Applying water efficiently is increasingly important for dairy farmers and other users of surface and groundwater resources to maintain sustainable production. However, irrigation is rarely monitored. We used a questionnaire survey and measurements of five spray irrigation systems working in normal farm conditions to make observations on how efficiently irrigation is being managed. Survey results from 93 dairy farmers showed that, although the farmers believe they know how much water is being applied during irrigation, only 60% make measurements, and about 18% measure irrigation uniformity. Catch-can measurement of irrigation application depth for the different spray systems indicated large variability in application depths during irrigation, and field distribution uniformity ranged greatly between the different systems, decreasing in the order of centre pivots >travelling irrigators> K-line. Changes in irrigation system settings were sometimes made without considering application depths or uniformity. If our five case studies are typical, they may explain the large range of seasonal irrigation amounts recorded in the survey. We recommend that farmers monitor irrigation application depths and uniformity to help manage irrigation water efficiently and to help them estimate the value of irrigation to their enterprise. Keywords: distribution uniformity, water use efficiency, catch cans

The Canterbury Plains are the major cropping area, and an increasingly important dairying area, in New Zealand. Irrigation is necessary on most of the Plains to achieve high productivity. However, competition for limited water resources means that farmers increasingly have to justify their demands for water. Farmers need to know when to irrigate, how much water to apply, and the yield penalty if the crop is not irrigated. For annual arable crops, we have used the maximum potential soil moisture deficit (MPSMD) model and a large rainshelter to answer these questions. Total or economic yield of nearly all the crops tested decreased linearly as MPSMD increased, regardless of the timing of drought. The slope of the regression line is the yield loss with increasing MPSMD. For ryegrass dairy pastures, which grow throughout the year, the model has been adapted to simulate actual soil moisture changes. As most farmers do not measure pasture production directly, we have developed a model for potential pasture production, and a simple relationship between grass growth restrictions and actual soil moisture deficits, that fits experimental farm data. Farmers can also reduce the amount of water they apply by improving the water application efficiency of their irrigation systems. Protocols had been developed to identify causes of poor performance in the field, and how these may be addressed. This information is being used by farmers to improve their irrigation decisions, and irrigation system performance, and has been included in decision support systems for arable and vegetable farmers, and is being developed for dairy farmers.