ALBERT

Description: Suspended sediment sources in the Upper Severn catchment are quantified using a composite fingerprinting technique combining statistically-verified signatures with a multivariate mixing model. Composite fingerprints are developed from a suite of diagnostic properties comprising trace metal (Fe, Mn, AI), heavy metal (Cu, Zn, Pb, Cr, Co, Ni), base cation (Na, Mg, Ca, K), organic (C, N), radiometric (137Cs, 210Pb), and other (total P) determinands. A numerical mixing model, to compare the fingerprints of contemporary catchment source materials with those of fluvial suspended sediment in transit and those of recent overbank floodplain deposits, provides a means of quantifying present and past sediment sources respectively. Sources are classified in terms of eroding surface soils under different land uses and channel banks. Eroding surface soils are the most important source of the contemporary suspended sediment loads sampled at the Institute of Hydrology flow gauging stations at Plynlimon and at Abermule. The erosion of forest soils, associated with the autumn and winter commercial activities of the Forestry Commission, is particularly evident. Reconstruction of sediment provenance over the recent past using a sediment core from the active river floodpiain at Abermule, in conjunction with a 137Cs chronology, demonstrates the significance of recent phases of afforestation and deforestation for accelerated catchment soil erosion.

Description: In a field experiment to investigate potential impacts of climatic change, a small area of flush wetland in Wales was subjected to three successive years of simulated summer drought/rewetting (autumn-spring) cycles (1992–94). Drought was simulated achieved by diverting stream water around the experimental wetland during the summer, so that the wetland received only precipitation inputs during that time. The effects on peat-water chemistry in the rhizosphere were monitored at regular intervals until spring 1996, and comparisons made with a control. Simulated summer drought decreased, significantly, the natural summer peaks in dissolved organic carbon (DOC) and iron, whilst subsequently increasing the natural autumn-winter peaks in sulphate concentrations and acidity in the peat water. The effects of simulated drought on SO4 concentrations in the peat water compared favourably with subsequent events monitored following a natural summer drought in 1995. Autumn-winter peaks in SO4 concentrations in the control wetland following the natural drought were of similar magnitude to those induced by the drought simulated in the experimental wetland in the previous three years.

Description: A summary of hydrochemical research in the Plynlimon catchments from 1983 to the present is related to identifying water flow pathways within catchments and the modelling of soil and stream water acidification. The study reveals a highly heterogeneous system that barely conforms with current understanding of hydrology and acidification theory. The role of hydrochemical mixing processes and groundwater flow routing is emphasised as is the need for maintaining long term monitoring studies and enhancing process based studies of water and chemical fluxes through catchments. The applicability of current environmental impact models for predictive and environmental management purposes is questioned and it is proposed that new hydrochemical modelling structures are needed to examine the highly heterogeneous systems being modelled.

Description: The concept of potential evaporation is defined on the basis of the following criteria: (i) it must establish an upper limit to the evaporation process in a given environment (the term "environment" including meteorological and surface conditions), and (ii) this upper limit must be readily calculated from measured input data. It is shown that this upper limit is perfectly defined and is given by the Penman equation, applied with the corresponding meteorological data (incoming radiation and air characteristics measured at a reference height) and the appropriate surface characteristics (albedo, roughness length, soil heat flux). Since each surface has its own potential evaporation, a function of its own surface characteristics, it is useful to define a reference potential evaporation as a short green grass completely shading the ground. Although the potential evaporation from a given surface is readily calculated from the Penman equation, its physical significance or interpretation is not so straightforward, because it represents only an idealized situation, not a real one. Potential evaporation is the evaporation from this surface, when saturated and extensive enough to obviate any effect of local advection, under the same meteorological conditions. Due to the feedback effects of evaporation on air characteristics, it does not represent the "real" evaporation (i.e. the evaporation which could be physically observed in the real world) from such an extensive saturated surface in these given meteorological conditions (if this saturated surface were substituted for an unsaturated one previously existing). From a rigorous standpoint, this calculated potential evaporation is not physically observable. Nevertheless, an approximate representation can be given by the evaporation from a limited saturated area, the dimension of which depends on the height of measurement of the air characteristics used as input in the Penman equation. If they are taken at a height of 2 m (the height of the meteorological observations), the dimension of the saturated surface in the direction of the wind ranges roughly from 50 to 200 m for a short green grass completely shading the ground.

Description: The increasing incidence of groundwater pollution has led to recognition of a need to develop objective techniques for designing reniediation schemes. This paper outlines one such possibility for determining how many abstraction/injection wells are required, where they should be located etc., having regard to minimising the overall cost. To that end, an artificial neural network is used in association with a 2-D or 3-D groundwater simulation model to determine the performance of different combinations of abstraction/injection wells. Thereafter, a genetic algorithm is used to identify which of these combinations offers the least-cost solution to achieve the prescribed residual levels of pollutant within whatever timescale is specified. The resultant hybrid algorithm has been shown to be effective for a simplified but nevertheless representative problem; based on the results presented, it is expected the methodology developed will be equally applicable to large-scale, real-world situations.

Description: Three upland Lake District Tarns, Scoat, Greendale and Burnmoor, have been evaluated using MAGIC (Model of Acidification of Groundwater In Catchments) to reconstruct past, present and future chemical behaviour. The modelled historical changes in acidity are compared with palaeoecological estimation of pH to demonstrate model validity. Chemistry as simulated for all anions and cations and two of the three lakes are shown to have undergone significant acidification. The effects of changing atmospheric pollution levels on lake chemistry is evaluated and 80-90% sulphur reduction levels are required to achieve zero alkalinity. The impacts of increased nitrogen deposition are assessed and are shown to further delay reversibility.

Description: Soil moisture gradients along hillslopes in humid watersheds, although indicated by vegetation gradients and by studies using models, have been difficult to confirm empirically. While soil properties and topographic features are the two general physio-graphic factors controlling soil moisture on hillslopes, studies have shown conflicting results regarding which factor is more important. The relative importance of topographic and soil property controls was examined in an upland forested watershed at the Coweeta Hydrologic Laboratory in the southern Appalachian mountains. Soil moisture was measured along a hillslope transect with a mesic-to-xeric forest vegetation gradient over a period spanning precipitation extremes. The hillslope was transect instrumented with a time domain reflectometry (TDR) network at two depths. Soil moisture was measured during a severe autumn drought and subsequent winter precipitation recharge. In the upper soil depth (0-30 cm), moisture gradients persisted throughout the measurement period, and topography exerted dominant control. For the entire root zone (0-90 cm), soil moisture gradients were found only during drought. Control on soil moisture was due to both topography and storage before drought. During and after recharge, variations in soil texture and horizon distribution exerted dominant control on soil moisture content in the root zone (0-90 cm). These results indicate that topographic factors assert more control over hillslope soil moisture during drier periods as drainage progresses, while variations in soil water storage properties are more important during wetter periods. Hillslope soil moisture gradients in southern Appalachian watersheds appear to be restricted to upper soil layers, with deeper hillslope soil moisture gradients occurring only with sufficient drought.

Description: Dambos, seasonally saturated wetlands, are widespread in headwater catchments in southern Africa and play an important role in the regional hydrological cycle. However, the processes influencing runoff from these catchments are poorly understood. This paper reports an isotopic investigation of runoff-generating mechanisms within a Zimbabwean catchment containing a dambo. Hydrograph separation using deuterium reveals that, once the dambo is saturated, up to 70% of total storm flow can be considered "new" water (i.e. derived directly from rainfall generating the runoff event). However, both the total proportion and the instantaneous maximum amount of "new" water in hydrographs are sensitive to rainfall characteristics and antecedent conditions. These results are (1) compatible with observations made in catchments in temperate climates when wetlands are present, and contrast with results obtained when wetlands are absent and (2) consistent with saturation overland flow, generated in saturated regions of the dambo, being the major storm runoff mechanism. To reconcile these observations with past perceptions that dambos attenuate flood flows, a dual role for dambos in storm flow production is postulated.

Description: The use of the nonparametric nearest-neighbour resampling technique is studied for generating time series of daily rainfall and temperature for seven stations in the German part of the Rhine basin. The emphasis is on the reproduction of extreme N-day precipitation amounts. The daily temperatures are used to determine snow accumulation and melt in winter. Two versions of the resampling method, conditional on the atmospheric circulation and unconditional, show comparable results. For precipitation, the autocorrelation properties are well reproduced, whereas for temperature the autocorrelation coefficients are systematically underpredicted. The distributions of the N-day annual maximum precipitation amounts are adequately preserved. Despite the systematic underprediction of the temperature autocorrelation, the distributions of N-day maximum snowmelt are well reproduced. A 1000-year simulation for the seven stations shows that unprecedented rainfall situations can be generated.

Description: By using topographic indices as derived from a Digital Terrain Models (DTM), it is possible to represent the heterogeneity within a landscape. This heterogeneity can reflect both long term evolutionary patterns seen in a landscape and the short term forcing of flow dynamics during storm events. By spatial analysis, the linkage between the geomorphological- hydrological-plant physiological phenomena can be examined. In this study, a direct link will be established between the topographically-driven hydrological phenomena and the eco-physiological response. The topographic distribution function of TOPMODEL is used to control the spatial and temporal flux of the channel flow and water table. The plant physiological model GAS-FLUX is used to give a spatially and temporally dissaggregated species-sensitive estimate of evapotranspiration flux. Evapotranspiration is sensitive to the vegetation phonology, to tundra community physiology and to the temperature regime. A simple linking of TOPMODEL and the GAS-FLUX model is applied to a summer snow-free period to the Imnavait catchment, Alaska (2.2 km2). A species-sensitive evapotranspiration model proved to give the highest quality results when validated against flow observations. Predicted dynamics of variable source area and the component hydrological processes are illustrated.

Description: A modification to the well-known water quality model "Quality Simulation Along River Systems" (QUASAR) is presented, extending its utility to real-time forecasting applications such as the management and control of pollution incidents. Two aggregated dead-zone (ADZ) parameters, namely time delay and dispersive fraction, are incorporated into the existing model formulation, extending the current continuously stirred tank reactor based model processes to account for advective and active mixing volume dispersive processes. The resulting river water quality model combines the strengths of the QUASAR model, which has proven non-conservative pollutant modelling capabilities, with the accurate advection and dispersion characterisation of the ADZ model. A discrete-time mathematical representation of the governing equations is developed that enables efficient system identification methods of parameter estimation to be utilised. The enhanced water quality model and associated methods of parameter estimation are validated using data from tracer experiments conducted on the River Mimram. The revised model produces accurate predictions of observed concentration-time curves for conservative substances.

Description: A one-dimensional model is used to investigate the relationship between land subsidence and compaction of basin sediments in response to sediment loading. Analysis of the model equations and numerical experiments demonstrate quasi-linear systems behaviour and show that rates of land subsidence due to compaction: (i) can attain a significant fraction (〉40%) of the long-term sedimentation rate; (ii) are hydrodynamically delayed with respect to sediment loading. The delay is controlled by a compaction response time τc that can reach values of 10-5-107 yr for thick shale sequences. Both the behaviour of single sediment layers and multiple-layer systems are analysed. Subsequently the model is applied to the coastal area of the Netherlands to illustrate that lateral variability in compaction-derived land subsidence in sedimentary basins largely reflects the spatial variability in both sediment loading and compaction response time. Typical rates of compaction-derived subsidence predicted by the model are of the order of 0.1 mm/yr but may reach values in excess of 1 mm/yr under favourable conditions.

Description: A simple distributed rainfall-runoff model, configured on a square grid to make best use of weather radar data, was developed in Part 1 (Bell and Moore, 1998). The simple form of the basic model, referred to as the Simple Grid Model or SGM, allows a number of model variants to be introduced, including probability-distributed storage and topographic index representations of runoff production and formulations which use soil survey and land use data. These models are evaluated here on three catchments in the UK: the Rhondda in south Wales, the Wyre in north-west England and the Mole in the Thames Basin near London. Assessment is initially carried out in simulation mode to focus on the conversion of rainfall to runoff as influenced by (i) use of radar or raingauge input, (ii) choice of model variant, and (iii) use of a lumped or distributed model formulation. Weather radar data, in grid square and catchment average form, and raingauge data are used as alternative estimates of rainfall input to the model. Results show that when radar data are of good quality, significant model improvement may be obtained by replacing data from a single raingauge by 2 km grid square radar data. The performance of the Simple Grid Model with optimised isochrones is only marginally improved through the use of different model variants and is generally preferred on account of its simplicity. A more traditional lumped rainfall-runoff model, the Probability-Distributed Moisture model or PDM, is used as a benchmark against which to assess the performance of the distributed models. This proves hard to better, although the distributed formulation of the Grid model proves more reliable for some storm and catchment combinations where spatial effects on runoff response are evident. Assessment is then carried out in updating mode to emulate a real-time forecasting environment. First, a state updating form of the Grid Model is developed and then assessed against an ARMA error-prediction technique. Both state updating and error prediction give much improved model performance when compared with simulation mode results. No one updating technique is superior, with the simulation model formulation having greatest impact on forecast accuracy. However, when the results from the different catchments are considered together it is apparent that in the rapidly responding Rhondda catchment state updating gives slightly better results, while in the slower responding Wyre and Mole catchments, error prediction is slightly superior. This is attributed to the greater difficulty of reliably adjusting states when there are significant time delays associated with the catchment response. In general, the influence of rainfall input type, model variant and distributed versus lumped model reflect the results obtained in simulation mode. Updating doesn't fully compensate for a poor rainfall input or a deficient rainfall-runoff model formulation, especially for longer forecast lead times.

Description: The possibility of modelling the temporal structure of rainfall in southern Sweden by a simple cascade model is tested. The cascade model is based on exact conservation of rainfall volume and has a branching number of 2. The weights associated with one branching are 1 and 0 with probability P(1/0), 0 and 1 with P(0/1), and Wx/x, and 1 - Wx/x, 0 〈 Wx/x, 〈 1, with P(x/x), where Wx/x is associated with a theoretical probability distribution. Furthermore, the probabilities p are assumed to depend on two characteristics of the rainy time period (wet box) to be branched: rainfall volume and position in the rainfall sequence. In the first step, analyses of 2 years of 8-min data indicates that the model is applicable between approximately 1 hour and 1 week with approximately uniformly distributed Wx/x values. The probabilities P show a clear dependence on the box characteristics and a slight seasonal nonstationarity. In the second step, the model is used to disaggregate the time series from 17- to 1-hour resolution. The model-generated data reproduce well the ratio between rainy and nonrainy periods and the distribution of individual volumes. Event volumes, event durations, and dry period lengths are fairly well reproduced, but somewhat underestimated, as was the autocorrelation. From analyses of power spectrum and statistical moments the model preserves the scaling behaviour of the data. The results demonstrate the potential of scaling-based approaches in hydrological applications involving rainfall disaggregation.

Description: This paper compares direct measurements of evaporation with the values predicted for reference transpiration. The measurements of actual evaporation were made using an eddy correlation device on a grass field adjacent to the river Thames. Measurements of soil moisture and the driving meteorological variables were also made. The results showed that, during a period with minimal rainfall but no water stress, the cumulative values of reference transpiration compared very well with the cumulative measured evaporation and changes in soil moisture content. However, the values on specific days did not compare well. Following significant rainfall, the measured evaporation increased for a few days, probably due to evaporation of free water from the canopy or soil. Reference transpiration fell consistently below the measured evaporation once the soil moisture deficits exceeded 140 to 150 mm.

Description: A survey and resurvey of 77 headwater streams in Wales provides an opportunity for assessing changes in streamwater chemistry in the region. The Model of Acidification of Groundwater In Catchment (MAGIC) has been calibrated to the second of two surveys, taken in 1994-1995, using a Monte-Carlo methodology. The first survey, 1983-1984, provides a basis for model validation. The model simulates a significant decline of water quality across the region since industrialisation. Agreed reductions in sulphur (S) emissions in Europe in accordance with the Second S Protocol will result in a 49% reduction of S deposition across Wales from 1996 to 2010. In response to these reductions, the proportion of streams in the region with mean annual acid neutralising capacity (ANC) 〉 0 is predicted to increase from 81% in 1995 to 90% by 2030. The greatest recovery between 1984 and 1995 and into the future is at those streams with low ANC. In order to ensure that streams in the most heavily acidified areas of Wales recover to ANC zero by 2030, a reduction of S deposition of 80-85% will be required.

Description: Two methods for modelling regional responses of lake water quality to changes in acidic deposition in southernmost Norway were examined. Both methods are based upon the MAGIC model but differ in mode of regional application; one uses site-specific while the other uses Monte-Carlo methods for model calibration. The simulations of regional responses from both methods were compared with observed responses based on data from three lake surveys in southernmost Norway conducted in 1974, 1986 and 1995. The regional responses of the two modelling approaches were quite similar and agreed well with the observed regional distributions of lakewater chemistry variables. From 1974 to 1986 the observed data indicated that despite a decline of approximately 10% in sulphate (SO4) deposition, the mean acid neutralizing capacity (ANC) of lakes in southernmost Norway declined by approximately 6 μeq l-1. Both modelling approaches simulated no change or a very small decline in mean ANC for that period. From 1986 to 1995 the observed data indicated that, in response to an approximate 40% decline in SO4 deposition, the mean ANC of lakes in southernmost Norway increased by 11-16 μeq l-1. The modelling approaches simulated increases of 9-10 μeq l-1 in mean ANC for the same period. Both simulations and observations indicate that 〉 65% of lakes in southernmost Norway were acidic in 1974 and 1995. Both simulation methods predict that 〉65% of the lakes in southernmost Norway will have positive ANC values within 10 years of reductions of SO4 deposition to 20% of 1974 levels. Of the two regionalization methods the site-specific method appears preferable, because whereas the Monte-Carlo method gives results for a region as a whole, the site-specific method also reveals patterns within the region. The maintenance of a one-to-one correspondence between simulated and observed systems means that simulation results can be mapped for a geographically explicit presentation of model results. The ability to examine geographic patterns of response is becoming increasingly important in regional assessments.

Description: The results from an application of MAGIC (Model of Acidification of Groundwater In Catchments) to 733 Scottish catchments are presented. The availability of representative, good quality soil data is frequently limiting factor for biogeochemical modelling, particularly those involving modelling at various spatial scales. This study tests the sensitivity of MAGIC to soil input data derived from two different methodologies; the "nearest neighbour method" considers the closest representative soil profile to a catchment, and the "spatial weighting method" of all soil types identified in a catchment, based on a soil physico-chemical classification of Scotland. Soil data (soil depth, density, cation exchange capacity and base saturation) calculated using the "nearest neighbour method" and the "spatial weighting method" were highly variable, although the range of upper and lower limits were greater for soil data produced using the nearest neighbour method. In contrast to the predominantly organic soil data calculated by the nearest neighbour method, the spatially weighted soil parameters included a greater proportion of mineral soils. With regard to simulated surface water Acid Neutralising Capacity (ANC) for 1851, 1997 and 2050, MAGIC predictions were similar irrespective of the methodology used to determine soil input parameters. However, soil input data derived from both methods resulted in variable base saturation predictions. It is concluded that the "nearest neighbour" methodology is most appropriate if the objective is to determine the predicted response of the most acid- sensitive sites within a region in line with the approach used in Critical Laod mapping. On the other hand, "spatial weighting" integrates catchment soils and represents a more robust methodology by which to determine changes in median soil and water response in a regional context. The anticipated reductions in S emissions associated with the Second S Protocol are predicted to have a marginal beneficial effect on the recovery of soils and surface waters of Scotland.

Description: A two-layer application of the catchment-based soil and surface water acidification model, MAGIC, was applied to 21 sites in the UK Acid Waters Monitoring Network (AWAMN), and the results were compared with those from a one-layer application of the model. The two-layer model represented typical soil properties more accurately by segregating the organic and mineral horizons into two separate soil compartments. Reductions in sulphur (S) emissions associated with the Second S Protocol and different forestry (land use) scenarios were modelled, and their effects on soil acidification evaluated. Soil acidification was assessed in terms of base saturation and critical loads for the molar ratio of base cations (CA2+ + MG 2+ + K+) to aluminium (Al) in soil solution. The results of the two-layer application indicate that base saturation of the organic compartment was very responsive to changes in land use and deposition compared with the mineral soil. With the two- layer model, the organic soil compartment was particularly sensitive to acid deposition, which resulted in the critical load being predicted to be exceeded at eight sites in 1997 and two sites in 2010. These results indicate that further reductions in S deposition are necessary to raise the base cation (BC):Al ratio above the threshold which is harmful to tree roots. At forested sites BC:Al ratios were generally well below the threshold designated for soil critical loads in Europe and forecasts indicate that forest replanting can adversely affect the acid status of sensitive term objectives of protecting and sustaining soil and water quality. Policy formulation must seek to protect the most sensitive environmental receptor, in this case organic soils. It is clear, therefore, that simply securing protection of surface waters, via the critical loads approach, may not ensure adequate protection of low base status organic soils from the effects of acidification.

Description: The theory of the geomorphological unit hydrograph (GUH) is examined critically and it is shown that the inherent assumption that the operation of the drainage network may be modelled by a corresponding network of linear reservoirs so restricts the instantaneous unit hydrograph (IUH) shape that the effects of further restrictions, reflecting the constraints imposed by the geomorphological laws of the channel network, cannot easily be identified. Without such identification, the geomorphological unit hydrograph theory is untestable and must remain only a plausible hypothesis providing an indication of a two-parameter IUH whose shape and scale factors must still be related empirically to appropriate catchment characteristics.

Description: The Model of Acidification of Groundwater In Catchments (MAGIC) has been calibrated to 94 catchments in Wales, enabling the quantification of regional response to reduced emissions under the second sulphur protocol and incorporating the potential impact of future changes in afforestation. To calibrate the model the methodology utilises, for each site, the best available data derived from databases which are consistent across the whole region but which vary in spatial resolution. The calibrated model closely matches observed stream water chemistry across the region. The predicted future response is a reduction in acidity of surface waters in the region. The recovery is less marked at forested catchments, and some forested sites undergo a further decrease in acid neutralising capacity (ANC). The predictions from the multiple-site method are statistically consistent with those produced by an alternative regional modelling method which is based upon Monte-Carlo procedures. The multiple-site method, however, has the advantage of providing future predictions which are also spatially distributed across the region.

Description: Major, minor and trace element chemistry of runoff at stormflow and baseflow from 67 catchments (2 to 5 ha in area) has been determined to investigate the effects of clear felling and replanting of conifers on stream water quality across Wales. Samples, collected by local forestry workers (Forest Enterprise staff) on a campaign basis on up to eight occasions, were for 16 mature first rotation standing forest: the remainder represented areas completely clear felled from less than one to up to forty years previously. As the waters drain acidic and acid sensitive soils, acidic runoff is often encountered. However, higher pH values with associated positive alkalinities and base cation enrichments are observed due to the influence of weathering reactions within the bedrock. There is little systematic variation in water quality between baseflow and stormflow for each site indicating a complex and erratic contribution of waters from the soil and underlying parent material. 80% or more of the data points show hardly any changes with felling time, but there are a few outlier points with much higher concentrations that provide important information on the processes operative. The clearest outlier felling response is for nitrate at five of the more recently felled sites on brown earth, gley and podzolic soil types. ANC, the prime indicator of stream acidity, shows a diverse response from both high to low outlier values (〉+400 to -300 μEq/l). In parallel to nitrate, aluminium, potassium and barium concentrations are higher in waters sampled up to 4 years post felling, but the time series response is even less clear than that for nitrate. Cadmium, zinc and lead and lanthanides/actinides show large variations from site to site due to localized vein ore-mineralization in the underlying bedrock. The survey provides a strong indication that forest harvesting can have marked local effects on some chemical components of runoff for the first four years after felling but that this is confined to a small number of sites where nitrate production and aluminium leaching are high. In general, deforestation leads to a reversal of acidification when the nitrate pulse is low. The variability in water quality from catchment to catchment is too high for generalized conclusions to be made over the extent of the potential changes from site to site. The value of an organised campaign of opportunistic sampling using an infrastructure of enthusiastic staff from regionally dispersed organisations associated with environmental matters (in this case the forestry industry) is highlighted.

Description: A dynamic, process-based model of surface water acidification, MAGIC, has been applied to over a thousand sites across the UK. The model is calibrated to surface water samples collected during a survey for the Critical Loads programme, and utilises the best available and consistent estimates of soil physical and chemical properties, rainfall and runoff volumes, and deposition chemistry. A total of 698 sites were calibrated successfully. At these sites, surface water chemistry was reconstructed from 1850 to the present day, and forecast to 2050 based on future decreases in sulphur (S) deposition in response to the Second S Protocol. Model outputs capture distinct regional patterns of acidification and recovery. the most acidic present-day conditions are found in acid-sensitive regions of Northern England (the Pennines, Lake District and North York Moors). Although a significant proportion of sites in these areas failed to calibrate, those that did are predicted to have experienced severe historic decreases in acidic neutralising capacity (ANC) in response to high levels of acidic deposition. The model also indicates significant acidification in the moderate deposition areas of Wales and Galloway, whereas in the low deposition region of northern Scotland, acidification has been minor even in areas of acid-sensitive geology. ANC is forecast to recover at virtually all sites, with the greatest recovery predicted for areas currently subject to high deposition. The model indicates that the Second S Protocol, however, will not be sufficient to produce full recovery, with average ANC increases to 2050 counteracting just 27% of the simulated decline from 1850 to present day. Acidic conditions (ANC 〈 0) are predicted to persist until 2050 at a significant number of sites in Northern England, Wales and Galloway.

Description: An electrical circuit analogue of a river catchment is described from which is derived an hydrological model of river flow called the River Electrical Water Analogue Research and Development (REWARD) model. The model is based upon an analytic solution to the equation governing the flow of electricity in an inductance-capacitance-resistance (LCR) circuit. An interpretation of L, C and R in terms of catchment parameters and physical processes is proposed, and tested for the River Irwell catchment in northwest England. Hydrograph characteristics evaluated using the model are compared with observed hydrographs, confirming that the modelling approach does provide a reliable framework within which to investigate the impact of variations in model input data.

Description: The MERLIN model was applied on the results of a field-scale manipulation experiment with decreased nitrogen (N) deposition in an N saturated forest ecosystem in the Netherlands. The aim was to investigate the mechanisms that could explain the observed rapid response of nitrate as a result of the decreased N input. Calibrating the model to pre-treatment data revealed that, despite the high atmospheric N input, the trees relied on N mineralised from refractory organic matter (ROM) for their growth. MERLIN could simulate only the fast response of nitrate leaching after decreased input if this ROM mineralisation rate was decreased strongly at the time of the manipulation experiment.

Description: Dynamics of C and N in forest soils in the Nutrient Cycling and Soil Acidification Model (NUCSAM) are described by the transformation and decomposition of three organic matter compartments, litter, fermented material and humic material. These three compartments are allocated to the morphological distinguishable L, F and H horizons of the organic layer. Changes in the pools of these organic compartments are described with first order equations for decomposition and transformation. Rate constants for decomposition and transformation were derived by calibrating the model to measured organic matter pools in organic layers of a chrono-sequence of five first succession Scots pine stands between 15 and 120 years old. Simulated pools of organic matter in the organic layers were in agreement with measured pools in the five pine stands, except for the first thirty years of the H-horizon. During this period, an increase in organic matter in the H horizon was simulated while no H horizons were observed in the field. The simulated total pool of organic matter in the organic layer agreed well with values from a field inventory in 20 other Scots pine stands, but the simulated distribution over the three horizons differed from the field measurements which varied among sites. For the Scots pine stands the model was able to simulate the organic matter accumulation in the top 40-cm of the mineral soil; derived almost completely from fine root turnover. The accumulated pool of nitrogen in the organic layer was in agreement with measured pools for the oldest Scots pine stand but was too high for the younger stands. Especially, the accumulation of N in the F-horizon was too fast, presumably due to an overestimated retention of nitrogen.

Description: A procedure to generate rainfall input for the EUROpean Soil Model is presented. To develop such a procedure, first of all the influence of rainfall event amount, rainfall event duration, and time to peak intensity of event rainfall on soil losses, calculated with EUROSEM, has been tested for several rainfall stations. Results revealed that every tested rainfall parameter had highly significant influence on computed soil loss. Therefore, distributions for each station of the dataset and for each of these rainfall parameters were calculated. To simulate rainfall event amounts, a mixed exponential distribution was applied. After transformation of rainfall event durations, their distribution could be simulated using a normal distribution. The location of the peak intensity was estimated using a kernel estimator, because no specific distribution characteristics could be identified. According to the respective distribution functions, parameter values for each of the tested rainfall event characteristic were then generated. These values were used to select rainfall events with identical parameter values out of the rainfall station-specific dataset. Computed soil losses for events selected this way were compared with soil losses calculated with available station specific rainfall event data. Comparisons for the respective means and medians generally revealed good agreement. A comparison of 75 % quartiles resulted in less good agreement, especially for test conditions with high soil losses. In general, the applied procedure was capable of simulation station-specific soil losses and of reflecting different environment conditions for the respective stations. Therefore, it seems possible to produce site specific appropriate rainfall input for EUROSEM, only with the knowledge of distributions for the investigated basic rainfall parameters. These are normally easier to obtain than long term rainfall information with high temporal resolution which would otherwise be necessary. In order to improve the procedure and make it practically useful, it will be necessary to account for seasonal changes of distributions of basic rainfall event parameters.

Description: A growth factor is the ratio of the T-year extreme value to an index extreme value such as the mean of annual maxima. Whereas a record length of ten or more years may suffice to estimate the index variable, it is generally necessary to blend data from several sites if estimates of exceptional extreme values are to be obtained. Methods of rainfall growth estimation are reviewed, including traditional methods which extend frequency curves to long return period by a distributional assumption, and methods which study spatial dependence in extreme rainfalls. It is desirable that estimates at neighbouring sites, and across different durations and return periods, are internally consistent. The review concludes that rather special techniques may be required if this goal of estimation extreme rainfall depth consistently is to be met. The motivation of the Focused Rainfall Growth Extension (FORGEX) method is presented.

Description: The Focused Rainfall Growth Extension (FORGEX) method produces rainfall growth curves focused on a subject site. Focusing allows the incorporation of rainfall extremes observed regionally while respecting local variations in growth rates. The starting point for the analysis is an extensive set of annual maximum rainfalls, with values at each gauged site standardized by the median. Following the philosophy of the earlier FORGE method, a strongly empirical approach is adopted. The rainfall growth curve is represented by linear segments on a Gumbel scale, and is fitted by a least-squares criterion. The selection of data points is intricate and includes both the traditional pooling of regional extremes and the incorporation of network maximum events. The latter comprise the largest events from successive hierarchical networks of gauges, focused on the site for which estimates are requires. Their treatment takes account of interdependence using the Dales and Reed model of spatial dependence in rainfall extremes.

Description: In crystalline basement regions of Africa, shallow weathered aquifers provide vital water resources for rural communities. To quantify evidence of the behaviour of these shallow aquifers, groundwater levels were observed at a network of 65 boreholes within the Romwe Catchment in southern Zimbabwe. Soil moisture was monitored at selected sites. Groundwater hydrographs showed considerable spatial and temporal variation. Where the soil profile was freely draining, groundwater levels typically responded within a few days of major rainstorms and large annual fluctuations in the water table of up to 7 m were recorded. In areas where a thick clay layer exists, annual fluctuations were smaller and groundwater levels rose more gradually in response to rainfall. In cultivated areas, vertical drainage was an important recharge mechanism. Groundwater hydrographs typically have an exponential recession and, by the end of the dry season in the years studied, levels were close to the base of the weathered aquifer. Variations in hydrograph response between years illustrate the importance of rainfall amount, intensity and distribution on groundwater recharge.

Description: Expressions for the upwelling and downwelling fluxes of optical and thermal radiation between soil, vegetation and the sky are derived, under certain simple assumptions. These are that interception of radiation by the vegetation is a purely geometric effect, while scattering is isotropic, with a strength given by a single-scattering albedo in the optical part of the spectrum, and by Kirchhoff's Law in the thermal. The soil is assumed to be a lambertian reflector, also scattering according to an albedo and Kirchhoff's Law. The model, called RM, conserves energy exactly. As part of a SVAT, it is driven by measured insolation instead of radiation, with little increase in computational cost and number of parameters.

Description: The reduction of chromate ions by Fe(OH)2 and the iron (II)-iron (III) hydroxysulphate green rust, GR(SO42-), was studied to evaluate whether such synthetic layered hydroxides and the corresponding natural green rust mineral could be involved in the natural attenuation of contaminated environments. The resulting Cr (III) bearing phases, which would govern the subsequent behaviour of chromium, were clearly characterised. Both compounds proved to be very reactive and oxidised instantaneously while chromate ions were reduced to Cr (III) as evidenced by X-ray photoelectron spectroscopy. Mass balance (ICP-AES) demonstrated that the Fe/Cr ratio inside the solid end product was equal to the initial Fe/Cr ratio. The solid phases, analysed by X-ray diffraction, Raman and Mossbauer spectroscopies were identified as Cr-substituted poorly crystallised iron (III) oxyhydroxides in both cases, more precisely δ-FeOOH when starting with Fe(OH)2 and ferrihydrite when starting with GR(SO42-).

Description: A semi-distributed conceptual model, HBV-SED, for estimation of total suspended sediment concentration and yield at the outlet of a catchment was developed and tested through a case study. The base of the suspended sediment model is a dynamic hydrological model, which produces daily series of areal runoff and rainfall for each sub-basin as input to the sediment routine. A lumped measure of available sediment is accumulated continuously based on a linear relationship between log-transformed values of rainfall and erosion, while discharge of suspended sediment at the sub-basin outlet is dependent on runoff and amount of stored available sediment. Four model parameter are empirically determined through calibration against observed records of suspended sediment concentration. The model was applied to a 200 km2 catchment with high altitude differences in the tropical parts of Bolivia, where recorded suspended sediment concentrations were available during a two-year period. 10,000 parameter sets were generated through a Monte Carlo procedure to evaluate the parameter sensitivity and interdependence. The predictability of the model was assessed through dividing the data record into a calibration and an independent period for which the model was validated and compared to the sediment rating curve technique. The results showed that the slope coefficients of the log-transformed model equations for accumulation and release were much stronger than the intercept coefficients. Despite and existing interdependence between the model parameters, the HBV-SED model gave clearly better results than the sediment rating curve technique for the validation period, indication that the supply-based approached has a promising future as a tool for basic engineering applications.

Description: This research quantified the role of topography and hydrological processes within and, hence, the development of, blanket bogs. Topographic characteristics were derived from digital elevation models (DEMs) developed for the surface and underlying substrate at three blanket bog sites on the southeastern lobe of the Avalon Peninsula, Newfoundland. A multinomial logit (MNL) model of the probability of bog occurrence was constructed in terms of relevant topographic characteristics. The resulting model was then used to investigate the probabilistic boundary conditions of bog occurrence within the landscape. Under average curvatures for the sites studied, substrate slopes up to 0.065 favoured blanket bog development. However, steeper slopes could, theoretically, be occupied by blanked bog where water is concentrated by convergent curvatures or large contributing areas. Near community boundaries, bog and heath communities both occupied similar topographic conditions. Since these boundary locations are capable of supporting the hydrological conditions necessary for bog development, the heath is likely to be encroached upon by bog.

Description: Fluid transmissivity (layer thickness times permeability) and electric transverse resistance (layer thickness time resistivity) are important parameter in groundwater and hydrocarbon exploration. Determination of these parameters provides a good knowledge of the potential of porous media, because they relate fluid flow to electric-current conduction, in terms of layer thickness, permeability and resistivity. In this study, both parameters were determined for shallow aquifers (Schleswig-Holstein, northern Germany) and deep reservoirs (Jeanne d'Arc Basin, offshore of eastern Canada), utilizing surface and well-log electric measurements. Direct relationships between both parameters, with coefficients of correlation of 0.99 (for the aquifers) and 0.94 (for the reservoirs), were obtained. The relationships suggest that an increase in both parameters indicate presence of zones of high fluid potential within the aquifers and the reservoirs.

Description: The matching of estimated to observed hydrograph shape is central to much hydrological analysis. This research note quantifies built-in biases that tend to inflate goodness of fit indicies, biases that arise from the similarity of geometry between observed and estimated hydrographs.

Description: Carbonate rocks form important aquifers in many parts of the world and in north-west Europe the Chalk is a primary source of potable water. When flushed with relatively fresh groundwaters, the Chalk may undergo significant diagenetic alteration at relatively shallow depths resulting in a physically and hydrogeochemically stratified aquifer. Diagenetic affects may have important implications for the effective exploitable thickness of the Chalk aquifer and for water quality. In order to assess the affects of diagenesis on the properties of carbonate aquifers, matrix porosity, permeability, pore water and rock chemistry profiles have been analysed for a 300 m deep borehole through the Chalk at the western end of the London Basin. An abrupt change in the matrix porosity profile at 155 mbgl indicates a change in dominant mode of historic diagenesis from mechanical compaction above 155 mbgl to predominantly pressure solution compaction below 155 mbgl. Pore water and rock chemistry profiles also change abruptly across this depth interval, suggesting that the present day hydrogeology is controlled by historic diagenetic trends. Below 155 mbgl, pore waters are relatively saline and there is no evidence for groundwater flow; above 155 mbgl pore waters are relatively fresh and geochemical evidence for incon-gruent carbonate dissolution indicates contemporary groundwater circulation. Possible physical and chemical evolution paths for the Chalk at Fair Cross are discussed. The results provide a hydrogeological context for other studies of the long-term response of carbonate aquifers to base-line changes in sea-level and pore water chemistry and also enable studies with relatively short time-frames or of localized phenomena to be placed in the broader context of the evolution of carbonate aquifers.

Description: This paper presents the background to the catchment studies at Plynlimon, through outlining the original water resource concerns which led to the establishment of the studies, the principal land use changes, land use impacts that were considered and the principal research findings. The "scene" is also set for the new areas of research, presented in this special issue, which are focused less on strictly water resource issues, but more on holistic concerns about the management of land use to protect and conserve the natural environment.

Description: The operational system known as MORECS which provides estimates of evaporation, soil moisture deficit and effective precipitation under British climatic conditions has been revised as version 2.0. An overview of the new system is described with emphasis on the new additions. The major changes from the older version (Thomson, Barrie and Ayles, 1981) include the introduction of the crop oil-seed rape, a revised treatment of soils and available water capacity and a land use data base which is representative of the 1990s.

Description: An analysis of hydrograph recessions can be used to identify the parameters of a conceptual catchment storage irnodel and, with the advent of large-scale digital data storage and automated logging systems, it has become desirable to automate recession curve analysis. Various studies have thus reported algorithms used to infer 'baseflow' storage models automatically from recession data. Such algorithms commonly operate by maximising the fit of measured recession data to some a priori function. Here, an alternative approach is taken in which the appropriate form for a catchment saturated zone store is investigated by combining observed recession data to form a Master Recession Curve (MRC). This is done within a software package that offers automated functions to help select recession periods suitable for inclusion within the MRC. These recession periods are combined automatically to form a "prototype" MRC, which can be modified interactively to overcome problems such as unrepresentative or sparse data. The master recession for a catchment is used to calculate an empirical catchment-averaged discharge-relative storage (QΔS) relationship. The method is considered to be general because the QΔS relationship may be of arbitrary form. Examples are given, showing the derivation for three catchments of different QΔS functions.

Description: Aggregation rules are derived for calculating the effective value of parameters that determine the exchange of momentum and energy between the land surface and the atmosphere at the length scales used in General Circulation Models (GCMs). The derivation involves starting from theories that link parameters relevant at grid scale and patch scale, and then imposing the limitations necessarily present when models are operated in a free-standing, predictive mode. The application of these rules is illustrated by example for the case of the Biosphere-Atmosphere Transfer Scheme (BATS). Remotely sensed global maps of land cover classes at 1 km x 1 km pixel scale for North America, South America, and Africa are used with these new aggregation rules to calculate area-average values of parameters for the 3° x 3° grid mesh used in the National Center for Atmospheric Research Community Climate Model. There are significant differences between the parameters calculated using aggregation rules and the values selected on the basis of the dominant vegetation cover in each grid, this being the selection procedure conventionally applied with BATS.

Description: Values of the momentum roughness length, z0, and displacement height, d, derived from wind profiles and momentum flux measurements, are selected from the literature for a variety of sparse canopies. These include savannah, tiger-bush and several row crops. A quality assessment of these data, conducted using criteria such as available fetch, height of wind speed measurement and homogeneity of the experimental site, reduced the initial total of fourteen sites to eight. These datapoints, combined with values carried forward from earlier studies on the parameterization of z0 and d, led to a maximum number of 16 and 24 datapoints available for d and z0, respectively. The data are compared with estimates of roughness length and displacement height as predicted from a detailed drag partition model, R92 (Raupach, 1992), and a simplified version of this model, R94 (Raupach, 1994). A key parameter in these models is the roughness density or frontal area index, λ. Both the comprehensive and the simplified model give accurate predictions of measured z0 and d values, but the optimal model coefficients are significantly different from the ones originally proposed in R92 and R94. The original model coefficients are based predominantly on measured aerodynamic parameters of relatively closed canopies and they were fitted `by eye'. In this paper, best-fit coefficients are found from a least squares minimization using the z0 and d values of selected good-quality data for sparse canopies and for the added, mainly closed canopies. According to a statistical analysis, based on the coefficient of determination (r2), the number of observations and the number of fitted model coefficients, the simplified model, R94, is deemed to be the most appropriate for future z0 and d predictions. A CR value of 0.35 and a cd1 value of about 20 are found to be appropriate for a large range of canopies varying in density from closed to very sparse. In this case, 99% of the total variance occurring in the d-data across 16 selected canopies can be explained, whereas the analogous value for the z0-data (24 datapoints available) is 81%. This makes the R94 model, with only two coefficients and its relatively simple equations, a useful universal tool for predicting z0 and d values for all kinds of canopies. For comparison, a similar fitting exercise is made using simple linear equations based on obstacle height only (e.g. Brutsaert, 1982) and another formula involving canopy height as well as roughness density (Lettau, 1969). The fitted Brutsaert equations explain 98% and 62% of the variance in the d and z0-data, respectively. Lettau's equation for prediction of z0 performs unsatisfactorily (r2 values

Description: Long term studies of fluvial sediment processes in the Plynlimon catchments have contributed to the assessment and quantification of plantation forestry impacts in British upland catchments, at all stages of the forest cycle. The results from the Plynlimon studies are placed in the context of the observed impacts of particular forest practices and studies of forestry effects on sediment transport elsewhere in the world. The effects associated with drain excavation, ploughing, track construction, ground and channel disruption are outlined for both bedload and, particularly, for suspended load. Finally, recent data on sediment yields from 1995 to 1997 at Plynlimon are reported and discussed in the light of longer-term sediment yield estimates. This paper also provides background information relevant to other sediment process studies which use data from the main Plynlimon sediment monitoring network.

Description: Stream water chemistry in the Cyff and Gwy subcatchments within the headwaters of the River Wye has been monitored regularly since 1980. In the Gwy, which is a predominantly semi-natural grassland catchment, land use has remained relatively static over the monitoring period, whilst the Cyff catchment is more buffered because of base cation inputs from agricultural improvement and ground water sources. Using a variety of statistical techniques, the long-term data are examined for evidence of trends after eliminating seasonal effects. The results highlight some of the difficulties associated with the analysis of longterm water quality data which show considerable variability over a variety of timescales. Some of this variability can be explained in terms of hydrochemical responses to climatic extremes and episodic events such as large atmospheric inputs of seasalts. The long-term fluctuations in solute concentration underline the continuing need for maintaining consistent long-term monitoring at sensitive upland sites if underlying trends related to gradual changes in pollutant deposition or climate are to be detected with any certainty.

Description: Water quality data spanning 13 years and covering an extensive range of major, minor and trace elements in rain and stream waters at Plynlimon in mid Wales, are presented. Rainfall water chemistry is highly variable due to varying proportions of marine and pollutant derived constituents associated with patterns of atmospheric circulation. Stream waters, being composed of different proportions of waters from three chemically distinct sources at any one time (atmospheric deposition, the soil system and deeper groundwaters), are also chemically highly variable. For example, components predominantly derived from deposition such as chloride change only in response to sea-salt deposition episodes. Solutes associated with bedrock weathering such as calcium, magnesium, and alkalinity decrease with increasing flow, those associated with the upper soil layers such as aluminium, many transition metals, dissolved organic carbon and hydrogen ions increase with increasing flow. The nutrients (e.g. nitrate, boron, bromide and iodine) exhibit strong seasonal cycles associated with cycles of vegetation growth and decay. The changes in stream water chemistry resulting from tree harvesting in the Afon Hore catchment are shown to have run their course within a period of eight years. Nutrient increases in the first few years following the commencement of felling have returned to or fallen below pre-felling values. Aluminium changes are shown to be complicated by changes in nitrate and calcium. Aluminium concentrations initially increased and have fallen below their pre-felling value. Data for chloride suggest a reduction in capture of dry and mist deposition; this indicates the importance of understanding reduced deposition as a result of felling. Felling has also affected the soil micro-climate which experiences greater fluctuations in temperature and an increase in the concentration of constituents associated with organic matter. Input-output mass balance estimates show that atmospheric inputs of many constituents are retained strongly by the catchment (e.g. ammonium, phosphate, barium, boron, lead and iodine). In contrast, many of the transition elements as well as divalent base cations, aluminium and alkalinity show a net release from the catchment. Conservative constituents such as chloride and sodium show a net input-output balance.

Description: Analyses of the chemical composition of rapidly percolating soil water were used to study the genesis of a shallow podzol in a Campinarana forest and a clayey ferralsol from a typical rainforest located in North Manaus (Amazonia, Brazil). The samples were collected in lysimeters and analysed for Ca2+, Na+, K+, NH4+, SO42-, NO3-, Fe, Si, and Al. A large percentage of the nutrients was recycled in the upper 40 centimetres of both soils. The soil water concentrations in nutrients were very similar for both environments but levels of Si, Fe and AI were higher in the podzol than in the ferralsol. In the podzolic environment, the waters were enriched in Si, Fe and AI when passing through the organic layer and the top 10 cm of the soil. The concentrations decreased between 10 and 40 cm depth due to variations in mineralogy of this soil. In the ferralsol, the Si concentrations increased considerably on reaching the soil top-horizons while small increases occurred for AI and Fe. Thermodynamic equilibrium calculations indicate that most of the dissolved AI and Fe in both soil environments were in the form of organometallic complexes and that the waters were under-saturated in respect to kaolinite and gibbsite.

Description: Acid Neutralization Capacity (ANC) data for ephemeral stream and shallow groundwater for the catchments of the upper River Severn show a highly heterogeneous system of within-catchment water flow pathways and chemical weathering on scales of less than 100m. Ephemeral streams draining permeable soils seem to be supplied mainly from shallow groundwater sources. For these streams, large systematic differences in pH and alkalinity occur due to the variability of the groundwater sources and variability in water residence times. However, the variability cannot be gauged on the basis of broad based physical information collected in the field as geology, catchment gradients and forest structure are very similar. In contrast, ephemeral streams draining impermeable soils are of more uniform chemistry as surface runoff is mainly supplied from the soil zone. Groundwater ANC varies considerably over space and time. In general, the groundwaters have higher ANCs than the ephemeral streams. This is due to increased chemical weathering from the inorganic materials in the lower soils and groundwater areas and possibly longer residence times. However, during the winter months the groundwater ANCs tend to be at their lowest due to additional event driven acidic soil water contributions and intermediate groundwater residence times. The results indicate the inappropriateness of a blanket approach to classifying stream vulnerability to acidification simply on the basis of soil sensitivity. However, the results may well indicate good news for the environmental management of acidic and acid sensitive systems. For example, they clearly indicate a large potential supply of weathering components within the groundwater zone to reduce or mitigate the acidifying effects of land use change and acidic deposition without the environmental needs for Aiming. Furthermore, the high variability of ephemeral stream runoff means that certain areas of catchments where there are specific problems associated with acidification can be identified for focused remediation work for the situation where liming is required. The case for focused field campaigns and caution against over reliance on blanket modelling approaches is suggested. The results negate the conventional generalizations within hydrology of how water moves through catchments to generate streamflow events (from Hortonian overland flow to catchment contributing areas).

Description: Soil water storage was monitored in three landscape elements in the forest (plateau, slope and valley floor) over a 3 year period to identify differences in sub-surface hydrological response. Under the plateau and slope, the changes of storage were very similar and there was no indication of surface runoff on the slope. The mean maximum seasonal storage change was 156 mm in the 2 m profile but it was clear that, in the dry season, the forest was able to take up water from below 3.6 m. Soil water availability was low. Soil water storage changes in the valley were dominated by the behaviour of a shallow water table which, in normal years, varied between 0.1 m below the surface at the end of the wet season and 0.8 m at the end of the dry season. Soil water storage changes were small because root uptake was largely replenished by groundwater flow towards the stream. The groundwater behaviour is controlled mainly by the deep drainage from beneath the plateau and slope areas. The groundwater gradient beneath the slope indicated that recharge beneath the plateau and slope commences only after the soil water deficits from the previous dry season have been replenished. Following a wet season with little recharge, the water table fell, ceasing to influence the valley soil water storage, and the stream dried up. The plateau and slope, a zone of very high porosity between 0.4 and 1.1 m, underlain by a less conductive layer, is a probable route for interflow during, and for a few hours after, heavy and prolonged rainfall.

Description: : Pedo-transfer functions are largely used in soil hydraulic characterisation of large areas. The use of physico-empirical approaches for the derivation of soil hydraulic parameters from disturbed samples data can be greatly enhanced if a characterisation performed on undisturbed cores of the same type of soil is available. In this study, an experimental procedure for deriving maps of soil hydraulic behaviour is discussed with reference to its application in an irrigation district (30 km2) in southern Italy. The main steps of the proposed procedure are: i) the precise identification of soil hydraulic functions from undisturbed sampling of main horizons in representative profiles for each soil map unit; ii) the determination of pore-size distribution curves from larger disturbed sampling data sets within the same soil map unit. iii) the calibration of physical-empirical methods for retrieving soil hydraulic parameters from particle-size data and undisturbed soil sample analysis; iv) the definition of functional hydraulic properties from water balance output; and v) the delimitation of soil hydraulic map units based on functional properties.

Description: A single-site Neyman-Scott Poisson cluster model of rainfall, with convective and stratiform cells, is fitted to data for 112 sites scattered throughout the UK using harmonic variables to account for seasonality. The model is regionalised by regressing the estimates of the harmonic variables on site dependent variables (e.g. altitude) to enable rainfall to be simulated at any ungauged site in the UK. An assessment of the residual errors indicates that the regression models can be used with reasonable confidence for urban sites. Furthermore, the regional variations of the model parameter estimates are found to be in agreement with meteorological knowledge and observation. Simulated I h extreme rainfalls are found to compare favourably with observed historical values, although some lack-of-fit is evident for higher aggregation levels.

Description: A series of boreholes of up to 50 m depth, drilled into Lower Palaeozoic mudstone, shale and greywacke bedrock in the headwater catchment areas of the River Severn at Plynlimon in Central Wales, shows an extensive chemically- and hydrologically-active shallow groundwater fracture flow system. Groundwater chemistry varies in space and time with lowest water levels and highest alkalinities occurring during the drier summer months. The groundwaters are enriched in base cations, silica, sulphate and alkalinity relative to surface waters indicating significant silicate weathering sources and sulphide oxidation. These sources provide important contributions to both stream water quality and flow. At one site, the introduction of a borehole near to the main river opened bedrock fractures which increased the amount of groundwater entering the river. This had a profound effect on the river water quality by increasing the pH, alkalinity and calcium concentrations. As well as pointing to the possibility of the wider availability of groundwater resources in upland areas, the results highlight (a) the potential value of groundwater as a acid neutralizing resource, (b) the importance of weathering processes and flow routing within the groundwater environment for stream water chemistry, (c) the potential for altering stream water quality by manipulation of groundwater routing and (d) the need to include groundwater characteristics in hydrochemical management models of surface water acidification.

Description: Water temperature data were collected from five stations in the upper Severn system. Temperatures were compared between a stream with a 335 ha catchment after it had flowed for c. 1.5 km through clear felled land and a stream with a 347 ha catchment after it had flowed for c. 2.5 km through coniferous forest. The results suggest that the effect of forest cover was to lower the annual mean water temperature by c. 0.4°C, mainly in summer and through depression of both daily maxima and daily minima, though mainly the former. There was no clear evidence of temperature elevation in the afforested stream in winter. It is important to note that these conclusions depend on several assumptions that cannot be substantiated objectively. There is some evidence that water temperatures in some parts of the upper Severn system may be influenced by groundwater inputs.

Description: Nutrient fluxes were investigated on a forested and a clearcut plot in a mixed conifer high elevation (2900 m) forest at the Fraser Experimental Forest in Fraser, Colorado, USA. Plots were located on a coarse loamy mixed Dystric Cryochrept with relatively high base saturation (30-90%) and underlain by an impermeable clay subsoil. Following harvest in late 1984, annual mean NO3 concentrations of 195 to 198 μmol l-1 were observed from 1988 through 1990 and concentrations were still above reference levels in 1993. Total nitrogen loss attributable to leaching following harvest was estimated at 48kg ha-1 over 8 years. Over this same period, atmospheric nitrogen inputs exceeded annual outflow of NH4 plus NO3 from the control plots by approximately 11 kg N ha-1. A slight enrichment Of SO4 and Cl was observed from the harvested plot in 1986 but concentrations later fell below control plot levels, apparently due to dilution by the increased discharge from the harvested plot which was three to four times that from the control plot. Elevated Ca, Mg, and Na concentrations followed a similar pattern to NO3 due to exchange reactions, while a depression in alkalinity of about one-third the amount of NO3 found was also observed. Enrichment of K occurred primarily in water collected at less than 1 m depth. Increases in base cation loss due to leaching after harvest were about twice the amount that can be accounted for by the increased flux of NO3, SO4, and Cl anions. The excess reflects the increased water flux and consequent leaching of base cations in association with HCO3 and organic anions.

Description: In Plynlimon streams, brown trout (Salmo trutta L.) are widespread in the upper Wye at population densities of 0.03 to 0.32 fish m-2 and show evidence of successful recruitment in most years. In the upper Severn, brown trout are found only in an area of c. 1670 -2 downstream of Blaenhafren Falls at densities of 0.03 to 0.24 fish -2 and the evidence suggests very variable year to year success in recruitment (Crisp & Beaumont, 1996). Analyses of the data show that temperature differences between afforested and unafforested streams may affect the rates of trout incubation and growth but are not likely to influence species survival. Simple analyses of stream discharge data suggest, but do not prove, that good years for recruitment in the Hafren population were years of low stream discharge. This may be linked to groundwater inputs detected in other studies in this stream. More research is needed to explain the survival of the apparently isolated trout population in the Hafren.

Description: The electrical conductivity (EC) of a medium invaded by TDR sensors can be estimated from the impedance of a TDR reflectogram. Four categories of sensor were tested in salt solutions and the impedances of the TDR pulse wave were correlated to the EC of the solution. The relation between the impedance and EC over a wide range of conductivities is non-linear but stable. Second- to fourth-degree polynomials can extend the measurement range to 44 dS m-1 (equivalent to a NaCl concentration of 28 g l-1 or 0.48 N) and result in better prediction of the conductivities than linear relations. For automatic measurement of EC with a datalogger, the method of Giese and Tiemann (1975, Adv. Mol. Rel. Processes, 7: 45-59) gives accurate measurement of conductivities lower than 10 dS m-1. Polynomial relations between EC and the datalogger's record provide an accurate estimate of the conductivity over a wide range. However, for both manual and automatic measurements, the sensors need to be calibrated individually. In particular, in the non-linear region, the differences between sensors are larger. Fortunately, the relation is sufficiently stable to eliminate significant error.

Description: Important questions concerning the resilience of current water management strategies have been raised by the recent volatility of climatic conditions across large parts of western Europe. The last decade, overall, has been exceptionally warm and there have been very large spatial and temporal variations in rainfall, river flows and aquifer recharge rates. Examination of historical rainfall and runoff records for parts of maritime western Europe confirms that there is no close modern parallel to the conditions experienced recently. Some-but far from complete-consistency with a number of favoured climate change scenarios may be recognised. Analyses of recent trends in lengthy rainfall and runoff series for the UK demonstrate significant regional differences and provide conflicting signals especially in relation to trends in catchment losses. Difficulties in reconciling the results from different areas may reflect both real hydroclimatological differences between catchments and variations in the precision of hydrometric time series-uncertainties in the assessment of areal precipitation in upland areas in particular. The dense monitoring networks at Plynlimon together with a rigorous data quality control programme underpins the value of the hydrometric datasets as important benchmarks against which to assess the significance of the very unusual patterns of rainfall and runoff which have characterised the recent past. This paper places the rainfall, runoff and losses data for Plynlimon in the perspective provided by a number of long hydrometric records for maritime western Europe. The representativeness of the Plynlimon base period is considered with particular reference to both the historical stability which typifies the great majority of European hydrometric time series and the recent extension in the recorded range of accumulated rainfall and runoff totals which has been identified in some regions (e.g. western Scotland and Norway). Particular attention is directed to changes in seasonal rainfall and runoff patterns and the recent increases in evaporative demands. Some of the implications for the overall water balance and for water resource management are considered.

Description: International agreements aimed at reducing sulphur deposition from the atmosphere are now focused on the "critical loads" concept. This provides a distributed link between sources and receptors of acidic oxides and also links the level of reduction with biological recovery. Methods for calculating critical loads include simple classification (Level 0), mass-balance equilibrium models (Level 1) and dynamic hydrochemical models (Level 2). Here, examples of each method are applied to headwater catchments in Plynlimon, mid-Wales. The critical loads derived are compared and the utility of the critical loads concept as a tool for local pollution control and land management is assessed. Differences in critical load methods for soils result from assumptions regarding biological receptors. Surface water critical loads are generally similar under all methods. As a local management tool, the dynamic model is most appropriate although it has a high data requirement. The utility of other methods depends upon detailed maps of soil and vegetation being available at an appropriate scale.

Description: Previous monitoring of the surface water chemistry of the forested catchments at Plynlimon has concentrated on weekly measurements. Hence, peak values and details of the rapid changes in chemistry occurring during events tend to be missed, particularly for smaller catchments. Knowledge of the detailed hydrochemical response is necessary to predict effects on stream ecology and to quantify processes occurring within the catchment. This information is vital if the likely impacts of changes in landuse and pollutant input on episodic hydrochemical variations are to be modelled. A summary of daily and three-hourly chemistry for the River Severn headwaters at Plynlimon, central Wales is presented. Data are from the summer and autumn of 1995 and the winter and summer of 1996. The effects of antecedent conditions, soiltype, landuse and groundwater and tributary inflow on stream episodic hydrochemical response are discussed for a range of spatial scales, including first, second and third order streams. Detailed information on the chemical changes due to a discrete groundwater inflow into the bed of one stream shows that, during episodes, its influence on stream chemistry is overridden by that of event water. A comparison of drainage ditch chemistry with main stream chemistry during storm events is used to consider the importance of tributary inflow at these times. The contrast in results between sites is assessed, to evaluate the effects of clearfelling and soil type on stream event chemistry.

Description: Over the last decade there has been some uncertainty over the calculation of "actual" catchment evaporation from the long-term difference between rainfall and runoff for the grassland Wye catchment at Plynlimon, compared to estimates made using the Penman formula on data from automatic weather stations. Hence, measurements of actual evaporation were made over a two month period in the late summer of 1992 using the eddy-correlation technique on a relatively flat but wetter than average site in this upland region. Although the site was rather restricted for this type of measurement, the quality of data proved typical of other eddy-correlation measurements made at more aerodynamically suitable sites. The ratio of actual evaporation to the Penman estimate on dry days in summer was 0.83. This compares with an average annual ratio, generated from the catchment data for the period 1978 to 1995, of 1.01. The catchment rainfall value has been improved recently by weighting each gauge using altitude domains.

Description: : Better understanding of field-scale unsaturated zone transport mechanisms is required if the fate of contaminants released at the surface is to be predicted accurately. Interpretation of results from direct tracer sampling in terms of operative hydraulic processes is often limited by the poor spatial coverage and the invasive nature of such techniques. Cross-borehole electrical imaging during progress of saline tracer migration is proposed to assist investigation of field-scale solute transport in the unsaturated zone. Electrical imaging provides non-destructive, high density and spatially continuous sampling of saline tracer transport injected over an area of the ground surface between two boreholes. The value of electrical imaging was tested at a field site on an interfluve of the UK Chalk aquifer. Improved understanding of active transport mechanisms in the unsaturated zone of the UK Chalk is required to predict its vulnerability to surface pollutants. In a tracer experiment in May 1996, a conductive saline tracer was infiltrated over 18 m2 at an average rate of 47 mm day-1 for 56 hours. Cross-borehole images obtained during and after infiltration show a large, homogenous, resistivity reduction in the top 3 m, no change between 3 m and 6 m depth, and smaller, inhomogeneous, resistivity reductions below 6 m depth. The resistivity has reduced at down to 15 m depth less than 2 days after tracer infiltration began. Hydrological interpretation of a sequence of electrical images obtained prior to, during, and up to three months after tracer injection suggests: (1) rapid tracer entry into the soil zone and upper 2 m of weathered Chalk, (2) intergranular transport of the bulk of the tracer, (3) a significant fissure flow component transporting tracer to at least 15 m depth in 31 hours, and (4) vertical changes in transport mechanisms possibly caused by interception of fissures by marl layers. The results of this experiment suggest that electrical imaging can assist the description of unsaturated zone hydraulic mechanisms through visual identification of spatial and temporal variations in transport processes.

Description: A simple dynamic soil model developed to analyse the effects of atmospheric deposition and nutrient cycling on terrestrial ecosystems, SMART 2, was applied to the Kangasvaara catchment in eastern Finland. Given the historical deposition and forest growth patterns and reasonable values for the input parameters, SMART 2 was calibrated successfully to reproduce present-day soil and Kangasvaara catchment on the soil and runoff water chemistry under a future deposition scenario (GRP scenario). These impacts were also compared to the effects of further reducing the deposition of sulphur and nitrate under the maximum feasible reduction (MFR) scenario. The model demonstrates the consequences of breaking the nutrient cycle, and predicts that final cutting results in increased leaching of inorganic nitrogen and base cations from the cut part of the catchment for about 10 years. The resulting concentrations in the stream will depend on the ability of the buffer zones surrounding the stream to capture and utilize these nutrients.

Description: A new probabilistic interpretation of at site rainfall sequences is introduced for the development of a stochastic model of rain. The model, is divided into two sub models; the first one describing the total number of rainfall spells within a window of time is described by a Pòlya process in order to reproduce better the variable probability of occurrence of rainfall during storm events (due to the presence of different numbers of rainfall cells); the second sub model, conditional on the first one, describes the total quantity of rainfall in the time window, given a number of rainfall spells. The probabilistic rainfall model, which has shown interesting properties in reproducing the probability distribution of observed data at time scales ranging from one hour to twenty-four hours, may be the basis for a number of applications which include the development of a conditional stochastic generator of rain, within the frame of real-time flood forecasting, and the derivation of a probabilistic distribution of rainfall extremes at the various time scales.

Description: The patterns of variation in water quality for an acidic stream draining plantation forest overlying acidic and acid sensitive gley soils with shale and slate bedrock changed following the introduction of a 45 m deep borchole near to the stream. During drilling, air flushing of debris from the borehole cleared fracture routes for groundwater penetration to the stream via the stream bed. Consequently, there were and there remain marked increases in pH, alkalinity and calcium concentrations in the stream water. The extent of this water quality improvement varies according to flow. Under extreme highfiow conditions, most of the stream water is supplied from near surface soil water sources and acidic stream waters (pH about 4.2) result. Under baseflow conditions, the stream water pH is about 7.0 upstream and about 7.5 downstream of the borehole. Under intermediate flow conditions, the improvement in pH is most marked and values increase from around 5 to around 6.3. For acid sensitive "hard rock" areas such as those studied here, the bedrock has frequently been assumed to be both impermeable and low in base cations. This study illustrates that this view may be incorrect, and that groundwater may provide an important modifier of streamwater quality, at least for slate and shale dominated hard rock areas. Indeed, the work demonstrates clearly the potential for water quality remediation through groundwater manipulation.

Description: Agricultural tile drainage lines have been implicated as a source of pesticide contamination of surface waters. Field experiments were conducted and a simple model was developed to examine preferential transport of applied chemicals to agricultural tile lines. The conceptual model consists of two linear reservoirs, one near the soil surface and one near the tile drain. The connection between the two reservoirs is via preferential flow paths with very little interaction with the soil matrix. The model assumes that only part of the field contributes solutes to the tile drain. The model was evaluated with data from the field experiments in which chloride, 2,4-D, and atrazine concentrations were measured on eight tile-drained plots that were irrigated twice. Atrazine was applied two months prior to the experiment, 2,4-D was sprayed just before the first irrigation, and chloride before the second irrigation. All three chemicals were found in the tile effluent shortly after the rainfall began. Generally, the concentration increased with increased flow rates and decreased exponentially after the rainfall ceased. Although the simple model could simulate the observed chloride concentration patterns in the tile outflow for six of the eight plots, strict validation was not possible because of the difficulty with independent measurement of the data needed for a preferential flow model applied to field conditions. The results show that, to simulate pesticide concentration in tile lines, methods that can measure field averaged preferential flow characteristics need to be developed.

Description: Near surface soil moisture measurements were recorded at hourly intervals at two contrasting sites within the Cyff sub-catchment using a prototype capacitance probe system. In a mire area within a valley bottom, over the twelve month recording period, very little change in moisture content occurred. At the other site, a well drained area on a steeply sloping hillside, major variations occurred with significant soil moisture deficits being generated during a particularly dry summer. Soil moisture on the slope responded rapidly to rainfall inputs during wet periods, with little response during particularly dry periods. A number of rainfall events was analysed to determine whether changes in soil moisture could be used to characterise storm hydrographs for the Cyff and the Gwy, two sub-catchments being composed of differing percentages of mire area and steep slopes. It was found that percentage runoff for the Cyff was correlated with antecedent soil moisture on the slope, though the agreements for peak flow and lag time were poorer. For the Gwy, poor agreements were obtained for all three hydrograph characteristics. A simple formulation, based on storm rainfall and antecedent soil moisture deficits in the slope and mire areas, gave good agreement with storm streamflow volumes.

Description: The catchment scale-experiments of the RAIN and CLIMEX projects conducted on boreal forest ecosystems at Risdalsheia, southernmost Norway, provide a unique set of data on the flux of nitrogen (N) in runoff following changes in N deposition, carbon dioxide (CO2) level and temperature. MERLIN (Model of Ecosystem Retention and Loss of Inorganic Nitrogen), a recently-developed model that focuses on N leaching, provides a means by which these data can be placed into a quantitative framework. The features of the N flux in runoff at Risdalsheia to be explained include (1) leaching of about 30-50 mmol m-2 yr-1 (30-40% of N deposition) during the period 1985-1997 at reference catchments, (2) rapid and dramatic reduction in N leaching following experimental reduction in N deposition in 1985 at KIM catchment, (3) increased flux of about 5 mmol m-2 yr-1 following onset of 3-5°C warming and increased CO2 in 1995 at KIM catchment, and (4) increased flux of about 12 mmol m-2 yr-1 following 3-5°C warming of soil in 1995 at EGIL catchment. One set of calibrated model parameters is sufficient to simulate the changes in N runoff at both experimental catchments for both of the manipulations. The model support the conceptual picture of the soil as the major sink for N inputs from deposition with N accumulating in both the forest floor (labile organic matter LOM) and the bulk soil (refractory organic matter ROM). As the molar carbon/nitrogen (C/N) ratio of LOM decreases to below 23, progressively less N is immobilised and more goes to runoff. The model also supports the conceptual picture of increased rate of decomposition of old soil organic matter in response to higher temperature. An increase of 5% is sufficient to produce the 5-12 mmol m-2 yr-1 increase in N flux in runoff observed at the 2 experimental catchments. The MERLIN simulations are consistent with measurements of increase in net mineralisation rates (per catchment area by 70 mmol m-2 yr-1) and N contents in foliage in treated and reference areas before and after onset of treatment. Runoff provides a very sensitive indicator of changes in N cycling within the ecosystem. Small changes in key processes such as N mineralisation give rise to large relative changes in N flux. Uncertainties in measurements are generally much larger than changes indicated by the model calibration.

Description: A simple mass balance has been used to estimate soil calcium depletion during the growth of a 50 year old Sitka spruce crop on acid, base-poor peaty podzol soils in upland Wales. Growth of the crop will deplete the soil calcium reserve by an amount (205 kg Ca ha-1) approximately equivalent to the exchangeable calcium pool to the bottom of the profile and equal to 14% of the total soil calcium reserve to the bottom of the B horizon. Despite these predictions, measurements of exchangeable calcium show no differences beneath mature forest and acid grassland, implying that i) weathering rates in forest soils are greater than long-term estimates and predictions by the PROFILE soil chemistry model ii) the trees can access other sources of calcium or iii) there are significant errors in the mass balance. Following stem-only harvesting, growth of a 50 year old second rotation crop will lead to further depletion of soil calcium, but this amount (79 kg Ca ha-1), is less than for a second rotation crop following whole-tree harvesting (197 kg Ca ha-1). After the first crop, stem-only harvesting would allow a further 18 rotations before depletion of the total calcium reserve to the bottom of the B horizon. Whole-tree harvesting would allow for seven rotations after the first crop. These calculations assume that all sources of calcium are equally available to the crop. This can only be resolved by dynamic modelling of the calcium cycle at the ecosystem scale based on appropriate field measurements. The potential for significant soil acidification is therefore greater following whole-tree harvesting and, in line with current recommendations (Nisbet et al., 1997), this technique should probably be avoided on acidic, nutrient-poor soils unless remedial measures are included to enhance the soil base cation status.

Description: An order-of-magnitude argument shows that the different surface temperatures of soil and sparse vegetation affect carbon assimilation and soil respiration significantly. However, regulation of assimilation through associated modulation of in-canopy carbon dioxide levels is weak. It is shown that for many vegetation types, the use of a two-layer representation of vegetation is essential to predict, accurately, terrestrial carbon fluxes, primarily through allowing different surface energy balances.

Description: Rainfall interception losses were monitored for twelve months and related to vegetation and rainfall characteristics at the Wanariset Sangai on the upper reaches of the Mentaya river, Central Kalimantan. The rainfall interception losses were quantified for one hectare each of unlogged and logged humid tropical rainforests. The results show that interception loss is higher in the unlogged forest (11% of total gross rainfall) than in the logged forest (6%). Interception loss was also simulated by the modified Rutter model and Gash's original and revised models. Both the Rutter and revised Gash models predicted total interception loss over a long period adequately, and resulted in estimates of the interception loss that deviated by 6 to 14% of the measured values, for both the unlogged and logged plots.

Description: The effects of ditch water management regimes on water tables are examined for two test sites in England, Halvergate in the Broads and Southlake Moor in the Somerset Levels and Moors Environmentally Sensitive Areas. It is observed that in some fields the effects of water management are only poorly transferred from the ditch to the field centre, especially where the hydraulic conductivity of the subsoil is small. Where there are large variations in the ditch water levels, reflecting the influence of major ditches subject to pump drainage, field soil water regimes differ significantly. Nevertheless, the effects of even quite small changes in the ditch regime cam be noticeable. Simple modelling studies show that much greater effects can be achieved by increasing the frequency of ditches within wetlands.

Description: Distance from the sea, proximity of mountains, continentality and elevation are all useful covariates to assist the mapping of extreme rainfalls. Regression models linking these and other variables calculated from a digital terrain model have been built for estimating the median annual maximum rainfall, RMED. This statistic, for rainfall durations between 1 hour and 8 days, is the index variable in the rainfall frequency analysis for the new UK Flood Estimation Handbook. The interpolation of RMED between raingauge sites is most challenging in mountainous regions, which combine the greatest variation in rainfall with the sparsest network of gauges. Sophisticated variables have been developed to account for the influence of topography on extreme rainfall, the geographical orientation of the variables reflecting the prevailing direction of rain-bearing weather systems. The different processes of short and long-duration extreme rainfall are accounted for by separate regression models. The technique of georegression combines estimates from regression models with a map of correction factors interpolated between raingauge locations using the geostatistical method of kriging, to produce final maps of RMED across the UK.

Description: Data on the water quality of streams draining a range of acidic and acid sensitive, mainly afforested, upland catchments in mid- and north-Wales and northern-England are described to investigate the acidification effects of conifer harvesting in relation to natural variability. Most sites show a large range in pH and major cation and major anion concentrations. The waters draining from the smaller catchments are more acidic and aluminium bearing reflecting a higher proportion of runoff from the acidic soils in each area. However, there is often a less acidic component of runoff under base-flow conditions due to ground-water contributions particularly within the larger streams. Higher concentrations of nitrate occur for sites which have been felled although declines in concentration occur several years after felling. Multiple regression analysis reveals the importance of cation exchange and within catchment acidification associated with sulphate and nitrate generation. Sulphate also has a component associated with weathering but the patterns vary from catchment to catchment. Analysis of the influence of changing anion concentrations associated with tree harvesting reveals that the acidification induced by increases in nitrate can be offset or reversed by the lowering of chloride and sulphate concentrations due to decreased atmospheric scavenging by the vegetation, reduced evapotranspiration and increased surface runoff diluting the acidity generated. It is concluded that contemporary UK forestry guidelines with an emphasis on phased harvesting of catchments over several years and careful harvesting methodologies can alleviate most problems of stream acidification associated with felling activities and in some cases can reverse the acidification pattern.

Description: Bacteriophages are increasingly used as tracers for quantitative analysis in both hydrology and hydrogeology. The biological particles are neither toxic nor pathogenic for other living organisms as they penetrate only a specific bacterial host. They have many advantages over classical fluorescent tracers and offer the additional possibility of multi-point injection for tracer tests. Several years of research make them suitable for quantitative transport analysis and flow boundary delineation in both surface and ground waters, including karst, fractured and porous media aquifers. This article presents the effective application of bacteriophages based on their use in differing Swiss hydrological environments and compares their behaviour to conventional coloured dye or salt-type tracers. In surface water and karst aquifers, bacteriophages travel at about the same speed as the typically referenced fluorescent tracers (uranine, sulphurhodamine G extra). In aquifers of interstitial porosity, however, they appear to migrate more rapidly than fluorescent tracers, albeit with a significant reduction in their numbers within the porous media. This faster travel time implies that a modified rationale is needed for defining some ground water protection area boundaries. Further developments of other bacteriophages and their documentation as tracer methods should result in an accurate and efficient tracer tool that will be a proven alternative to conventional fluorescent dyes.

Description: The UP (Upscaled Physically-based) hydrological modelling system to the Arkansas-Red River basin (USA) is designed for macro-scale simulations of land surface processes, and aims for a physical basis and, avoids the use of discharge records in the direct calibration of parameters. This is achieved in a two stage process: in the first stage parametrizations are derived from detailed modelling of selected representative small and then used in a second stage in which a simple distributed model is used to simulate the dynamic behaviour of the whole basin. The first stage of the process is described in a companion paper (Ewen et al., this issue), and the second stage of this process is described here. The model operated at an hourly time-step on 17-km grid squares for a two year simulation period, and represents all the important hydrological processes including regional aquifer recharge, groundwater discharge, infiltration- and saturation-excess runoff, evapotranspiration, snowmelt, overland and channel flow. Outputs from the model are discussed, and include river discharge at gauging stations and space-time fields of evaporation and soil moisture. Whilst the model efficiency assessed by comparison of simulated and observed discharge records is not as good as could be achieved with a model calibrated against discharge, there are considerable advantages in retaining a physical basis in applications to ungauged river basins and assessments of impacts of land use or climate change.

Description: Calibration of conceptual hydrological models is frequently limited by a lack of data about the area that is being studied. The result is that a broad range of parameter values can be identified that will give an equally good calibration to the available observations, usually of stream flow. The use of total stream flow can bias analyses towards interpretation of rapid runoff, whereas water quality issues are more frequently associated with low flow condition. This paper demonstrates how model distinctions between surface an sub-surface runoff can be used to define a likelihood measure based on the sub-surface (or baseflow) response. This helps to provide more information about the model behaviour, constrain the acceptable parameter sets and reduce uncertainty in streamflow prediction. A conceptual model, DIY, is applied to two contrasting catchments in Scotland, the Ythan and the Carron Valley. Parameter ranges and envelopes of prediction are identified using criteria based on total flow efficiency, baseflow efficiency and combined efficiencies. The individual parameter ranges derived using the combined efficiency measures still cover relatively wide bands, but are better constrained for the Carron than the Ythan. This reflects the fact that hydrological behaviour in the Carron is dominated by a much flashier surface response than in the Ythan. Hence, the total flow efficiency is more strongly controlled by surface runoff in the Carron and there is a greater contrast with the baseflow efficiency. Comparisons of the predictions using different efficiency measures for the Ythan also suggest that there is a danger of confusing parameter uncertainties with data and model error, if inadequate likelihood measures are defined.

Description: Springs are important groundwater discharge points on the high altitude (〉800m) plateaux of the Cairngorm mountains, Scotland and form important wetland habitats within what is often a dry, sub-arctic landscape. The hydrogeochemistry of a typical spring in the Allt a'Mharcaidh catchment was examined between 1995-98 in order to characterise its chemical composition, identify the dominant controls on its chemical evolution and estimate groundwater residence time using 18O isotopes. Spring water, sustained by groundwater flow in shallow drift deposits and fractured bedrock, was moderately acidic (mean pH 5.89), with a very low alkalinity (mean 18 μeq l-1) and the ionic composition was dominated by sea-salts derived from atmospheric sources. Geochemical modelling using NETPATH, predicted that the dissolution of plagioclase mainly controls the release of Si, non-marine Na, Ca, K and Al into spring water. Hydrological conditions influenced seasonal variations in spring chemistry, with snowmelt associated with more rapid groundwater flows and lower weathering rates than summer discharges. Downstream of the spring, the chemistry of surface water was fundamentally different as a result of drainage from larger catchment areas, with increased soil and drift cover, and higher evaporation rates. Thus, the hydrogeochemical influence of springs on surface waters appears to be localized. Mean δ18O values in spring water were lower and more damped than those in precipitation. Nevertheless, a sinusoidal seasonal pattern was observed and used to estimate mean residence times of groundwater of around 2 years. Thus, in the high altitude plateau of the Cairngorms, shallow, coarse drift deposits from significant aquifers. At lower altitudes, deeper drift deposits, combined with larger catchment areas, increase mean groundwater residence times to 〉5 years. At high altitudes, the shallow, permeable nature of the drifts dictates that groundwater is vulnerable to impacts of environmental changes that could be usefully monitored at spring sites.

Description: The UP modelling system has been applied to the 570,000 km2 Arkansas-Red River Basin (ARRB) as part of the UK NERC Terrestrial initiative in Global Environmental Research (TIGER). The model can be run as a stand-alone basin hydrology model or be linked to existing climate and weather forecasting models. It runs on a grid comprising 1923 UP elements, each 17km by 17km in area, and each containing five water storage compartments: one each for the snowpack, vegetation canopy, surface water, root zone and groundwater. All the main transfers and processes of the terrestrial phase of the hydrological cycle are represented, including river network routing of the runoff from the UP elements. The parameters of the ARRB model are physically-based, being derived either from fine-scale, sub-grid, data on the topography and physical properties of the soils, aquifers and vegetation of the basin, or from the results of fine-scale physically-based simulations. With the approach, the parameters account for the effects of sub-grid variations in moisture status and spatial distribution and are sensitive to changes in the fine-scale property data. This sensitivity is either absent or less directly represented in existing large-scale hydrology models, yet it plays a central role in studies of the impact of changes in climate and land-use. The ARRB model, as described here and in Kilsby et al. (1999), is a first attempt at large-scale physically-based hydrological modelling of the type outlined in the "blueprint" for the UP system (Ewen, 1997), and gives a clear, positive, indication of the nature and quality of what is currently practical with the approach.

Description: Time domain reflectometry [TDR] was used to investigate the spatial and temporal variation in surface soil water dynamics under a number of types of vegetation, including both trees and crops grown in isolation, and grown together as an agroforestry system. The installation and operation of this technique are presented, and discussed in terms of its suitability to monitor rapid fluctuations in soil-water content in a spatially heterogeneous system such as that described in this experiment. The relatively small sampling volume of each of the TDR waveguides permitted discrete measurements to be made of soil water content (θv). In the tree-only and tree+crop treatments, this revealed considerable variation in θv resulting from spatial redistribution of rainfall under the tree canopies, with a significant input to soil close to the base of the trees being made by stemflow, i.e. water intercepted by the tree canopy and channelled down the stem. Over the experimental period (one rainy season) the TDR data suggested that net recharge to the soil profile in the sole crop system was 53 mm, almost 75% more than occurred in either of the two treatments containing trees, reflecting greater rainfall interception by the tree canopies.

Description: This paper illustrates the performance of the FORGEX method of rainfall growth estimation. Results are presented for three regions of the United Kingdom: the East Midlands, north-west England and south-west England. Focused rainfall growth curves are compared between regions and between different sites within each region. Typical growth curve shapes are discussed with reference to the climate of each region. Daily growth curves are derived from a large number of records of annual maximum rainfalls. A smaller number of hourly annual maximum series is available for estimation sub-daily rainfall growth curves. Rainfall growth rates are compared with the results of a widely used method. The present method allows more local and regional variation in growth rates. The new growth rates are higher for durations of 1 and 2 days in parts of south-west England, but lower for moderate return periods at some focal points in the north-west. In the East Midlands, the new 1-hour growth rates are considerably higher for long return periods. Confidence limits for growth rates are derived by bootstrapping. This is accomplished by fitting a large number of growth curves to resampled sets of rainfall data.

Description: In this study, it is shown that the complexity of Soil Vegetation Atmosphere Transfer (SVAT) models leads to an equifinality of functional behaviour - many parameterizations from many areas of the parameter space lead to very similar responses. Individual parameters derived by calibration (i.e. model inversion) against limited measurements are, therefore, highly uncertain. Due to the non-linear internal behaviour of SVAT models, aggregation of uncertainly known parameter fields to parameterize landscape scale variability in surface fluxes will yield highly uncertain predictions. A disaggregation approach suggested by Beven (1995) requires that the land surface be represented by a linear sum of a number of representative parameterizations or functional types. This study explores the nature of the parameter space in terms of a simple definition of functional behaviour. Parameter interactions producing similar predicted behaviours are investigated through application of Principal Component Analyses. These reveal the lack of a dominant global interaction indicating the presence of highly complex parameter interactions throughout the feasible parameter space.

Description: A transient, mixed analytical-numerical model of hillslope hydrological behaviour is used to study the patterns of infiltration, evapotranspiration, recharge and lateral flow across hillslopes. Computational efficiency is achieved by treating infiltration and phreatic surface movement analytically. The influence of dynamic coupling of the saturated and unsaturated zones on the division of hillslopes into units of distinct hydrological behaviour is analyzed. The results indicate the importance of downhill groundwater flow on the lateral distribution of soil moisture and hydrological fluxes; unsaturated lateral flow is shown to be of relatively minor importance. For most conditions, the hillslope organizes itself into three distinct regions; an uphill recharge and a downhill discharge zone separated by a midline zone over which there is, on average, no recharge or discharge. A temporal perturbation analysis of the phreatic surface, made to quantify the deviations between the equivalent-steady water table derived by Salvucci and Entekhabi (1995) and the long-term mean water table, shows that the equivalent-steady water table effectively couples the unsaturated and saturated zone dynamics across storm and interstorm periods and divides the hillslope into distinct hydrological regions. The second order closure terms in the perturbation analysis, expressed as the gradient of water table variance, quantify the deviations and tend to make the hydrological zones relatively less distinct.

Description: Geological fault zones are usually assumed to influence hydrocarbon migration either as high permeability zones which allow enhanced along- or across-fault flow or as barriers to the flow. An additional important migration process inducing along- or across-fault migration can be associated with dynamic pressure gradients. Such pressure gradients can be created by earthquake activity and are suggested here to allow migration along or across inactive faults which "feel" the quake-related pressure changes; i.e. the migration barriers can be removed on inactive faults when activity takes place on an adjacent fault. In other words, a seal is viewed as a temporary retardation barrier which leaks when a fault related fluid pressure event enhances the buoyancy force and allows the entry pressure to be exceeded. This is in contrast to the usual model where a seal leaks because an increase in hydrocarbon column height raises the buoyancy force above the entry pressure of the fault rock. Under the new model hydrocarbons may migrate across the inactive fault zone for some time period during the earthquake cycle. Numerical models of this process are presented to demonstrate the impact of this mechanism and its role in filling traps bounded by sealed faults.

Description: A Bayesian approach is described for dealing with the problem of infilling and generating stochastic flow sequences using rainfall data to guide the flow generation process, and including bounded (censored) observed flow and rainfall data to provide additional information. Solutions are obtained using a Gibbs sampling procedure. Particular problems discussed include developing new procedures for fitting transformations when bounded values are available, coping with additional information in the form of values, or bounds, for totals of flows across several sites, and developing relationships between annual flow and rainfall data. Examples are shown of both infilled values of unknown past river flows, with assessment of uncertainty, and realisations of flows representative of what might occur in the future. Several procedures for validating the model output are described and the central estimates of flows, taken as a surrogate for historical observed flows, are compared with long term regional flow and rainfall data.

Description: The spatial distribution of stream water composition, as determined by the Geochemical Baseline Survey of the Environment (G-BASE) conducted by the British Geological Survey (BGS) can be successfully related under baseflow conditions to bedrock geochemistry. Further consideration of results in conjunction with site-specific monitoring data enables factors controlling both spatial and temporal variability in major element composition to be highlighted and allows the value of the survey to be enhanced. Hence, chemical data (i) from streams located on Lower Silurian (Llandovery) bedrock at 1 km2 resolution collected as part of the G-BASE survey of Wales and the West Midlands and (ii) from catchment monitoring studies located in upland mid-Wales (conducted by Institute of Terrestrial Ecology), have been considered together as an example. Classification of the spatial survey data set in terms of potentially controlling factors was carried out so as to illustrate the level of explanation they could give in terms of observed spatial chemical variability. It was therefore hypothesised that on a geological lithostratigraphic series of limited geochemical contrast, altitude and land-use factors provide better explanation of this variability than others such as lithology at sampling site and stream order. At an individual site, temporal variability was also found to be of considerable significance and, at a monthly time-step, is explicable in terms of factors such as antecedent conditions and seasonality. Data suggest that the degree of this variability may show some relationship with stream order and land-use. Monitoring data from the region also reveal that relationships between stream chemistry and land-use may prove to be strong not only at base flow but also in storm flow conditions. In a wider context, predictions of the sensitivity of stream water to acidification based on classifications of soil and geology are successful on a regional scale. However, the study undertaken here has shown that use of such classification schemes on a catchment scale results in considerable uncertainty associated with prediction. Uncertainties are due to the large degree of variability in stream chemistry encountered both spatially within geological units and temporally at individual sampling sites.

Description: In Zimbabwe during the 1980s and early 1990s, a sequence of severe droughts caused widespread food shortages and great hardship to rural communities. The droughts exacerbated the problems of environmental degradation in communal lands and highlighted the lack of understanding of the links between the climate, land use and hydrology of dryland regions. The Romwe Catchment Study addresses these issues, and has led to the establishment of the first fully-instrumented research catchment in a communally-managed dryland environment in southern Africa. The key objectives were (a) to improve the understanding of hydrological processes in communal land areas, mostly underlain by crystalline basement aquifers, and (b) to investigate the impacts of variations in climate and changes in land use and management on the hydrology and water resources. In this introductory paper, the physical characteristics of the catchment are described together with the instrumentation to monitor hydrological processes and quantify the catchment water balance.

Description: Two contrasting models are used to account for the effects of vegetation on microwave emission from the soil. These are: a simple model which requires two empirically derived parameters as input data (optical depth and single scattering albedo); and a complex discrete model which requires a detailed description of all of the components of the vegetation canopy. Both models account effectively for the vegetation, although the simple model takes a fraction of the computation time compared to the discrete model. However, the simple model was fitted to the data, whereas the discrete model used measured parameters as input. In addition to predicting the microwave brightness temperature, the discrete model also calculates the optical depth and single scattering albedo. These calculated values were in agreement with those fitted using the simple model. Therefore, it is suggested that the discrete model could be used to calculate the input parameters for the simple model.

Description: This paper describes the application and testing of a method for deriving spatial estimates of albedo from multi-angle remote sensing data. Linear kernel-driven models of surface bi-directional reflectance have been inverted against high spatial resolution multi-angular, multi- spectral airborne data of the principal cover types within the HAPEX-Sahel study site in Niger, West Africa. The airborne data are obtained from the NASA Airborne Solid-state Imaging Spectrometer (ASAS) instrument, flown in Niger in September and October 1992. The maps of model parameters produced are used to estimate integrated reflectance properties related to spectral albedo. Broadband albedo has been estimated from this by weighting the spectral albedo for each pixel within the map as a function of the appropriate spectral solar irradiance and proportion of direct and diffuse illumination. Partial validation of the results was performed by comparing ASAS reflectance and derived directional-hemispherical reflectance with simulations of a millet canopy made with a complex geometric canopy reflectance model, the Botanical Plant Modelling System (BPMS). Both were found to agree well in magnitude. Broadband albedo values derived from the ASAS data were compared with ground-based (point sample) albedo measurements and found to agree extremely well. These results indicate that the linear kernel-driven modelling approach, which is to be used operationally to produce global 16 day, 1 km albedo maps from forthcoming NASA Earth Observing System spaceborne data, is both sound and practical for the estimation of angle-integrated spectral reflectance quantities related to albedo. Results for broadband albedo are dependent on spectral sampling and on obtaining the correct spectral weigthings.

Description: The land-surface parameters required as input to a GCM grid box (typically a few degrees) are often set to be those of the dominant vegetation type within the grid box. This paper discusses the use and effect of aggregation rules for specifying effective values of these land cover parameters by taking into account the relative proportion of each land-cover type within each individual grid box. Global land-cover classification data at 1 km resolution were used to define Biosphere Atmosphere Transfer Scheme (BATS) specific aggregate (using aggregation rules) land-cover parameters. Comparison of the values of the aggregate parameters and those defined using the single dominant vegetation type (default parameters) shows significant differences in some regions, particularly in the semi-desert and in forested regions, e.g. the Sahara Desert and the tropical forest of South America. These two different sets of parameters were used as input data for two 10-year simulations of the NCAR CCM3 model coupled to the BATS land-surface scheme. Statistical analyses comparing the results of the two model runs showed that the resulting effects on the land-surface diagnostics are significant only in specific regions. For example, the sensible heat flux in the Sahara Desert calculated for the aggregate parameter run increased due to the marked increase in the minimum stomatal resistance and the decrease in fractional vegetation cover in the aggregate parameters over the default parameters. The modelled global precipitation and surface air temperature fields were compared to observations: there is a general improvement in the performance of the aggregate parameter run over the default parameter run in areas where the differences between the aggregate and default parameter run are significant. However, most of the difference between the modelled and observed fields is attributable to other model deficiencies. It can be concluded that the use of aggregation rules to derive land-surface parameters results in significant changes in modelled climate and in some improvements in the land-surface diagnostics in selected regions. There is also some evidence that there is a response in the global circulation pattern, which is a focus of further work.

Description: There are at least two needs to be met by the current research efforts on large scale hydrological modelling. The first is for practical conceptual land-surface hydrology schemes for use with existing operational climate and weather forecasting models, to replace the overly simple schemes often used in such models. The second is for models of large scale hydrology which are properly sensitive to changes in physical properties and inputs measured (or predicted) over a wide range of scales, from the point-scale upwards, yet are simple enough in structure to be coupled to climate and weather forecasting models. Such models of large scale hydrology are needed for studying the environmental impact of pollution and changes in climate and land-use, especially the impact On water resources. The UP system (name derived from Upsealed Physically-based) is an attempt to satisfy the second need. It uses a physically-based approach and has a simple structure, yet incorporates sufficient information on sub-grid behaviour to make it a useful tool for the study of environmental impacts over a wide range of scales. The system uses a new approach to large scale modelling, giving physically-based predictions of hourly flows, storages, saturated areas, etc., for regions covering hundreds of thousands of square kilometres. The basic component of the system is the UP element. This has seven water storage compartments (one each for the snowpack, vegetation canopy, surface water, root zone, unsaturated percolation, interflow and groundwater) and allows all the main processes of the terrestrial phase of the hydrological cycle to be represented. A region is modelled as a collection of UP elements, linked by a river routing scheme. Each compartment represents a fixed zone within the area covered by the UP element, and each is related to a physical process such as groundwater flow. Most of the parameterizations for the compartments are in the form of look-up tables, linking the outputs from the compartments to state variables such as the current storage in the compartment. These parameterizations are, in the main, derived from results from physically-based, distributed models applied to the zones (e.g. a groundwater compartment is parameterized using a groundwater model). For large regions modelled using many UP elements, the UP parameters are regionalized using a classification scheme, thus reducing the overall effort spent in parameterization. The development of the UP system is a long-term project involving research into physically-based parameterization of large scale hydrology models, including the effects of sub-grid spatial variations. The first stage involved developing a "blueprint" for the UP element, based on experience with physically-based, distributed river basin modelling and reviews of existing techniques and modelling approaches for large scale and linked atmosphere-hydrology modelling. This paper describes the UP element and the concepts and ideas behind the development of the UP system and, briefly, describes some of the research and development work currently in progress on UP and its parameterization.

Description: The SAFE model has been applied to an acid grassland site, located on base-poor stagnopodzol soils derived from Lower Palaeozoic greywackes. The model predicts that acidification of the soil has occurred in response to increased acid deposition following the industrial revolution. Limited recovery is predicted following the decline in sulphur deposition during the mid to late 1970s. Reducing excess sulphur and NOx deposition in 1998 to 40% and 70% of 1980 levels results in further recovery but soil chemical conditions (base saturation, soil water pH and ANC) do not return to values predicted in pre-industrial times. The SAFE model predicts that critical loads (expressed in terms of the (Ca+Mg+K):Alcrit ratio) for six vegetation species found in acid grassland communities are not exceeded despite the increase in deposited acidity following the industrial revolution. The relative growth response of selected vegetation species characteristic of acid grassland swards has been predicted using a damage function linking growth to soil solution base cation to aluminium ratio. The results show that very small growth reductions can be expected for "acid tolerant" plants growing in acid upland soils. For more sensitive species such as Holcus lanatus, SAFE predicts that growth would have been reduced by about 20% between 1951 and 1983, when acid inputs were greatest. Recovery to c. 90% of normal growth (under laboratory conditions) is predicted as acidic inputs decline.

Description: Time domain reflectometry (TDR) was used to monitor soil water conditions and to evaluate infiltration characteristics associated with rainfall into a volcanic-ash soil (Hydric Hapludand) with a low bulk density. Four 1 m TDR probes were installed vertically along a 6 m line in a bare field. Three 30 cm and one 60 cm probes were installed between the 1 m probes. Soil water content was measured every half or every hour throughout the year. TDR enabled prediction of the soil water content precisely even though the empirical equation developed by Topp et al. (1980) underestimated the water content. Field capacity, defined as the amount of water stored to a depth of 1 m on the day following heavy rainfall, was 640 mm. There was approximately 100 mm difference in the amount of water stored between field capacity and the driest period. Infiltration characteristics of rainfall were investigated for 36 rainfall events exceeding 10 mm with a total amount of rain of 969 mm out of an annual rainfall of 1192 mm. In the case of 25 low intensity rainfall events with less than 10 mm h-1 on to dry soils, the increase in the amount of water stored to a depth of 1 m was equal to the cumulative rainfall. For rain intensity in excess of 10 mm h-1, non-uniform infiltration occurred. The increase in the amount of water stored at lower elevation locations was 1.4 to 1.6 times larger than at higher elevation locations even though the difference in ground height among the 1 m probes was 6 cm. In the two instances when rainfall exceeded 100 mm, including the amount of rain in a previous rainfall event, the increase in the amount of water stored to a depth of 1 m was 65 mm lower than the total quantity of rain on the two occasions (220 mm); this indicated that 65 mm of water or 5.5% of the annual rainfall had flowed away either by surface runoff or bypass flow. Hence, approximately 95% of the annual rainfall was absorbed by the soil matrix but it is not possible to simulate soil water movement by Darcy's law over a long period at farm level due to the local differences in rainfall intensity.

Description: Boron concentrations in rainfall, throughfall and stemflow for Spruce stands, mist, streamwater and groundwater are compared with chloride to assess atmospheric sources and catchment input-output balances for the Plynlimon catchments. In rainfall, boron concentration averages about 4.5 μg-B l-1 and approximately two thirds of this comes from anthropogenic sources. In through-fall and stemflow, boron concentrations are approximately a factor of ten times higher than in rainfall. This increase is associated with enhanced scavenging of mist and dry deposition by the trees. As the sampling sites were close to a forest edge, this degree of scavenging is probably far higher than in the centre of the forest. The throughfall and stemflow concentrations of boron show some evidence of periodic variations with time with peak concentrations occurring during the summer months indicating some vegetational cycling. In mist, boron concentrations are almost twenty times higher than in rainfall and anthropogenic sources account for about 86% of this. Within the Plynlimon streams, boron concentrations are about 1.4 to 1.7 times higher than in rainfall. However, after allowance for mist and dry deposition contributions to atmospheric deposition, it seems that, on average, about 30% of the boron input is retained within the catchment. For the forested catchments, felling results in a disruption of the biological cycle and a small increase in boron leaching from the catchment results in the net retention by the catchment being slightly reduced. Despite the net uptake by the catchment, there is clear evidence of a boron component of weathering from the bedrock. This is shown by an increased boron concentration in a stream influenced by a nearby borehole which increased groundwater inputs. The weathering component for boron is also observed in Plynlimon groundwaters as boron concentrations and boron to chloride ratios are higher than for the streams. For these Goundwaters, increases in boron concentrations are matched linearly by increases in the concentration of the principal ase cation weathering component in the bedrock, calcium. However, the bedrock weathering term is not uniform as the ratio of boron to calcium concentration varies for the different boreholes sampled.

Description: The composition of cloudwater samples collected at Plynlimon, Mid Wales by the Institute of Hydrology is described based on one of the most comprehensive chemical records for deposition in the UK. Comparison with bulk rainwater samples for the same area demonstrates a tenfold enrichment of most elements in cloudwater. Large variations in cloudwater composition occur due to variations in marine and terrestrial/anthropogenic sources, general weather patterns, atmospheric circulation and seasonal effects. All trace metal concentrations are associated with anthropogenic contamination. The lanthanides La, Cc and Pr, and Y are highly correlated in ratios associated with lithogenic sources and fossil fuel combustion. Outliers suggest the influence of catalysts used in the petrochemical cracking process. Plantation forestry significantly enhances the annual deposition of solutes from cloudwater; between 15 and 40% of most constituents to upland forested catchments and around 50% for NO3, B and Cd come from cloud deposition. In upland moorland areas, only 10% of the annual deposited load of inorganic constituents comes from cloudwater.

Description: As identified by the detailed long-term monitoring networks at Plynlimon, increased sediment supply to upland fluvial systems is often associated with forestry land-use and practice. Literature is reviewed, in the light of recent results from Plynlimon sediment studies, to enable identification of the potential ecological impacts of fluvial particulate inputs to upland gravel bed rivers draining forested catchments similar to the headwaters of the River Severn. Both sediment transport and deposition can have significant impacts upon aquatic vertebrates, invertebrates and plants.

Description: Field evidence of finger formation and reformation during Successive rain events over an eight months' observation period from June 1994 until January 1995 is presented. Fingered flow pathways were monitored in a no-tilled, grass-covered water repellent sandy field soil using an automated, stand-alone TDR device. Within a 2 m long and 0.7 m deep transect, 98 three-wire probes were installed horizontally at depths of 4, 12, 20, 30, 40, 55, and 70 cm. The horizontal distance between two adjacent probes was IS cm. Finger formation occurred during distinct rainy periods and was most pronounced under heavy rainfall with initially wet topsoil conditions. The percentage of water infiltrated and transported preferentially through the fingers to the deep subsoil varied between 0 and 80%, depending on the wetting history of the soil and the rainfall characteristics.