Increases in crop water use efficiency have been found achieved through greater precision in

irrigation scheduling and the use of irrigated crop management strategies such as regulated deficit and

deficit irrigation. However, limitations exist in the use of soil moisture sensors and/or the water

balance approach method of irrigation scheduling. A key limitation with using either of these

approaches for irrigation scheduling is that they do not provide a measure of actual plant water status.

Crop growth and response to irrigation is a function of plant water status and depends on soil water

status, evaporative demand, the rate of water flow through the plant and the corresponding hydraulic

flow resistance between the bulk soil and the appropriate plant tissue. Hence, this project investigated

the potential to use plant based measurements for commercial irrigation scheduling of cotton.

The first year (2005/06) of this project evaluated the potential to use stem diameter sensors for

irrigation scheduling in cotton under a lateral move machine near Leyburn. While the first season

results were encouraging, the second season (2006/07) conducted on furrow irrigation at Nandi across

a range of irrigation schedules and three crop varieties found weaker relationships (i.e the technique

lacks robustness). There was also significant plant to plant variation in sensor responses. The key

recommendation from this work is that stem diameter sensors can be used to identify plant stress

responses associated with irrigation. However, their benefits over traditional irrigation scheduling

technologies are marginal and these sensors will continue to have limited application as an irrigation

scheduling and assessment tool in cotton unless appropriate threshold levels can be identified which

take into account varietal differences and crop conditioning.

During 2006/07, the project evaluated the relationships between hyperspectral canopy reflectance data

and that of plant water status and identified band widths correlated to plant water status when

measured during a normal commercial irrigation cycle. The 2007/08 trial used two sites (Pampas and

Cecil Plains) to evaluate remote methods of plant based sensing (i.e. satellite imagery and groundbased NDVI) and to test the robustness of the relationships between the hyperspectral bandwidth data to changes in crop water status. The active sensor technology (i.e. NDVI) was able to identify varietal differences and varying yield responses under all climatic conditions encountered.

However ground rig speed and sensor distance above the canopy was found to be critical. The passive sensors (i.e. satellite or handheld radiometric) were found to be hampered by poor atmospheric conditions and high variability in localised sunlight intensity. No significant relationships were identified between leaf water potential and either the NDVI or hyperspectral data.

Other outputs include an analysis of historical yield monitor data across five cotton regions to identify

field scale trends in spatial variability. A grower guide to plant based sensing for irrigation

scheduling has also been produced.