FORESTRY & SURFACE WATER ACIDIFICATION (FORWATER) 2008 MARY KELLY-QUINNI, R OBERT C RUIKSHANKSI, JAMES JOHNSONI , RONAN MATSONI, JAN-R OBERT BAARSI & MICHAEL BRUEN2, 1 –School of Biology and Environmental Science 2 - School Of Architecture, Landscape & Civil Engineering TABLE OF CONTENTS 1. 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 3 3.1 3.2 3.3 4 5 6 2 SUMMARY INTRODUCTION 4 The source of the Pressure Pathway Susceptibility Receptor Sensitivity 5 6 6 SITE SELECTION & METHODOLOGY 8 SITE SELECTION HYDROCHEMISTRY MACROINVERTEBRATES FISH GEOGRAPHICAL IMAGING & SITE CHARACTERISATION STATISTICAL ANALYSES QUALITY CONTROL 8 11 13 14 17 18 20 RESULTS 21 HYDROCHEMISTRY Conductivity PH Alkalinity, Sodium Dominance Index and Aluminium Sources of Acidity Comparison of Source Streams – Upstream and Downstream of Forestry MACROINVERTEBRATES Community Composition Selection of Macroinvertebrates Metrics Taxon Richness Ephemeropteran Richness Ephemeropteran Abundance EPT Metric BMWP & ASTP Metrics Community Diversity and Species Evenness Clustering of Biological, Chemical and Physical Metrics Multivariate Analyses Evaluation of the Degree of Biological Impairment Comparison of Source and Downstream Communities Longitudinal Variation in Macroinvetebrate Metrics Further Evaluation of Potential Longitudinal Patterns in Macroinvertebrate Recovery from Forest Effects FISH 21 21 22 28 32 34 DISCUSSION REFERENCES APPENDIX A 37 37 38 40 41 44 46 48 50 50 53 55 56 57 58 62 67 71 76 1 SUMMARY This project set-out to investigate the presence and extent of acidification associated with coniferous forestry in Ireland and to assess the risk of impact with respect to different geological settings. In the selection of forested sites it was aimed to represent a combination of the risk factors in terms of catchment forest cover and acid-sensitive geology that were perceived to have the greatest potential for acidification. This study was designed to allow comparisons of the hydrochemical and ecological quality of two groups of sites, forested and non-forested, control or reference sites in two geological settings (igneous/metamorphic and sedimentary) with four dominant soil types (peat, podzolic/lithosolic, poorly drained gleys and well drained mineral soil). The 239 control and forested sites were categorised to reflect a gradient in catchment forest cover (control (<5% forest cover) – 73 sites, and three coniferous forestry bands (5-25, 25-50 and >50% coniferous cover with 27, 41 and 98 sites, respectively). Water samples were collected from all sites on three separate dates and covered a range of flow conditions. Macroinvertebrate samples were collected in spring 2007 and electrofishing was carried out at 19 paired sites in summer 2007. The pH results analyses suggested that most of the streams were episodically acidic with a small group more likely to be circum-neutral. Overall, the pH results indicated increased acidity at some sites associated with forestry on peat and podzolic/lithosoilic soils on both igneous/metamorphic and sedimentary geology and to some extent on poorly drained gleys. Furthermore, the frequency of low pH readings was substantially higher among some groups of forested sites than the control sites. Certainly the minimum pH for both peat and podzolic/lithosolic sites on igneous/metamorphic geology began to fall below the lower limit of the control sites when forest cover exceeded values in the region of 25-30%. The same applied to peat sites draining sedimentary geology. Sites on podzolic/lithosolic soils on sedimentary geology did not have minimum pH values below the lower limit of the control sites until forest cover exceeded 60%. A similar threshold might be applied to sites on poorly drained gleys but the level of replication was too low for this decision. The presence of forestry tended to depress site pH and alkalinity. Calculations suggested that dilution makes a variable contribution to loss of alkalinity and in many cases the forested sites showed a slightly higher % value. Anion titration was detected in all events examined. The principal contributors were organic acids and excess sulphate, particularly in the east. 2 Overall, the biological data largely mirrored the trends for the acidity variables. Several macroinvertebrate metrics (taxon richness, ephemeropteran richness, abundance of baetids, EPT richness, diversity indices), which showed a strong relationship with pH, were also shown to vary significantly across the forest cover bands or to correlate with % forest cover. The analyses on the individual metrics highlighted similar % forest thresholds for risk of impact as described for the hydrochemistry. When a selection of non-correlated metrics were combined it was clear that a large proportion of sites in the >50% cover band, and a smaller number of the 25-50% band, had some degree of impairment. However, not all forested sites were impaired and further research must target these sites to better understand the mechanisms governing responses to acid impact. Finally, the length of stream impacted by forest-mediated acidification is likely to vary depending on geological and other catchment characteristics. The fish analyses was limited to 19 paired sites with similar habitat but did highlight significant differences in fish catch and density between the control and forested groups. This difference was mainly attributed to low numbers of fry (salmon and trout) in the forested streams. In terms of identified knowledge gaps we need to determine the acidification risk associated with each of the key forestry practices from site preparation to felling. More detailed spatial and temporal analyses of the chemical characteristics of waters draining sedimentary geology is required for more precise mapping of acid sensitivity. The contribution of organic acids to the acid pulses in control and forested sites on both igneous and sedimentary geology and the process contributing to their release needs to be further explored. The influence of flowpath and drainage on buffering potential and its interaction with acid inputs also requires further research. Finally, with respect to the biota it is critical to understand seasonal and longitudinal changes in the community in response to acid inputs together with the functional significance of any impairment. 3 1 INTRODUCTION The Water Framework Directive (WFD) (2000/60/EC), which came into force in December 2000, requires EU member states to implement the necessary measures to prevent deterioration in the status of all bodies of surface and groundwater and where necessary, restore all waters to good ecological and chemical status by 2015. As part of the characterization process, the first risk assessment of the anthropogenic pressures on water resources was undertaken to identify the pressures present in each river basin and the threat they pose to the chemical and ecological status of waterbodies. The resulting “National Characterisation Report for Ireland” (Anon, 2005) identified forestry (SD4) as one of the land-use activities posing a potential risk in terms of diffuse pollution. Among the pressures highlighted as arising from forestry were increased acidification from plantations in acidsensitive catchments (SD4a), sedimentation from clear fell, harvesting, new plantations, road construction and erosion on steep catchments (SD4b) and eutrophication from fertilisation on steep catchments and forest harvesting on peat soils (SD4c). The Western River Basin District was given, inter alia, the task of further characterisation of the risks from plantation forests and forest related activities on surface waters and to subsequently developing a programme of measures to address any significant risks. The present project was commissioned to address some of the knowledge gaps pertaining to the acidification risk. The keys questions addressed by this research were: 1. Is there any evidence for a forest effect on hydrochemistry and biology? 2. What is the impact on the aquatic biota? 3. What conditions pose the greatest risks The research sought to identify patterns that indicate risk and inform a ‘precautionary approach’ through the first programme of measures. The work was prefaced with a literature review which clarified the state of knowledge on forest-mediated acidification. The literature review (Johnson et al. unpublished) followed the Source–Pathway–Receptor model of the risk assessment and the key findings are summarised here. 4 The Source of the Pressure The role of forestry in the acidification of surface waters is primarily attributed to the interception of atmospheric pollutants coupled with the inability, in sensitive areas, of the substrate soils and geology to buffer the acidity (Jenkins et al., 1990; Ormerod et al., 1991). The magnitude of the pressure exerted by the scavenging effect of forests depends primarily on (a) the pollutant load and (b) the percentage of catchment forest cover. The extent of the pressure is likely to vary with tree species with some species, such as Sitka spruce (P. sitchensis), being more effective scavengers of pollutants than others. The pollutant load at a site is further influenced by emission levels, climatic conditions such as the frequency and magnitude of rainfall events, the amount of annual rainfall, prevailing wind direction and air mass circulation patterns as well as site characteristics such as elevation and aspect, tree species, stand age and structure. Few studies have attempted to clarify the relationship between the extent of catchment afforestation and surface water chemistry across a range of catchment characteristics but there is general agreement in the literature that an increase in forest cover has the potential to increase the acidification pressure. However, as yet there is no guidance on the threshold above which adverse impacts are apparent on stream chemistry or biology in acid-sensitive areas. Sea salt driven acid pulses can occur in coastal catchments. Forests capture marine ions as wet or dry deposition (Farrell, 1995; Harriman, Anderson and Miller, 1995). During storm events, high inputs of Na+ can displace other cations. The associated chloride ion is largely conservative and most of it is quickly leached. As it passes through acidic soil to associated drainage water it can be accompanied by H+ and Al3+. Other processes with potential to contribute to acidification include (a) uptake of base cations by trees and subsequent removal by harvesting, (b) oxidation and mineralization of organic matter producing organic acids and (c) alterations to site hydrology resulting in the reduced residence time of water and (d) the short-term release of nitrate following the large-scale felling of forest sites in acid-sensitive catchments. Certainly increased drying of soil and altered drainage increases the oxidation of organic matter and generates carboxylate anions, increases mineralization of organic matter and consequently potentially increases in losses of ammonium and/or nitrate as well as sulphate to drainage waters (Hornung et al., 1995). However, the contribution of these to acidification processes in acid-sensitive areas has not been adequately assessed. 5 Pathway Susceptibility The pathway susceptibility is primarily controled by catchment geology. The available literature indicates that water bodies susceptible to acidification are located in catchments dominated by slow weathering bedrock such as granite and quartzite with shallow carbonate free soils as well as areas of sandy, siliceous soils and highly weathered old leached soils (Hornung et al., 1990). In Ireland, granitic areas in the west and northwest (Allott et al., 1990; Bowman, 1991; Allott et al., 1997) as well as the east have been shown to be acidsensitive (Kelly-Quinn et al., 1996a; Kelly-Quinn, Tierney & Bracken, 1997). The potential for acidification on Old Red Sandstone is likely to be less but this is not fully established and is addressed in the current project. Catchment size and hydrology/drainage also have a bearing on the susceptibility of running waters to acidification. High drainage rates and steep topography of small upland catchments reduces the contact time for runoff with bedrock and soil and consequently the time for soils to impart buffering capacity to the runoff water (Waters and Jenkins, 1992). As a result, waters draining smaller catchments may be more acidic and have higher concentrations of inorganic monomeric aluminium reflecting a higher proportion of runoff from the acidic mineral soils in the catchment. In larger catchments, the overall residence time of water in soil system is longer and it is therefore likely to be more effectively neutralized. The Sensitivity of the Receptor The sensitivity of the receptor shows as an increases in acidity and in many cases elevated aluminium concentrations (Ormerod et al., 1991). The response of the biota to forest- mediated acidification has been well documented (examples Clenaghan et al., 1998; Harriman and Morrison, 1982; Stoner, Gee & Wade, 1984; Ormerod et al., 1991, Ormerod & Wade, 1990; Allott et al., 1997; Tierney, Kelly-Quinn & Bracken, 1998). Biological impacts associated with acidification in streams include 1) reductions in or total elimination of fish populations, 2) reductions in taxon richness and elimination of some acid-sensitive macroinvertebrate groups (particularly the Ephemeroptera) and 3) changes in the quality of primary producers (Stoner, Gee & Wade, 1984; Ormerod, Wade and Gee, 1987; Rees and Ribbens, 1995; Tierney, Kelly-Quinn & Bracken, 1998). The studies carried out in Ireland to date have highlighted some impact in areas of Wicklow (Tierney, Kelly-Quinn and Bracken, 1998) and Galway and south Mayo (Allott et al., 1997) but no acidification-related impacts of aquatic fauna were detected for sites on Old Red Sandstone in Cork (Giller et al., 1997). 6 However, the latter authors noted that some macroinvertebrate communities at medium altitude (200-300m) with medium to high levels of forestry (25 to> 50%) seemed to resemble communities at higher altitudes (>300m) than sites with low levels of forest cover at a medium altitude. The present study set out to undertake more extensive sampling of Old Red Sandstone to further assess the potential for acidification impacts. A further issue arising from the AQUAFOR and indeed more recent WATERAC projects was that the occurrence of detectable impact in rivers (as evidenced by loss of macroinvertebrate taxa and salmonids) appeared to be rather patchy. Therefore, the current study set out to target good numbers of sites with a combination of perceived high risk factors, so that a better measure of the extent of impact could be achieved. 7 2. SITE SELECTION AND METHODOLOGY 2.1 SITE SELECTION This study was initially designed to allow comparisons of the hydrochemical and ecological quality of two groups of sites, forested and non-forested, control or reference sites. This approach was adopted following consultation with the Forest and Water National Steering Committee members (EPA, Cóillte, Irish Forest Service, Marine Institute, Central Fisheries Board and National Parks and Wildlife) and is the current approach required by the WFD and widely applied in freshwater studies. The forested sites were to represent a combination of the risk factors perceived to have the greatest potential to facilitate acidification impacts. Percentage forest cover in the catchment and geographical location were considered to influence the magnitude of the acidification pressure. Factors influencing pathway susceptibility were geology and soil type. A total of 239 sites (Appendix A), both control and forested, were therefore selected to provide wide geographical coverage within acid-sensitive geologies (igneous/metamorphic geology and Old Red Sandstone) and to encompass combinations of geology and soil type (peaty and mineral). Four categories of soil were targeted, well drained mineral, poorly drained gleys, podzolic-lithosols and peats [categories followed consultation with Cóillte, Irish Forest Service and WRBD]. The forested sites were selected to have >25% catchment forest cover, the majority of which was closed canopy representing mature forests in the forest cycle. Catchment is here defined and applied throughout this study as the drainage basin to the study site, alternatively referred to as site watershed. Control sites were chosen within each of the regions where the forested sites were located. To ensure comparability, where possible control and forested sites were within the same larger river catchment, although adjacent catchments were selected in some instances. However, the geographic spread of the control and forested catchments were similar. Control sites were initially chosen to represent catchments with no forestry. However, during the course of the study access to updated forest inventory information revealed variable amounts of forest in some of the control catchments. Consequently, all 239 control and forested sites were re-categorised to reflect the gradient in forest cover (control (<5% forest cover) – 73 sites, and three coniferous forestry bands (5-25, 25-50 and >50% coniferous cover) with 27, 41 and 98 sites, respectively). The numbers of sites in each forestry/soil/forest cover category are given in Table 1 and their location is indicated in Fig. 1. Broadleaf forestry represented a small percentage of the total forest cover and was not considered in the 8 analyses. Therefore, the forest cover values referred to throughout this report represent coniferous plantation. There was no significant land-use pressure in the control catchments apart from some rough grazing, although historical influences could not be ruled out entirely. Table 1: Numbers of site sampled in each geology/soil category/forest cover category Geology/Soils Igneous/Metamorphic Forest Cover <5% 5-25% 25-50% >50% Totals Peat 22 9 6 17 54 Podzolic/Lithosolic 11 4 9 8 32 Peat 20 4 11 30 65 Podzolic/Lithosolic 13 2 2 22 39 Gleys 6 3 7 11 27 Well Drained Mineral 2 5 6 9 22 Totals 74 27 42 97 239 Sedimentary In each region the sites were selected on first to third order streams. A number of additional sites were located further downstream to examine longitudinal changes in hydrochemistry and aquatic biota. Every effort was made to control for slope, elevation and catchment size assessed using channel length and stream complexity as indicated on OSI maps (1:50000) and by restricting the majority of the sites to sub-catchments. Access to the sites was facilitated by Coillte/Irish Forest Service, many of which were remote with the only access by foot. Each sites was represented by a 50 (macroinvertebrate sampling) to100 (fish sampling) metre stretch. Physical descriptions for each of the sites were derived from GIS and from onsite measurements of stream width (four measurements), depth (four measurements taken in randomly selected pools in the stretch), substrate composition, mesohabitat cover and flow condition (low, elevated and flood). Substrate was estimated as the percentage cover of bedrock, boulder (26-200cm), cobble (6-26cm), gravel (0.4-6cm), sand (0.06-0.2cm) and silt 9 (<0.06cm). Mesohabitats were assigned to three categories, including riffles, glides and pools. • Control Sites • 5-25% Forested • 25-50% Forested • >50% Forested Fig. 1: Distribution of hydrochemical and macroinvertebrate sampling sites (colour coded for forest-cover categories). 10 2.2 HYDROCHEMISTRY Water samples were collected from all sites in clean one-litre and 250ml polypropylene bottles (pH). Readings of conductivity (µS/cm), and oxygen (% saturation and mg/l O2) were taken on site using automatic field meters and probes. All samples were sent to the Aquatic Services Unit at the Environmental Research Institute (ERI, UCC) for analysis within 24hours of collection. A full suite of hydrochemical analyses were carried out using the methodologies outlined in Table 2. Three rounds of water sampling were undertaken, 2/5/076/6/07, 13/11/07-21/12/07 and 18/3/08-16/4/08. The aim was to sample each site at variable flow conditions, from low flow to flood. samples for all sites. It was however not possible to obtain flood Furthermore, it was often difficult to ascertain the stage in the hydrograph represented on any one date and therefore flow condition was simply recorded as low, elevated or flood. Additional samples were taken upstream and downstream of the forestry block on selected source streams. Samples were collected at comparable locations on control streams at similar distances from the source (as indicated on OSI maps). Sources of acidity and those potentially responsible for any observed differences between forested and control sites were evaluated by examination of the results for sulphate, nitrate, chloride and organic carbon. The potential loss of alkalinity during elevated flow due to dilution by precipitation was assessed using the following formula applied by Kowalik et al. (2007): Alkalinity Dilution = ((( BC low - BC high)/ BC low) Alk. low) ____________________ ______________________________________________________ (Alk. low - Alk. high) BC=Base Cations, Alk.=Alkalinity All concentrations are entered in units of µeq/l. High percentage values close to 100% indicate that dilution is strongly affecting buffering. Lower values indicate reduced likelihood of dilution and possible titration by an acid anion. 11 Titration Ratio Loss of alkalinity due to titration by an acidic anion is evidenced by changes in the following ratio. Titration Ratio = Alkalinity / BC This can be confirmed with the results from the titration ratio equation. The proportional contribution of acid anions to any titration processes was calculated as Anion/ Acid Anions (Kahl et al., 1992). Losses in ANC due to sea salt effects were evaluated from changes in the concentrations of Cl- and Na+ between low and high flow as applied by Evans et al (2008). Table 2: Methods applied in the chemical analyses Parameter Method Unit pH Conductivity Alkalinity Total Hardness Colour WTW pH330i pH meter WTW LF330 Conductivity meter @ 25°C Gran Titration ETDA Titration Colorimetric method using platinum/cobalt solution as colour standard SHIMADZU TOC-VCPH TOC analyzer Automated Molybdate method using Lachat™ Quikchem FIA Manual molybdate method after sample digestion Automated salicylate method using Lachat™ Quikchem FIA Automated colourimetric method using Lachat™ Quikchem FIA after cadmium reduction Subtraction nitrite from TON Manual colourimetric method Graphite furnace AAS Graphite furnace AAS after Amberlite™ Resin fractionation Automated IC method using Lachat™ Quikchem IC Automated IC method using Lachat™ Quikchem IC Automated IC method using Lachat™ Quikchem IC Automated IC method using Lachat™ Quikchem IC Automated IC method using Lachat™ Quikchem IC Automated IC method using Lachat™ Quikchem IC Gravimetric method after filtering through GF/C filter paper and dried at 104°C Manual colourimetric method Calculated µS/cm @ 25ºC mg/l CaCO3 mg/l CaCO3 Hazen Units Dissolved Total Organic Carbon Soluble Reactive Phosphorus Total Phosphorus Ammonia Total Organic Nitrogen Nitrate Nitrite Total Monomeric Aluminium Inorganic Aluminium Calcium Magnesium Potassium Sodium Chloride Sulphate Suspended Solids Silicate Sodium Dominance Index (SDI) mg/ l DTOC mg/l SRP mg/l TP mg/l Ammonia mg/l TON mg/l Nitrate mg/l Nitrite µg/l Aluminium µg/l Aluminium mg/l Ca2+ mg/l Mg2+ mg/l K+ mg/l Na+ mg/l Clmg/l SO42+ mg/l SS mg/l Si % 12 2.3 MACROINVERTEBRATES Benthic macroinvertebrate samples were collected over a six week period in 2007 from the beginning of April until the second week in May at the 239 sites. Additional samples were collected on source streams above and below forest blocks and at similar points on comparable streams. A multi-habitat sampling approach was employed involving kick samples of 1-minute duration taken using a standard pond net (mesh – 1mm). The time spent sampling each mesohabitat (riffle, pool glide) was proportional to its percentage representation in the study site ( Wright, 1995). Habitats contributing less than 5% of the stable habitat in the reach were not sampled (Barbour et al., 1997). An additional one minute was spent carrying out hand searches for attached invertebrates. Sampling was initiated downstream of the reach and proceeded upstream. To avoid the confounding effects of shading the forested sites were downstream of the forest within open, un-shaded reaches at least 20m downstream of the forest block. Six samples were collected at each site and preserved using 70% alcohol (IMS). These were sorted in the laboratory and the macroinvertebrates were removed and identified to the lowest taxonomic unit possible using FBA keys (Table 3). Identified samples were stored in 70% alcohol (IMS). Table 3: Level of identification for macroinvertebrate groups. Taxon Plecoptera Ephemeroptera Trichoptera Coleoptera Chironomidae Simuliidae Other Diptera Odonata Hemiptera Mollusca Hirudinea Oligochaeta Level of Taxonomic Identification Species Species Genus/species Genus/species Subfamily Genus/ Family/genus/species Genus/species Genus/species Species Species Order 13 2.4 FISH Streams were selected on a paired catchment basis (one non-forested catchment, one forested catchment) to represent similar physical characteristics including catchment area, elevation and slope. With few exceptions paired streams were selected on the same main channel, in close proximity so that both streams had the same source fish population within the main channel. Fig. 2 illustrates this approach. In total, 38 sites were fished (19 non-forested and 19 forested sites) (Table 4, Fig. 3). Each site was fished using backpack electrofishing equipment (Safari Research 550D backpack model). Site habitat characteristics including numbers of riffles, glides and pools, stream width (four measurements) and pool depth (four measurements) were recorded on site, along with on-site measurements of oxygen and conductivity. The time taken to fish each site was recorded to compute fishing effort so that comparisons could be made between forested and non-forested sites and corrected if necessary for differences in effort. A single-pass approach was adopted for the electrofishing sampling. Several investigations have evaluated the accuracy and usefulness of single-pass electrofishing to estimate abundance or relative abundance of salmonids in streams. These studies have indicated that there is a significant relationship between number of fish caught in the first pass and the total population size estimated from three or more passes (Hayes & Baird, 1994; Jones and Stockwell, 1995; Kruse, Hubert & Rahel, 1998; Mitro and Zale, 2000; Arnason, Antonsson & Einarsson, 2005; Bertrand, Gido & Guy, 2006) and it is therefore a sensitive method for detecting differences in relative abundance. The FAME protocol recommends at least 10-20 times the wetted width be fished (Economou et al., 2002). As the majority of the selected sites were approximately 2m wide, the 100m stretch fished in the present study was more than adequate to satisfied this condition. To avoid the problem of shading sampling in the forested catchments took place outside of the forest, usually immediately downstream (circa 20m). All species encountered were captured and identified. The salmonids were measured (fork length) and weighed. Scales were removed from a representative sample and retained for age analysis. After capture all fish were held in keep-nets to ensure their complete recovery before being returned to the river. 14 Fig. 2: Illustration of the paired site selection approach in the King’s River catchment, Co. Wicklow. Site DWW2 was paired with DWW15 and DWW5 was paired with DWW15. • Control Sites • Forested Sites n = 19 paired sites Fig. 3: Location of electrofishing site pairs 15 Table 4: Location of paired fishing sites Main System River Site Code Control/Forest Easting Northing Main System River Site Code Control/Forest Easting Northing Kings Ballinagee BALLIN1 DWW1 Control 304462 204045 Kings Annalecka ANNA1 DWW13 25-50% 306426 202755 Kings Glencreemore GLEEN1 DWW5 Control 302788 200283 Kings Glashaboy GLASH1 DWW26 25-50% 306535 201611 Kings Knickeen KNICK1 DWW6 Control 299726 195072 Oilitigh Oilitigh OILI1 DWW17a 25-50% 299286 196067 Cloghoge Cloghoge CLOG1 DWW8 Control 312761 207455 Inchavore Inchavore INCH1 DWW19 25-50% 311004 206075 Srahmore Srahmore SRAH1a DM11a Control 096560 305240 Srahmore Srahmore SRAH2 DM23 >50% 095227 306980 Srahmore Glenamong GAMON4 DM8 Control 093918 304178 Srahmore Glenamong GAMON2 DM10 25-50% 092809 303819 Srahmore Glenamong GAMON5 DM9 Control 094080 303977 Srahmore Glenamong GAMON6 DM9a 25-50% 094019 303524 Owengarve Callowswallagh CALLOW1 DM22 Control 093465 298191 Owengarve Glendahurk CALLOW1 DM3 25-50% 091103 300931 Glenamoy Glenamoy GMOY3 DM19 Control 093801 332628 Glenamoy Glenamoy GMOY4 DM20/M8 25-50% 095106 335854 Glenamong Fiddaunatoreen FREEN1 DM6 Control 095099 301909 Glenamong Glenamong GAMON1 DM7 25-50% 094472 302777 Glenumerra/Glendavock Glenumerra GLENU1 DM24 Control 085739 267708 Glenumerra/Glendavock Glenumerra GLENU2 DM25 25-50% 089975 267651 Owenree Owenree OREE1 DG11 Control 101586 246870 Owenwee Owenwee OWEE2 DG22 >50% 103160 245498 Maumwee L. Inflow Maumwee MAUM1 DG24 Control 097255 248472 Owenwee Owenwee OWEE3 DG23 >50% 103292 245896 Owenriff Owenriff ORIFF1a DG33 Control 105151 242453 Owenriff Glengawbeg GBEG1a DG27 25-50% 106686 240525 Owenboliska Owenboliska OLISKA3 DG7 Control 111162 234916 Owenboliska Owenboliska OLISKA1 DG13 >50% 114582 235506 Owenboliska Owenboliska OLISKA3 DG7 Control 111162 234916 Owenboliska Owenboliska OLISKA6 DG15 >50% 108519 232725 Glenleheen Glenleheen GHEEN1 DD18 Control 190732 404350 Gweebara Gweebara GBAR3 DD11 25-50% 185974 402744 Deele Deele DEEL1 DD2 Control 211261 403208 Deele Deele DEEL2 DD19 5-25% 208901 405512 Elatagh Elatagh ELAT3 DD16 Control 202238 405218 Elatagh Elatagh ELATA5 DD14a >50% 204295 403973 16 2.5 GEOGRAPHICAL IMAGING AND SITE CHARACTERISATION Site co-ordinate readings were recorded from a GPS handset at all sites. Arcview™ 3.3 was used to plot site distributions and delineated catchment basins for all sites. The Geoprocessor™ extension program allowed the calculation of various catchment characteristics including geology, soil (and sub-soil) coverage composition, percentage catchment forestry, catchment land-use and catchment area. Catchment delineation was performed by Compass Informatics™. Catchment characteristics were derived using this delineation, the EPA geology database and the most up-to date FIPS database. FIPS 07 was under development during this process and represented the best available data in March 2008. The key information extracted from FIPS 07 included species composition, forest cover, and felling history. The age of the tress was not available. Daniel McInerney, SBES, UCD, undertook in part the geo-processing as outlined. As mentioned previously two broad geology categories were assigned on the basis of the dominant rock grouping, igneous/metamorphic or sedimentary. The igneous/metamorphic category was predominantly composed of granite, but also included mica schist, quartzite, Diorite, Gabbro and Dolerite while the sedimentary group included mainly Old Red Sandstone, Coal Measures and some Carbonate geology. Soils were assigned to four groups. This categorisation followed an agreement on interpretation of Irish Forest Soils and Teagasc Soil Categories. The initial catchment soil mapping was based on the Teagasc Soils Map database distributed by the EPA. However, the accuracy of this was unclear. Subsequently, the Forest Service undertook to cross reference the IFS site specific data from 1,732 sites (re-categorised as per the four soil categories) with that held on the Teagasc National Soil cover data. The highest (73%) agreement was within peats, followed by well drained minerals/gleys (62%). Lower percentage agreement was obtained for poorly drained gleys (46%), podzolic/lithosolics (37%) and peaty gleys (21%). Additional soil surveying in a selection of the study catchments was deemed necessary to validate the soil designation. Priority was given to catchments where sites showed variable responses in the aquatic biota to the presence of forestry. A total of 106 catchments were targeted for soil surveying. Using a series of systematic grids (250, 300 and 500m), 1,196 sample points were created and sampled. The attributes data captured was based on the NFI methodology. The results again highlighted 17 the highest confidence in the designation of peats and well drained mineral soils (report by John Redmond to WRBD). When the catchment soil allocations were finalised the dominant soil type was used in all subsequent analyses as it was considered to have the greatest potential for influencing the stream hydrochemistry. 2.6 STATISTICAL ANALYSES Extensive databases for biological and physico-chemical parameters were generated in Excel™. Univariate and multivariate analyses were performed using SPSS™ v. 12.0.1, STATISITICA™ v. 7.1, Community Analysis Package (CAP™ v. 3.1) and Ecological Community Analysis (ECOM™ v. 2.0). The AQEM Project (ASTERICS 3.10™) program was used to generate over 40 water quality and macroinvertebrate metrics [using Europe version]. Impairment in terms of the various biological metrics and hydrochemical parameters was detected using metric values outside of two standard deviations (or 95% confidence interval) of the control site values as expressed by Resh et al. (1988). A similar approach was used by Johnson et al. (2005) to develop a clearfelling impact metric. Data from 1 st and 2nd order sites were combined following preliminary analyses which indicated no significant relationship between catchment size and taxon richness. These catchment sizes ranged from 21.4 to 661.8ha. Sites with catchment sizes greater than 700ha were excluded but were included in the analyses of longitudinal patterns. Sites with catchments less than 18ha were also excluded. The hydrochemical data were used to derive means and minimum/maximum values for each parameter. The minimum/maximum values were considered to represent the worst case scenarios and were used to test relationships with forest cover, other catchment descriptors and hydrochemical variables as well as the biological metrics. Cluster analyses was carried out on the hydrochemical and biological datasets. Clustering is the process of finding groups of objects (or data) such that those in a group are similar (or related) to one another and different from (or unrelated to) the objects in other groups. Some defined distance measure such as the Euclidean distance is often used to determine proximity of the data in a cluster. The k-means clustering algorithm (Hartigan and Wong, 1979) is one of the simplest unsupervised learning algorithms for this partitioning when the number of 18 clusters (k) is known or specified a priori. A good method will produce high quality clusters with high intra-class similarity and low inter-class similarity (see figure below). The quality of a clustering method is measured by its ability to discover some or all of the hidden patterns. The quality of a clustering result also depends on both the similarity measure (like Simpsons, Bray Curtis of Jaccards) used by the method and its implementation. In regression analysis or modelling, the clustering helps determine if there are groups of similar data that might exhibit a similar response ( which might require a specific model or set of parameters) and also if the available data do not cover or span the region of interest. When applied to response variables it clusters those with a similar pattern of responses (which may or may not have a specific physical interpretation). The k-means algorithm (Hartigan & Wong, 1979) used in the present study is one of the simplest numerical methods used to implement clustering and works as follows: 1. The number of clusters required must be chosen in advance and a significance tolerance for stopping the iterations. 2. An initial position in the data space is chosen for each cluster. These should be as far apart as possible and should cover the range of the data space as well as possible. 3. The Euclidian distance from each point in the data set to all cluster centroids is calculated and each data point is then associated with the nearest centroid. Thus a cluster of data points is associated with each centriod. 4. The actual centroid of the points associated with each cluster is calculated and replaces the previous centriod of that cluster. 19 5. Steps 3 and 4 are repeated until the change in the centroid positions is less than some specified tolerance. 6. The solution is the set of clusters when the tolerance is satisfied. Clustering was carried out separately using metrics to describe the chemical signature (hydrogen (max), alkalinity (max) monomeric aluminium (max) cations (min), DOC (max) and organic acids), physical descriptors that may affect the magnitude of the pressure and run-off potential (% coniferous forest, catchment area, slope & area) and selected biological response metrics (taxon richness & ephemeroptera richness/abundance). 2.7 QUALITY CONTROL Quality control procedures were employed for macroinvertebrate sorting and identification. Previously sorted samples were re-checked for missed specimens to check for % accuracy. At most 10 individuals were recovered representing well below 3% of the total macroinvertebrates initially sorted from the samples. A number of specimens from each identified taxon were checked by an independent taxonomist, Dr Gustavo Becerra Jurado. Quality control of data inputting to the physico-chemical and biological databases was also undertaken. The macroinvertebrate databases from UCD and UCC were reviewed for inconsistencies in taxonomy. 20 3. RESULTS 3.1 HYDROCHEMISTRY Conductivity The river sites examined were typically low conductivity waters. In fact, over 80% of the sites recorded maximum conductivity reading below 150 S/cm (Fig. 4). The highest value recorded was 295 S/cm in a tributary of the River Loobagh which drains sedimentary geology. Overall, there was no significant differences in the mean readings across geology and soil site groupings (Fig. 5) although the sedimentary sites on well drained mineral soils had marginally higher values. In terms of a forest effect significant differences were detected across the forest cover bands only in sites draining igneous/metamorphic geology and peaty soils (Kruskall Wallis-H(3,65)=13.328, P=0.004). Counts 100 100% 80 80% 60 60% 40 40% 20 20% 0 0% 0-50 50-100 100-150 150-200 200-250 250-300 Conductivity Range ( S/cm) Fig. 4: Frequency distribution and accumulative percentage of conductivity readings from all dates and sites. The differences between low and high flow readings were highly variable and were typically less than 100 S/cm but one sedimentary site on peat recorded a difference of 205 S/cm, the highest value was associated with low flow. In other sites the highest conductivity readings were associated with high flow. 21 Granite/Metamorphic Sedimentary Mean Conductivity ( S/cm) 160 140 120 100 80 60 40 20 0 Peat Podzolic/Lithosolic Peat Podzolic/Lithosolic Gleys Well Drained Mineral Fig. 5: Variation in mean conductivity reading across geology and soil groups. pH The pH readings for the various sampling dates were highly variable (Figs. 6 and 7). Most of the sites appeared to be episodically acidic. Some, especially those draining well drained mineral soils, were more circum-neutral in character. Much of the variation within sites could be related to differences in flow conditions, the low pH values were generally associated with elevated flow. However, as previously mentioned it was difficult to know the stage of the hydrograph represented and full flood conditions were not encountered at many of the sites. It is therefore possible that the highest acidity levels were not captured by the sampling programme. 8.5 8 7.5 Peat Podzolic/Lithosolic 7 pH 6.5 6 5.5 5 4.5 4 3.5 DG20 DWW16 DM12 M3 DWW20 DWW22 DWW15 DWW21 DWW17 DWW26 DM3 M2 G16 DM10 DM11 G15 DM7 DD7 DWW10 DM5 DWW12 DM8 DM2 DM17 DM1 DG1 DG3 DG24 DWW7 DM4 DD10 DM6 G7 DG21 DG22 DG12 DD9 DG16 DG19 DG18 DD13 DG15 G9 DG14 DD5 G8 DG13 DD14 DG17 DG23 DD16 DM15 DWW19 DWW13 DD4 DD15 M8 G11 DG30 M11 DD8 DG6 DG7 DD6 DWW23 G6 DG8 DG31 DG29 DG25 G18 M6 DD2 M7 DG9 G5 DD3 DG28 G4 DM16 DWW6 DG11 DWW4 DWW2 DD11 DWW8 DWW9 Fig. 6: Distribution of pH readings from sites draining peat and podzolic/lithosolic soils on granite/metamorphic geology. Sites within each soil group are ordered according to increasing forest cover as indicated by the green arrow. The various colours represents the three sampling dates. 22 8 7.5 7 6.5 pH 6 5.5 5 Gleys Peat Well-drained Mineral 4.5 Podzolic/Lithosolic 4 3.5 Fig. 7: Distribution of pH readings from sites draining various soils on sedimentary geology. Sites within each soil group are ordered according to increasing forest cover as indicated by the green arrow. The various colours represents the three sampling dates. Despite the uncertainty relating to flow conditions it should be noted that a good number of control and forested sites were sampled in any one area under the same weather/flow conditions. The randomised sampling should permit assessment of pH changes in relation to forest cover. The data were initially analysed across the forestry bands. Minimum pH was selected for analysis of the worst case condition. On igneous/metamorphic geology minimum pH was significantly different across the forest bands (Soil Type: Peats Minimum pH: Kruskall-Wallis-H(3,55) = 15.8426, p = 0.0012; Soil Type: Podzolic Lithosolic Minimum pH: Kruskall-Wallis -H(3,31) = 9.228, p = 0.0264 – Fig. 7). Some of the lowest values were associated with high forest cover, particularly on peats. The results were similar when maximum hydrogen ion concentrations were analysed. 8.0 7.5 7.0 6.5 6.0 5.5 4.5 SoilTypes: Peats Forested(>50%) Forested(25-50%) Forested(5-25%) Control (<5%) Forested(>50%) Forested(25-50%) 3.5 Forested(5-25%) 4.0 Control (<5%) Minimum pH 5.0 Median 25%-75% Non-Outlier Range Outliers Extremes SoilTypes: Podzolic Lithosolic Fig. 8: Box plots of minimum pH values for sites within four forest cover bands draining granite/metamorphic catchments with different dominant soil types. 23 A similar pattern was recorded on sedimentary geology but none of the differences was statistically significant. Although the pH of sites on well drained mineral soils decreased Forested(>50%) Forested(25-50%) Forested(5-25%) Control (<5%) Forested(>50%) Forested(25-50%) Poorly Drained Gleys Forested(>50%) Forested(25-50%) Control (<5%) Forested(>50%) Forested(25-50%) Podzolic Lithosolic Forested(5-25%) Forested(5-25%) Forested(5-25%) Peats 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 Control (<5%) Minimum pH 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 Control (<5%) across the forestry bands the streams remained circum-neutral (Fig. 9). Median 25%-75% Non-Outlier Range Outliers Extremes Well Drained Mineral So Fig. 9: Box plots of minimum pH values for sites within four forest cover bands draining sedimentary geology with different dominant soil types. The relationship between minimum pH and % forest cover was also examined for both geological settings. On igneous/metamorphic geology streams draining peat showed a significant decrease (r = -0.6834, p = 0.0000002) in pH with increasing forest cover (Fig. 10). Although control and sites with low forest cover had some pH reading as low as the more heavily forested sites the latter sites had fewer readings in the circum-neutral range. The minimum pH for both peat and podzolic/lithosolic sites fell below the lower limit of the control sites when forest cover exceeds values in the region of 25-30%. The relationship on peat on sedimentary geology was also significant (r = -0.2515, p = 0.0505) and largely similar to that on igneous/metamorphic geology except that more readings were in the circum-neutral 24 range, except when forest cover exceeded 80% (Fig. 10). The podzolic/lithosolic sites did not show a significant correlation between minimum pH and % forest cover. However, it should be noted that the minimum pH values fell below the lower limit for the control sites when forest cover exceeded 60% (Fig. 11). 8.0 7.5 7.0 Minimum pH 6.5 6.0 5.5 5.0 4.5 4.0 3.5 -20 0 20 40 60 80 100 120 -20 0 SoilTypes: Peats 20 40 60 80 100 120 SoilTypes: Podzolic Lithosolic % Coniferous Cover Fig. 10: Relationship between minimum pH and % forest cover for sites draining granite/metamorphic geology with different dominant soil type. Minimum pH 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 -20 0 20 40 60 80 100 120 -20 0 20 Peats 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 -20 0 20 40 40 60 80 100 120 80 100 120 Podzolic Lithosolic 60 80 100 120 -20 Poorly Drained Gleys 0 20 40 60 Well Drained Mineral % Coniferous Cover Fig. 11: Relationship between minimum pH and % forest cover for sites draining sedimentary geology with different dominant soil types. To further analyse these pH data for a possible forest effect it was hypothesised that the number of pH readings below 5.0 would increase across the forest cover bands. This was 25 based on previous research in upland Wicklow streams which suggested that the duration of low pH values in some forested streams exceeded that in control moorland streams (KellyQuinn et al., 1996a). Tables 5 and 6 present the results for each geological setting. In igneous/metamorphic catchments the number of pH readings <5.0 was substantially higher than the result for the control sites in the 25-50% and >50% forestry bands for peat sites and the 25-50% band for the podzolic/lithosolic sites. Table 5: Numbers of pH readings 5 in each of the samplings rounds 1-3 and as % percentage of overall samples for sites draining granite/metamorphic geology. Soil Type % Forest Cover Band 1 2 3 Total Samples % total samples with pH 5 <5 4/22 3/22 4/22 66 17 5-25. 1/9 1/8 0/7 24 8.3 25-50 4/6 2/7 3/8 21 43 >50 10/17 4/16 12/16 49 53 <5 0/11 2/8 1/8 11 7 5-25. 0/4 0/4 0/4 16 0 25-50 6/9 5/9 4/8 26 57 >50 2/8 0/8 1/8 24 8 Peat soil Podzolic/Lithosolic The result was similar for peat sites on sedimentary geology. On podzolic/lithosolic soils pH values <5.0 were only encountered when forest cover exceeded 50%. The same applied to sites on the poorly drained gleys. No pH readings <5 were recorded in catchments dominated by well drained mineral soils (Table 6). 26 Table 6: Numbers of pH readings 5 in each of the samplings rounds 1-3 and as % percentage of overall samples for sites draining sedimentary geology. Soil Type % Forest Cover Bands 1 2 3 Total Samples % total samples with pH 5 <5 1/17 1/16 0/15 32 4 5-25. 1/4 1/4 0/4 12 17 25-50 2/10 2/10 2/9 29 21 >50 7/27 3/27 3/27 81 16 <5 0/13 0/13 0/13 39 0 5-25. 0/2 0/1 0/2 6 0 25-50 0/2 0/2 0/2 6 0 >50 0/22 3/22 4/22 66 11 <5 0/6 0/6 0/6 18 0 5-25. 0/3 0/3 0/3 9 0 25-50 0/7 0/7 0/5 19 0 >50 0/11 1/11 0/11 33 3 <5 0/2 0/4 0/2 6 0 5-25. 0/3 0/3 0/6 13 0 25-50 0/7 0/7 0/6 20 0 >50 0/9 0/9 0/9 27 0 Peat Podzolic/Lithosolic Poorly drained Gleys Well-drained Mineral 70 60 Peat Podzolic/Lithosolic % Records. 50 40 30 20 10 0 2 4 6 8 10 12 14 16 18 20 >20 Alkalinity (mg/l CaCO3) Fig. 12: Distribution of alkalinity readings from sites on predominately peat and podzolic/lithosolic soils in igneous/metamorphic catchments. 27 Alkalinity, Sodium Dominance Index and Aluminium The majority of the alkalinity readings from sites draining igneous/metamorphic geology with either peat or podzolic/lithosolic soil cover fell below 2 mg/l CaCo3 (Fig. 12). A few sites had values >20 mg/l CaCO3, these catchments were influenced by variable amounts of carbonate geology in the catchment. On the same soils in catchments dominated by sedimentary geology readings were more evenly distributed across the alkalinity range with over 20% higher than 20 mg/l CaCO3 (Fig. 13). % Records. 70 60 Peat 50 Podzolic/Lithosolic 40 30 20 10 0 2 4 6 8 10 12 14 16 18 20 >20 Alkalinity (mg/l CaCO3) Fig.13: Distribution of alkalinity readings from sites on predominately peat and podzolic/lithosolic soils on sedimentary geology. As expected, sites draining well drained mineral soil were more buffered and the majority of readings were greater than 20 mg/l CaCO3 (Fig. 14). 70 60 Gleys % Records. 50 Well Drained Mineral 40 30 20 10 0 2 4 6 8 10 12 14 16 18 20 >20 Alkalinity (mg/l CaCO3) Fig. 14: Distribution of alkalinity readings for sites on predominately peat and podzolic/lithosolic soils in igneous/metamorphic catchments. 28 In terms of minimum alkalinity the highest number of values <8mg/l CaCO3 were associated with the following geological setting; podzolic/lithosolic/igneous/metamorphic peat/igneous/metamorphic (95%), peat/sedimentary (96%), (72%), podzolic/lithosolic/sedimentary (52%). The presence of forestry tended to depress the site alkalinity as can be seen from Fig.15 which compares the distribution of alkalinity readings from control and heavily forested sites on peat and igneous geology. The effects of forestry were most obvious when minimum alkalinity was considered. Minimum alkalinity decreased significantly across the forest cover bands on peat/igneous/metamorphic (Kruskall-WallisH(3,55) = 14.6122, p = 0.0022), podzolic/lithosolic/ igneous/metamorphic (Kruskall-Wallis H(3,31) = 8.0601, p = 0.0448), but this trend was not statistically significant in the other geological settings. However, in the case of peat on granite the occurrence of negative minimum alkalinity increased across the forest cover range. Only the forested sites on podzolic/lithosolic soils recorded negative alkalinity values, some were detected on igneous/metamorphic geology when forest cover exceeded circa 25%. On sedimentary geology negative alkalinity values were recorded at some sites when forest cover exceeded % Records. 60%. It should be noted that not all forested sites exhibited this loss of buffering capacity. 50 45 40 35 30 25 20 15 10 5 0 Peat on Granite/Metamorphic - <5% Forest Cover Peat on Granite/Metamorphic - >50% Forest Cover 0 2 4 6 8 10 12 14 16 Alkalinity (mg/l CaCO3) Fig. 15: Distribution of alkalinity readings for sites on predominately peat soils in igneous/metamorphic catchments with and without forest cover. The relationship between maximum and minimum alkalinity values illustrates the level of change in buffering capacity between the two extremes. The majority of the sites that recorded zero or negative alkalinity had maximum alkalinity below 8 mg/l CaCO3 (Fig. 16a). 29 In contrast, several forested sites with maximum alkalinity up to 20 mg/l CaCO3 and higher exhibit zero or negative minimum alkalinity values (Fig. 16b). 70 60 50 2 R = 0.6422 40 30 20 . Min. Alkalinity (mg/l CaCO3) 80 10 0 -20 -10 0 20 40 60 80 100 120 140 160 Max. Alkalinity (mg/l CaCO3) (a) 80 Min. Alkalinity (mg/l CaCO3) 70 2 60 R = 0.6893 50 40 30 20 10 0 -20 -10 0 20 40 60 80 100 120 140 160 Max. Alkalinity (mg/l CaCO3) (b) Fig. 16: Relationship between maximum and minimum alkalinity for (a) control sites,(b) sites with >25% forest cover. Over 90% of the sites dominated by peat or podzolic/lithosolic soil on igneous/metamorphic geology had Sodium Dominance values >60% supporting their acid-sensitive designation. On sedimentary geology the results were more variable, with an increasing proportion of the sites falling below 50% SDI as one moved from peat, through podzolic/lithoslic soils to the gleys and well drained mineral soils. In addition, there were greater differences between maximum and minimum SDI values. 30 Total aluminium concentrations were highest at sites draining predominantly peat (Figs. 17 & 18). In the two geological settings total aluminium increased significantly with increasing forest cover. The trend for sites influenced by podzolic/lithosolic soils was only significant on sedimentary geology. 600 500 Maximum Total Aluminium µg/l) ( 400 300 200 100 0 -100 -20 0 20 40 60 80 100 120 -20 0 SoilTypes: Peats 20 40 60 80 100 120 SoilTypes: Podzolic Lithosolic % Coniferous Cover Fig. 17: Relationship between forest cover and maximum total aluminium concentrations on Maximum Total Aluminium (µg/l) granite/metamorphic geology with peat and podzolic/lithosolic soils. 800 700 600 500 400 300 200 100 0 -100 -20 0 20 40 60 80 100 120 -20 0 20 Peats 800 700 600 500 400 300 200 100 0 -100 -20 0 20 40 40 60 80 100 120 80 100 120 Podzolic Lithosolic 60 80 100 120 -20 Poorly Drained Gleys 0 20 40 60 Well Drained Mineral % Coniferous Cover Fig. 18: Relationship between forest cover and maximum total aluminium concentrations on sedimentary geology with various dominant soil types. 31 Few measurement of labile monomeric aluminium were made and most were for forested sites. Maximum aluminium concentrations ranged from 17.0 – 348 ug/l. No significant correlation with forest cover was detected which may be a factor of the sample size. Sources of Acidity Dissolved Total Organic Carbon (DTOC) As expected the variation in background DTOC concentrations, as illustrated by the control sites, reflected the organic nature of the dominant catchment soils. The highest concentrations were recorded from sites draining peat on igneous/metamorphic geology (Fig. 19) followed by peat on sedimentary geology (Fig. 20). In these two settings DTOC concentrations showed a significant increase across the forest cover bands (Peats on Igneous/metamorphic-Kruskall Wallis-H(3,55) = 19.7422, P = 0.0002; Peats on SedimentaryKruskall Wallis-H(3,61) = 12.5833, p = 0.0056). . 45 40 35 Maximum DOC (mg/l) 30 25 20 15 10 5 0 -5 Control (<5%) Forested(25-50%) Control (<5%) Forested(25-50%) Fores ted(5-25%) Forested(>50%) Forested(5-25%) Fores ted(>50%) SoilTypes: Peats Median 25%-75% Non-Outlier Range Outliers Extremes SoilTypes: Podzolic Lithosolic Fig. 19: Box plots of maximum DTOC values from sites within four forest cover bands draining igneous/metamorphic catchments with different dominant soil types. 32 60 50 40 30 20 10 Maximum DOC (mg/l) 0 -10 Control (<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Control (<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Peats Podzolic Lithosolic 60 50 40 30 20 10 0 -10 Control (<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Control (<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Poorly Drained Gleys Well Drained Mineral Median 25%-75% Non-Outlier Range Outliers Extremes Fig. 20: Box plots of maximum DTOC values from sites within four forest cover bands draining sedimentary geology with different dominant soil types. Excess or Non-Marine Sulphate Maximum non-marine sulphate differed significantly across forestry bands for sites draining peat in both geological settings (Igneous/metamorphic – Kruskall-Wallis-H(3,55) = 8.1725, p = 0.0426; Sedimentary Geology – Kruskall-Wallis-H(3,62) = 25.0755, p = 0.00001) The trend was similar for podzolic/lithosolic soils but was only significant on sedimentary geology. Interestingly on mineral soils non-marine sulphate decreased significantly across the forest-cover bands (Kruskall-Wallis-H(3,34) = 9.2423, p = 0.0262). Nitrate Sites draining peats and podzolic/lithosolic soils in both geological settings recorded maximum nitrate concentrations largely below 0.5 mg/l NO3 and there was no significant correlation with forest cover. Concentrations were higher at sites on gleys (0.05-2.75 mg/l NO3) and well drained mineral soils (0.18-6.22 mg/l NO3). The latter sites recorded a significant decrease in nitrate concentration across the forest cover gradient. 33 Chloride Maximum chloride concentrations ranged from 6.10 to 44.96 mg/l at sites draining peat on igneous/metamorphic geology. Values were up to 10mg/l lower at peat sites on sedimentary geology and marginally lower at sites draining podzolic/lithosolic soils. The relationship with forest cover was only significant for sites on peat in both geological settings. Calculations suggested that dilution makes a variable contribution to loss of alkalinity and in many cases the forested sites showed a slightly higher % value. Anion titration was detected in all events examined. The principal contributors were organic acids and sulphate. Excess sulphate only made a contribution in the Wicklow sites and at one site in Galway. The contribution of nitrate across all sites was insignificant. The contribution of sea salts to acidification was also low and only one significant sea-salt event was detected at one site in Galway. Comparison of Source Streams – Upstream and Downstream of Forestry On igneous geology, no significant differences for any of the chemical variables were found between control sources and the sources sampled above forests (both 5-25% and >25% forest bands), (Mann-Whitney, P>0.05). A similar result was noted on sedimentary geology (Mann-Whitney, P>0.05). However, it should be noted that on sedimentary geology, only two sites were sampled above the forestry in the 5-25% and >25% categories. Therefore, both of these forestry bands had to be combined into a single forest category and compared to the control sources. This result suggests that all sources (control and above forests) had no significant differences. Control sources did not differ significantly from the downstream sites on the same stream in terms of pH and alkalinity (Wilcoxon Ranked Sign Test, P>0.05). However, chloride, NM sulphate and sodium were significantly higher (Wilcoxon; Chloride: Z = -2.757, P = 0.006; NM Sulphate: Z = -2.114, P = 0.034; Sodium: Z = -3.371, P = 0.001) downstream. Sites downstream of the 5-25% forested band on igneous/metamorphic geology differed significantly from their sources in terms of pH, SDI, chloride, sulphate, NM sulphate, sodium, magnesium, calcium, NM calcium, total hardness and non-marine hardness (Wilcoxon, P<0.05). Sites downstream of >25% forest cover had significantly higher total 34 monomeric aluminium, chloride, sulphate, NM sulphate and sodium (Wilcoxon, P<0.05) than their respective sources. On sedimentary geology sites, the downstream control sites recorded significant differences from their corresponding sources for chloride, sulphate, NM sulphate, sodium, NM sodium and NM magnesium (Wilcoxon, P<0.05). On sedimentary geology only one site pairing represented the 5-25% forest cover category however, a difference in NM Ca was noted (Wilcoxon, P<0.01). The >25% forested category on sedimentary geology presented significant differences between downstream and source for pH, hydrogen, alkalinity, SDI, NM sodium, magnesium, NM magnesium, calcium, NM calcium, total hardness and NM hardness (Wilcoxon, P<0.05 and P<0.001). Results are presented in Table 7. The higher sodium levels at downstream sites were accompanied by higher magnesium and calcium values. This maintained the SDI values as the overall ratio of cations remained quite similar at the source and downstream sites. Table 7: Significant results from paired analysis for selected chemical variables at downstream and source sites (Wilcoxon Ranked Sign Test). Igneous Metamorphic Sites Control <5% Forest 5-15% Forest Cover >25% Forest cover Parameter Chloride NM Sulphate Sodium Wilcoxon (Z) 2.757 -2.114 -2.371 P value 0.006 0.034 0.001 pH Chloride -1.503 -3.11 0.028 0.002 SDI Sulphate NM Sulphate Sodium Magnesium Calcium NM Calcium Total Hardness -2.062 -2.97 -3.11 -3.18 -2.551 -2.831 -2.481 -2.9 0.039 0.003 0.001 0.001 0.011 0.005 0.013 0.004 Total Aluminium Chloride Sulphate -3.068 0.002 -2.425 -2.516 0.001 0.012 NM Sulphate Sodium -2.0 -3.555 0.012 <0.001 Sedimentary Sites Parameter Wilcoxon (Z) -3.516 -3.206 -2.999 -2.999 -3.154 -2.223 P value <0.001 0.001 0.003 0.003 0.002 0.026 Control <5% Forest Chloride Sulphate NM Sulphate Sodium NM Sodium NM Magnesium 5-15% Forest Cover NM Calcium -2.201 0.028 >25% Forest cover pH Hydrogen Alkalinity SDI NM Sodium Magnesium NM Magnesium Calcium -2.971 -2.621 -2.345 -2.342 -2.201 -2.622 -2.271 0.003 0.009 0.019 0.019 0.028 0.009 0.023 -2.622 0.009 NM Calcium Total Hardness NM Hardness -2.411 -2.691 0.016 0.007 -2.621 0.009 Although the above analyses highlighted only two site grouping that recorded lower downstream pH than at the sources there were several individual sites within other groups 35 that followed this pattern. Several of the igneous sites in Co. Wicklow were more acidic downstream than their corresponding sources. These sites included those on the Annalecka, Lugduff and Glashaboy rivers. These sites had ~40-70% catchment cover of coniferous forest. 36 3.2 MACROINVERTEBRATES Community Composition In total, over 318,000 individual specimens were sorted and identified to the lowest possible taxonomic level from the 239 study sites. These yielded a total of 204 distinct taxa. The most diverse group was the Trichoptera followed by the Coleoptera (Table 8). Table 8: Taxon richness in the major taxonomic groups Taxon Trichoptera Coleoptera Diptera Ephemeroptera Plecoptera Gastropoda Odonata Crustacea Hirudinea Hemiptera Neuroptera Lamellibranchia Richness 61 53 29 18 17 8 5 3 4 2 1 1 Some of these taxa were highly localised, such as the mayfly species, Ameletus inopinatus Eaton, found only in samples collected in Wicklow and Donegal. Other mayfly, such as Baetis rhodani (Pictet.) and Leptophlebia vespertina (Linn.) were more ubiquitous. Species such as Caenis rivulorum Eaton and the caddis-fly, Sericostoma personatum (Kirby & Spence) were considered acid-sensitive as they were located in more buffered regions on sedimentary geology. More acid-tolerant species including, Ameletus inopinatus, Siphlonurus lacustris (Eaton), Leptopheblia vespertina and Plectrocnemia conspersa (Curtis) were present in higher abundances in areas of weathering tolerant, acid-sensitive, igneous geologies. 37 In general the mean abundances of macroinvertebrates was significantly higher at the sedimentary sites (One-way ANOVA; F(1,5) = 59.058, P = 0.002; Fig. 21). 700 Sedimentary Igneous/Metamorphic 600 Abundance 500 400 300 200 100 0 Control 5-25% Forest Cover 25-50% Forest Cover >50% Forest Cover Fig. 21: Mean macroinvertebrate abundances at the sedimentary and igneous/metamorphic sites across the four forest cover bands. The higher total macroinvertebrate abundances at the sedimentary sites could be largely attributed to the Ephemeroptera and Chironomidae (Fig. 22). In both geological settings the Ephemeroptera was reduced in abundance at sites in the two highest forest cover bands. At the igneous/metamorphic sites the reduction in ephemeropteran abundance was largely balanced by an increase in the numbers of Plecoptera. This did not occur at the sites draining sedimentary geology and consequently overall abundance declined gradually across the forest cover bands. Selection of Macroinvertebrates Metrics Approximately 45 different water quality and diversity metrics were generated for the dataset using the AQEM (ASTERICS 3.10™) computer software. Those which were most appropriate for Ireland and which showed a significant correlation with pH were selected to detect impacts due to acidification. These included taxon richness (Fig. 23), ephemeropteran richness, ephemeropteran abundance, trichopteran richness, Baetis abundance, %EPT. 38 250 Control 5-25% Forest Cover 25-50% Forest Cover >50% Forest Cover Abundance 200 150 100 50 Ot he rs Ch iro no m id ae Si m ul iid ae Cr us ta ce a Tr ich op te ra Pl ec op te ra Ep he m er op te ra 0 (a) 250 Control 5-25% Forest Cover 25-50% Forest Cover >50% Forest Cover Abundance 200 150 100 50 Ot he rs Si m ul iid ae Ch iro no m id ae Cr us tac ea Tr ich op ter a Pl ec op te ra Ep he m er op te ra 0 (b) Fig. 22: Mean abundances of the major taxonomic groups at sites on (a) igneous/metamorphic and (b) sedimentary geology. Standard error bars are included. pH(min):Taxon Richness: r2 = 0.2437; r = 0.4937, p = 0.0000 60 Taxon Richness 50 40 30 20 10 0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 pH (minimum) Fig. 23: Relationship between taxon richness and minimum pH across all sites. 39 Taxon Richness On igneous/metamorphic geology taxon richness ranged from 11 to 52, the lower value was from a forested catchment (DWW22, 67% forest cover). The range of values was similar on sedimentary geology (16 at L5, 37% forest cover to 55 at one of the control sites). There was a significant decline in taxon richness with increasing coniferous forest cover on peat and well drained mineral sites on sedimentary geology (Kruskall-Wallace; Sedimentary/Peat – H(2,65) = 10.4252, P<0.05; Sedimentary/Well Drained Mineral – H(2,34) = 9,4919, P<0.05, Fig. 24). The differences were not statistically significant for igneous/metamorphic sites. However, the number of sites with taxon richness less than 30 was highest in the >50% forest cover band compared to the control group (Fig. 25) draining peats, and the pattern was retained when sites on podzolic/lithosolic soils were added to the analysis (Fig. 26). Taxon Richness 60 55 50 45 40 35 30 25 20 15 10 Control(<5%) Forested(25-50%) Control(<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Forested(5-25%) Forested(>50%) Peats Podzolic Lithosolic 60 55 50 45 40 35 30 25 20 15 10 Control(<5%) Forested(25-50%) Control(<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Forested(5-25%) Forested(>50%) Poorly Drained Gleys Median 25%-75% Non-Outlier Ran Outliers Extremes Well Drained Mineral Fig. 24: Box plots of taxon richness for sites on sedimentary geology with different soil categories. 40 Control >50% Forest Cover ............... % Sites 45 40 35 30 25 20 15 10 5 0 20 25 30 Taxon Richness >30 . Fig. 25: Comparison of the distribution of taxon richness at sites draining peat on igneous/metamorphic geology. 70 60 50 >50% Forest Cover 25-50% Forest Cover 5-25% Forest Cover Control % Sites 40 30 20 10 0 10 15 20 25 30 >30 Taxon Richness Fig. 26: Distribution of taxon richness counts at sites draining peat and podzolic/lithosolic soils on igneous/metamorphic geology. Ephemeropteran Richness Ephemeropteran richness reached a maximum of 8 species on igneous/metamorphic geology with one additional species on the sedimentary geology. In both settings some of the forested sites recorded a low diversity of Ephemeroptera. Ephemeropteran richness was significantly negatively correlated with % forest cover at sites draining peats (r=-0.4640, P<0.001) and podzolic lithosols (r=-0.3884, P<0.05) on igneous/metamorphic geology. As forest cover increased an increasing number of sites 41 recorded low taxon richness (Fig. 27). On sedimentary geology (Fig. 28) ephemeropteran richness was again significantly correlated to percentage conifer cover on peat (r = -0.5378, P<0.001) and well drained mineral soils (r = -0.4855, P<0.001). 9 8 Ephemeropteran Richness 7 6 5 4 3 2 1 0 -1 -20 0 20 40 60 80 100 120 -20 0 20 SoilTypes: Peats 40 60 80 100 120 SoilTypes: Podzolic Lithosolic % Coniferous Cover Fig. 27: Relationship between ephemeropteran richness and forest cover at sites draining peat and podzolic/lithosolic soils on igneous/metamorphic geology. 10 8 6 4 Ephemeropteran Richness 2 0 -2 -20 0 20 40 60 80 100 120 -20 0 20 Peats 40 60 80 100 120 80 100 120 Podzolic Lithosolic 10 8 6 4 2 0 -2 -20 0 20 40 60 80 100 120 -20 0 Poorly Drained Gleys 20 40 60 Well Drained Mineral % Coniferous Cover Fig. 28: Relationship between ephemeropteran richness and forest cover at sites draining different soils on sedimentary geology. 42 As was highlighted for total taxon richness, an increasing number of sites showed a reduction in ephemeropteran richness (Table 9) along the forest cover gradient. Some 13% of sites draining peat on igneous/metamorphic geology in the >50% coniferous cover band were devoid of Ephemeroptera, while a further 75% only had one species present. On podzolic/lithosilic soils there relatively few Ephemeroptera at sites in the three forestry bands. A decline in the occurrence of Baetis spp. occurred across the forestry bands. For example 9% of the control sites on peat had low numbers of Baetis spp. compared to 71% of sites in the >50% forest cover band. The replacement of Baetis by more acid-tolerant species (e.g. Siphlonurus lacustris) was a feature of the latter group of sites. Table 9: Percentage distribution of ephemeropteran taxon richness counts for sites on (a) peats and (b) podzolic/lithosolic soils on igneous/metamorphic geology (a) Richness 0 1 2 3 4 5 6 7 8 (b) Control 0 5 18 36 9 23 5 0 4 Forest 5-25% 0 22 33 11.5 0 11.5 22 0 0 Cover 25-50% 0 14 14.5 14 14.5 29 14 0 0 >50% 13 75 0 6 0 6 0 0 0 Richness 0 1 2 3 4 5 6 7 8 Control 0 18 9 9.5 36 9.5 9 0 9 Forest 5-25% 25 0 0 25 25 0 25 0 0 Cover 25-50% 11 44 22 11.5 0 0 0 11.5 0 >50% 25 25 12.5 12.5 25 0 0 0 0 The pattern was similar for peat sites on sedimentary geology where a low number of Ephemeroptera was recorded in the two top forestry cover bands (Table 10). On podzolic/lithosolic soils only sites in the >50% band recorded no mayfly and a large proportion of the sites in the >50% band had just one or two species present. Baetis spp. were only absent from sites (10%) on podzolic/lithosolic soils in the >50% forest cover band. At sites on poorly drained gleys and well drained mineral soils there was little differences in the distribution of ephemeropteran counts across the forest cover bands (Table 11). 43 Table 10: Percentage distribution of ephemeropteran taxon richness counts for sites draining (a) peats and (b) podzolic/lithosolic soils on sedimentary geology. (a) Richness 0 1 2 3 4 5 6 7 8 9 (b) Control 0 0 0 15 10 40 15 10 5 5 Forest 5-25% 0 50 0 25 0 0 0 25 0 0 Cover 25-50% 18.5 9 9 18.5 9 9 9 9 9 0 >50% 7 43 20 3.5 13 10 0 3.5 0 0 Richness 0 1 2 3 4 5 6 7 8 9 Control 0 0 0 0 23 23 31 15 8 0 Forest 5-25% 0 0 0 0 0 50 50 0 0 0 Cover 25-50% 0 0 0 0 50 0 0 0 50 0 >50% 8 8 8 8 17.5 25 17.5 8 0 0 Table 11: Percentage distribution of ephemeropteran taxon richness counts for sites on (a) poorly drained gleys and (b) well drained mineral soils on sedimentary geology. (a) (b) Richness 0 1 2 3 4 5 6 7 8 9 Control 0 33.3 0 33.3 33.3 0 0 0 0 0 Forest 5-25% 0 0 0 33.3 0 33.3 33.3 0 0 0 Cover 25-50% 0 0 14.5 14 14.5 14 29 14 0 0 >50% 0 33 0 0 22.5 22 22.5 0 0 0 Richness 0 1 2 3 4 5 6 7 8 9 Control 0 0 0 0 0 50 50 0 0 0 Forest 5-25% 0 0 20 0 40 20 0 20 0 0 Cover 25-50% 0 0 20 0 40 20 0 20 0 0 >50% 0 0 44.5 0 0 44.5 11 0 0 0 Ephemeropteran Abundance A reduction in abundance of indicator taxa can often highlight environmental stress and it is considered to be a useful early warning indicator of impact. No significant correlation was detected between ephemeropteran abundance and forest cover for sites draining either peat or podzolic/lithosolic soil types on igneous/metamorphic geology. The same applied to these soil types on sedimentary geology. However, a significant relationship was detected for sites located on well drained mineral soils on sedimentary geology (r = -0.6358, P>0.001). 44 700 600 500 400 300 200 Ephemeropteran Abundance 100 0 -100 -20 0 20 40 60 80 100 120 -20 0 20 Peats 40 60 80 100 120 80 100 120 Podzolic Lithosolic 700 600 500 400 300 200 100 0 -100 -20 0 20 40 60 80 100 120 -20 Poorly Drained Gleys 0 20 40 60 Well Drained Mineral % Coniferous Cover Fig. 29: Relationship between % forest cover and ephemeropteran abundance on sedimentary geology The cluster of high mayfly abundance noted on sedimentary peats between pH levels 6.5 and 7.5 (Fig. 29) corresponds to a cluster of highly buffered, high pH sites in Co. Cork. Despite high levels of coniferous forest cover at these sites, the occurrence of variable amounts of mineral soils among the peats improved buffering capacity and allowed for higher abundances of mayfly (particularly Baetis rhodani). Despite the lack of a strong correlations between abundance and % forest cover it was clear that on peat and podzolic soil in both geological settings the number of sites with low numbers (zero and <5) of ephemeropteran specimens increased across the forest cover bands. The results are illustrated for sites on (a) peat and (b) podzolic/lithosolic soils in Fig. 30. 45 (a) 70 Control 60 5-25% Forest Cover % Sites 50 25-50% Forest Cover 40 >50% Forest Cover 30 20 10 0 0 5 10 15 20 >20 Ephemeropteran Abundance % Sites . 70 (b) 60 Control 50 5-25% Forest Cover 40 25-50% Forest Cover >50% Forest Cover 30 20 10 0 0 5 10 15 20 >20 Ephemeropteran Abundance Fig. 30: Frequency distribution of ephemeropteran abundance counts assigned to 6 abundance categories (0; ≤5; ≤10; ≤15; ≤20; >20) at sites draining (a) peat and (b) podzolic/lithosolic soils on igneous/metamorphic geology. EPT Metric Ephemeropteran (E), plecopteran (P) and trichopteran (T) richness values are used to calculate EP and EPT metric. In the present study EPT richness correlated significantly (R2= 0.735, P<0.0001) with EP richness and therefore only one of these was applied in the analyses. The variability plot of EPT richness indicated a shift in distribution towards the lower end of the scale as one moved across the forest cover bands (Fig. 31). While no significant differences in median EPT was detected across forestry bands on igneous/metamorphic geology (P>0.05), significant declines in both metrics were found for peat sites on sedimentary geology (Kruskall-Wallis- EPT – H(3,65) = 10.0914, P = 0.0178, Fig. 32). A similar trend occurred on well drained mineral soil but the relationship was not significant (P>0.05). It is worth noting that the median EPT of sites draining podzolic/lithosolic soils fell well below the control median in the >50% forest cover band. 46 35 EPT (Richness) 30 25 20 15 10 Sedimentary Forested(>50%) Forested(25-50%) Control(<5%) Forested(5-25%) Forested(>50%) WellDrainedMineral Poorly Drained Gleys Forested(25-50%) Control(<5%) Forested(5-25%) Forested(>50%) Forested(25-50%) Control(<5%) Forested(5-25%) Podzolic Lithosolic Forested(>50%) Forested(25-50%) Peats Control(<5%) Igneous-metamorphic Forested(5-25%) Forested(>50%) Control(<5%) WellDrainedMineral Forested(25-50%) Podzolic Lithosolic Control(<5%) Forested(5-25%) Forested(>50%) Forested(25-50%) Peats Control(<5%) 0 Forested(5-25%) 5 Control SoilTypes Geology Fig. 31: Variability plot of EPT across coniferous forest bands for each geology and soil category. 35 30 25 20 15 10 5 EPT (Richness) 0 Control(<5%) Forested(25-50%) Control(<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Forested(5-25%) Forested(>50%) Peats Podzolic Lithosolic 35 30 25 20 15 10 5 0 Control(<5%) Forested(25-50%) Control(<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Forested(5-25%) Forested(>50%) Poorly Drained Gleys Median 25%-75% Non-Outlier Range Outliers Extremes Well Drained Mineral Fig. 32: Box plots for EPT richness on sites draining sedimentary geology. 47 The relationship between EPT richness and % forest cover is further explored in the correlation plots (Fig. 33). The correlation was significant for peat sites in both geological SoilTypes: Podzolic Lithosolic SoilTypes: Poorly Drained Gleys EPT (Richness) 35 30 25 20 15 10 5 0 SoilTypes: WellDrainedMineral 35 30 25 20 15 10 5 0 SoilTypes: Peats Scatterplot (MetricsUCDAnalysis_29_4_08_JRB 241v*239c) 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 -20 0 20 40 60 80 100 120 -20 Geology: Sedimentary 0 20 40 60 80 100 120 Geology: Igneous-metamorphic %Coniferous Fig. 33: Relationship between EPT richness and % forest cover in the various geological settings. settings and site EPT richness began to fall below the lower limit of the control sites when forest cover exceeded circa 25-30% . This also applies to sites on podzolic/lithosolic soils on igneous/metamorphic geology. On sedimentary geology the podzolic sites recorded low EPT above 50% forest cover. BMWP & ASTP Metrics While Kruskall-Wallis tests on BMWP did not show a significant difference between forest bands on peats and podzolic/lithpsolic sites on igneous/metamorpgic geology there were nonetheless strong trends of decreasing BMWP across the forest cover bands. Peat sites on sedimentary geology demonstrated a significant decrease in BMWP with increasing forest cover bands (H(3,65) = 10.3406, P = 0.0159). While the trend was only statistically significant 48 for peats some sites on podzolic/lithosols and poorly drained gleys showed a distinct decrease in BMWP scores in the >50% forest cover band. The Biological Monitoring Working Party (BMWP) Score was significantly correlated with % coniferous cover for sites draining peat soils on both igneous/metamorphic (r = -0.3209, P = 0.03, Fig. 34 ) and sedimentary rock (r = -0.4195, P = 0.0009) geology (Fig. 35). The relationship was not significant for podzolic/lithosolic soils but here again it should be noted that scores for sites on igneous/metamorphic and sedimentary geology began to fall below the control BMWP scores when forest cover exceed 25% and 60%, respectively. Average Score Per Taxa (ASTP) significantly correlated with increasing % coniferous cover at sedimentary peats sites only (r = -0.3679, P = 0.0041). A significant difference was also detected between forest cover bands for sedimentary peat sites, with ASTP decreasing as forest bands increased (H(3.65) = 7.9399, P = 0.0473) 200 180 160 BMWP Score 140 120 100 80 60 40 20 -20 0 20 40 60 80 100 120 -20 SoilTypes: Peats 0 20 40 60 80 100 120 SoilTypes: Podzolic Lithosolic BMWP Score Fig. 34: Relationship between BMWP and % coniferous cover for soils on igneous/metamorphic geology. 220 200 180 160 140 120 100 80 60 40 20 -20 0 20 40 60 80 100 120 -20 0 20 Peats 220 200 180 160 140 120 100 80 60 40 20 -20 0 20 40 40 60 80 100 120 80 100 120 Podzolic Lithosolic 60 Poorly Drained Gleys 80 100 120 -20 0 20 40 60 Well Drained Mineral Fig. 35: Relationship between BMWP scores and % forest cover for sites draining sedimentary geology. 49 Community Diversity and Species Evenness A trend of decreasing scores for both the Simpson and Margaley diversity indices was noted across the forest cover bands. However, the differences between bands was only significant for peat sites in both geological settings. The correlation with % forest cover was highly significant (P=0.001) for sites draining peat on sedimentary geology (Fig. 36). Species evenness varied greatly across sites and settings and none of the trends was statistically significant at P<0.05. However, the pattern, previously discussed, whereby some forested sites fell below the minimum values of the control sites at forest cover was repeated but the number of sites involved was much fewer. 10 9 8 7 6 5 Diversity (Margalef Index 4 3 2 Control(<5%) Forested(25-50%) Fores ted(5-25%) Forested(>50%) Control(<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Peats Podzolic Lithos olic 10 9 8 7 6 5 4 3 2 Control(<5%) Forested(25-50%) Fores ted(5-25%) Forested(>50%) Control(<5%) Forested(25-50%) Forested(5-25%) Forested(>50%) Poorly Drained Gleys Median 25%-75% Non-Outlier Range Outliers Extremes Well Drained Mineral Fig. 36: Box plot of the Margalef diversity Index scores at sites on various soil types on sedimentary geology Clustering of Biological, Chemical and Physical Metrics The k-means algorithm was used to implement separate clustering of the 239 sites based on selected chemical, physical and biological (macroinvertebrate response) metrics. In each case the algorithm was asked to select four clusters. 50 The four biological clusters represented a gradient in the three metrics, group 1 having the highest richness and cluster 4 the lowest. The latter contained a large proportion of the impaired sites (31%, Table 12a). Thirty-three of these sites (44.5%) also grouped into the chemical cluster (cluster 1, Table 13), the centroids of which represented the most acidic conditions (Table 12b). Finally, 52 of the sites in the impoverished biological cluster 4 (66.6%, Table 13) also appeared in the physical cluster with the highest levels of percentage coniferous cover (cluster 3) (Table 12c). The location of the biological cluster is shown in Fig. 37. Sites within cluster 4 occurred in all regions. Table 12: Centroid values for a) biological, b) chemical and c) physical clusters (a) Site cluster1 cluster2 cluster3 cluster4 Taxon_Richness 42.53 38.61 33.23 25.59 Ephemeropteran_Abundance 108.13 465.11 55.91 37.75 Ephemeropteran_Richness 6.14 4.61 3.93 1.14 (b) (c) Site cluster1 cluster2 cluster3 cluster4 %Coniferous 8.7 13.83 71.24 11.31 Slope 0.06 0.16 0.06 0.04 Elevation 297.36 292.02 187.81 113.4 51 Table 13: Numbers of sites loading into both a) biological and chemical clusters and b) biological and physical clusters. (a) Biological 1 2 3 4 Biological (b) 1 2 3 4 Chemical 1 2 1 19 33 Physical 1 11 4 12 8 2 10 9 7 5 2 6 3 5 5 3 11 1 14 16 3 17 6 20 52 4 29 6 20 20 4 20 5 21 13 Fig. 37: Location of the biological site clusters. 52 Multivariate Analyses Much of the analyses presented thus far deals with trends in individual metrics. The multivariate analyses examined the relationships between the invertebrate communities at each site. To better visualise the similarities between sites in terms of community composition NMDS plots, with Bray Curtis as the similarity measure, were generated using the key indicator groups, Ephemeroptera, Plecoptera and Trichoptera, on site groups within the selected geological settings. Plot were prepared for sites on peat and podzolic/lithosolic soils draining granite and peat sites on sedimentary geology. These were highlighted earlier as showing responses to forest cover. Fig 38 is a plot of the sites draining peat on igneous/metamorphic geology. The control sites are positioned largely to the left of the plot. The three most acidic control sites (DWW4,8 & 9 –Co. Wicklow; DG9-Co. Galway) with low taxon richness sit on the right side of the group outline. At the other end of the plot sites M6, M7 and DWW6 (Knickeen, Co. Wicklow) represent sites with high total taxon richness and good representation of Ephemeroptera (richness and abundance) and Plecoptera. While there is, as expected, some overlap with the most acidic controls, sites in the highest forest cover bands fall largely on the right side of the plot. Those on the extreme right show impairment in a number of metrics. The only unimpaired sites within this region (DG21, 22 & 23-Owenwee River, Co. Galway) lie well within the main grouping of control sites. Fig. 38: NMDS plot of Bray Curtis similarity measure of EPT community composition at sites draining peat on igneous/metamorphic geology. Outline of control sites excluded the acidic outliers. 53 A similar picture emerged with respect to sites draining podzolic/lithosolic on igneous geology. Impaired sites within the three forestry bands plot on the right side of Fig. 39. In contrast, the unimpaired afforested sites (DWW15 , 16 and 17) with high EPT taxon richness and abundances plot to the left. Fig. 39: MDS plot of Bray Curtis similarity measure of EPT community composition at sites draining podzolic/lithosolic soils on igneous/metamorphic geology. The control peat sites on sedimentary geology form a closer cluster than seen on igneous geology (Fig. 40). Here sites with low EPT richness and abundance largely plot outside the grouping of control sites, most of which show some impairment in the metrics applied earlier. The frequency of impacted sites occurring outside of the control site grouping increases with increasing forest cover. Fig. 40: MDS plot of Bray Curtis similarity measure of EPT community composition at sites draining peat on igneous/metamorphic geology. . 54 Evaluation of the Degree of Biological Impairment Five metrics (ephemeropteran richness, abundance Baetis spp., trichopteran richness, evenness and diversity indices) which were not autocorrelated were selected to evaluate potential impairment. The metrics selected target known indicator taxa as well as abundance, evenness and diversity aspects of the community. As outlined in the methods section, sites with metric values below two standard deviations of the mean of the control site values were considered impaired. Table 14 shows the number of sites in each geological setting that shows impairment for 1 to 5 metrics and gives an overall estimate of the number of sites that fail on two or more metrics. Table 14: Estimation of the number of impaired sites as indicated by low metric scores for 5 metrics in each geological and forest cover setting. % Sites with >2 impacted sites 1 2 3 4 5 Total Sites 22.2 33.3 5.6 22.2 16.7 22.2 0 16.7 16.7 33.3 16.7 11.1 0 0 27.8 9 7 17 55.6 50.0 77.8 0 11.1 25.0 75.0 11.1 25.0 0 33.3 37.5 0 22.2 0 0 0 0 4 9 8 * 66.7 62.5 0 27.3 6.7 25.0 18.2 10.0 50.0 9.1 20.0 0 27.3 33.3 0 9.1 20.0 4 11 30 * 63.6 83.3 Podzolic/Lithosolic 5-25% 25-50% >50% 0 0 22.7 0 0 22.7 50.0 50.0 18.2 0 0 9.1 0 0 18.2 2 2 22 * * 68.2 Poorly drained Gleys 5-25% 25-50% >50% 0 14.3 9.1 33.3 28.6 18.2 0 0 0 0 0 18.2 0 0 18.2 3 7 11 * 28.6 54.5 Geological setting Igneous/Metamorphic Peat 5-25% 25-50% >50% Podzolic/Lithosolic 5-25% 25-50% >50% Sedimentary Peat 5-25% 25-50% >50% • low replication 55 The % impaired sites increased across the forest cover bands on igneous/metamorphic geology. The same applied for peat sites on sedimentary geology. Podzolic/lithosolic sites on sedimentary geology recorded significant impairment in the >50 forest band. It should however be noted that replication was low in the other two forest bands. Some 55% of sites on poorly drained gleys in the >50 forest band failed on more than two metrics. None of the well drained mineral sites failed on more than two metrics. Four sites recorded low ephemeropteran abundances, however there was low replication of the control sites. Comparison of Source and Downstream Communities As outlined in the methods a number of forested streams were sampled above and below the forest. Control sites were sampled at equivalent points. No significant differences were detected between source and downstream sites in pairwise comparison statistics using macroinvertebrate metrics (richness and abundance data for total taxa, Ephemeroptera, Plecoptera, Trichoptera and Coleoptera). However, community differences were revealed by the multivariate analyses. Fig. 41 shows the NMDS plot based on Sorensen similarity measure. It clearly shows that the control downstream sites (blue) were distinctly different from their sources (red). It also indicates that the macroinvertebrate communities from forested downstream sites (green) were similar to the sources from both forested (yellow) and control sites (red). Fig. 41: MDS plot of Sorensen similarity measure of community composition at sites upstream and downstream of forestry and at similar locations on control sites. 56 Longitudinal Variation in Macroinvetebrate Metrics Several catchments were sampled at several sites to illustrate longitudinal changes in macroinvertebrate community and to evaluate the potential distance downstream that a forest effect might be detected. The example presented here is the King’s catchment, Co. Wicklow. Some of the forested headwater sites had poor total taxon richness and ephemeropteran richness compared to the control sites (Table 15) and this was maintained well down the system to Site Kings1 (Fig. 42). Fig. 42 Sites sampled along the length of the King’s River, Co. Wicklow. Table 15: Locaton of sites sampled in the King’s catchment and recorded metric scores for total taxon and ephemeropteran richness River % Forest Taxon Ephem. Cover Richness Richness ANNA3 0 25 2 Annalecka Brook ANNA1 34.71 26 2 BALLIN2 0 24 5 Ballinagee River BALLIN1 0 34 5 GLASH2 40.9 35 2 Glashaboy River GLASH1 65.72 35 4 KINGS2 36.63 22 3 King’s River KINGS1 28.26 25 2 Ephem.=Ephemeroptera 57 Further Evaluation of Potential Longitudinal Patterns in Macroinvertebrate Recovery from Forest Effects Forested sites in counties Wicklow (Vartry stream, DWW20 / VART1) and Cork (Foherish river, DK26 / FOHER1) which recorded a paucity of Ephemeroptera during the spring 2007 sampling season were re-visited in spring 2008. In each case a nearby control/reference site was also sampled, the Bohill river (DK23 / BOHIL1) in Co. Cork and the nearby nonforested tributary of the Vartry catchment, Co. Wicklow (DWW10 / VAR1). The controls were selected to be comparable in terms of geology, soil type, elevation, catchment area, slope, aspect and catchment size. The two paired streams were then sampled approximately 800 metres from source and every 500 metres thereafter over a two kilometre stretch of the streams. The Cork sites drained areas of Old Red Sandstone, while all but one of the Wicklow sites were situated on Palaeozoic sediments. Soils types within the catchment of each paired stream were also comparable. The sampling sites where located between 200m360m asl. Sites 1 in all cases were located at >300m; Sites Nos. 2 between 275-300m; Sites Nos. 3 between 250-275m; Sites Nos. 4 below 250m. The forested sites in Wicklow were coded WKF1, WKF2, WKF3 and WKF4, while open (control) sites were labelled WKO1, WKO2, WKO3 and WKO4. A similar site designation was used for the Cork sites (Figs. 43 & 44). Fig. 43: Location of Wicklow forested and control sites Fig. 44: Location of Cork forested and control sites. 58 The community composition was dominated by Chironomidae and other dipteran larva in the Wicklow (WKO-64%; WKF-87%) and Cork sites (CKO-57%; CKF-63%). The Ephemeroptera represeneted a lower percentage of the fauna at the Wicklow forested sites (2%) compared to the control sites (10%). The Tricoptera varied little across sites ranging from 4-7%. The Plecoptera accounted for 15% (CKF) to 11% (CKO) of the total abundances in the Cork sites compared to 2% (WKF) and 3% (WKO) sites. Crustaceans were particularly abundant in non-forested sites; WKO and CKO accounting for 9% and 8% of the total abundances. However, they accounted for only 4% of the fauna in forested sites in Cork and where absent from forested sites in WKO. In Wicklow, non-forested sites supported between 7 and 25 taxa, while forested sites had between 4 and 14. In Cork, nonforested sites recorded between 14 and 24 taxa, while between 8 and 21 taxa were found at forested sites. The differences between the paired control and forested sites was significant (Wicklow – Wilcoxon test: Z=-2.598, P<0.01; Cork – Wilcoxon test: Z=-2.096, P<0.05) (Fig. 45). Taxon Richness 25 20 15 10 5 0 1 2 3 4 Site Number Non-Forested Wicklow Sites Forested Wicklow Sites Taxon Richness 25 20 15 10 5 0 1 2 3 4 Site Number Non-Forested Cork Sites Forested Cork Sites Fig. 45: Mean taxon richness at forested and non-forested sites in Wicklow and Cork. 59 Low ephemeropteran richness was a common feature of the first 2 sites in each forested catchment (Figs. 46 & 47). The only species present was Baetis rhodani. While this indicated a slight recovery in terms of species richness in comparison to the original sampling period of the study (April-May 2007 - in which no mayfly were found), the abundances were far lower than those for the corresponding controls in both Cork and Wicklow. Wilcoxon paired Test, showed there was a significant differences in the mean ephemeropteran richness between forested and non-forested sites in Cork (Z=-0.2366; P<0.05) and Wicklow (Z= 3.781; P<0.001). However, in Cork ephemeropteran richness differed significantly only between the first three forested and control site pairs. The fourth, located almost 2.5km from 45). . the source, and still within close proximity to the forestry, was not showing impairment (Fig. Ephemeropteran Richness 8 6 4 2 0 1 2 3 4 Site Number Non-Forested Cork Sites Forested Cork Sites Fig. 46: Mean ephemeropteran richness at forested and non-forested sites in Cork In Wicklow the significantly lower ephemeropteran richness at the forested sites persisted down to Site 4. (Kruskall-Wallace ANOVA, P<0.05) (Fig. 47). Trichopteran richness was also significantly higher at the control sites in Wicklow (Z= -3.776’ P<0.001) and Cork (Z= 3.530, P <0.001) but here again the differences between the site pairs was eliminated by Site 4 in Cork. 60 Ephemeropteran Richness 8 7 6 5 4 3 2 1 0 1 2 3 4 Site Number Non-Forested Wicklow Sites Forested Wicklow Sites Fig. 47: Mean ephemeropteran richness at forested and non-forested sites in Wicklow. The longitudinal pH profile at the time of sampling is illustrated (Fig. 48). Although all sites were circum-neutral there was at least a 0.5 unit of differences between the control and forested site pairs and the differences was greatest for the first two sites. 8 7.5 Cork Non-Forested 7 Cork Forested 6.5 pH Wicklow NonForested 6 Wickow Forested 5.5 5 1 2 3 4 Site No. Fig. 48: Longitudinal pH profile for sites sampled in Wicklow and Cork. 61 3.3 FISH As outlined in the methods sites were selected on a paired-catchment basis. To further ensure that all site pairings were comparable, the coverage of each habitat type (riffle (P=0.38), glide (P=0.402), pool (P=0.175)), depth (P=0.822), width (P=0.705), wetted area (P=0.812), conductivity (P=0.492) and time fishing (effort – m2/min, P=0.12) were examined between pairs. In each case there were no significant differences (Wilcoxon Ranked Sign Test, P>0.05) detected, making all pairings comparable for further analysis. The results of the Wilcoxon Signed Ranked Test on abundance of salmonids at the two site grouping are given (Table 15). Overall the catch of trout was higher than salmon (Figs. 49 & 50), few salmon were caught in the Wicklow sites. The Galway sites recorded the highest catches of salmonids. Total salmonid catch differed significantly between the paired control and forested sites with the lowest numbers at the forested sites. Significant differences were also detected for total salmon, trout and salmon fry (Table 16). In all cases, there were fewer individuals at forested sites (Figs. 49 & 50). The differences were not significant for adult trout or adult salmon. Table 16: Wilcoxon Ranked Sign Test results for comparison of salmonid abundances between non-forested and forested sites. Total Salmonid Abundance ZValue -2.939 PValue 0.003* Total Trout Abundance Trout Adult Abundance Trout Fry Abundance -1.731 -1.168 -2.049 0.083 0.243 0.041* Total Salmon Abundance Salmon Adult Abundance Salmon Fry Abundance -2.194 -1.55 -2.366 0.028* 0.121 0.018* * Significant to P = 0.05 Wilcoxon Sign Ranked Tests 62 Abundance 50 45 40 35 30 25 20 15 10 5 0 Total Salmon Non-Forested Sites Adult Salmon Salmon Fry Forested Sites Abundance Fig. 49: Mean abundances of salmon captured at forest and non-forested sites. Standard error bars included. 50 45 40 35 30 25 20 15 10 5 0 Total Trout Non-Forested Sites Adult Trout Trout Fry Forested Sites Fig. 50: Mean abundances of trout captured at forest and non-forested sites. Standard error bars included. Salmonid densities (fish/m2) were also compared between the paired control and forested sites. The results were similar to those described for fish catch. Total trout density (Z = 2.45, P = 0.014) and trout fry (Z = 2.50, P = 0.0122) density differed between the two site groups. 63 The differences were significant for adult trout. Total salmon (Z = 2.73, P = 0.006) and fry (Z = 3.54, P = 0.0003) density was significantly lower in the forested sites. The length frequency distribution of salmonids across all sites is shown in Fig. 51. The populations were generally dominated by 1+ fish ranging in length from 9 to 15 cm. The numbers of larger fish were highly variable across sites. Fry numbers were generally lower at the forested sites. An evaluation of growth differences between control and forested sites was undertaken for fry. The data were examined on a regional basis to incorporate any natural variation in growth. No significant differences were detected in any of the regions or across the total dataset (Mann-Whitney, P>0.05, Table 17). Table 17: Mann-Whitney test statistic results (P>0.05) for a comparison of the length of trout fry in control and forested sites in four regions. Region Wicklow Mayo Galway Donegal Mann-Whitney Test Statistic 0.260 0.279 0.314 0.082 Several sites were noted for their low abundances or paucity of salmonids (Table 18). These sites were all highly forested with the exception of one control site on the Cloghoge river in Co. Wicklow (CLOG1). Kelly-Quinn et al. (1996b) suggest that the combination of site elevations >400 m.a.s.l along with steep slopes can exclude salmonid fish from streams. However, while several of the sites in Table 18 were approaching the elevation cut-off, the slopes were not steep. None of the sites had any known barriers to fish movement. 64 a) c) b) d) Fig. 51: Fish population structure of the salmonids caught in the control (blue) and forested (green) sites in the Galway (a), Mayo (b), Donegal (c) and Wicklow (d) regions. The stream pairs are maintained alongside each other for comparative purposes. 65 Table 18: Sites with absences of salmonid fish. All Fish Absent ANNA1 (25-50% Forest) Fry Absent ANNA1 (25-50% Forest) GAMON5 (>50% Forest) INCH1 (25-50% Forest) GMOY1 (>50% Forest) GLASH2 (25-50% Forest) CORRIB1 (>50% Forest) CORRIB2 (25-50% Forest) SRAG1 (5-25% Forest) CLOG1 (Control <5% Forest) Adult Fish Absent ANNA1 (25-50% Forest) CROE1 (5-25% Forest) 66 4. DISCUSSION This project set-out to investigate the presence and extent of any acidification associated with coniferous forestry in Ireland and to assess the risk of impact with respect to different geological settings. In the selection of forested sites it was aimed to represent a combination of the risk factors in terms of catchment cover and acid-sensitive geology that were perceived to have the greatest potential for acidification. The large number of sites selected allowed for good spatial coverage but it did limit the amount of water sampling that could be undertaken. The aim was to sample each site at variable flow conditions, from low flow to flood. It was however not possible to obtain flood samples for all sites as a result of their geographic spread and remote locations. Furthermore, it was often difficult to ascertain the stage in the hydrograph represented on any one date. Nevertheless, within any one region a good representation of control and forested sites were sampled within the same timeframe and usually under the same flow/weather conditions. The pH results analyses suggested that most of the streams were episodically acidic with a small group more likely to be circum-neutral. Overall, the pH results indicated increased acidity at some sites associated with forestry on peat and podzolic/lithosoilic soils on both igneous/metamorphic and sedimentary geology and to some extent on poorly drained gleys. Two components of these results require clarification. Firstly, while a small number of the control sites, especially on peat/granite, recorded minimum pH values as low as some of the forested sites the frequency of low pH readings was substantially higher among some groups of forested sites. So the critical issue may be that the frequency and duration of acid pulses can be higher in some forested catchments. Previous intensive monitoring of acid pulses in a heavily afforested stream in the Wicklow mountains alluded to this (Kelly-Quinn, Tierney & Bracken, 1997). The current dataset, unfortunately, does not have sufficient data to further test this hypothesis. This association emerged when forest cover in the site catchment exceeded 25-30%. Factors controlling the severity and duration of acid pulses require further research to better target measures. The second issue to consider is whether the current results suggest a forest-cover threshold above which the risk of acidification increases. Certainly the minimum pH for both peat and podzolic/lithosolic sites on igneous/metamorphic geology began to fall below the lower limit of the control sites when forest cover exceeded values in the region of 25%. The same applied to peat sites draining sedimentary geology. Sites on podzolic/lithosolic soils on sedimentary geology did not have 67 minimum pH values below the lower limit of the control sites until forest cover exceeded 60%. A similar threshold might be applied to sites on poorly drained gleys but the level of replication is too low for this decision. It is important to point out that not all sites within the high forest-cover bands had low minimum pH. It may be that the sampling did not capture the peak of the acidity or alternatively the sites are buffered against pH change. Indeed, when alkalinity was examined, many sites, particularly those on peat/podzolic/lithosolic on sedimentary geology, had alkalinity values well above 20 mg/l CaCO3. Further analyses of 57 of these sites confirmed that 78% have some alkaline sub-soils or carbonate geology within the catchment. Interestingly, a number of sites with maximum alkalinity >20 mg/l CaCO3 recoded high flow alkalinity values of close to zero. This was also mirrored in the SDI results for these sites. Evaluation of the differences in the flow pathway between base and flood conditions in forested catchments is clearly required to better understand factors controlling buffering potential. Overall, the greatest variation in alkalinity was recorded on sedimentary geology which may relate to more complex geology with, as already mentioned, occurrences of some carbonate soils or rocks among the largely acid-sensitive geology. More detailed spatial and temporal analyses of the chemical characteristics of waters draining sedimentary geology is required for more precise mapping of acid sensitivity and this should be an element of future research projects. The presence of forestry tended to depress site pH and alkalinity. Calculations suggested that dilution makes a variable contribution to loss of alkalinity and in many cases the forested sites showed a slightly higher % value. Anion titration was detected in all events examined. The principal contributors were organic acids and sulphate. Excess sulphate only made a contribution in the Wicklow sites and at one site in Galway. The contribution of nitrate across all sites was insignificant. The contribution of sea salts to acidification was low and only one significant sea-salt event was detected at one site in Galway. Similar variability in contributing variables has been reported by Kowalik et al. (2007). Overall, it is likely that a combination of dilution and higher organic acidity concentrations, and occasionally excess sulphate, contributed mostly to the differences in acidity between control and heavily afforested catchments. Reasons for the differences in organic acidity are unclear and may relate to the effects of patterns of drying and wetting and other climatic factors associated with forest soils (Raveh and Avnimelech, 1978; Worrall, Burt and Adamson, 2004). This 68 represents another key knowledge gap and if addressed may help to develop focused forestry practices that minimize the risk of acidification. Indeed, given that the focus of the current study was on mature forestry, we need to determine the acidification risk associated with each of the key forestry practices from site preparation to felling. In terms of the macroinvertebrtae analyses the control sites draining sedimentary sites were more productive in terms of biomass than the igneous/metamorphic sites but the ranges for taxon richness were similar. The higher total macroinvertebrate abundances at the sedimentary sites could be largely attributed to the Ephemeroptera and Chironomidae. This may relate to slightly higher pH and cation concentrations in the former. In both geological settings the Ephemeroptera were reduced in abundance at sites in the two highest forest cover bands, a factor of increasing pH. The sensitivity of the Ephemeroptera to acidification is well established through field observations as well as stream microcosm experiments (e.g. Courtney &Clements, 1998) Interestingly, at the igneous/metamorphic sites the reduction in ephemeropteran abundance was largely balanced by an increase in the numbers of Plecoptera. This did not occur at the sites draining sedimentary geology and consequently overall abundance declined gradually across the forest cover bands. Most the analyses in relation to forestry effects was performed separately for the two geological settings to avoid any confounding effects of differences in taxon abundances. Overall, the biological data largely mirrored the trends for the acidity variables. In fact pH was the key variable structuring the community, a feature that is commonly reported in the literature (Ormerod and Edwards, 2006; Sutcliffe and Carrick, 1973). The study on the source communities illustrated two key findings. Firstly, the macroinvertebrate communities found in control source streams differ substantially from those at sites located further downstream and therefore one must be cautious about making comparisons between sites upstream and downstream of forestry if source sites are included. Secondly, the study showed that the macroinvertebrates communities downstream of the forest were more similar to the source sites than the downstream counterpart on the non-forested stream. This implies some impact of forestry inputs, most likely relating to acidity. Several metrics (taxon richness, ephemeropteran richness, abundance of baetids, EPT richness, diversity indices), which showed a strong relationship with pH, were also shown to vary significantly across the forest cover bands or to correlate with % forest cover bands. 69 Ephemeroptera were absent from several sites in the >50 forest bands on peat and podzolic/lithosolic soils. A striking finding was that the number of sites with low ephemeropteran richness and abundance increased across the forest cover bands. The analyses on the individual metrics highlighted similar % forest thresholds for risk of impact as described for the hydrochemistry. When a selection of non-correlated metrics were combined it was clear that a large proportion of sites in the >50% cover band, and a smaller number of the 25-50% band, had some degree of impairment. These same sites were shown to have a different invertebrate community to the control sites as indicated on the NMDS plots. The implications of the detected impairment for overall ecological health and functioning requires further research. It should be highlighted that, as for pH, not all sites within the high forest cover bands showed impact. Most of those that were impacted recorded minimum pH values well below 5.0 and alkalinity values below or close to zero but this was not consistent across the soil/geological categories. In fact, some of the non-impacted sites also became substantially acidic. Further research must target these sites to better understand the mechanisms governing responses to acid impact under naturally acidic conditions. It is likely that both the degree and duration of acidity are important factors. Some limited analyses of the season and longitudinal extent of forest impact was undertaken. The results suggest that impact may be seasonal and that recovery in some of the metrics (e.g. baetid numbers) takes places. The potential for seasonal recovery may be dependent on climatic factors viz. severity, duration and frequency of precipitation leading to acid pulses; but also life history patterns of the biota (e.g. baetids). The limited data also suggest that recovery may occur over a shorter distance on sedimentary geology than on igneous/metamorphic geology but this is likely to be controlled by many interacting factors that change with distance from the source, such as catchment size, forest cover, inputs from other sub-catchments as well as geology/soils. The longitudinal responses to forestry as a land-use activity needs to be addressed by further research. The fish analyses was limited to 19 paired sites with similar habitat but did highlight significant differences in fish catch and density between the control and forested groups. This difference was mainly attributed to low numbers of fry (salmon and trout) in the forested streams. Low recruitment in forested streams is most likely to be related to pH as previously 70 highlighted by Kelly-Quinn, Tierney and Bracken (1993) in upland streams in Co. Wicklow. However, there may be other contributing factors such as discharge and availability of food. In conclusion this study has addressed the objectives as set out in the introduction. It has detected increases in acidity and biological impairment associated with forest plantations. The risk appears to be a factor of soil geology/soil and increasing catchment cover with the greatest impact occurring above a coniferous forest cover of 50%. The sources and pathways of acid inputs needs to be better clarified and related to forest activities to allow further refinement of the programme of measures. 5. REFERENCES Allott, N., Mills, W.R.P., Dick, J.R.W., Eacrett, A.M., Brennan, M.T., Clandillon, S., Phillips, W.E.A., Critchley, M., and Mullins, T.E. (1990) Acidification of Surface Waters in Connemara and South Mayo. Current Status and Causes, Dublin: duQuesne Limited, Economic and Environmental Consultants, pp 61. Allott, N., Brennan, M., Cooke, D., Reynolds, J., and Simon, N. (1997) Stream chemistry, hydrology and biota, Galway-Mayo region. In: A study of the effects of stream hydrology and water chemistry in forested catchments on fish and macroinvertebrates, AQUAFOR Report 4, COFORD, Dublin, Ireland. Anon. (2005) The characterisation and analysis of Ireland’s river basin districts in accordance with Section 7(2 & 3) of the European Communities (Water Policy) Regulations 2003 (SI 722 of 2003). National Summary Report (Ireland). Arnason, F., Antonsson, T. and Einarsson, S.M. (2005) Evaluation of single-pass electric fishing to detect changes in population size of Atlantic salmon (Salmo salar L.) juveniles. Icelandic Agricultural Sciences, 18: 67-73. Barbour, M.T., Gerritsen, J. and Snyder, B.D. (1997) Revision to Rapid Bioassessment Protocols for Use in Streams and Rivers. EPA 841-D-97-002, Office of Water, United States Environmental Protection Agency, Washington, D.C., USA. Bertrand, K.N. Gido, K.B. and Guy, C.S. (2006) An evaluation of single-pass versus multiple-pass backpack electrofishing to estimate trends in species abundance and richness in prairie streams. Transactions of the Kansas Academy of Science, 109: 131138. 71 Bowman, J. (1991) Acid sensitive surface waters in Ireland. Environmental Research Unit, Dublin. pp 321. Clenaghan, C., Giller, P.S., O'Halloran, J., and Hernan, R. (1998) Stream macroinvertebrate communities in a conifer-afforested catchment in Ireland: relationships to physicochemical and biotic factors. Freshwater Biology 40: 175-193. Courtney, L.A. and Clements, W.H. (1998) Effects of acidic pH on benthic macroinvertebrate communities in stream microcosms. Hydrobiologia 379: 135-145. Economou, A.N., Schmutz, S., Melchior, A. and D. Heidvogel (2002) Development, evaluation & implementation of a standardised fish-based assessment method for the ecological status of European rivers: A contribution to the Water Framework Directive (FAME): Defining reference conditions (D3), Final Report. FAME, Austria. Evans, C.D., Reynolds, B., Hinton, C., Hughes, S., Norris, D., Grant, S. and Williams, B. (2008) Effects of decreasing acid deposition and climate change on acid extremes in an upland stream. Hydrology and Earth System Sciences 12: 337-351. Farrell, E.P. (1995) Atmospheric deposition in maritime environments and its impact on terrestrial ecosystems. Water, Air, & Soil Pollution, 85: 123-130. Giller, P., O'Halloran, J., Kiely, G., Evans, J., Clenaghan, C., Hernan, R., Roche, N., and Morris, P. (1997) An Evaluation of the Effects of Forestry on Surface Water Quality and Ecology in Munster. In A study of the effects of stream hydrology and water quality in forested catchments on fish and invertebrates AQUAFOR Report 2, COFORD, Dublin, Ireland: COFORD, pp 77. Harriman, R., and Morrison, B.R.S. (1982) Ecology of streams draining forested and nonforested catchments in an area of central Scotland subject to acid precipitation. Hydrobiologia, 88: 251-263. Harriman, R., Anderson, H., and Miller, J.D. (1995) The role of sea-salts in enhancing and mitigating surface water acidity. Water, Air, & Soil Pollution, 85: 553-558. Hartigan, J.A., Wong, M.A., (1979) Algorithm AS 139: A k-means clustering algorithm. Applied Statistics 28: 100-108. Hayes. J.W. and Baird, D.B. (1994) Estimating relative abundance of juvenile brown trout in rivers by underwater census and electrofishing. New Zealand Journal of Marine and Freshwater Research, 28: 243-253. Hornung, M., Le-Grice, S., Brown, N., and Hughes, S. (1990) The role of geology and soils in controlling surface water acidity in Wales. In Acid Waters in Wales (R.W. 72 Edwards, J.H. Stoner, and A.S. Gee, eds.), Dordrecht: Kluwer Academic Publishers, pp 55-66. Hornung, M., Bull, K.R., Cresser, M., Ullyett, J., Hall, J.R., Langan, S., Loveland, P.J., and Wilson, M.J. (1995) The sensitivity of surface waters of Great Britain to acidification predicted from catchment characteristics. Environmental Pollution, 87: 207-214. Jenkins, A., Cosby, B.J., Ferrier, R., Walker, T.A.B., and Miller, J.D. (1990) Modelling stream acidification in afforested catchments. An assessment of the relative effects of acid deposition and afforestation. Journal of Hydrology, 120: 163-181. Jones, M.L. and Stockwell, J.D. (1995) A rapid assessment procedure for the enumeration of salmonine populations in streams. North America Journal of Fisheries Management, 15: 551-562. Johnson, M.J., Giller, P.S., O’Halloran, J., O’Gorman, K. and Gallagher, M.B. (2005) A novel approach to assess the impact of landuse activity on chemical and biological parameters in river catchments. Freshwater Biology, 50: 1273–1289. Kahl, J.S., Norton, S.A., Haines, T.A., Rochette, E.A., Heath, R.H. and Nodvin, S.C. (1992) Mechanisms of episodic acidification in low-order streams in Maine, USA. Environmental Pollution 78: 37-44. Kelly-Quinn, Tierney, D. and Bracken, J.J. (1993) Survival of salmon eggs planted in upland streams. Aquaculture and Fisheries Management 24: 791-796. Kelly-Quinn, M., Tierney, D., Coyle, C. and Bracken, J.J. (1996a) Factors affecting the susceptibility of Irish soft-water streams to forest-mediated acidification. Fisheries Management and Ecology 3, 287-301. Kelly-Quinn, M., Tierney, D., Roche, W. and Bracken, J.J. (1996b) Distribution and abundance of trout populations in moorland and afforested upland nursery streams in Co. Wicklow. Biology and Environment, Proceeding of the Royal Irish Academy. 96B, No. 3 127-140. Kelly-Quinn, M., Tierney, D. and Bracken J.J. (1997) Stream Chemistry, Hydrology and Biota, Wicklow Region. In: A study of the effects of stream hydrology and water quality in forested catchments on fish and invertebrates. AQUAFOR Report 3, COFORD, Dublin. 92pp. ISBN:0 9523938 4 0 Kowalik, R.A., Cooper, D.M., Evans, C.D. and Ormerod, S.J. (2007) Acidic episodes retard the biological recovery of upland British streams from chronic acidification. Global Change Biology 13: 2439-2452. 73 Kruse, C.G., Hubert, W.A. and Rahel, F.J. (1998) Single-pass electrofishing predicts trout abundance in mountain streams with sparse habitat. North American Journal of Fisheries Management, 18: 940-946. Mitro, M.G. and Zale, A.V. (2000) Predicting fish abundance using single-pass removal sampling. Canadian Journal of Fisheries and Aquatic Sciences, 57: 951-961. Ormerod, S.J., Wade, K.R., and Gee, A.S. (1987) Macrofloral assemblages in upland Welsh streams in relation to acidity, and their importance to invertebrates. Freshwater Biology, 18: 545–557. Ormerod, S.J., and Wade, K.R. (1990) The role of acidity in the ecology of Welsh lakes and streams. In Acid Waters in Wales (R.W. Edwards, J.H. Stoner, and A.S. Gee, eds.), Dordrecht: Kluwer Academic Publishers, pp 95-109. Ormerod, S.J., Rutt, G., Weatherley, N.S., and Wade, K. (1991) Detecting and managing the influence of forestry on river systems in Wales: Results from surveys, experiments and models. In: Irish Rivers: Biology and Management (M.W. Steer, ed.) Royal Irish Academy, Dublin. pp 163-184. Ormerod, S.J. and Edwards, R.W. (2006) The ordination and classification of macroinvertebrate assemblages in the catchment of the River Wye in relation to environmental factors. Freshwater Biology, 17: 533-546. Raveh, A. and Avnimelech, Y. (1978) The effect of drying on the colloidal properties and stability of humic compounds. Plant Soil, 50: 545-552. Rees, R.M., and Ribbens, J.C.H. (1995) Relationships between afforestation, water chemistry and fish stocks in an upland catchment in south west Scotland. Water, Air, & Soil Pollution, 85: 303-308. Resh, V.H., Brown, A.V., Covich, A.P., Gurtz, M.E., Li, H.W., Minshall, G.W., Reice, S.R., Sheldon, A.L., Wallace, J.B. and Wissmar, R.C. (1988) The role of disturbance in stream ecology. Journal of the North American Benthological, 7: 433-455. Stoner, J., Gee, A., and Wade, K. (1984) The effects of acidification on the ecology of streams in the upper Twyi catchment in West Wales. Environmental Pollution, A, 36: 125-157. Sutcliffe, D.W. and Carrick, T.R. (1973) Studies on mountain streams in the English Lake District I. pH, calcium and the distribution of invertebrates in the River Duddon, Freshwater Biology, 3:, 437-462. 74 Tierney, D., Kelly-Quinn, M., and Bracken, J.J. (1998) The faunal communities of upland streams in the eastern region of Ireland with reference to afforestation impacts. Hydrobiologia, 389: 115-130. Waters, D., and Jenkins, A. (1992) Impacts of afforestation on water quality trends in two catchments in mid-Wales. Environmental Pollution, 77: 167-172. Worrall, F., Burt, T. and Adamson, J. (2004) Can climate change explain increases in DOC flux from upland peat catchments? Science of the Total Environment, 326: 95-112. Wright, J.F. (1995) Development and use of a system for predicting the macroinvertebrate fauna in flowing waters. Australian Journal of Ecology, 20: 181-197. 75 APPENDIX A Location of the 239 sampling sites 76 Main system Trib of River Blackwater Caher River Dripsey River Dripsey River Trib of Caher River Glennaharee River Glengarriff Stream Fermoyle River Trib of River Blackwater Ownagluggin River Ownagluggin River Ownagluggin River Ownagluggin River Carrigduff River Carrigduff River Crinnaloo River Aghalode River Trib of River Laney Trib of River Blackwater Bregoge River Castlepook River Trib of Bregoge River Fluckane Stream Trib of River Funshion Trib of Sheep River Garrane River Trib of Ross River Trib of River Bride Bunnaglanna River Ballycorban River Trib of Scarriff River Bow River Bow River Corlea River Trib into Lough Atorick Muchnagh Douglas Muchnagh Araglin Araglin Araglin Geeragh River Geeragh River Burncourt River Trib Araglin Glenakeefe Glennandaree Sheep Invertebrate Site Code CK1 CK10 CK12 CK13 CK14 CK15 CK17 CK19 CK2 CK21 CK22 CK23 CK24 CK25 CK26 CK27 CK28 CK29 CK3 CK30 CK31 CK32 CK33 CK34 CK35 CK36 CK4 CK6 CK8 CL2 CL3 CL4 CL5 CL6 CL8 MUCH1 DOUG3 MUCH2 ARAG1 ARAG2 ARAG3 GEER1 GEER2 BURN1 ARA1 GKEEF1 GLENN1 SHEEP1 Water Chemistry Site Code CK1 CK10 CK12 CK13 CK14 CK15 CK17 CK19 CK2 CK21 CK22 CK23 CK24 CK25 CK26 CK27 CK28 CK29 CK3 CK30 CK31 CK32 CK33 CK34 CK35 CK36 CK4 CK6 CK8 CL2 CL3 CL4 CL5 CL6 CL8 DC1 DC10 DC12 DC13 DC14 DC15 DC16 DC17 DC18 DC2 DC4 DC5 DC6 County Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Cork Clare Clare Clare Clare Clare Clare Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Cork/Tipperary Grid Reference W 717 958 W 452 863 W 416 857 W 415 858 W 444 869 W 459 889 W 454 922 W 394 919 W 671 973 W 370 872 W 377 875 W 377 875 W 384 877 W 357 884 W 354 883 W 370 892 W 384 851 W 353 855 W 666 970 R 595 134 R 606 137 R 620 133 R 631 131 R 693 140 R 720 149 R 594 179 W 647 968 W 740 926 W 709 928 R 635 891 R 639 891 R 663 904 R 669 917 R 616 938 R 641 940 R 882 071 R838 057 R 868 069 S 007 066 S 006 067 S 006 068 R 824 178 R 829 193 R 937 192 R 904 062 S 062 060 S 040 077 R 910 178 Easting 171785 145243 141683 141503 144463 145995 145465 139459 167196 137010 137788 137749 138438 135710 135486 137044 138427 135399 166688 159550 160600 162013 163110 169333 172068 159451 164725 174075 170926 163510 163978 166353 166934 161652 164190 188278 183827 186850 200701 200602 200662 182482 182988 193779 190438 206269 204056 191095 Northing 95834 86306 85758 85862 86909 88996 92232 91961 97330 87248 87521 87501 87701 88412 88331 89243 85148 85576 97055 113454 113785 113334 113186 114043 114983 117961 96815 92612 92860 189122 189115 190427 191795 193837 194005 107108 105741 106913 106648 106743 106800 117850 119382 119290 106297 106014 107794 117840 77 Main system Sheep Sheep Douglas Gweebarra River Gweebarra River Elatagh River Elatagh River Elatagh River Elatagh River Trib Deele River Stranagoppoge Gweebarra River Gweebarra River Sruhanboy River Trib of Strachashell River Cloghroe River Stranagoppoge Trib Owenmore Trib. Owenree Owenboliska Owenboliska Owenboliska Owenboliska Owenboliska Lough More Owenboliska Sruffaunanulra River Trib Bunowen River Trib Lough Corrib Owenwee River Owenwee River Owenwee River Trib to Maumwee Lough Trib to Owenriff Glengawbeg River Glengawbeg River Trib Owenmore Owenakilla River Gowlaun River Trib Owenwee Trib Owenwee Loughanillaunmore Owenboliska Owendunnakilla Knockbane river River Loo Clydagh Clydagh Clydagh Clydagh Invertebrate Site Code SHEEP3 SHEEP2 DOUG2 GBAR1 GBAR3 ELATA2 ELATA4 ELATA1 ELAT3 DEEL1 STRAN2 GBAR2 GBAR4 SRUHA1 STRAC1 CROE1 STRAN1 OWEN2 OREE1 OLISKA2 Oliska1 OLISKA7 OLISKA6 OLISKA5 MORE1 OLISKA4 SRUFF1 BOWEN1 CORRIB1 OWEE1 OWEE2 OWEE3 MAUM1 ORIFF2 GBEG2 GBEG3 OWEN3 OWENK1 GLAUN1 OWENN2 OWENN1 LOUGH1 OLISKA3 OKILLA1 KBANE1 LOO1 CLYDA9 CLYDA10 CLYDA6 CLYDA1 Water Chemistry Site Code DC7 DC8 DC9 DD10 DD11 DD13 DD14 DD15 DD16 DD2 DD3 DD4 DD5 DD6 DD7 DD8 DD9 DG1 DG11 DG12 DG13 DG14 DG15/G7 DG16 DG17 DG18 DG19 DG2 DG20 DG21 DG22 DG23 DG24 DG25 DG28 DG29 DG3 DG30 DG31 DG4 DG5 DG6 DG7/G6 DG8 DG9 DK1 DK11 DK12 DK13 DK14 County Cork/Tipperary Cork/Tipperary Cork/Tipperary Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Donegal Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Galway Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Grid Reference R 906 201 R 893 203 R 850 056 B 857 044 B 861 026 C 043 039 C 045 041 C 043 039 C 022 052 C 112 031 G 912 976 B 854 038 B 840 005 C 048 017 G887 965 C102 009 G 924 992 L 913 731 M 015 468 M 142 360 M 145 355 M 084 322 M 085 327 M 080 332 M 068 310 M 104 345 M 090 379 L 837 757 M 057 485 M 025 452 M 032 455 M 033 458 L 973 484 M 083 423 M 056 409 M 053 410 L 929 728 M 097 465 M 089 471 L 951 771 L 945 771 M 098 283 M 112 345 M 165 364 M 171 353 W 048 796 W 176 843 W165 842 W 159 829 W 183 833 Easting 190631 189384 185087 185721 186163 204371 204500 204307 202224 211241 191262 185497 184082 204803 188777 210272 192448 91317 101571 114238 114582 108464 108519 108003 106835 110492 109007 83779 105770 102550 103259 103301 97370 108363 105672 105371 92970 109748 108947 95168 94573 109823 111290 116552 117126 104813 117667 116554 115954 118389 Northing 120168 120327 105620 404452 402664 403953 404109 403947 405222 403177 397661 403876 400536 401716 396441 400935 399263 273196 246870 236010 235506 232252 232725 233271 231025 234516 237978 275767 248510 245210 245508 245877 248412 242381 240995 241083 272821 246534 247150 277166 277143 228341 234524 236440 235361 79671 84397 84248 82944 83368 78 Main system Clydagh Roughty River Roughty River Roughty River Clydagh River Loo Clydagh Owgarriv River Bohill River Clydagh Foherish River Inchamore River Aughboy River Slievenaneav River Roughty River Trib to Flesk River Roughty River Inchamore River Kealgorm Kealgorm Trib to Slaheny River Garrrow River Glenthomas River Glennamong River Trib Srahmore River Trib Skerdagh River Trib Crumpaun River Trib Crumpaun River Fiddaungrave Fiddaungal Glennamong River Glenthomas River Glendahurk River Glendahurk River Glendahurk River Fiddaunatoreen Glennamong River Glennamong River Vartry Derrybaun River Annalecka Brook Glashaboy Garryknock Oiltiagh Brook Inchavore River Ballinagee River Vartry Lugduff Brook Lugduff Brook Invertebrate Site Code CLYDA2 ROUGH3 ROUGH1 ROUGH2 CLYDA3 LOO2 CLYDA8 OWGAR1 BOHIL1 CLYDA7 FOHER1 INCHMR1 AUGHB1 SLIEVE1 ROUGH5 FLESK1 ROUGH4 INCHMR2 KEAL1 KGORM1 SLAH1 GARW1 GTHOM1 GAMON2 SRAH1 SKERD1 CRUM2 CRUM1 FIDD2 FIDD1 GAMON3 GTHOM2 GHURK3 GHURK1 GHURK2 FREEN1 GAMON1 GAMON4 VAR1 DERRY1 ANNA1 GLASH1 GARRY1 OILI1 INCH1 BALLIN2 VART1 LUG3 LUG1 Water Chemistry Site Code DK15 DK16 DK17 DK18 DK19 DK2 DK20 DK22 DK23 DK24 DK26 DK27 DK28 DK29 DK3 DK30 DK31 DK32 DK4 DK5 DK6 DK8 DM1 DM10 DM11 DM12 DM13 DM14 DM15 DM16 DM17 DM2 DM3 DM4 DM5 DM6 DM7 DM8 DWW10 DWW12 DWW13 DWW15 DWW16 DWW17 DWW19 DWW2 DWW20 DWW21 DWW22 County Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Cork/Kerry Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Grid Reference W 222 845 W 068 751 W 072 712 W 065 710 W 210 865 W 045 782 W 201 844 W 100 800 W 195 806 W 206 845 W 241 807 W117 775 W 125 783 W 089 800 W 038 743 W 102 842 W 097 730 W 120 773 W 012 771 W 007 782 W 029 700 W 081 753 L 889 999 F 928 038 F 965 054 G 012 023 G 046 021 G 073 047 G 061 069 G 055 076 F 941 037 F 887 003 F 910 009 L 909 985 F 912 007 F 950 019 F 944 027 F938 304 O 204 092 T 133 946 O 067 026 O 053 013 O 026 022 S 991 958 O 110 060 O 047 047 O 190 069 T 110 955 T 111 957 Easting 122212 106817 107266 106572 121066 104554 120194 110094 119594 120620 124186 111751 115953 108969 103836 110254 109706 112095 101263 100724 102958 108180 88932 92869 96560 101267 104638 107325 106136 105540 94103 88783 91083 90982 91275 95086 94493 93899 320426 313316 306721 306599 302636 299141 311100 304702 319044 311096 311123 Northing 84523 75194 71289 71000 86565 78231 84468 80028 80696 84561 80767 77576 78398 80033 74388 84282 73061 77334 77185 78269 70085 75351 299925 303830 305240 302346 302162 304771 306948 307619 303724 300376 300953 298562 300780 301904 302799 304237 209227 194649 202647 201610 202250 195835 206200 204744 206912 195572 195749 79 Main system Sraghoe Brook Glashaboy Annalecka Brook Knickeen River Slaney River Cloghoge River Cloghoge River Trib of Owendalulleegh River * Trib of Boleyneendorrish River Trib of Boleyneendorrish River Trib into Derryclare Lough Trib into Derryclare Lough Trib of Owenglin River Trib of Owendalulleegh River Owenaglanna River * Owendunnakilla Trib of Owenboliska River * Trib of Owenboliska River Trib into Seecon Lough Trib of Smearlagh River Dromaddamore River Trib of Smearlagh River Barranahown River Trib of Ahaphuca River Trib of Awbeg River Trib of Awbeg River Trib of Awbeg River Trib of Awbeg River Trib of Assaroola River Trib of River Loobagh Trib of Keale River Trib of River Ogeen Trib of River Ogeen Trib of Keale River Trib of River Loobagh Barranahown River Delour River Trib of River Barrow Trib of Delour River Trib of Delour River Trib of Delour River Delour River Delour River Invertebrate Site Code SRAG1 GLASH2 ANNA3 KNICK1 SLAN1 CLOG1 CLOG2 Water Chemistry Site Code DWW23(f) DWW26 DWW4 DWW6 DWW7 DWW8 DWW9 County Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Wicklow Grid Reference O 097 135 O 065 016 O 065 033 S 998 952 S 995 937 O 130 074 O 126 076 Easting 309741 306500 306512 299854 299593 313029 312627 Northing 213564 201700 203325 195214 193771 207418 207642 G1 G11 G1 G11 Galway Galway R 627 996 M 548 106 162760 154844 199671 210646 G12 G12 Galway M 565 052 156586 205222 G13 G15 G16 G18 G13 G15 G16 G18 Galway Galway Galway Galway M 565 052 L 830 498 L 827 493 L 740 513 156556 83047 82718 74043 205208 249800 249356 251332 G2 G3 G4 G5 G6 G7 G8 G9 K1 K2 K3 L10 L12 L13 L14 L15 L16 L17 L2 L3 L4 L5 L6 L8 L9 LS1 LS10 LS12 LS13 LS14 LS2 LS3 G2 G3 G4 G5 G6 G7 G8 G9 K1 K2 K3 L10 L12 L13 L14 L15 L16 L17 L2 L3 L4 L5 L6 L8 L9 LS1 LS10 LS12 LS13 LS14 LS2 LS3 Galway Galway Galway Galway Galway Galway Galway Galway Kerry Kerry Kerry Longford Longford Longford Longford Longford Longford Longford Longford Longford Longford Longford Longford Longford Longford Laois Laois Laois Laois Laois Laois Laois M 566 015 M 612 065 M 563 105 M 160 363 M 112 350 M 085 327 M 099 328 M 086 359 Q 969 170 Q 982 182 Q 990 204 R 702 244 R 721 237 R 618 188 R 616 187 R 597 190 R 597 190 R 817 227 R 638 204 R 650 186 R 638 168 R 646 171 R 660 175 R 694 217 R 684 242 N 285 029 N 331 054 S 246 970 S 237 966 S 225 967 N 281 031 N 285 032 156641 161215 156391 116027 111216 108534 109902 108638 96915 98243 99014 170264 172168 161896 161672 159776 159757 181764 163838 165078 163859 164626 166024 169474 168480 228597 233181 224699 223754 222549 228188 228549 201559 206594 210551 236399 235073 232725 232848 235904 117058 118241 120483 124474 123782 118844 118760 119073 119065 122760 120431 118651 116884 117113 117505 121787 124258 202935 205441 197009 196687 196740 203180 203287 80 Main system Delour River Delour River Delour River Trib of Mountrath River Trib of River Barrow Trib of Glenummera River Trib of Glenora River Glennafrankagh Trib of Altderg River Coolin River Sruffaunmuinganierin Trib of Glenummera River * Trib of Glencullin River Trib of Glencullin River * Trib of Glenamoy River Trib of Glenamoy River Bellanaminnaun River Trib of Sheep River Trib of River Aherlow Trib of River Aherlow Trib of River Aherlow Trib of River Aherlow Trib of River Aherlow Trib of River Aherlow Trib of River Aherlow Clydagh River Clydagh River Trib of River Aherlow Trib of Sheep River Trib of River Aherlow Trib of Burncourt River Trib of Burncourt River Trib of Burncourt River Trib of Burncourt River * Trib of River Suir Trib of River Aherlow Trib of Licky River Trib of Licky River Trib of Licky River Trib of Licky River Trib of Goish River Trib of Goish River Goish River Invertebrate Site Code LS4 LS5 LS6 LS7 LS8 M1 M10 M11 M12 M13 M14 M2 M3 M4 M5 M6 M7 M8 M9 T1 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T2 T20 T3 T4 T5 T6 T7 T8 T9 W1 W2 W4 W6 W7 W8 W9 Water Chemistry Site Code LS4 LS5 LS6 LS7 LS8 M1 M10 M11 M12 M13 M14 M2 M3 M4 M5 M6 M7 M8 M9 T1 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T2 T20 T3 T4 T5 T6 T7 T8 T9 W1 W2 W4 W6 W7 W8 W9 County Laois Laois Laois Laois Laois Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Mayo Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Tipperary Wexford Wexford Wexford Wexford Wexford Wexford Wexford Grid Reference N 276 038 N 297 033 N 295 019 S 348 997 N 366 079 L 905 676 G 046 339 G 028 347 G 010 322 G 047 283 G 072 288 L 896 674 F 855 178 F 911 255 F 898 261 F 878 309 F 909 331 F 951 358 G 066 365 R 893 204 S 016 279 S 002 280 R 993 279 R 988 280 R 979 284 R 948 263 R 945 263 R 887 255 R 884 253 R 901 279 R 906 201 R 891 279 R 907 220 R 908 220 R 922 205 R 955 200 R 991 247 S 029 270 S 017 279 X 212 855 X 224 878 X 174 848 X 190 859 X 170 885 X 159 904 X 171 894 Easting 227675 229712 229594 234839 236680 90547 104608 102811 101088 104744 107280 089611 085545 091155 089800 087822 090976 95114 106695 189359 201647 200240 199374 198875 197987 194857 194527 188710 188495 190180 190631 189140 190702 190815 192276 195582 199159 202978 201719 221260 222457 217494 219066 217047 215946 217105 Northing 203807 203334 201996 199782 207933 267687 333981 334720 332253 328324 328874 267419 317826 325560 326190 330950 333125 335885 336507 120401 127960 128048 127905 128086 128490 126398 126387 125536 125348 127915 120157 127993 122059 122056 120535 120028 124754 127038 127950 85579 87880 84858 85973 88586 90467 89485 81