An innovative research study was established at IGER, North Wyke, Devon, to investigate preferential flow through a poorly structured relatively impermeable soil. Macropore channels were added by a mole plough in order to investigate soil water pathways and chemical transport in a soil in which preferential flow was guaranteed. The investigation focused on water and solute movement through specific flowpaths, namely macropores and mesopores, and the interaction between mobile and immobile zones within the soil. Two large soil blocks (1 m2 by 0.85 m) of the Hallsworth series were removed from the field and placed on sand tables so that a suction could be induced at the base of the soil block. The edge was sealed using paraffin wax. Eight tensiometers and suction cup lysimeters were installed in each block together with fifteen pairs of time-domain reflectometry wave guides. A regular spacing pattern was employed so that spatial variations could be easily identified. Samples were collected from suction cup lysimeters every 4 hours. Soil water status was observed from the TDR probes daily and from tensiometers every 10 minutes. Five tracer experiments were conducted; three involved the miscible displacement of chloride at concentrations of 100 and 250 mg I"' and two used nitrate (500 mg l ') and chloride (2500 mg 1') applied as a pulse. Tracer and irrigation water was applied through a misting system at an irrigation rate of 2.76 mm h-1. Three techniques were used to examine soil structure in the macropore and mesopore pore size range to investigate potential flowpaths in more detail. The profile tracing method (PTM). binary transect method (BTM) and resinated core section method (RCSM) provided useful quantitative structural information. Soil water status averaged over a large sampling volume (TDR, 1540000 mm3) was considered to be stable through time. Detailed observations of soil water suction using tensiometers showed that soil water conditions remained unsaturated, at approximately 10 to 20 cm H2O, and varied by 3 cm H2O throughout the experiment. Suction varied depending on the location of each tensiometer with respect to position within or between aggregates. Results based on Poiseuille's law and suction data showed that the flowpaths were predominantly mesopores. This result was supported by breakthrough curve analysis for the bulk of the soil although macropore flow was observed towards the mole drain. Flow rates observed from tracer movement varied throughout the soil regardless of depth. Chloride moved quickly towards the mole drain and the arrival of tracer was recorded within 4 hours. Time to breakthrough monitored at the suction cups varied from 4 to 76 hours. When the concentration gradient between applied solute and antecedent solute was large, reduced time to attain peak concentration was noted. As the concentration gradient reduced, speed of rise to peak concentration increased. An advection-dispersion model (CLEARY) fitted change in observed solute concentration through time at the suction cup lysimeters well. The study concluded that although water moved rapidly through the soil, the tortuous nature and increased contact with soil particles encountered as water moved through the mesopores resulted in water with matrix flow solute characteristics.

The Rowden Moor experimental site (A.F.R.C. I.G.E.R., North Wyke) provided an opportunity to characterise discharge regimes, elucidate runoff generation mechanisms and to consider implications for solute movement under natural and drained conditions. Research was conducted on a heavy clay grassland soil in an area of high rainfall (1053 mm a ˉ¹) in South West England. A combined hydrometric and tensiometric study was undertaken within a nested experimental design (1 m² to 1 ha) on one undrained and one drained site throughout a drainage season (October to March). Results at the hectare scale demonstrated that drainage did not substantially alter the volume of field runoff ( ~ 400 mm) but did change the dominant flowpaths. Drainage diverted water from surface/near surface routes to depth so that drain storm runoff was lagged by some 30 minutes over undrained site discharge. The drained site also exhibited a more peaky regime, with a maximum daily discharge of 45 mm being almost twice that for the undrained field. At the field and plot scale, the significance of macropore flow was noted. To investigate this in more detail, a tracer experiment was performed on an isolated soil block which had been mole drained and so had enhanced macroporosity. Macropore flow was generated under unsaturated conditions (little matric potential response and no water table was identified). Stable oxygen concentrations were δ18O +3.5 and -5.8 in tracer and background water respectively. Drainflow indicated that there was rapid interaction between applied tracer and soil water (peak flow δ18O -1.1). Thus, the matrix-macropore interface was not a boundary between two separate domains of old and new water, high and low conductivity but a site of rapid interchange and mixing. Temporal variability of soil status and malric water composition, also indicated that limited areas of the matrix were capable of transmitting rapid flow. It became clear that even in a heavy clay soil such as that found at Rowden, where macropore flow was promoted by drainage operations, the matrix still had an important role to play. On the basis of potential, soil moisture and observation of tracers, it is proposed that discrete (finger-like) volumes of the matrix are capable of rapid water transmission. Although it was frequently impossible to relate moisture content and soil water potential because instrumentation monitored different volumes of soil, hysteretic soil moisture behaviour over the drainage season was evident in both data sets. This study confirmed the importance of rapid subsurface runoff generation mechanisms on drained soils, but noted that discontinuous translatory flow in the matrix and macropore flow occurred and that the two ‘domains’ were inextricably linked. Further work should be undertaken at the detailed scale to elucidate the soil characteristics which promote rapid runoff mechanisms and the consequences for water quality, especially where the soil subsurface represents a major reservoir (e.g. nitrates).