This paper investigates the influence of rainfall intensity, and its components drop size and impact frequency, on steady state flow through sealed soils. Infiltration rates are often observed to increase with greater rainfall intensity. There appears to be no consensus in the literature on the reasons for this behaviour, making it difficult to incorporate in infiltration models. Seals were formed on 2 swelling Black Vertosols and a contrasting non-swelling Brown Sodosol, under simulated rainfall treatments with a range of rainfall intensities, drop sizes, and drop size distributions. For Vertosols, aggregate size composition of the seal, sub-seal moisture content, and suction were similar across rain treatments and between soils. Steady infiltration rate and conductivity increased 3-fold as rainfall intensity increased from 30 to 120 mm/h for Vertosols. The component of intensity responsible for this trend was isolated, with a strong positive linear relationship found between conductivity and raindrop impact frequency (R2 = 0.87). This trend was attributed to increased water entry by hydraulic penetration of drops through pores in the seal as the frequency of drop impacts increased, on more porous Vertosol seals. In contrast, rain intensity and drop impact frequency had little effect on steady infiltration rates and conductivity for the Sodosol. Abundant fine particles in the seal decreased porosity, apparently leaving few pores of large enough diameter to allow hydraulic penetration. While hydraulic penetration was suppressed, other seal properties for the Sodosol were influenced by rainfall characteristics. Larger, more energetic drops produced seals with lower moisture content and higher subseal suction than seals formed under other treatments. Results of this study indicate that water entry through sealed surfaces may consist of an actively driven component associated with hydraulic penetration of water through larger pores (jetting), as observed elsewhere in early stages of surface sealing. Water entry by this mechanism continues through to steady state on soils where significant numbers of larger pores remain open at the soil surface. This mechanism would act alone, or in addition to other mechanisms such as erosive stripping or microrilling of seals, to explain widely observed increases in infiltration rates with increasing rainfall intensity.

This paper reports comparisons between aggregate breakdown on wetting by rainfall with breakdown measured by a range of alternative methods. It also reports correlations between measured breakdown and steady infiltration rates of simulated rain of high and low energy, and hydraulic conductivities of surface seal layers formed under high energy rain. A wide range of soils in eastern Australia were studied. Highly significant correlations were found between measurements of aggregate breakdown to < 125 �m caused by rainfall wetting and both steady infiltration rates and hydraulic conductivities. Significant, but poorer correlations were found between steady infiltration rates and breakdown resulting from immersion wetting. Deletion of swelling soils from the data set greatly improved correlations between steady infiltration rates of high energy rain and breakdown measured by both immersion and tension wetting, showing that these methods of wetting ace particularly inappropriate for swelling soils. No correlation was found between infiltration rates and measured clay dispersion. Different relationships between the proportion of particles (%) < 125 �m at the soil surface (P125) and steady infiltration rates of low and high energy rain indicated that compaction of the soil surface layer, rather than increased aggregate breakdown, is a major cause of surface sealing by raindrop impacts. Measurements of fall cone penetration confirmed that drop impacts had compacted the surface layer. Suctions across the surface seal were related to P125 in that layer, and the relationship obtained was used in calculating hydraulic conductivities. The results confirm that measurement of aggregate breakdown under rainfall wetting produces results of much greater relevance to soil behaviour under field conditions than do tests based on immersion and tension wetting.

No abstract

The Australian cotton industry and governments have funded research into the deep-drainage component of the soil–water balance for several decades. Cotton is dominantly grown in the northern Murray–Darling and Fitzroy Basins, using furrow irrigation on cracking clays. Previously, it was held that furrow irrigation on cracking clays was inherently efficient and there was little deep drainage. This has been shown to be simplistic and generally incorrect. This paper reviews global and northern Australian deep-drainage studies in irrigation, generally at point- or paddock-scale, and the consequences of deep drainage. For furrow-irrigated fields in Australia, key findings are as follows. (i) Deep drainage varies considerably depending on soil properties and irrigation management, and is not necessarily ‘very small’. Historically, values of 100–250 mm year–1 were typical, with 3–900 mm year–1 observed, until water shortage in the 2000s and continued research and extension focussed attention on water-use efficiency (WUE). (ii) More recently, values of 50–100 mm year–1 have been observed, with no deep drainage in drier years; these levels are lower than global values. (iii) Optimisation (flow rate, field length, cut-off time) of furrow irrigation can at least halve deep drainage. (iv) Cotton is grown on soils with a wide range in texture, sodicity and structure. (v) Deep drainage is moderately to strongly related to total rainfall plus irrigation, as it is globally. (vi) A leaching fraction, to avoid salt build-up in the soil profile, is only needed for irrigation where more saline water is used. Drainage from rainfall often provides an adequate leaching fraction. (vii) Near-saturated conditions occur for at least 2–6 m under irrigated fields, whereas profiles are dry under native vegetation in the same landscapes. (viii) Deep drainage leachate is typically saline and not a source of good quality groundwater recharge. Large losses of nitrate also occur in deep drainage. The consequences of deep drainage for groundwater and salinity are different where underlying groundwater can be used for pumping (fresh water, high yield; e.g. Condamine alluvia) and where it cannot (saline water or low yield; e.g. Border Rivers alluvia). Continuing improvements in WUE are needed to ensure long-term sustainability of irrigated cropping industries. Globally there is great potential for increased production using existing water supplies, given deep drainage of 10–25% of water delivered to fields and WUE of <50%. Future research priorities are to further characterise water movement through the unsaturated zone and the consequences of deep drainage.

Past studies have shown that soil-specific calibrations are required to attain a higher level of accuracy when measuring soil water content with ThetaProbe and ECHO probe soil water sensors, particularly in swelling clay soils. Both probes were assessed for their capacity to accurately monitor soil water in a deep drainage study on a Black Vertosol. Probes were trialled in situ and calibrated against hand-sampled volumetric measurements. The generic calibrations given by the manufacturers resulted in significant errors in water content estimates for both probes. Using the generic calibration, ECHO probes under-estimated water content by 0.10–0.2 m3/m3, whereas ThetaProbes under-estimated by 0.04 m3/m3 at the wet end and over-estimated by 0.08 m3/m3 at the dry end. The soil-specific calibrations significantly improved the accuracy of both probes. ThetaProbes were chosen for the drainage study. The calibration allowed for accuracy across the full wet–dry range to within 0.001–0.004 m3/m3 of volumetric measurements. ECHO probes were less accurate at the wet end, but still determined soil water content to within 0.02–0.05 m3/m3 of volumetric measurements.