Catchment-scale peatland restoration benefits stream ecosystem biodiversity



1. Drainage of peat-dominated catchments across the world has caused widespread degradation of peat and freshwater ecosystem services. In the UK, an estimated £500 million has been spent over the last decade blocking drains to reverse these changes. The practice raises water-tables to induce rewetting and promote peat aggradation. However, the potential benefits for impacted ecosystems such as streams remain unknown.

2. This study examined stream physicochemistry and benthic macroinvertebrates across peatland catchments with artificial drainage networks, or drains that have recently been blocked, and compared these with intact peatland sites having no history of drainage.

3. Streams in artificially drained catchments were characterised by more benthic fine particulate organic matter (FPOM), higher suspended sediment concentrations and finer bed sediments (D50) than in drain-blocked and intact catchments.

4. Drained sites had higher abundance of Diptera (Simuliidae and Chironomidae) larvae, and lower abundance of Ephemeroptera, Plecoptera and Trichoptera larvae, than drain-blocked sites. In contrast, streams in drain-blocked catchments had macroinvertebrate communities broadly similar to intact sites in terms of taxon richness, overall species composition and community structure. These changes were associated with lower suspended sediment and benthic FPOM concentrations following drain-blocking.

5.Synthesis and applications. This study has shown changes in the structure of stream benthic macroinvertebrate assemblages linked to increases in benthic particulate organic matter and suspended sediment following peatland drainage. However, these effects seem to be reversible following catchment-scale restoration by drain-blocking. Drain-blocking therefore appears to benefit not only peatland soil, vegetation and hydrological ecosystem services but also stream water quality and biodiversity. The numerous agencies undertaking peatland restoration should consider implementing detailed pre- and post-blocking monitoring of streams to further improve our understanding of the mechanisms through which peatland management affects stream biodiversity and biological recovery dynamics, refine drain-blocking practices, and inform aquatic conservation and management strategies.


River catchments world-wide have been altered significantly as a consequence of land-use change (e.g. Harding et al. 1998; Allan 2004). Peatland covers c. 500 m ha globally but, historically, many countries such as the Netherlands, Finland, Russia, Ireland, UK, Canada and Ecuador have suffered artificial peatland drainage. However, the full catchment-scale effects of these drainage practices have rarely been examined (Ramchunder, Brown & Holden 2009). In the UK, some drainage pre-dates Roman times but many peatlands were drained during the 1960s and 1970s to increase grouse, sheep and timber production, provide peat for horticulture and fuel, and with the intention of reducing downstream flooding (Holden, Chapman & Labadz 2004). However, as Stewart & Lance (1983) noted, there was no evidence of economic benefits from moorland drainage and little research was undertaken into the wider environmental impacts.

Artificial peatland drainage is considered to have negative effects including alterations to baseflow and stormflow run-off regimes with some evidence of increased flow peaks and lower baseflows (e.g. Conway & Millar 1960; Holden et al. 2006). Impacts have also included changes in solute (e.g. calcium, magnesium, potassium) and nutrient (e.g. carbon-bound nitrogen and sulphur and organically bound phosphorus) concentrations of streams (e.g. Ramchunder, Brown & Holden 2009), and erosion of exposed peat resulting in increased sediment delivery to streams (e.g. Holden, Gascoign & Bosanko 2007). In an attempt to restore peatlands following such degradation, the practice of installing drain dams was initiated in the UK in the late 1980s (Armstrong et al. 2009).

Drain-blocking has so far been implemented over >3200 ha of England (Holden 2009), with common practice being a series of dams rather than total infilling. The principal aim of drain-blocking is to increase water-table height, encouraging re-vegetation by peat forming species such as Sphagnum. The restorative capabilities of drain-blocking on terrestrial vegetation and soil structure/chemistry are reviewed by Ramchunder, Brown & Holden (2009), and the efficacy of different blocking methods is detailed in Armstrong et al. (2009). Soil water dissolved organic carbon concentrations have been shown to be lower in drain-blocked peatland areas compared with drained areas (Wallage, Holden & McDonald 2006). However, questions remain as to whether catchment-scale restoration by drain-blocking has any benefits for stream biodiversity. Such knowledge would be useful given the general need to increase understanding of the ecological benefits of catchment-scale remediation schemes (Hillman & Brierley 2005), which have received comparatively less attention than restoration schemes focused on river sections or reaches (e.g. Bernhardt et al. 2005; Palmer et al. 2005).

This study investigated stream macroinvertebrate communities from nine headwater (2–4 km2) peatland catchments (three intact, three with widespread artificial drainage networks and three with blocked drains), on five occasions from 2007 to 2008. Based on knowledge from hydrological studies of peatland drainage and drain-blocking, it was hypothesised that (i) streams in catchments that possess artificial drainage would have higher suspended sediment concentrations, finer bed sediments and more benthic fine particulate organic matter (FPOM) when compared with streams from intact catchments (Prévost, Plamondon & Belleau 1999; Holden, Gascoign & Bosanko 2007). These changes to the stream environment were expected to: (ii) lead to macroinvertebrate communities containing taxa more associated with in-stream fine sediment deposition and benthic particulate organic matter (POM) (e.g. Ramchunder, Brown & Holden 2009) compared with stream systems from intact catchments. In contrast, (iii) drain-blocked stream ecosystems were expected to be in a similar condition to those at intact sites, because drain dams effectively reduce sediment flux to the stream network (Holden, Gascoign & Bosanko 2007), and restore hydrological and hydrochemical aspects of the terrestrial ecosystem. The findings of this study are subsequently considered in the context of more general literature discussing the effects of land management and catchment-scale remediation schemes on stream ecosystems.

Materials and methods

Study areas

The study was undertaken at Moor House National Nature Reserve, Teesdale and Wensleydale, Wharfedale, Geltsdale and Weardale in northern England (Table 1). All sites have blanket peat cover, with vegetation dominated by Sphagnum spp. (mosses), Eriophorum spp. (cotton grasses) and Calluna vulgaris (heather). The climate of all sites is typical of the UK uplands; data were not available for all sites but mean annual precipitation of 2012 mm (1951–1980 and 1991–2006) at Moor House (Holden & Rose 2011) and 1817 mm (1981–2008) at Oughtershaw Beck (Wallage, Holden & McDonald 2006) can be considered broadly representative. Mean annual air temperature at Moor House is 5·3 °C (1931–2006; Holden & Rose 2011) and 6 °C at Oughtershaw Beck (2007–2009; Brown et al. 2010a).

Table 1.  Summary information for the nine stream study sites
Stream nameManagement typeSoil typesCatchment area (km2)aGrid reference
  1. aMeasured using the hydrology tool in ArcGIS (Esri, Redlands, CA, USA).

Moss BurnIntactBlanket peat2·1554°41′1″N
Snaizeholme BeckIntactBlanket peat, stagnogley, stagnohumic gley, humic gley, fine loam, alluvial gley2·2354°15′7″N
Unnamed 2nd order tributary of the River TeesIntactBlanket peat, alluvial floodplain1·1254°41′8″N
Crook BurnDrain-blockedBlanket peat, alluvial floodplain2·9854°42′2″N
South TyneDrain-blockedBlanket peat, alluvial floodplain3·5254°43′6″N
Blea GillDrain-blockedBlanket peat, humic gley, humic loam,1·6154°14′2″N
Cam BeckDrainedBlanket peat, stagnohumic gley, seasonally waterlogged loam with peaty surface2·9154°13′9″N
Old WaterDrainedBlanket peat, stagnohumic gley, seasonally waterlogged loam with peaty surface1·2954°52′4″N
Unnamed 2nd order tributary of Killhope BurnDrainedBlanket peat, slowly permeable loam over clay, stagnohumic gley1·0854°46′7″N

Potential study sites were identified with second order streams based on 1 : 25 000 Ordnance Survey maps. Catchment (1·08–3·52 km2; Table 1) and stream size were standardised as far as possible following Ramchunder et al. (2011) who suggested catchment size (and stream order) can have a significant effect on peatland stream macroinvertebrate community composition. Candidate drained and drain-blocked (3–11 years post-blocking) sites were identified from aerial photographs and following discussions with land managers. Sites were selected with no confounding effects of wildfire, rotational heather burning, mining, major erosion or forest cover. At each catchment outlet, a representative 15-m reach was selected randomly for study with subsequent sampling undertaken in riffle areas of those reaches.

Field sampling

Streams were sampled seasonally across 2–4 days per quarter (2007: September 3–7; December 4–8; 2008: March 5–9; June 4–8; September 2–6). On each occasion, 16 environmental variables were measured (Table 2). Water temperature, pH and electrical conductivity (EC) were measured using Mettler Toledo MP120 and MP126 handheld probes (Mettler-Toledo Ltd, Leicester, UK). Dissolved oxygen (DO) concentration was measured using a Hanna Instruments HI9412 probe (Hanna Instruments Ltd, Bedfordshire, UK). Additionally, 120 mL of stream water were passed through a 0·45-μm filter and subsequently analysed in the laboratory for major anions (Cl, SO4 and NO3), dissolved organic carbon (DOC) and dissolved metals (Al and Fe). A further 500 mL of unfiltered stream water were collected for determination of suspended sediment concentration (SSC) by filtration. Streambed sediments were characterised by measuring b-axis length and median size (D50) of 100 randomly sampled clasts. To provide a relative indication of flow differences between sites and over time, stream discharge (Q) was calculated using an open channel flow meter (Valeport, Devon, UK) and the velocity–area method.

Table 2.  Descriptive statistics and RM-anova results for the stream environmental variables
 Cl (mg L−1)NO3 (mg L−1)SO4 (mg L−1)Al (mg L−1)Fe (mg L−1)DOC (mg L−1)DO (mg L−1)EC (μS cm−1)pHSSC (mg L−1)D50 (cm)CPOM (mg m−2)FPOM (mg m−2)POM (mg m−2)Water temperature (°C)Discharge (m3 s−1)
  1. DOC, dissolved organic carbon; EC, electrical conductivity; SSC, suspended sediment concentration.

All streams
Artificially drained
Stream (F6,44) F = 3·62
P = 0·011
F = 8·96
P < 0·001
F = 10·23
P < 0·001
F = 2·38
P = 0·081
F = 1·33
P = 0·283
F = 3·44
P = 0·014
F = 0·74
P = 0·622
F = 6·35
P = 0·001
F = 15·42
P < 0·001
F = 5·17
P = 0·002
No replicates F = 4·37
P = 0·004
F = 11·87
P < 0·001
F = 6·50
P < 0·001
F = 1·45
P = 0·239
F = 4·61
P = 0·004
Season (F4,44) F = 5·18
P = 0·004
F = 54·51
P < 0·001
F = 16·46
P = <0·001
F = 6·74
P = 0·002
F = 3·75
P = 0·017
F = 2·34
P = 0·084
F = 18·41
P < 0·001
F = 6·10
P = 0·002
F = 4·83
P = 0·005
F = 2·21
P = 0·098
No replicates F = 1·07
P = 0·395
F = 5·09
P = 0·004
F = 2·93
P = 0·042
F = 21·45
P < 0·001
F = 5·02
P = 0·005
Land management (F2,44) F = 2·12
P = 0·201
F = 1·69
P = 0·261
F = 2·79
P = 0·139
F = 0·15
P = 0·868
F = 1·10
P = 0·392
F = 0·18
P = 0·836
F = 1·12
P = 0·384
F = 0·64
P = 0·560
F = 1·00
P = 0·423
F = 5·73
P = 0·041
F = 47·05
P < 0·001
F = 0·28
P = 0·768
F = 5·23
P = 0·048
F = 2·75
P = 0·142
F = 0·94
P = 0·443
F = 1·70
P = 0·260
Season*Land management (F8,44) F = 0·79
P = 0·614
F = 4·20
P = 0·003
F = 2·19
P = 0·066
F = 1·19
P = 0·366
F = 1·78
P = 0·131
F = 1·43
P = 0·236
F = 0·23
P = 0·982
F = 1·13
P = 0·384
F = 0·52
P = 0·829
F = 1·65
P = 0·164
No replicates F = 1·07
P = 0·420
F = 1·89
P = 0·110
F = 1·03
P = 0·439
F = 0·91
P = 0·522
F = 0·42
P = 0·896

Five replicate 0·05 m2 benthic macroinvertebrate samples were collected from riffle habitats using a modified Surber sampler (250 μm mesh). Sediment was disturbed to a depth of c. 10 cm for c. 3 min per sample. All samples were preserved immediately in 70% ethanol and then transported back to the laboratory for sorting and identification. Where possible, macroinvertebrates were identified to species level under a light microscope (×40 magnification) but other taxa were identified to higher levels (e.g. Diptera [Family/Genus], Oligochaeta [Class]) using standard UK freshwater macroinvertebrate identification keys. POM retained within the samples was sorted into fine (<1 mm; FPOM) and coarse particulate organic matter (>1 mm; CPOM) fractions and ashed to determine dry mass.

Data analysis

Principal component analysis (PCA) helped examine interrelationships between 16 environmental variables across all sampling dates. Principal components (PCs) with Eigenvalues >1 were retained and % variance of each recorded. Repeated Measures anova (season as repeated measure) with Bonferroni correction was used to determine significant differences in stream environmental variables and PC scores as a function of site, land management type and season. Site was a random factor in the model while land management and season were fixed.

Prior to analysis, the five replicate invertebrate data sets per site/date were pooled (e.g. Brown, Milner & Hannah 2007) to enable clearer elucidation of land management impacts as opposed to patch-scale variability. Macroinvertebrate community structure was summarised using five measures: (i) log10 (total abundance +1) expressed as the total number of individuals per m2; (ii) taxonomic richness; (iii) relative abundance (%) of Ephemeroptera, Plecoptera, Trichoptera, Chironomidae, Simuliidae and Other taxa; (iv) 1/Simpson’s diversity index (1/S):


where N is the total number of individuals in a sample and ni is the number of individuals of taxon i; (v) taxonomic dominance (D) was estimated using the Berger–Parker index:


where Nmax is the number of individuals in the most abundant species and N is the total abundance.

RM-anova was repeated for the macroinvertebrate summary measures using the same methods as outlined above. All environmental and macroinvertebrate data sets were tested for normality and, where necessary, log10 (x+1), arcsin or square root transformed to improve normality and homogeneity of variance prior to statistical tests. All tests were undertaken in spss v17.0 (IBM Corporation, Armonk, New York, USA) or Minitab v15 (Minitab Inc., State College, PA, USA) and considered significant where P < 0·05.

Species–habitat relationships were assessed using ordination in canoco v4.5 (Plant Research International, Wageningen, The Netherlands). Macroinvertebrate data were log10 (x + 1) transformed prior to analysis. A preliminary detrended correspondence analysis showed axis 1 gradients were 3·0 SD; thus, subsequent analyses used direct gradient analysis (Redundancy Analysis; RDA) to test for linear trends in species compositional change (Lepš & Šmilauer 2003). Forward selection determined which physicochemical variables accounted for a significant proportion of the species variance. An initial RDA included a dummy variable ‘Time’ (no. days from start of sampling) to determine whether there were significant temporal trends within the stream macroinvertebrate communities. Thereafter, a partial RDA (pRDA) was carried out to remove variance accounted by Time, to provide a better indication of the spatial component of the data set (Borcard, Legendre & Drapeau 1992). One-way analysis of similarity (anosim) tested the null hypothesis that differences in stream macroinvertebrate taxa abundance between peatland management types were not different to those within types. anosim was undertaken on all samples combined owing to the small number of replicates per quarterly sample collection, and because spatial dynamics (linked to management type) were the key focus of this analysis. anosim was undertaken using both the Bray–Curtis (BC) dissimilarity index (based on taxa relative abundance) and the Jaccard’s coefficient of similarity (based on taxa presence–absence), with 10 000 permutations and Bonferroni corrections using past 2.05 (Hammer, Harper & Ryan 2001).


Mean SO4, FPOM and CPOM, SSC and Al were all highest in drained streams but mean DOC concentrations and D50 were the lowest (Table 2). Average EC and pH were the highest in drain-blocked sites, although the highest pH was from an intact peatland stream. Average Cl, NO3 and SO4 concentrations were all the lowest at intact sites. The RM-anova showed significant differences in SSC, FPOM, D50, and NO3 (with seasonal interaction) between peatland management types (Table 2).

The PCA generated five PCs with Eigenvalues >1 (Table 3 and Fig. 1). PC1 had strong positive loadings (>0·5) of NO3, EC, FPOM and POM and a strong negative loading of Al. PC1 scores were significantly different between management types (F2,44 = 6·3; P = 0·004). The Tukey’s post hoc test showed that PC1 scores were significantly different between streams in intact and artificially drained catchments (P = 0·003). Streams with artificial drainage were characterised by more negative PC1 scores but these were not significantly different to those for drain-blocked streams (P > 0·05). PC2 had strong positive loadings of CPOM, FPOM, POM and Fe, and strong negative loadings of pH and D50. PC2 scores were also significantly different between management types (F2,44 = 15·8; P < 0·001). The Tukey’s post hoc test showed that sites with artificial drainage were characterised by more positive PC2 scores (Fig. 1), and there were significant differences between intact and drain-blocked (P = 0·022), intact and artificially drained (P = 0·018), as well as between drain-blocked and artificially drained sites (P < 0·001).

Table 3.  Loading scores, Eigenvalues and % variance explained for the five principle components produced from the environmental variables data set
VariablePrinciple component
  1. Values >0·5 or <−0·5 are highlighted to aid interpretation.

  2. DOC, dissolved organic carbon; EC, electrical conductivity; SSC, suspended sediment concentration.

Cl0·284−0·1410·102 0·577 0·392
NO3 0·795 −0·310−0·2510·165−0·017
SO40·365−0·451 0·535 0·023−0·397
EC 0·529 −0·4160·3770·0890·283
Water temperature−0·338−0·080 0·822 −0·2620·005
CPOM0·422 0·526 0·064−0·3420·399
FPOM 0·641 0·534 −0·046−0·225−0·130
POM 0·671 0·618 −0·014−0·3050·040
Fe−0·343 0·524 0·274−0·028 0·520
D50−0·341 −0·726 −0·187−0·2740·301
% variance explained21·419·515·26·96·8
Figure 1.

 PC scores for (a) axis 1 and (b) axis 2, for the three peatland management types (Error bars show ±1 SD from the mean). See Table 1 for individual variable loadings, Eigenvalues and % variance of each PC.

Mean invertebrate abundance and richness were the highest in drain-blocked and intact sites, respectively, while mean 1/S and dominance were the highest in intact and drain-blocked sites, respectively. Lowest abundance, richness, 1/S and dominance were documented in the drained sites (Table 4). The RM-anova revealed a borderline significant difference in taxonomic richness between peatland management types (Table 4 and Fig. 2). The relative abundance of Ephemeroptera and Trichoptera was typically higher in intact and drain-blocked sites compared with drained sites (Fig. 3), whereas Plecoptera and other taxa relative abundance were similar across the management types. The relative abundance of Chironomidae and Simuliidae was typically the highest in artificially drained systems (Fig. 3).

Table 4.  Descriptive statistics and RM-anova results for the macroinvertebrate community metrics
 Log10 (total abundance+1) (per m2)RichnessSimpson’s Diversity (1/S)Dominance (D)
All streams
Artificially drained
Stream (F6,44) F = 6·67
P < 0·001
F = 9·92
P < 0·001
F = 7·34
P < 0·001
F = 6·78
P < 0·001
Season (F4,44) F = 4·61
P = 0·007
F = 3·42
P = 0·024
F = 1·17
P = 0·351
F = 1·40
P = 0·264
Land management (F2,44) F = 0·42
P = 0·676
F = 5·03
P = 0·05
F = 1·11
P = 0·390
F = 0·71
P = 0·530
Season*Land management (F8,44) F = 1·15
P = 0·365
F = 1·66
P = 0·159
F = 2·02
P = 0·088
F = 2·01
P = 0·089
Figure 2.

 Effects of management on (a) log10 (abundance +1); (b) Richness; (c) 1/S; and (d) Dominance (Error bars show ±1 SD from the mean).

Figure 3.

 Seasonal changes in the mean relative abundance of Ephemeroptera, Plecoptera, Trichoptera, Chironomidae, Simuliidae and Other taxa in relation to management types.

Axis 1 of the RDA accounted for a total of 16·2% of the total variance while the second axis accounted for 7%. Taxa–environment correlations were 0·790 and 0·876 for axis 1 and 2, respectively. Time accounted for 10·1% of the species variance, thus a pRDA was undertaken. Axis 1 and 2 accounted for a total of 17·0% and 7·2% of the total variance while taxa–environment correlations were 0·791 and 0·853 for axis 1 and 2, respectively. Forward selection showed pH, SSC, Al, Fe, FPOM, SO4 and water temperature were associated with a significant proportion of the variance.

The analysis showed that the intact and drain-blocked sites were associated with low suspended sediment, FPOM and SO4 concentrations (Fig. 4a). The taxon–environmental variables biplot showed some Ephemeroptera species (e.g. Baetis rhodani, Ecdyonurus torrentis, Ecdyonurus dispar and Rhithrogena semicolorata), Plecoptera (e.g. Perlodes microcephala and Isoperla grammatica) and caseless Trichoptera larvae (e.g. Polycentropus flavomaculatus and Hydropsyche pellucidula) were more abundant in streams from intact and drain-blocked catchments. In contrast, some dipterans (e.g. Simuliidae and Chironomidae), the ephemeropteran Ephemera danica and the cased Trichoptera larvae Sericostoma personatum were more strongly associated with drained sites with higher SSC, FPOM and SO4 concentrations (Fig. 4b). A diverse assemblage of Ephemeroptera species was found in the intact and drain-blocked sites, while only E. danica was documented in the drained sites (Fig. 4b).

Figure 4.

 Redundancy Analysis (RDA) ordination diagrams of (a) management types and environmental variables, and (b) macroinvertebrate taxa (with selected taxa highlighted). Species abbreviations follow those provided in the main text.

anosim based on macroinvertebrate relative abundance data revealed significant differences in community composition between land management types (R2 = 0·39; P < 0·001), with pair-wise comparisons highlighting drained site macroinvertebrate assemblages were different to those at intact and drain-blocked sites (R2 = 0·43 and 0·65, respectively; both P < 0·001). Similarly, community composition at the intact and drain-blocked streams was significantly different (R2 = 0·10; P = 0·032). anosim based on presence–absence data also showed significant differences between land management types (R2 = 0·36; P < 0·001), with pair-wise comparisons showing macroinvertebrate communities were different at drained sites compared with intact and drain-blocked sites (R2 = 0·43 and 0·59, respectively; P < 0·001 in both cases). In contrast to the relative abundance analysis, the intact and drain-blocked streams were not significantly different (R2 = 0·07; P = 0·105).


Peatland management effects on stream environmental variables

This study has provided the first detailed insight into the spatial and temporal dynamics of stream physicochemical environmental variables and macroinvertebrate communities in peatland rivers influenced by artificial drainage and drain-blocking. Stream systems in catchments that possess artificial drainage were hypothesised to have higher SSC and FPOM compared with stream systems from intact catchments. Indeed, the findings of this study showed artificial drainage of peatlands was linked to changes in several stream physicochemical variables (e.g. increases in SSC, FPOM and NO3), allowing hypothesis 1 to be upheld. These findings are supported in part by evidence from other studies, where 50-fold (Robinson & Blyth 1982) and 12-fold (Ahtiainen & Huttunen 1999) increases in stream sediment loads have been observed because of peat drainage. Erosion of artificial drains can promote fine particulate deposition in nearby aquatic systems (Holden 2006; Holden, Gascoign & Bosanko 2007). This study showed significantly higher concentrations of FPOM in artificially drained sites compared with intact sites. Lower concentrations of NO3 in the intact sites compared with drain-blocked sites could be related to seasonal water-table drawdown, which may be greater with drainage and not fully restored following drain-blocking (Holden et al. 2011). Leached ammonia and dissolved organic nitrogen may be mineralised and nitrified under aerobic conditions, providing a greater source of NO3 to streams (Williams & Wheatley 1988).

Stream benthic FPOM concentrations in drain-blocked catchment streams were lower than those with artificial drainage most probably because, even in poorly dammed drains, fine particulate transport can be markedly lower when compared with unblocked drains (Holden, Gascoign & Bosanko 2007). Following drain-blocking, the area of exposed bare peat, and the connectivity of bare areas with the stream, is reduced. Thus, peat erosion declines and less FPOM reaches the stream and is deposited. Following drain-blocking, excess FPOM in streambed sediments is likely to be removed during high flows faster than it is re-supplied (e.g. Bilby & Likens 1979), leading to significantly lower concentrations over time. Further study is necessary to document the rapidity of this likely recovery. However, we recognise that our samples were taken mainly during baseflow conditions and further sampling is required across a range of flows to properly characterise the suspended sediment and FPOM within the fluvial system. Nevertheless, peat streams spend the majority of time at low flow (Holden & Burt 2003) and so baseflow conditions are an important component of stream habitat.

Peatland management effects on stream macroinvertebrate communities

The results supported the second hypothesis of significant differences in stream macroinvertebrate communities between intact, drain-blocked and artificially drained catchments. The third hypothesis, that macroinvertebrate communities in intact and drain-blocked sites would be similar reflecting successful catchment-scale restoration, was also supported by the results. Taxonomic richness was higher in intact and drain-blocked sites compared with drained sites, and higher SSC in the latter could have led to greater deposition on benthic habitats and reduced macroinvertebrate diversity. These inferences are supported from work by Vuori & Joensuu (1996) and Vuori et al. (1998), who investigated forestry drainage in Finland and observed reduced species richness linked to suspended and deposited sediment. Deposited fine sediments can alter substrate composition and suitability (Richards & Bacon 1994), affect respiration by deposition on respiratory surfaces/appendages/nets (Lemly 1982), prevent feeding by grazers (Aldridge, Payne & Miller 1987) and reduce the nutritional value and abundance of periphyton (Vuori et al. 1998). Additionally, reduced diversity and richness could be related to higher levels of benthic FPOM. The usual consequence of increased FPOM is an increase in a few detritivorous species with a decrease in overall richness and diversity (Moss 1973; Perry & Sheldon 1986) because excess FPOM can alter the physical environment in a manner similar to fine inorganic sediment.

The low relative abundances of Plecoptera and Trichoptera observed at the drain-blocked sites could be due to prolonged effects of stream ecosystem stressors linked to earlier artificial drainage. Historic land-use can continue to affect stream diversity and communities over decades, thus long-term studies are often required to quantify recovery (e.g. Harding et al. 1998; Foster et al. 2003). In this study, stream biota were collected at sites where drain-blocking had been implemented between three and 11 years prior to sampling, hence exact recovery times/rates were not measured. Nonetheless, low relative abundance of Plecoptera and Trichoptera in drain-blocked sites suggests earlier drainage impacts may continue. In contrast, relative abundance of Ephemeroptera was similar in drain-blocked sites to intact streams. Ephemeroptera are often among the first macroinvertebrates to recolonise stream habitats following disturbance, because adults are relatively strong fliers and the group has a propensity for drift as nymphs from any undisturbed upstream populations (Williams 1980). Recolonisation by drift seems unlikely in these artificially drained peatland streams because drains covered the entire catchment and so all parts of the stream network would be likely to have been affected. Further research investigating pre- and post-blocking would help establish the role of dispersal constraints in the temporal sequence of drain-blocked peatland stream recolonisation.

The ordination analysis indicated that several macroinvertebrate species found in the intact sites were also documented in drain-blocked sites (e.g. R. semicolorata, B. rhodani, E. dispar, P. microcephala, I. grammatica, H. pellucidula and Hydroptila spp.) but not in artificially drained sites. These taxonomic differences were associated with higher SSC and benthic FPOM. Most species within the families Baetidae and Heptageniidae are algal scrapers/grazers, and so their feeding could be quickly impaired on sediment smothered periphyton (Larsen & Ormerod 2010). Hydropsychiids (Trichoptera) are sensitive to increases in sediment loads as their retreats and nets can become embedded under excess particles, thus reducing oxygen levels in interstitial spaces (Runde & Hellenthal 2000). Nets can also clog, rip or be buried, resulting in decreased food acquisition (Strand & Merritt 1997), interference with respiration (Lemly 1982) and increased energy expenditure because of net cleaning (Runde & Hellenthal 2000). Furthermore, many predatory Plecoptera such as Perlidae and Perlodidae are normally found living in coarse sediments. Fine sediment can limit oxygen availability by reducing flow velocities in clogged interstices, reduce interstitial water exchange and constrict the movement of these invertebrates in the substrata (e.g. Beauger et al. 2006; Bo et al. 2007).

Higher abundance of E. danica, Chironomidae and Simuliidae was recorded in drained peatland streams compared with intact and drain-blocked sites. Previous work by Beisel, Usseglio-Polatera & Moreteau (2000) in north-east France also found higher abundance of E. danica in habitats with excess fine sediment, while Vuori & Joensuu (1996) found Finnish forest drainage encouraged increased Chironomidae and Simuliidae abundance. Simuliidae larvae filter FPOM effectively from the stream water column (e.g. Vuori & Joensuu 1996; Ciborowski, Craig & Fry 1997), which could explain their increased abundance in streams from drained peatland catchments characterised by higher suspended and deposited organic particulates.

The anosim results further supported the finding of significant community and taxonomic differences between artificially drained catchment streams and intact and drain-blocked streams. The analysis based on Jaccard’s coefficient of similarity indicated that taxa presence–absence was similar in drain-blocked and intact streams, whereas there were significant differences between these peatland management types in the analysis based upon BC distances. Drain-blocked sites have similar taxonomic composition but there are differences in relative abundance among constituent taxa. This could indicate that drain-blocked streams have not completely recovered to an ‘intact’ state or it may represent the development of an alternative endpoint (e.g. Bradshaw 1996; Ormerod 2003) in restored peatland streams.

Catchment-scale restoration effects on stream ecosystems

In many regions of the world, stream ecosystem services and biodiversity have been compromised because of catchment degradation (Harding et al. 1998; Allan 2004). In response, catchment-scale rehabilitation programmes have become more common (see Hillman & Brierley 2005) but there is growing recognition that many schemes lack adequate pre- and post-restoration monitoring, a problem similar to that for reach-scale river rehabilitation schemes (e.g. Bernhardt et al. 2005; Palmer et al. 2005; Woolsey et al. 2007). This study of peatland catchment remediation and stream ecosystem response is one of only a few to consider stream ecosystem responses to catchment-scale restoration. Stream ecosystems of catchments, where there was recent drain-blocking, appeared to have improved water quality, thus sustaining broadly similar macroinvertebrate communities to those in catchments with no peat drainage. These findings illustrate that such intervention may promote positive effects for in-stream biodiversity. However, this progress is undoubtedly being missed by high profile and costly monitoring schemes, such as those tracking attempts to improve ecological status under the EU Water Framework Directive (WFD), because catchments <10 km2 are rarely appraised (Logan & Furse 2002). More detailed consideration of small headwater systems may be informative with regard to improving estimates of the number of watercourses in the different WFD ecological status classes.

Conservation of stream biodiversity has received increasing recent attention as human modification and disturbance of ecosystems increase (Harding et al. 1998; Sponseller, Benfield & Valett 2008). Although rapid recovery of biotic communities following short-term catastrophic disturbances (e.g. pulse disturbances) can often happen as a result of immigration from nearby unimpacted streams (e.g. Doeg & Koehn 1994), impacts of sustained catchment-scale disturbances may profoundly affect all streams in a catchment, eliminating local refugia and meaning that recovery following any subsequent restoration may take a long time because of dispersal constraints (Harding et al. 1998). Although artificial drains studied here have only been blocked for a short period of time (3–11 years), recolonisation of streams appears to have been relatively rapid, perhaps as a result of the close proximity to intact peatland streams (cf. Mackay 1992). To ensure the quickest possible recovery of peatland stream ecosystems following drain-blocking, land managers should consider proximity to potential sources of recolonisers when planning future restoration schemes. Priority locations for drain-blocking are typically those moorlands producing high water colour or with rapid rates of erosion. From a biological viewpoint, recovery may be aided by prioritising the blocking of drains on moorlands that are adjacent to intact moorland.

It is acknowledged that this study examined the spatiotemporal interactions between physiochemical habitat variables and macroinvertebrate communities over only a short period of time using a space-for-time approach. This design was necessitated by the lack of any pre- or post-restoration stream ecosystem data for drain-blocked catchments, a common problem afflicting river rehabilitation schemes (Bernhardt et al. 2005). To measure the efficacy of drain-blocking and in-stream ecological recovery, detailed pre- and post-blocking stream ecosystem monitoring is sorely needed. The importance of, and need for, long-term monitoring to investigate impacts of past river system alterations cannot be understated (Palmer et al. 2005); such information would help deduce disturbance level, post-restoration recovery times and re-assembly mechanisms of biotic communities following drain-blocking. To achieve such an aim, there is a need to improve knowledge exchange between upland stakeholders/government agencies managers and freshwater scientists (e.g. Brown et al. 2010b), particularly when planning drain-blocking schemes. Overcoming these issues will be necessary to establish true baselines against which to judge the response of stream ecosystems to future peatland restoration.


This research was funded by a NERC studentship NER/S/A/2006/14151 with CASE support from Yorkshire Water, and additional funding from the North Pennines AONB Peatscapes project (ED1113347) and Natural England (SAE03-02-051). Steve Ormerod and an anonymous reviewer provided insightful comments on the manuscript.