In South Africa (SA) approximately 30% of sugarcane is grown under irrigation and there is increasing pressure to demonstrate efficient use of limited water resources. Agronomic practices such as the use of a crop residue layer, changed row spacing, growing suitable varieties and accurate irrigation scheduling could potentially increase water use efficiency (WUE) by saving water and/or increasing yield. The aim of the study was to investigate to what extent WUE of irrigated sugarcane production in SA can be improved by better agronomic practices, and to gain a better understanding of the mechanisms involved in crop response to these factors.An overhead irrigated field experiment was conducted near Komatipoort, South Africa on a shallow, well-drained, sandy clay loam over a four year period (one plant (P) and three ratoon crops (R1, R2 and R3)). Treatments consisted of factorial combinations of variety (N14 and N26), row spacing (single rows spaced at 1.5m and dual rows spaced at 1.8m) and soil surface cover (bare soil and crop residue layer). Measurements included tiller population, interception of photosynthetically active radiation (FIPAR), soil water content, and cane yield at harvest. Crop water use (CWU) was estimated using the water balance approach.This study showed that significant reductions in water use and irrigation requirements, and increases in WUE, are possible by using a crop residue layer to cover the soil. Water savings were largest in P (26% in CWU and 32% in irrigation requirement) but substantial savings were also achieved in R crops (about 15%). It is essential to practice accurate irrigation scheduling to realise these savings, taking into account soil cover and cultivar effects, especially during the period of partial canopy. Although the residue layer caused small reductions in yield in the P, R1 and R2 crops (on average 9%) these were not statistically significant. The combined effect of large CWU reductions and small changes in cane yield resulted in increased WUE (on average 18%).These responses to a residue layer were achieved through a reduced rate of canopy development due to delayed emergence of tillers, causing less green canopy cover and reduced CWU, especially during the period of partial canopy cover when stalk growth has not yet commenced. CWU and FIPAR were affected much less during the subsequent period of stalk growth, thus affecting cane yield minimally, provided irrigation scheduling was adjusted.Variety N14 consistently developed a canopy more rapidly, intercepted more radiation and achieved a higher yield than N26. Row configuration had a significant impact on canopy development, seasonal FIPAR, final stalk population but did not affect cane yield or WUE.The study produced quantitative data for parameterising crop models which will improve their reliability in irrigation management and yield prediction applications.