projects considering community-level responses at different sites and systems over ecologically relevant timescales. This will help inform on the generality and longevity of responses and provide the evidence needed to develop sound restoration practice.
Citation: Dolman LA, Vowles AS, Kemp PS (2024) Chalk stream restoration: Physical and ecological responses to gravel augmentation. PLoS ONE 19(11): e0313876. https://doi.org/10.1371/journal.pone.0313876
Editor: Miguel Cañedo-Argüelles Iglesias, Universitat de Barcelona, SPAIN
Received: November 15, 2023; Accepted: November 2, 2024; Published: November 20, 2024
Copyright: © 2024 Dolman et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data files are available from the University of Southampton repository at https://eprints.soton.ac.uk/491661/
Funding: This project was funded by the Engineering and Physical Sciences Research Council and Environment Agency as a part of the Centre for Doctoral Training in Sustainable Infrastructure Systems (EP/L01582X/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
For centuries, humans have intensively modified rivers for agriculture, navigation, flood defence, industrial and domestic water supply and energy generation [1, 2]. As a consequence, rivers have been dammed, channelised, dredged, and cut-off from their flood plains, disrupting natural hydrogeomorphological regimes [3, 4], reducing fluvial connectivity [5] and homogenising habitat [6]. Extensive river modification has reduced biodiversity [7], modified communities of fauna and flora [8] and degraded ecological quality [9], contributing to global threats to fresh waters and associated potential for humanitarian crisis [10].
Localised depletion of river sediments through direct (e.g. dredging and mining [11, 12]) and indirect (e.g. interruption of sediment regime by dams [13]) mechanisms threatens morphological and ecological fluvial integrity [14, 15]. Morphologically, sediment depletion can promote channel incision, bank instability and armouring of the river bed [14, 15], reduce habitat heterogeneity [16] and inhibit channel dynamics [17, 18]. Ecologically, it can negatively impact biodiversity, communities and food web dynamics [11, 19], habitat suitability for substrate-spawning taxa [20, 21], and promote the establishment of non-native species [19]. Globally, efforts have increased to mitigate the negative impacts associated with sediment depletion through rehabilitation measures, including gravel augmentation, one of the more common substrate restoration techniques [22–24].
The artificial addition of gravels to degraded rivers has been widely employed in several regions, with examples from North America (United States [25], Canada [26]), Europe (United Kingdom [27], Germany [28], Norway [29]) and Asia (Japan [30]). This practice can help facilitate the regeneration of natural processes and be used to restore riverbed structure (e.g. by redistributing silt) and features over a range of scales [24]. Those operating at the reach-scale tend to focus on re-naturalising riverbed structure and features, e.g. to create riffles [31] and gravel bars [32], and improve ecological utility, e.g. salmonid spawning habitat quality [33]. At larger scales, efforts are directed at re-establishing dynamic processes, e.g. the formation and evolution of geomorphic features [34]. However, logistical constraints (e.g. land ownership) and the costs of implementation and monitoring can result in bias (e.g. in publication [29, 35]) towards the more common reach-scale gravel augmentation projects [36].
Despite the widespread use of gravel augmentation to help restore rivers [23], the ecological responses observed are often variable and remain poorly understood [35]. On one hand, the addition of gravel can enhance fish habitat, particularly for gravel spawning taxa [29, 33, 35, 37], macroinvertebrate diversity [38] and density [32, 35], and macrophyte species richness [39]. On the other, several studies have indicated limited (e.g. fish [31], macroinvertebrates [40, 41]), temporary (e.g. fish [28], vegetation [42]), mixed (e.g. fish [43]) and negative effects (e.g. reduced macroinvertebrate abundance and biomass [44], reduced macrophyte diversity, biomass and cover [35]). A lack of understanding of the mechanisms that underpin the ecological responses observed may be attributed to multiple factors, including limited monitoring [45], insufficient time-scales of appraisal [46], and the potential for catchment scale processes to overshadow reach scale outcomes limiting or resulting in short-term responses [31, 47]. Moreover, the general focus on single ecological indicators, especially salmonids [29], and the bias to certain regions and river types [23, 32, 44, 48] may constrain understanding of the value of gravel augmentation as an effective restoration strategy.
The majority of groundwater fed chalk streams are found in southern and eastern England, and typically exhibit low power and relatively stable flow and temperature regimes [49]. Chalk streams have been highly modified over many centuries (e.g. for agriculture, mechanical power and flood water conveyance [50, 51]), including dredging and impoundment, which has contributed to their widespread degradation [52]. Indeed, 85% of English chalk streams failed to achieve good ecological status under the European Union Water Framework Directive (2000/60/EC) in 2019, with physical modification contributing to 35% of these failures [53, 54]. Physical habitat degradation, such as that due to infrastructure, alongside intensive water abstraction and land use practices, have widely disrupted natural sediment regimes of chalk streams [51, 55]. Due to their stable hydrogeomorphology and naturally low levels of sediment recruitment and bedload mobilisation [56, 57], alongside impoverished base flows and highly modified channels, the natural regeneration of sediment is difficult without appropriate intervention [52]. Consequently, the replenishment of gravels has become an important management strategy in the restoration of chalk streams [58]. However, few studies have evaluated the ecological consequences of gravel augmentation in these systems, and those that have tend to focus on single ecological indicators (e.g. macroinvertebrates [27]).
This study quantified physical and ecological responses to gravel augmentation at two case study chalk stream restoration sites in southern England. Restoration was predicted to have: (1) an immediate effect on physical habitat and that this would be maintained throughout the duration of the study period (two years post restoration). Specifically, when compared with the conditions prior to restoration and at a control site, the restored sections were expected to be characterised by greater velocity and depth variability, higher maximum velocities and shallower water depth, and higher coarse-substrate cover. These physical responses were predicted to result in: (2) changes to the stream ecology, considered in terms of enhanced richness, abundance and/ or cover of aquatic fauna and flora, and shifts in community composition towards those considered characteristic of chalk stream environments (e.g. greater presence of Ephemeroptera, Plecoptera and Trichoptera [EPT] and rheophilic and silt-intolerant taxa such as Ranunculus spp. [50, 52]). Finally, it was expected that: (3) time would play an influential role in the ecological response, with limited or potentially negative changes associated with the immediate disturbance of habitat in the year following restoration (especially for benthic groups), followed by a shift to the aforementioned outcomes thereafter. To test these predictions, this study focused on two ecological indicators: (a) macrophytes (total macrophyte cover and cover of different macrophyte types); and (b) macroinvertebrates (abundance, taxon richness, percentage of abundance [EPTA] and taxon richness [EPTN] of the community comprised of EPT, Proportion of Sediment-sensitive Invertebrates [PSI] [59] and Lotic Index for Flow Evaluation [LIFE] [60]). To investigate the influence of time, physical and ecological responses were monitored throughout the year immediately after (immediate response) and between one and two years (short-term) post restoration.
2. Materials and methods
2.1 Study sites
This study monitored the physical and ecological response of two gravel augmentation restoration projects on the Rivers Test (Home Stream [HS]) and Itchen (East Lodge [EL]; Fig 1), Hampshire, United Kingdom. Both rivers (catchment area: 1,260 km2 and 470 km2, respectively) rise in the north of the county and flow southwards to join Southampton Water. In the upper catchment the geology is dominated by chalk, transitioning to sands and clays in the lower reaches [61]. Land use is predominantly arable and pasture, with dispersed settlements in the upper catchment and becoming increasingly urbanised as the rivers flow downstream [61]. Both rivers contribute to the local economy, including through recreational fly fishing and aquaculture [62]. They have also been designated as Sites of Special Scientific Interest, a conservation designation under the United Kingdom’s Wildlife and Countryside Act (1981) which protects the best wildlife and geological sites [63], for distinct habitats (e.g. fen meadows) and species (e.g. Southern damselfly [Coenagrion mercuriale]; [64, 65]). The River Itchen receives additional protection as a Special Area of Conservation under Directive 92/43/EEC on the Conservation of Natural Habitats and of Wild Fauna and Flora (i.e. Habitats Directive) [66], due to its faunal (e.g. European bullhead [Cottus gobio]) and floral (e.g. Ranunculus fluitantis) communities [67].
(A) Location of study sites on the Rivers Test and Itchen in Hampshire, southern England, United Kingdom, with greater detail provided for (B) Home Stream (HS) 51.07296, -1.51685, and (C) East Lodge (EL) 51.00047, -1.32551. The arrows, black lines and crosses respectively indicate flow direction, extent of control and restored reaches and positions where felled trees were secured. Country and county shapefiles and river network were supplied from [68, 69], respectively. Detailed maps (B and C) were supplied from [70].
Both the River Test and Itchen have been heavily modified over many centuries. Reasons for this include the generation of mechanical power, navigation, and agriculture [62]. As a result the rivers host a high density of low-head weirs and leats and many reaches are heavily channelised, unnaturally incised and widened, and characterised by low uniform velocity regimes and fine sediment deposition [62, 71]. Additionally, catchment scale land-use has resulted in high diffuse pollutant input and associated nutrient enrichment [72], fine sediment run-off [56, 73] and over abstraction [74]. Ecologically, both rivers are considered highly degraded and suffering from “Chalk Stream Malaise”, a deterioration from the classic chalk stream habitat [75]. Indeed, in 2019, 35% of the River Test and 14% of the River Itchen Water Framework Directive monitoring sites were considered heavily modified (physically altered by human activity and substantially changed in character [76]), and 52% and 50% of sites failed to achieve ‘good ecological status’, respectively [54]. Physical modification of the channel was the primary driver of these failures (Test = 75%, Itchen = 42% [53]).
At each study site, a reach of river approximately 200 m in length was restored in October 2019 with the goal of increasing in-river habitat heterogeneity and enhancing biodiversity and conditions for Ranunculus spp. and gravel spawning brown trout and Atlantic salmon (Salmo salar). Approximately 1500 and 3000 tonnes of washed gravel substrates were deposited along the HS (natural gravel excavated from a “borrow pit” on site > 10 m away from the river to reduce costs and ensure suitability of substrate; predominantly 2–64 mm but with some finer and coarser grain sizes) and EL (predominantly 16–30 mm imported washed river gravel) reaches, respectively. This was intended to reduce depth, remobilise silt, reprofile the planform, recover a more naturalised riverbed and create geomorphic features. Localised cobble sized substrate was also placed at EL. Additionally, at each site, a limited number of felled trees (approximately 2) were secured at the riverbank to form low velocity, sheltered habitats and Ranunculus spp. were sparsely ‘seeded’ (approximately 1 plant per 5 m2) throughout the reaches via translocation from unrestored areas nearby.
A single restored reach paired with a control (both 150 m in length) were monitored at both restoration sites (Fig 1). Control reaches were selected based on their proximity and similarity to the restored sites prior to the rehabilitation work. As such, we expected restored and control reaches to be similar prior to restoration, but that the restored reach would change according to the aforementioned predictions following restoration. To prevent downstream effects of the interventions the control sites were located 200 m and 150 m upstream of the restored sections at HS and EL, respectively. At HS, the only suitable control was located upstream of a 1.1 m high weir. A before-after-control-impact (BACI) study design was used to quantify changes in physical habitat, macrophytes and macroinvertebrates. Access permission was granted by the relevant property manager prior to each data collection period.
2.2 Physical habitat
Physical habitat was measured at five equidistant points across 16 transects located at 10 m intervals during five surveys per restoration project (see Fig 2 for a summary of data collection periods and the flow conditions under which surveys were conducted). For EL, physical data were collected during July 2019 [pre-restoration], December 2019 and July 2020 [immediate response] and December 2020 and July 2021 [short-term response]. For HS, data were collected during September 2019 [pre-restoration], January and September 2020 [immediate response] and January and September 2021 [short-term response]. At each point, depth (cm), velocity (m s-1; average taken over 10 secs at 1 Hz and 60% depth using a Valeport Model 801 flow meter; it was not possible to measure velocity at one point in the EL restored reach in December 2019) and the dominant surface substrate (silt [0.0039–0.123 mm], sand [0.125–2 mm], gravel [2–64 mm], cobble [> 64 mm]) within a 0.5 m2 quadrat was recorded. Wetted widths (m) were measured at each transect, although for some sampling periods and transects this was not possible due to excessive riparian growth and safety concerns. Depth (DCSV) and velocity (VCSV) cross sectional variability, a measure of habitat heterogeneity, was calculated by taking the standard deviation of the depth and velocity measurements taken across each transect.
Symbols in grey shaded area show the months during which data was collected at HS (River Test, grey symbols) and EL (River Itchen, black symbols). Flow data was obtained from Chilbolton (River Test) and Highbridge (River Itchen) gauging stations, approximately 10 km upstream and 2 km downstream of the restored sites, respectively [77]. Contains public sector information licensed under the Open Government Licence v3.0.
2.3 Macrophytes
Total macrophyte cover and submerged macrophyte group cover were quantified on five and seven occasions at HS and EL, respectively. Surveys took place at the same time as the physical data collection, with the exception that September 2020 [immediate response] and 2021 [short-term response] were opportunistically also surveyed in EL. Using the same transects as for the physical habitat metrics, total macrophyte cover was assessed by estimating the percentage cover (including emergent and submerged plants) within a 0.5 m2 quadrat at each point (five per transect). Additionally, at the same transects, the cover of five submerged macrophyte groups (Table 1) was calculated (eight transects were conducted at 20 m intervals in the EL restored reach due to time constraints during the first data collection period). The overall macrophyte cover within a prescribed area (1 m length spanning the width of the river) was estimated using the mean of the five total macrophyte cover and wetted width measurements at each transect. The amount of overall cover comprising each macrophyte group was estimated for each transect (S1 Appendix for example of calculation). To control for limited accuracy of estimates of group cover in HS when turbidity was high, further analysis was restricted to data obtained during September when turbidity was low. Where wetted widths were not recorded (EL September 2020 and 2021, HS control September 2019), widths measured during the other sampling periods under similar flow conditions were used.
2.4 Macroinvertebrates
Macroinvertebrate samples were collected at the same times as for physical habitat data at eight transects in both restored and control sites per sampling date (i.e. at every second transect, at 20 m intervals) using a standard 3-minute kick sample (250 mm2 net, 1 mm mesh), or sweep sample from a boat if depth restricted wading [79]. This was followed by a 1-minute hand search for macroinvertebrates not likely to be found in the kick sample, such as those residing under larger substrate or woody material [79]. To ensure standardised sampling, the technique (i.e. kick or sweep sample) used during the first sampling period was maintained throughout the study. Over the 3-minute kick sampling period, substrates and all habitat types (e.g. macrophytes) were sampled in estimated proportion to their occurrence in the respective transect to provide appropriate representation of invertebrates from each microhabitat. Following collection, samples were placed in 1.2 L sampling pots and fixed with 70% methylated spirit.
Samples were sorted and macroinvertebrates identified within one month of collection. Each sample was poured through a 30 μm sieve and washed lightly with tap water to remove methylated spirit, before being placed into a tray with water and evenly distributed throughout. Samples were sub-sampled by dividing the area of the tray in half and sorting a random half by hand, collecting any macroinvertebrates present. These were identified by a single Freshwater Biological Association trained practitioner to the family level (Oligochaeta was classified as such) using a compound microscope (Motic ST-30) and identification guide [80]. Following identification, the abundance, taxon richness, EPTA, EPTN, LIFE [60] and PSI [59] was calculated (see S2 Appendix for description and calculation). These metrics were specifically selected as they provide information on community responses to changes in physical habitat (e.g. LIFE and PSI provide information on flow and silt pressures, with lower values suggesting a more lentic, silt-tolerant community, respectively [59, 60]) or encompass an important component of chalk stream ecology and food chains (e.g. EPT taxa [52, 81]). No permits were required to sample macroinvertebrates.
2.5 Statistical analysis
To evaluate the influence of time since restoration on physical habitat and ecological responses, the analysis was divided into two periods: immediate (0–1 years) and short-term (1–2 years) post restoration. Separate statistical models were created for EL and HS as a direct comparison between the two sites was not the focus of this study. Each dominant substrate type was summed across each transect and VCSV was square-root transformed. Analyses were carried out in R Studio [82] and the packages Lme4 [83], LSmeans [84], performance [85], ggplot2 [86], patchwork [87], nparLD [88], nparcomp [89] and MuMIn [90].
Physical habitat (excluding substrate cover) and macroinvertebrate metrics were analysed using generalised linear mixed models (GLMM; macroinvertebrate abundance [HS] and taxon richness [HS and EL]) with a Poisson error distribution and log-link function and linear mixed models (LMMs; all other physical and macroinvertebrate metrics). Models encompassed the terms ‘sampling period’ (pre-restoration, 0–1, 1–2 years post-restoration), ‘location’ (restored or control site) and the ‘interaction between sampling period x location’. ‘Sampling date’ and ‘sampling point/transect’ were considered random effects to account for pseudoreplication. Random effect structures were automatically assigned in Lme4 and were nested in sampling period and site, respectively. All metrics were categorical. In these models, significant sampling period x location interactions suggest an effect of the restoration. The significance of each variable was assessed using likelihood ratio tests by comparing full models with one with the focal terms removed. For all LMMs and GLMMs, multicollinearity and model assumptions were checked using variance inflation factor and QQ and fitted vs residual plots, respectively. Random effect variance and R² values were recorded and are available in S3 Appendix. TukeyHSD pairwise comparisons were performed post-hoc when sampling period or sampling period x location interactions were significant.
For macrophyte and substrate metrics, diagnostic plots (e.g. QQ and fitted vs residual plots) suggested model assumptions (e.g. normality of residuals) were not met by mixed models. Therefore, a non-parametric rank-based repeated measures approach was conducted using ‘nparLD’ [88]. This approach does not require any distribution assumptions, is considered robust to small and variable samples sizes and outliers, and provides an ‘ANOVA-type’ statistic (ATS [88]). Where multiple samples were taken within a sampling period, a single value for each sampling point/transect was calculated by taking the mean across each period, a necessity of the statistical test. Models were created with the same terms as for the parametric analysis, although ‘sampling date’ was not included due to averaging across each sampling period. Where sampling period or sampling period x location interactions were significant for macrophyte and substrate metrics, post-hoc pairwise comparisons were conducted in ‘nparcomp’ and using the function ‘mctp.rm’ with TukeyHSD corrections [89]. For significant interactions, pairwise comparisons were used to test differences between sampling periods on separate models for restored and control reaches.
3. Results
3.1. Physical habitat
Interactions between sampling period and location were observed for depth, VCSV and the cover of sand and gravel substrates in HS (Figs 3 and 4 and Table 2). Depth and sand cover was lower, and gravel cover higher during both post-restoration periods in the restored reach compared to pre-restoration (see S4 Appendix for post-hoc results). Compared to pre-restoration, VCSV was higher 1–2 years post-restoration in the restored reach. No difference in metrics between periods was observed for the control site.
Black bars and boxes indicate median and 25th and 75th percentiles, respectively. Whiskers represent minimum and maximum values excluding outliers. Dots show outliers (values > 1.5 x the interquartile range).
In EL, interactions between sampling period and location were observed for depth, velocity, and cover of silt, sand, gravel and cobble. In the restored reach, silt cover was lower, and gravel and cobble cover higher, during both post-restoration periods compared to pre-restoration. Gravel cover was higher 0–1 years compared to 1–2 years post-restoration. Compared to pre-restoration, sand cover and depth was lower 0–1 years but not 1–2 years post-restoration in the restored reach. Velocity in the restored reach was lower than the control 0–1 years post-restoration. Aside from higher gravel cover 1–2 years post-restoration than the pre-restoration period, there was no change in physical habitat metrics over time in the control site.
3.2. Macrophytes
Total macrophyte cover in HS and EL was influenced by sampling period, and in the case of HS also by location (Table 2 and Fig 5). There was no interaction between sampling period and location (Table 2). Total macrophyte cover was higher prior to restoration compared to 0–1 and 1–2 years post-restoration in HS, and higher 1–2 years post-restoration compared to pre- and 0–1 years post-restoration in EL (S4 Appendix). In HS, total macrophyte cover was higher overall in the restored compared to the control reach.
Black bars and boxes indicate median and 25th and 75th percentiles, respectively. Whiskers represent minimum and maximum values excluding outliers. Dots show outliers (values > 1.5 x the interquartile range).
In HS, an interaction between sampling period and location for filamentous green algae indicated lower cover in the restored reach during both post-restoration periods compared to pre-restoration, while there was no difference over time for the control reach. In EL, an interaction between sampling period and location was observed for broad leaved macrophytes and tape grass (Table 2). However, this was marginal and post-hoc comparisons indicated the overall responses of these groups did not differ between the restored and control reach.
3.3. Macroinvertebrates
In total, 11,963 and 18,835 individual macroinvertebrates representing 66 and 63 families were sampled in HS and EL, respectively. Prior to restoration, dominant families were Gammaridae (mean abundance per sample ± SD = 22.3 ± 18.4), Ephemeridae (8.6 ± 6.0) and Aphelocheiridae (7.6 ± 9.6) in HS, and Gammaridae (28.5 ± 31.5), Ephemerellidae (28.5 ± 18.3) and Bithyniidae (21.5 ± 16.5) in EL. One year after the initial survey in HS, Gammaridae (51.4 ± 30.1) and Ephemeridae (25.6 ± 15.0) remained the most abundant families, but Valvatidae (19.4 ± 17.3) exhibited increasing numbers. Likewise, in EL, Gammaridae (113.1 ± 47.8) and Ephemerellidae (90.3 ± 55.7) remained most abundant, but Baetidae (31.4 ± 31.4) became increasingly more frequent. The most abundant taxa found two years following the initial surveys remained similar at the end of the study period, although the values did vary. Indeed, Gammaridae (65.1 ± 22.9) and Valvatidae (17.8 ± 9.4) remained abundant, but Baetidae (38.9 ± 24.7) replaced Ephemeridae as the second most abundant taxa in HS. Gammaridae (40.5 ± 34.1), Ephemerellidae (36.0 ± 25.8) and Baetidae (38.9 ± 20.8) were most abundant in EL.
In HS and EL, interactions between sampling period and location were observed for taxon richness and PSI (Table 2 and Fig 6). Compared to pre-restoration values, the restored sites had a higher PSI in both post-restoration periods and a higher taxon richness 1–2 years post-restoration (S4 Appendix). Taxon richness was also lower for both sites 0–1 years compared to 1–2 years post-restoration. In HS, an interaction between sampling period and location was also observed for abundance, LIFE and EPTN. Abundance in the restored reach was higher 1–2 years post-restoration compared to pre-restoration. Post-hoc comparisons did not indicate a change in LIFE or EPTN across periods in the restored or control reach, although LIFE was lower in the restored compared to the control reach prior to restoration. In both HS and EL, there was no change in macroinvertebrate metrics between periods in the control reach.
Black bars and boxes indicate median and 25th and 75th percentiles, respectively. Whiskers represent minimum and maximum values excluding outliers. Dots show outliers (values > 1.5 x the interquartile range).
4. Discussion
Although gravel augmentation is commonly used to improve the physical and ecological condition of sediment depleted rivers [23], the impact on different ecological groups remains poorly understood. This study investigated the physical and ecological responses to gravel augmentation at two chalk stream restoration sites in southern England. Unsurprisingly, and in support of our first prediction, changes in physical habitat metrics were often immediate (e.g. increased gravel cover) and remained so for the duration of the study. However, other factors (e.g. velocity) showed little change. Likewise, and in support of our second prediction, the consequent ecological response could be considered mostly positive in light of typical restoration goals (e.g. to improve biodiversity), but also varied between sites and ecological groups. For example, while there was little change in macrophytes following restoration, clearer responses were observed for macroinvertebrates (e.g. enhanced dominance of silt-intolerant taxa). Furthermore, in support of our third prediction the ecological response was often influenced by time. For example, an increase in macroinvertebrate abundance (HS only) and taxon richness was only observed 1–2 years post-restoration.
When applied at the reach scale, as in this study, the addition of gravel to a degraded river represents a feature-based restoration approach intended to enhance localised physical habitat [24]. Such interventions are hoped to result in a desired outcome related to improvement in some measure of ecological condition (e.g. ecological status: Water Framework Directive [91]; biodiversity net gain: United Kingdom Environment Act [92]). In chalk streams, which tend to be highly engineered, managed and exhibit stable hydrogeomorphology, we predicted that changes to physical habitat would be rapid, and predominantly relate to a reduction in depth, and increased velocity, habitat heterogeneity and coarse substrate cover. This expectation was partially realised. At both sites, restored reaches became shallower (EL 0–1 years post-restoration only), more gravel-dominated, and velocity became more heterogeneous in HS (1–2 post-restoration only), possibly aided by a small number of felled trees. Conversely, mean velocity showed little change. Although unlikely, the lower velocity in the EL restored reach compared to the control 0–1 years post-restoration may have resulted from slight variation in flow between sampling restored and control reaches. Furthermore, although gravel cover increased following restoration, at EL cover was lower 1–2 compared to 0–1 years post-restoration, possibly because fine sediment input [62] confounded the influence of gravel addition the year following restoration. Continued fine sediment deposition might smother introduced coarse substrates and compromise the long-term success of such an approach [28, 93], illustrating the potential for the benefits of reach-scale interventions to be confounded by catchment-scale processes [94, 95]. This problem is especially relevant in chalk streams that typically exhibit stable flow regimes and limited stream power to redistribute fine sediments [56]. This emphasises the need to understand the underlying root causes of degradation and for restoration actions to take place over commensurate scales [94]. Unexpectedly, the proportion of gravel in the control site at EL increased over time and was higher 1–2 years post-restoration compared to pre-restoration, possibly a result of difference in flow across sampling periods. It would be of interest to also monitor natural changes in sediment grainsize and underlying causes of any fluctuations observed. Due to differences in the perception of river restoration of those involved in the United Kingdom context (e.g. key stakeholders can range from pro- to anti-restoration), and consequent willingness of riparian landowners to participate, holistic catchment-scale targets are more likely to be achieved if the planning process is based on pragmatic opportunism (undertaking restoration when possible, e.g. when stakeholder agreement and funding align) that embraces what may first appear to be a rather piecemeal approach. As such, reach-scale interventions that may represent feature based approaches when viewed in isolation, may also enable the reestablishment of more natural hydrogeomorphological processes over greater spatial and temporal scales [94], perhaps reversing processes of ecological degradation that occurred over centuries of river engineering.
Gravel augmentation was predicted to bring about desirable ecological change in line with the restoration and overarching catchment management aims (e.g. River Test and Itchen Restoration Strategy [62]). For the most part this prediction was realised, although responses varied considerably between ecological groups and sites. Indeed, macroinvertebrates responded strongly to restoration. At HS they became more abundant, and at both sites became more taxon rich and increasingly dominated by silt-intolerant taxa. This may have been due to an increase in suitable habitat within the interstitial spaces of the coarser substrate added [96, 97], and the redistribution of fine sediments. In contrast, the response of macrophytes to restoration was comparatively weak, despite translocation of Ranunculus spp. Only a reduction in the cover of filamentous green algae was observed at HS, potentially due to the redistribution of nutrients within fine sediment [98] or because Cladophora are mid to late successional species [99] and require a longer colonisation time than provided in this study. Responses to restoration can be complex and vary between systems and ecological groups due to a variety of interacting factors. For example, catchment size [100], watershed degradation [28], substrate mobility [44] and recolonisation potential (e.g. species pools [101]) can all influence recovery, whilst differences in organism group (e.g. benthic versus open water groups), species (e.g. salmonid versus lamprey substrate requirements [29, 102]) and life-stage [93] habitat preferences can further complicate the prediction of outcomes. Gaining a deeper comprehension of the factors that contribute to variation in response is important for reducing uncertainty and advancing restoration practices. To achieve this, there is a need to promote the wider adoption of monitoring and its timely dissemination as a key part of the restoration process [103]. This includes the reporting of results that contradict study predictions or projects aims, which can provide valuable insights when implementing new projects.
The manipulation of river habitat through restoration can lead to localised ecological disturbance that may delay the realisation of desirable outcomes as communities recover [32, 104, 105]. For this reason, many have argued that the time-scales over which monitoring should occur must be sufficient to capture the response investigated (e.g. [46]); in many studies this is not the case [106]. We predicted that the ecological response may be limited or even negative (e.g. decline in abundance and taxon richness) when measured within one-year of restoration and that desirable outcomes (e.g. increased abundance and taxon richness) may take longer to accrue, i.e. between one and two years post-restoration. In our study, negative impacts were limited, with most metrics showing either positive responses (e.g. a shift towards more silt-intolerant macroinvertebrate taxa) or no change. As predicted, positive responses in several metrics (e.g. macroinvertebrate abundance [HS only] and taxon richness) became apparent only between one and two years after restoration, reflecting the time required for colonisation of the restored sites [107].
Despite an increase in river restoration over recent decades, robust and rigorous monitoring of effectiveness against stated aims remains relatively rare [108]. The lack of sufficient monitoring, independent of time-scale, is perhaps unsurprising considering the costs of doing so relative to the overall budget available for individual projects, particularly for those that are relatively small [109]. One approach to resolve this challenge is to develop collaborative and co-ordinated river restoration networks to share resources and strategically plan programmes in which representative (flag-ship) projects are selected for robust monitoring over appropriate spatial and temporal scales [110]. Likewise, best practice guidance for monitoring should be updated considering the technological advance that has occurred over recent times [111], including in genetics, filming, telemetry, remote sensing, and artificial intelligence. Doing so will advance the collection of appropriate evidence on which future decision making can be based, allowing lessons to be learnt from both successes and failures.
This study indicates that gravel augmentation can positively benefit degraded chalk stream habitat and ecology, at least over limited spatial and temporal scales. However, responses to such interventions can be variable due to a multitude of factors, including physical and chemical attributes of the site, the ecological group considered, and the time since restoration. Whilst not assessed in this study, the responses to gravel augmentation will likely vary across river types, emphasising the need for robust monitoring to facilitate the development of system-specific best practice. For example, in chalk streams and other low power groundwater dominated systems, augmented gravels may be less prone to redistribution and more susceptible to sedimentation compared with more powerful surface-fed systems [28, 44, 93]. This will likely reduce the long-term benefits of such projects and emphasises the need for gravel augmentation to be carried out alongside a range of interventions to reinvigorate natural processes (e.g. reduction of catchment-scale sediment inputs, recolonisation of gravel mobilising ecosystem engineers, addition of woody materials [52]). In addition to ensuring strategically selected projects are rigorously monitored using appropriate scientific techniques and field study design over appropriate spatial and temporal scales, there is also a need to adopt more holistic approaches that consider community level response, moving away from a single target species bias that typically focuses on fish [23]. The use of feature-based approaches, not in isolation, but as part of a spatially and temporally broader and integrated strategy to reinstate process-based river re-naturalisation is likely to yield considerable benefits in the long-term. Just as rivers were degraded through engineering practices over many centuries, their regeneration will take time and likely depend on opportunistic reach-scale projects apparently conducted in a piecemeal fashion in a resource limited environment. However, the effectiveness of such may be enhanced if applied within a strategic catchment scale management framework that recognises the false dichotomy between the relative values of feature versus process-based restoration.
Supporting information
S1 Appendix. Schematic and example calculation of the cover of macrophyte groups across each transect.
Thick lines show river boundary. Hashed area shows the 1 m area across the transect which was assessed. Boxes show the five percent total macrophyte cover (TMC) estimates made across each transect. Letters and coloured areas show different macrophyte types.
https://doi.org/10.1371/journal.pone.0313876.s001
(TIF)
S2 Appendix. Details of the macroinvertebrate metrics used in a study to investigate ecological response to chalk stream gravel augmentation and how they were calculated.
https://doi.org/10.1371/journal.pone.0313876.s002
(DOCX)
S3 Appendix. Random effect variances, standard deviations (SD) and marginal and conditional R2 values for full linear and generalised linear mixed models.
https://doi.org/10.1371/journal.pone.0313876.s003
(DOCX)
S4 Appendix.
Results of Tukey pairwise comparison post-hoc tests following significant before-after x control-impact interactions (A and C) and before-after (B) terms. Significant comparisons are boldened. FGA = filamentous green algae; TMC = total macrophyte cover.
https://doi.org/10.1371/journal.pone.0313876.s004
(DOCX)
Acknowledgments
We thank the managers of Bossington Estate and East Lodge for granting access to carry out this research and all those who assisted with fieldwork. We also thank the three anonymous reviewers for their helpful comments on this manuscript.
References
- 1. Lenders HJ, Chamuleau TP, Hendriks AJ, Lauwerier RC, Leuven RS, Verberk WC. Historical rise of waterpower initiated the collapse of salmon stocks. Scientific Reports. 2016; 6: 29269. pmid:27435118.
- 2. Gibling M. River Systems and the Anthropocene: A Late Pleistocene and Holocene Timeline for Human Influence. Quaternary. 2018; 1: 21.
- 3. Piqué G, Batalla RJ, Sabater S. Hydrological characterization of dammed rivers in the NW Mediterranean region. Hydrological Processes. 2016; 30(11): 1691–707.
- 4. Galia T, Škarpich V, Ruman S. Impact of check dam series on coarse sediment connectivity. Geomorphology. 2021; 377: 107595.
- 5. Foster I, Boardman J, Evans J, Copeland-Phillips R, Vadher A, Wright S, et al. Anthropogenic sediment traps and network dislocation in a lowland UK river. Earth Surface Processes and Landforms. 2021; 47(1): 143–58.
- 6. Im RY, Kim JY, Nishihiro J, Joo G-J. Large weir construction causes the loss of seasonal habitat in riverine wetlands: a case study of the Four Large River Projects in South Korea. Ecological Engineering. 2020; 152: 105839.
- 7. Wang J, Ding C, Heino J, Jiang X, Tao J, Ding L, et al. What explains the variation in dam impacts on riverine macroinvertebrates? A global quantitative synthesis. Environmental Research Letters. 2020; 15: 124028.
- 8. Graf W, Leitner P, Hanetseder I, Ittner LD, Dossi F, Hauer C. Ecological degradation of a meandering river by local channelization effects: a case study in an Austrian lowland river. Hydrobiologia. 2016; 772(1): 145–60.
- 9. Grizzetti B, Pistocchi A, Liquete C, Udias A, Bouraoui F, van de Bund W. Human pressures and ecological status of European rivers. Scientific Reports. 2017; 7(205): 1–11. pmid:28302999
- 10. Albert JS, Destouni G, Duke-Sylvester SM, Magurran AE, Oberdorff T, Reis RE, et al. Scientists’ warning to humanity on the freshwater biodiversity crisis. Ambio. 2021; 50(1): 85–94. pmid:32040746.
- 11. Freedman J, Carline R, Stauffer J. Gravel dredging alters diversity and structure of riverine fish assemblages. Freshwater Biology. 2013; 58(2): 261–74.
- 12. Rentier ES, Cammeraat LH. The environmental impacts of river sand mining. Science of The Total Environment. 2022; 838(1): 155877. pmid:35569654.
- 13. Brenna A, Surian N, Mao L. Response of a Gravel Bed River To Dam Closure: Insights From Sediment Transport Processes And Channel Morphodynamics. Earth Surface Processes and Landforms. 2020; 45(3): 756–70.
- 14. Kondolf GM. Hungry Water: Effects of Dams and Gravel Mining on River Channels. Environmental Management. 1997; 21(4): 533–51. pmid:9175542
- 15. Koehnken L, Rintoul MS, Goichot M, Tickner D, Loftus AC, Acreman MC. Impacts of riverine sand mining on freshwater ecosystems: A review of the scientific evidence and guidance for future research. River Research and Applications. 2020; 36(3): 362–70.
- 16. Brown AV, Lyttle MM, Brown KB. Impacts of Gravel Mining on Gravel Bed Streams. Transactions of the American Fisheries Society. 1998; 127(6): 979–94. .
- 17. Draut AE, Logan JB, Mastin MC. Channel evolution on the dammed Elwha River, Washington, USA. Geomorphology. 2011; 127(1): 71–87.
- 18. Marren PM, Grove JR, Webb JA, Stewardson MJ. The potential for dams to impact lowland meandering river floodplain geomorphology. Scientific World Journal. 2014; 2014: 309673. pmid:24587718.
- 19. Paukert C, Schloesser J, Fischer J, Eitzmann J, Bouska K, Thornbrugh D. Effect of Instream Sand Dredging on Fish Communities in the Kansas River USA: Current and Historical Perspectives. Journal of Freshwater Ecology. 2008; 23: 623–33.
- 20. Cote IM, Vinyoles D, Reynolds JD, Doadrio I, Perdices A. Potential impacts of gravel extraction on Spanish populations of river blennies Salaria fluviatilis (Pisces, Blenniidae). Biological Conservation. 1999; 87(3): 359–67.
- 21. Mingist M, Shewit G. Could sand mining be a major threat for the declining endemic Labeobarbus species of Lake Tana, Ethiopia?: Sand mining as a threat for the endemic fishes. Singapore Journal of Tropical Geography. 2016; 37: 195–208.
- 22. Kondolf GM, Gao Y, Annandale GW, Morris GL, Jiang E, Zhang J, et al. Sustainable sediment management in reservoirs and regulated rivers: Experiences from five continents. Earth’s Future. 2014; 2(5): 256–80.
- 23. Staentzel C, Kondolf GM, Schmitt L, Combroux I, Barillier A, Beisel JN. Restoring fluvial forms and processes by gravel augmentation or bank erosion below dams: A systematic review of ecological responses. Science of The Total Environment. 2020; 706: 135743. pmid:31841838.
- 24. Mörtl C, De Cesare G. Sediment Augmentation for River Rehabilitation and Management—A Review. Land 2021; 10(12): 1309.
- 25. Sellheim KL, Watry CB, Rook B, Zeug SC, Hannon J, Zimmerman J, et al. Juvenile Salmonid Utilization of Floodplain Rearing Habitat After Gravel Augmentation in a Regulated River. River Research and Applications. 2016; 32(4): 610–21.
- 26. Kasahara T, Hill AR. Instream restoration: Its effects on lateral stream-subsurface water exchange in urban and agricultural streams in Southern Ontario. River Research and Applications. 2007; 23(8): 801–14. PMID: 000251254500001.
- 27. England J, Wilkes MA. Does river restoration work? Taxonomic and functional trajectories at two restoration schemes. Science of The Total Environment. 2018; 618: 961–70. pmid:29126643
- 28. Pulg U, Barlaup B, Sternecker K, Trepl L, Unfer G. Restoration of Spawning Habitats of Brown Trout (Salmo trutta) in a Regulated Chalk Stream. River Research and Applications. 2013; 29.
- 29. Pulg U, Lennox RJ, Stranzl S, Espedal EO, Gabrielsen SE, Wiers T, et al. Long-term effects and cost-benefit analysis of eight spawning gravel augmentations for Atlantic salmon and Brown trout in Norway. Hydrobiologia. 2022; 849(2): 485–507.
- 30. Matsushima K, Hyodo M, Shibata N, Shimizu Y. Effectiveness of Flexible Dam Operation and Sediment Replenishment at Managawa Dam, Japan. Journal of Disaster Research. 2018; 13: 691–701.
- 31. Pretty JL, Harrison SSC, Shepherd DJ, Smith C, Hildrew AG, Hey RD. River rehabilitation and fish populations: assessing the benefit of instream structures. Journal of Applied Ecology. 2003; 40(2): 251–65.
- 32. Merz JE, Ochikubo Chan LK. Effects of gravel augmentation on macroinvertebrate assemblages in a regulated California River. River Research and Applications. 2005; 21(1): 61–74.
- 33. Zeug SC, Sellheim K, Watry C, Rook B, Hannon J, Zimmerman J, et al. Gravel augmentation increases spawning utilization by anadromous salmonids: a case study from California, USA. River Research and Applications. 2014; 30(6): 707–18.
- 34. Gaeuman D. High flow gravel injection for constructing designed in-channel features. River Research and Applications. 2014; 30(6): 685–706.
- 35. Mueller M, Pander J, Geist J. The ecological value of stream restoration measures: An evaluation on ecosystem and target species scales. Ecological Engineering. 2014; 62: 129–39.
- 36.
Bannister N, Mant J, Janes M. A review of catchment scale river restoration projects in the UK. Bedfordshire, United Kingdom: River Restoration Centre; 2005. https://www.therrc.co.uk/pdf/Publications/Catchment_report_Jan05.pdf
- 37. Palm D, Brännäs E, Lepori F, Nilsson KN, Stridsman S. The influence of spawning habitat restoration on juvenile brown trout (Salmo trutta) density. Canadian Journal of Fisheries and Aquatic Sciences. 2007; 64: 509–15.
- 38. Staentzel C, Arnaud F, Combroux I, Schmitt L, Trémolières M, Grac C, et al. How do instream flow increase and gravel augmentation impact biological communities in large rivers: A case study on the Upper Rhine River. River Research and Applications. 2018; 34(2): 153–64.
- 39. Staentzel C, Combroux I, Barillier A, Schmitt L, Chardon V, Garnier A, et al. Responses of biological communities to restoration actions in large rivers (Old Rhine, France). La Houille Blanche. 2018; 104(2): 99–106.
- 40. Harrison SSC, Pretty JL, Shepherd D, Hildrew AG, Smith C, Hey RD. The effect of instream rehabilitation structures on macroinvertebrates in lowland rivers. Journal of Applied Ecology. 2004; 41(6): 1140–54.
- 41. McManamay RA, Orth DJ, Dolloff CA. Macroinvertebrate Community Responses to Gravel Addition in a Southeastern Regulated River. Southeastern Naturalist. 2013; 12(3): 599–618.
- 42. Bauer M, Woellner R, Strobl K, Kollmann J. Resilience of riparian vegetation after restoration measures on River Inn. River Research and Applications. 2018; 34: 451–60.
- 43. Romanov A, Hardy J, Zeug S, Cardinale B. Abundance, size structure, and growth rates of Sacramento pikeminnow (Ptychocheilus grandis) following a large-scale stream channel restoration in California. Journal of Freshwater Ecology. 2012; 27: 495–505.
- 44. Albertson LK, Cardinale BJ, Zeug SC, Harrison LR, Lenihan HS, Wydzga MA. Impacts of Channel Reconstruction on Invertebrate Assemblages in a Restored River. Restoration Ecology. 2011; 19(5): 627–38.
- 45. Pander J, Geist J. Ecological indicators for stream restoration success. Ecological Indicators. 2013; 30: 106–18.
- 46. Kail J, Brabec K, Poppe M, Januschke K. The effect of river restoration on fish, macroinvertebrates and aquatic macrophytes: A meta-analysis. Ecological Indicators. 2015; 58: 311–21.
- 47. Polvi LE, Lind L, Persson H, Miranda-Melo A, Pilotto F, Su X, et al. Facets and scales in river restoration: Nestedness and interdependence of hydrological, geomorphic, ecological, and biogeochemical processes. Journal of Environmental Management. 2020; 265: 110288. pmid:32421567
- 48. Albertson L, Koenig L, Lewis B, Zeug S, Harrison L, Cardinale B. How Does Restored Habitat For Chinook Salmon (Oncorhynchus Tshawytscha) In The Merced River In California Compare With Other Chinook Streams? River Research and Applications. 2013; 29: 469–82.
- 49. Kemp PS, Vowles AS, Sotherton N, Roberts D, Acreman MC, Karageorgopoulos P. Challenging convention: the winter ecology of brown trout (Salmo trutta) in a productive and stable environment. Freshwater Biology. 2017; 62(1): 146–60.
- 50.
Mainstone CP. Chalk rivers nature conservation and management. Peterborough, United Kingdom: English Nature and Environmental Agency; 1999. http://publications.naturalengland.org.uk/publication/5981928
- 51. Mondon B, Sear DA, Collins AL, Shaw PJ, Sykes T. The scope for a system-based approach to determine fine sediment targets for chalk streams. CATENA. 2021; 206: 105541.
- 52.
CaBA. Chalk stream restoration strategy 2021 main report. Catchment Based Approach; 2021. https://catchmentbasedapproach.org/wp-content/uploads/2021/10/CaBA-CSRG-Strategy-MAIN-REPORT-FINAL-12.10.21-Low-Res.pdf
- 53.
Environment Agency [dataset]. Reasons for not achieving good status dataset. 2021 [cited 29/07/2021]. Available from: https://environment.data.gov.uk
- 54.
Environment Agency [dataset]. The Water Framework Directive 2000/60/EC (WFD) Classification Status Cycle 2. 2021 [cited 29/07/2022]. Available from: https://data.gov.uk/dataset
- 55. Bickerton M, Petts G, Armitage P, Castella E. Assessing the ecological effects of groundwater abstraction on chalk streams: Three examples from Eastern England. Regulated Rivers: Research & Management. 1993; 8(1–2): 121–34.
- 56. Acornley RM, Sear DA. Sediment transport and siltation of brown trout (Salmo trutta L.) spawning gravels in chalk streams. Hydrological Processes. 1999; 13(3): 447–58. .
- 57. Sear DA, Armitage PD, Dawson FH. Groundwater dominated rivers. Hydrological Processes. 1999; 13(3): 255–76. .
- 58.
River Restoration Centre [dataset]. NRRI database: custom dataset containing rivers with chalk geology (provided on request). 2023 [cited 25/03/2023]. Available from: https://www.therrc.co.uk/national-river-restoration-inventory-nrri
- 59. Extence C, Chadd R, England J, Dunbar MJ, Wood P, Taylor ED. The Assessment of Fine Sediment Accumulation in Rivers Using Macro-Invertebrate Community Response. River Research and Applications. 2010; 29(1): 17–55.
- 60. Extence C, Balbi D, Chadd R. River Flow Indexing Using British Benthic Macroinvertebrates: A Framework for Setting Hydroecological Objectives. Regulated Rivers Research & Management. 1999; 15: 545–74. .
- 61.
AEDA [Internet]. Test and Itchen catchment; 2013 [cited 17/09/2022]. Available from: http://www.environmentdata.org/archive/geopref:3493
- 62.
Skinner K. Test and Itchen River Restoration Strategy. Oxford, United Kingdom: Atkins; 2013. http://www.therrc.co.uk/sites/default/files/files/Designated_Rivers/Test_Itchen/technical_report_issue_5_final.pdf
- 63.
Natural England [Internet]. Sites of Special Scientific Interest and historical monuments; 2013 [cited Available from: https://www.gov.uk/guidance/sites-of-special-scientific-interest-and-historical-monuments
- 64.
Natural England [Internet]. River Test SSSI; 1996 [cited 30/07/2022]. Available from: https://designatedsites.naturalengland.org.uk/SiteDetail.aspx?SiteCode=S2000170&SiteName=test&countyCode=19&responsiblePerson=&SeaArea=&IFCAArea=
- 65.
Natural England [Internet]. River Itchen SSSI and SAC; 2000 [cited 30/07/2022]. Available from: https://designatedsites.naturalengland.org.uk/SiteList.aspx?siteName=itchen&countyCode=19&responsiblePerson=&DesignationType=All
- 66. Communities European. Council directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora. Official Journal of the European Union,. 1992; L206: 7–50.
- 67.
JNCC [Internet]. River Itchen; 2023 [cited 11/03/2023]. Available from: https://sac.jncc.gov.uk/site/UK0012599
- 68.
Ordnance Survey [dataset]. Boundary-Line™. 2023 [cited 13/11/2023]. Available from: https://osdatahub.os.uk/downloads/open/BoundaryLine
- 69.
Ordnance Survey [dataset]. OS Open Rivers dataset. 2022 [cited 04/01/2023]. Available from: https://osdatahub.os.uk/downloads/open/OpenRivers
- 70.
Ordnance Survey [Internet]. OS Open Grey Tile Layer; 2016 [cited 01/09/2023]. Available from: https://hub.arcgis.com/maps/eba4e059903747f7b0aa9e92e85ff716/about
- 71.
Grieve N. River Itchen SSSI channel unit condition assessment. 2014. http://www.hwa.uk.com/site/wp-content/uploads/2017/12/1.-Grieve-2014-River-Itchen-SSSI-Condition-Assessment-UCL.pdf
- 72. Fones GR, Bakir A, Gray J, Mattingley L, Measham N, Knight P, et al. Using high-frequency phosphorus monitoring for water quality management: a case study of the upper River Itchen, UK. Environmental Monitoring and Assessment. 2020; 192(3). pmid:32072347
- 73. Zhang Y, Collins AL, McMillan S, Dixon ER, Cancer-Berroya E, Poiret C, et al. Fingerprinting source contributions to bed sediment-associated organic matter in the headwater subcatchments of the River Itchen SAC, Hampshire, UK. River Research and Applications. 2017; 33(10): 1515–26.
- 74.
Environment Agency. Test and Itchen Abstraction licensing Strategy version 3. Bristol, United Kingdom: Environment Agency; 2019. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/793438/Test_and_Itchen_Abstraction_Licence_Strategy.pdf
- 75. Jarvie HP, Neal C, Jürgens MD, Sutton EJ, Neal M, Wickham HD, et al. Within-river nutrient processing in Chalk streams: The Pang and Lambourn, UK. Journal of Hydrology. 2006; 330(1): 101–25.
- 76.
Environment Agency [Internet]. River Basin Management Plan South East River Basin District Annex I: Designating artificial and heavily modified water bodies; 2009 [cited 05/01/2024]. Available from: https://assets.publishing.service.gov.uk/media/5a750d0fe5274a59fa717027/geso0910bstj-e-e.pdf
- 77.
DEFRA [Internet]. Hydrology Data Explorer; 2021 [cited 15/07/2021]. Available from: https://environment.data.gov.uk/hydrology/explore
- 78.
Poynter A. Impacts of environmental stressors on the River Itchen Ranunculus community [PhD thesis]. University of Birmingham; 2013. Available from: https://etheses.bham.ac.uk/id/eprint/5112/
- 79.
WFDUK. Biological Status Methods Rivers ‐ Benthic Invertebrates. WFD UK TAG; 2022. https://www.wfduk.org/sites/default/files/Media/Characterisation%20of%20the%20water%20environment/Summaries%20of%20Biological%20Status%20Methods/river%20acidification%20summary.pdf
- 80.
Dobson M, Pawley S., Fletcher M. and Powell A. Guide to freshwater invertebrates. 1st ed. Ambleside: Freshwater Biological Association; 2012.
- 81. Montori A, Tierno De Figueroa JM, Santos X. The Diet of the Brown Trout Salmo trutta (L.) during the Reproductive Period: Size-Related and Sexual Effects. International Review of Hydrobiology. 2006; 91(5): 438–50.
- 82.
R Studio Team. RStudio: Integrated Development for R. RStudio, Inc. Boston, MA. 2020. Available from: http://www.rstudio.com
- 83. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software. 2015; 67(1): 1–48.
- 84. Lenth RV. Least-Squares Means: The R Package lsmeans. Journal of Statistical Software. 2016; 69(1): 1–33.
- 85. Lüdecke D, Ben-Shachar M, Patil I, Waggoner P, Makowski D. performance: An R Package for Assessment, Comparison and Testing of Statistical Models. Journal of Open Source Software. 2020; 6(10): 3139.
- 86.
Wickham H. ggplot2: Elegant Graphics for Data Analysis. 2nd ed. New York: Springer-Verlag 2016.
- 87.
Pedersen T, L. Patchwork: The Composer of Plots. R package version 1.1.1. https://CRAN.R-project.org/package=patchwork.
- 88. Noguchi K, Gel YR, Brunner E, Konietschke F. nparLD: An R Software Package for the Nonparametric Analysis of Longitudinal Data in Factorial Experiments. Journal of Statistical Software. 2012; 50(12): 1–23.
- 89. Konietschke F, Placzek M, Schaarschmidt F, Hothorn L. nparcomp: An R Software Package for Nonparametric Multiple Comparisons and Simultaneous Confidence Intervals. Journal of Statistical Software. 2015; 64: 1–17.
- 90.
Bartoń K. MuMIn: Multi-Model Inference (version 1.48.4). https://CRAN.R-project.org/package=MuMIn.
- 91.
EU Parliament Council. Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in the field of water policy. Official Journal of the European Union. 2000; L327: 1–73.
- 92.
UK Government [Internet]. Environment Act 2021; 2021 [cited 01/09/2023]. Available from: https://www.legislation.gov.uk/ukpga/2021/30/contents/enacted
- 93.
Mitchell LTN. An assessment of rehabilitation gravels for Salmo trutta spawning: a case study from a small chalk stream, the River Stiffkey, Norfolk, UK [PhD thesis]. University College London; 2016. Available from: https://discovery.ucl.ac.uk/id/eprint/1478280/
- 94. Beechie T, Sear D, Olden J, Pess G, Buffington J, Moir H, et al. Process-Based Principles for Restoring River Ecosystems. BioScience. 2010; 60.
- 95. Bernhardt ES, Palmer MA. River restoration: the fuzzy logic of repairing reaches to reverse catchment scale degradation. Ecological Applications. 2011; 21(6): 1926–31. pmid:21939034
- 96. Duan X, Wang Z, Tian S. Effect of streambed substrate on macroinvertebrate biodiversity. Frontiers of Environmental Science & Engineering in China. 2008; 2: 122–8.
- 97. Duan X, Wang Z, Xu M-z, Zhang K. Effect of streambed sediment on benthic ecology. International Journal of Sediment Research. 2009; 24: 325–38.
- 98. Jones J, Collins A, Naden PS, Sear D. The relationship between fine sediment and macrophytes in rivers. River Research and Applications. 2012; 28: 1006–18.
- 99. Dodds W, Gudder D. The ecology of Cladophora. Journal of Phycology. 1992; 28: 415–27.
- 100. Carlson PE, Donadi S, Sandin L. Responses of macroinvertebrate communities to small dam removals: Implications for bioassessment and restoration. Journal of Applied Ecology. 2018; 55(4): 1896–907.
- 101. Sundermann A, Stoll S, Haase P. River restoration success depends on the species pool of the immediate surroundings. Ecological Applications. 2011; 21(6): 1962–71. pmid:21939037
- 102. Aronsuu K, Virkkala P. Substrate selection by subyearling European river lampreys (Lampetra fluviatilis) and older larvae (Lampetra spp). Ecology of Freshwater Fish. 2014; 23(4): 644–55.
- 103. Palmer MA, Bernhardt ES, Allan JD, Lake PS, Alexander G, Brooks S, et al. Standards for ecologically successful river restoration. Journal of Applied Ecology. 2005; 42(2): 208–17.
- 104. Biggs J, Corfield A, Grøn P, Hansen HO, Walker D, Whitfield M, et al. Restoration of the rivers Brede, Cole and Skerne: a joint Danish and British EU-LIFE demonstration project, V—short-term impacts on the conservation value of aquatic macroinvertebrate and macrophyte assemblages. Aquatic Conservation: Marine and Freshwater Ecosystems. 1998; 8(1): 241–55. .
- 105. Hansen J, Hayes D. Long-term implications of dam removal for macroinvertebrate communities in Michigan and Wisconsin rivers, United States River Research and Applications. 2012; 28(9): 1540–50.
- 106. Wohl EE, Angermeier PL, Bledsoe BP, Kondolf GM, MacDonnell L, Merritt DM, et al. River restoration. Water Resources Research. 2005; 41.
- 107. Tonkin JD, Stoll S, Sundermann A, Haase P. Dispersal distance and the pool of taxa, but not barriers, determine the colonisation of restored river reaches by benthic invertebrates. Freshwater Biology. 2014; 59(9): 1843–55.
- 108. dos Reis Oliveira PC, van der Geest HG, Kraak MHS, Westveer JJ, Verdonschot RCM, Verdonschot PFM. Over forty years of lowland stream restoration: Lessons learned? Journal of Environmental Management. 2020; 264: 110417. pmid:32217312
- 109. Roni P, Åberg U, Weber C. A Review of Approaches for Monitoring the Effectiveness of Regional River Habitat Restoration Programs. North American Journal of Fisheries Management. 2018; 38(5): 1170–86.
- 110. England J, Angelopoulos N, Cooksley S, Dodd J, Gill AB, Gilvear D, et al. Best Practices for Monitoring and Assessing the Ecological Response to River Restoration. Water. 2021; 13: 3352.
- 111. England J, Naura M, Mant J, Skinner K. Seeking river restoration appraisal best practice: supporting wider national and international environmental goals. Water and Environment Journal. 2020; 34(S1): 1003–11.