The present study was conducted from 1998 to 2000, to evaluate seasonal water use and soil-water extraction by Kabuli chickpea (Cicer arietinum L.). The response of three cultivars to eight irrigation treatments in 1998/99 and four irrigation treatments in 1999/2000 at different growth stages was studied on a Wakanui silt loam soil in Canterbury, New Zealand. Evapotranspiration was measured with a neutron moisture meter and water use efficiency (WUE) was examined at crop maturity. Water use was about 426 mm for the fully irrigated treatment and at least 175 mm for the non-irrigated plants. There was a significant correlation (P<0·001) between water use and biomass yield (R2=0·80) and water use and seed yield (R2=0·75). There were also highly significant (P<0·001) interacting effects of irrigation, sowing date and cultivar on WUE and the trend was similar to that for seed yield. The estimated WUE ranged from 22–29 kg DM/ha per mm and 10–13 kg seed yield/ha per mm water use.

The three chickpea cultivars were capable of drawing water from depths greater than 60 cm. However, most of the water use (0·49–0·93 mm/10 cm soil layer per day) came from the top 0–30 cm, where most of the active roots were concentrated. The study has shown that using actual evapotranspiration and water-use efficiency, the biomass yield and seed yield of Kabuli chickpeas can be accurately predicted in Canterbury. Soil water shortage has been identified as a major constraint to increasing chickpea production. Drought was quantified using the concept of maximum potential soil moisture deficit (Dpmax) calculated from climate data. Drought responses of yield, phenology, radiation use efficiency and yield components were determined, and were highly correlated with Dpmax. The maximum potential soil moisture deficit increased from about 62 mm (irrigated throughout) to about 358 mm (dryland plots). Chickpea yield, intercepted radiation and the number of pods per plant decreased linearly as the Dpmax increased. Penman's irrigation model accurately described the response of yield to drought. The limiting deficit for this type of soil was c. 165 and 84 mm for the November and December sowings in 1998/99 and 170 mm in 1999/2000. Beyond these limiting deficits, yield declined linearly with maximum potential soil moisture deficits of up to 358 mm. There was little evidence to support the idea of a moisture sensitive period in these Kabuli chickpea cultivars. Yield was increased by irrigating at any stage of crop development, provided that the water was needed as determined by the potential soil moisture deficit and sowing early in the season.