Contract EVK1-1999-00087 - RECOVER:2010

Part of the 'Sustainable Management and Quality of Water'

Ecosystem Functioning

Directorate General Research

Macaulay Land Use Research Institute logo CEH Logo BIOTEK Logo NIVA Logo CNR Logo Czech Logo IIASA Logo University of Virginia IVL Logo



Long-term changes in the water quality of rainfall, cloud water and stream water for moorland, forested and clear-felled catchments at Plynlimon, mid-Wales

C. Neal1, B. Reynolds2, M. Neal1, B. Pugh2, L. Hill1 and H. Wickham1
1Centre for Ecology and Hydrology, Wallingford, Oxon OX10 8BB,UK
2Centre for Ecology and Hydrology, Bangor, Gwynedd, LL57 2UP,UK

Full Reference

Neal, C., Reynolds, B., Neal, M., Pugh, B., Hill, L. and Wickham, H. (2001). Long-term changes in the water quality of rainfall, cloud water and stream water for moorland, forested and clear-felled catchments at Plynlimon, mid-Wales. Hydrology and Earth System Sciences Vol. 5, No. 3, 459-476.

Summary of Research

Long term changes in the water quality of rainfall, cloud water and stream waters draining acidic and acid sensitive moorland and forested catchments at Plynlimon, mid Wales, are examined for the period 1983 to 2001. Atmospheric inputs of chloride and sulphate are influenced by the relative inputs of clean maritime and polluted land based air masses. There is no systematic increase or decrease over time for chloride and non-sea-salt sulphate. Rather, there is a decadal scale process possibly representative of the influence of the North Atlantic Oscillation that affects the maritime and pollution climate of the Atlantic seaboard of the UK. Over 17 years of study, there may be a small decrease in non-sea-salt sulphate of about 10 µeq l-1 and a small improvement in acid neutralising capacity of about 20 to 30 µeq l-1 in rainfall. There is a clear improvement in cloud water chemistry with respect to pollutant components (ammonium, nitrate, non-sea-salt sulphate) and acidity (acid neutralising capacity improved by about 300 µeq l-1) through the study period. Many of the changes in cloud water chemistry are similar to rainfall over the same period except the magnitude of change is larger for the cloud water. Within the streams, there is some evidence for reductions in acidity as reflected by acid neutralising capacity becoming less negative. For one stream, deforestation occurred during the sampling period and this led to large increased in nitrate and smaller increases in aluminium midway through the study period. However, the climate and hydrological variability largely masked out other changes. The current analysis provides only a start to identifying trends for such a complex and variable environmental system. The need for strong statistical tools is emphasised to resolve issues of: (a) hydrological induced water quality variability, (b) changing soil and groundwater "endmember" chemistry contribution to the stream and (c) the non-linear patterns of change. Nonetheless, the analysis is enhanced by examining trends in chemistry for yearly averages and yearly average low catch and high catch rainfall and cloud water events as well as low and high flow stream chemistry. This approach allows trends to be examined within the context of endmember mixing.

Fig.1. The Plynlimon catchments.The study relates to two main tributaries of the upper River Severn, the Afon Hafren and the Afon Hore, that drain the moorland areas of the plateau top and forested southeastern slopes of Pumlumon Fawr in mid-Wales (Figure 1). The bedrock geology comprises lower Palaeozoic mudstones, greywackes and sandstones that are overlain by thin (typically 70 cm thick) acidic soils comprised mainly of peaty podzols with some deep peats, brown earths and gleys. On the upper parts of the Hafren and Hore catchments, vegetation is dominated by acid grassland and acid heath land. The Hafren Forest, comprising plantation forestry (mainly Sitka spruce Picea sitchensis), introduced in various phases between 1937 and 1964, dominates the lower parts of these catchments. Apart from minor thinning, the standing forest within the Afon Hafren catchment remained intact during the first 8 to 10 years of the study, although in recent years the forest has been progressively harvested. Within the lower half of the Hore catchment, however, the standing crop was clear felled in one operation between spring 1985 and autumn 1989 leaving extensive brash piles and tree stumps to decompose in situ. The slopes were replanted with Sitka spruce seedlings (<0.3 m high) within two years of felling. The short harvesting timescale employed in the lower Hore cannot be considered as standard forestry practice but was an experimental treatment requested by the Forestry Commission to determine the maximum effects of felling on water resources, sediment transport and water quality (Kirby et al., 1991). The altitude range for the upper Severn (from its source to the confluence of the Hafren and the Hore) is about 740 to 330 m.a.s.l. Rainfall averages about 2500 mm yr-1 for the two catchments and evaporation losses (canopy/soil evaporation plus transpiration) are typically between 500 and 800 mm yr-1 (Hudson et al., 1997). Evaporation rates vary from year to year and the moorland parts of the catchment have lower evaporation losses (Hudson et al., 1997). The stream flow response to rainfall is rapid and there is a strong "spiky pattern" to the hydrograph. Being adjacent to each other, the hydrograph responses are very similar with flows varying typically between 0.01 and 4.5 m3 s-1 (Kirby et al., 1991).

Fig.2. Ammonium,NO3,Cl,non-sea-saltSO4,ANC and DOC in rainfall:a time series.
There are clear trends in the data over time for the low-catch samples with increases in concentration from the starting year of 1990 through to about 1995 and a subsequent decline thereafter up to 2001 (Figure 2). For average and high-catch concentrations there seem to be clear tends:

· pH increases by about a 0.3 unit.

· Gran Alkalinity increases by about 70 µeq l-1 for the average catch and there is a much smaller increase for high catch values (the first two lower values for the record may distort the pattern away from no significant change).

· ANC increases by about 300 µeq l-1 for the average concentration through the period with the major change occurring during the last three years of record when NH4 in baseflow declines the most. There is a much smaller trend for the high catch values and the first two lower data points for the record may distort the pattern away from no significant change.

· Sulphate and non-sea-salt sulphate concentrations have decreased by about 200 µeq l-1 for both determinands through the period (an approximate halving of concentrations) with a much smaller trend for high-catch concentrations.

· Chloride concentrations have approximately halved during the study period from about 200-300 and down to about 120 µeq l-1 in line with the rainfall trend over the same period of record. The trends for ANC, SO4, non-sea-salt SO4 and Cl in cloud water are similar to those in rainfall, over the corresponding time period, except that they are much greater in size.

Fig.3.Aluminium,NO3,Cl,non-sea-saltSO4,ANC and DOC in the upper Hafren:a time series.
The streamwaters show a wide range in chemistry (Figures 3 to 6) reflecting the differences between low- and high-flow within-catchment and atmospheric sources. The salient features are as follows. Chloride, sulphate and non-sea-salt sulphate concentrations show much lower fluctuations within the streams with respect to data scatter than the rainfall and cloud water and there is no marked separation between low- and high-flow values. This is expected as described above in relation to the low chemical reactivity of these components within the catchment and the physical dampening of the rainfall signal due to water storage. Chloride provides the dominant anion averaging around 200 µeq l-1 with SO4 and non-sea-salt SO4 concentrations averaging about 78 and 61 µeq l-1, respectively. About 78% of the SO4 is present in non-sea-salt form. For all three determinands, the lowest concentrations occur for the upper Hafren and this reflects differences in the scavenging of sea-salts and pollutants by moorland and forest vegetation (trees with their higher surface area, are more effective at scavenging mist and aerosols relative to shorter vegetation), as well as the larger evaporation losses from the trees.

Fig.4. Aluminium,NO3,Cl,non-sea-saltSO4,ANC and DOC in the Afon Hafren:a time series.

Nitrate shows a strong seasonality reflecting seasonally changing biological activity within the catchment, with peak concentrations during the winter months when uptake from the soil solutions into the growing vegetation is lowest. Average NO3 concentrations in the streams are similar to NO3 concentrations in rainfall at around 20 µeq l-1 with moderately lower values for the upper Hafren and upper Hore. In general, high-flow NO3 concentrations are higher than their baseflow counterparts by up to a factor of two. This links in part to the timing of high-flow events, which predominate during the winter months when NO3 uptake by the vegetation is least as opposed to the low-flow periods during the summer when groundwater inputs from below the soil zone dominate.

Fig.5.Aluminium,NO3,Cl,non-sea-saltSO4,ANC and DOC in the upper Hore:a time series.

Ammonium concentrations are low (typically about 1 µeq l-1) relative to rainfall as the atmospheric input is immobilised within the soil, assimilated into the biomass and/or oxidized to NO3.

Fig.6.Aluminium,NO3,Cl,non seasaltSO4,ANC and DOC in the Afon Hore:a time series.


Hudson, J.A., Crane, S.B. and Blackie, J.R., 1997. The Plynlimon water balance 1969-1995: the impact of forest and moorland vegetation on evaporation and streamflow in upland catchments. Hydrol. Earth Syst. Sci., 1, 743-754.

Kirby, C., Newson, M.D. and Gilman, K., 1991. Plynlimon research: the first two decades. Institute of Hydrology Report No. 109. Institute of Hydrology, Wallingford, OX10 8BB, UK. 188 pp.