Drained and undrained grassland lysimeter plots were established in 1982 on a clay loam of the Hallsworth series at a long-term experimental site in south-west England. The plots were continuously grazed by beef cattle, and received fertilizer at either 200 or 400 kg N ha -I per annum to the existing permanent sward, or at 400 kg N ha-' to a new sward, reseeded to perennial ryegrass following cultivation. Drainage water was monitored at V-notch weirs and sampled daily for the analysis of nitrate-N. Seven years of data are presented (five years for the reseeded swards). On the drained plots a large proportion of the rainfall was routed preferentially down large pores to the mole drains, whilst on the undrained plots, drainage was mainly by surface runoff. The average quantities of nitrate N leached per year were 38.5,133.8 and 55.7 kg ha-' from the old sward that received 200 and 400 kg N ha-', and from the reseed that received 400 kg N ha-' fertilizer, respectively. Ploughing and reseeding resulted in a two-fold reduction in leaching, except during the first winter after ploughing, and twice as much leaching occurred after a hot, dry summer as after a cool, wet one. Nitrate concentrations in drainage from either drained or undrained plots were rather insensitive to rainfall intensity, such that concentration was a good predictor of nitrate load for a given drainage volume. The drainage volume determined the proportion of the leachable N that remained in the soil after the winter drainage period. Initial (peak) concentrations of nitrate N ranged, on average, from 55 mg dm-3 for the drained old sward that received 400 kg N ha-' fertilizer, to 12 mg dm-3 for the undrained sward at 200 kg N ha-' fertilizer input. Concentrations of nitrate N in drainage from similar, unfertilized plots rarely exceeded 1 mg dm-3. The results suggest that manipulating the nitrate supply can lessen leaching and that the route of water through soil to the watercourse determines the maximum nitrate concentration for a given load.

Soil water regimes and water balances are presented for a series of drained and undrained experimental grassland plots, intended to examine the agronomic consequences of drainage. Although drainage has lowered the watertables and reduced the duration of waterlogging in the drained plots, its effects in terms either of the total water quantities leaving the site or of peak flows is quite small. The major effect of drainage is to alter the route of water loss from the site. In its undrained state, the soil is waterlogged for the majority of the winter, incident rainfall cannot infiltrate, and water leaves as surface runoff or near-surface flow. The introduction of fissures by mole drainage both provides an outlet and enhances the macroporosity, so that the rain moves rapidly through the soil and appears as drainflow. Consequently, the additional delay in generating peak runoff through the drainage system is only of the order of 30 minutes on this site.

Spatial variation of soil properties can be studied by use of the concept of fractals, and the overall roughness of a surface characterized by its fractal dimension. Soil surface strength was recorded by both cone penetrometer and field shear vane, along a series of transects, at 10 cm spacings. In the natural state, such data gave fractal dimensions very close to 2, implying an almost totally random data set, and only the deliberate action of cultivations introduced an order of persistence. Microtopography, measured by a profile meter at 1 cm intervals, had fractal dimensions of 1.7 on two separate sites. Problems are identified specifically in the choice of a suitable model for fitting the semi-variogram, and in the collection of an efficient data set.

No abstract

The potential impact of climate change by the year 2050 on British grazing livestock systems is assessed through the use of simulation models of farming systems. The submodels, consisting of grass production, livestock feeding, livestock thermal balance, the thermal balance of naturally ventilated buildings and a stochastic weather generator, are described. These are integrated to form system models for sheep, beef calves and dairy cows. They are applied to scenarios representing eastern (dry) lowlands, western (wet) lowlands and uplands. The results show that such systems should be able to adapt to the expected climatic changes. There is likely to be a small increase in grass production, possibly allowing an increase in total productivity in some cases

Slope profile sequences are as much influenced by their boundary conditions as they are by the processes operating, The development of a simple slope under a single process can adopt a variety of forms depending on initial and boundary conditions, and a given form can be derived from several different processes. For example, convexoconcave forms develop under soil creep with basal accumulation, under wash with an initial step form, and under mixed wash and creep processes.

The potential impact of climate change by the year 2050 on intensive livestock systems in Britain is assessed through the use of simulation models of farming systems. The submodels comprise livestock feeding, livestock thermal balance and the thermal balance of controlled environment buildings and a stochastic weather generator. These are integrated to form system models for growing pigs and broiler chickens. They are applied to scenarios typical of SE England, which is the warmest region of the country and represents the worst case. For both species the frequency of severe heat stress is substantially increased, with a consequent risk of mortality. To offset this, it would be necessary to reduce stocking densities considerably, or to invest in improved ventilation or cooling equipment. Other effects on production are likely to be small

No abstract