Table Of ContentsNext Page

Wheat and canola water requirements and the effect of spring irrigation on crop yields in the Central Murray Valley

Sam North1, Phil Eberbach2 and John Thompson1

1 NSW DPI, PO Box 736, Deniliquin, NSW 2710. Email
CSU, School of Agriculture & Wine Sciences, Locked Bag 588, Wagga Wagga, NSW 2678.


Grain yield and water use data were obtained from field irrigation experiments in the Central Murray Valley. Mid-season crop factors (Kc mid) were found from this data to be 1.2 and 1.15 for the wheat and canola respectively. Yield (Y) and applied water (WA) production functions were derived for flood irrigated wheat and canola crops growing on red-brown earth soils in the southern Murray-Darling Basin:

Ycanola = -3.40 + 2.5910-2.WA – 2.4410-5.WA2 (R2 = 0.92)
= -2.25 + 2.3010-2.WA – 1.2510-5.WA2 (R2 = 0.94).

This data was used to determine profit maximising irrigation strategies for these crops.

Key Words

canola, wheat, production function, deficit irrigation, crop factor


The current focus to improve water use efficiency in Australia’s irrigation industries is generally interpreted as encouragement for a shift to higher value commodities or to more capital intensive irrigation systems. Such a shift would require irrigation farmers in the NSW Central Murray Valley (CMV) to increase the level of capital invested in their farms without any guarantee of improved reliability in their water supply, and they would have to raise this capital using their current (relatively) low value enterprise mix. Few farmers are willing to accept this risk, particularly as their soils are generally unsuited to higher value horticultural crops and/or pressure application systems. The dilemma they face is how to remain profitable as water becomes increasingly scarce and costly, without needing to raise large amounts of capital and exposing themselves to greater economic risk for an uncertain outcome.

This study examined the potential of deficit irrigation strategies for improving farm profitability for the dominant water user group (i.e. rice growers) in the NSW CMV. The study aimed to determine:

  • the relative responses of wheat and canola to a range of spring irrigation frequencies
  • local crop factors (Kc) and the shape of a generalised Kc curve for wheat and canola
  • crop production functions for flood irrigated canola and wheat in the CMV to indicate the profit maximising irrigation strategy when water is limited


Field experiments were conducted in 1998 and 1999 at Deniliquin, NSW to compare the effects of 4 spring irrigation treatments on wheat var. Janz and canola var. Oscar. All plots were irrigated prior to sowing and the 4 spring irrigation strategies were: no spring irrigation (I0); one spring irrigation (I1); irrigation at an evaporation minus rainfall (E-R) deficit of 90 mm (IDef); and full irrigation at an E-R deficit of 60 mm (IFull).

One site was used in 1998 and 2 sites in 1999. Both sites were in border check, flood irrigation bays on red-brown earth soils (Stace, et al. 1968). Site 1 was a Subnatric Calcic Red Sodosol (Isbell 1996), known locally as a Birganbigil Loam (Smith 1945). It had only fair internal drainage and was classified suitable for growing rice. Site 2 was a Brown Chromosol (Isbell 1996), known locally as a Cobram Loam (Smith 1945). It was lighter textured, with a slightly deeper A horizon, a fine sandy clay B2 horizon and a higher proportion of coarse sand at all depths. It had good internal drainage and was classified unsuitable for growing rice because of excessive deep percolation.

Agronomic measurements were made to assess treatment effects on crop growth and yield. Hydrological data was collected in 1998 and used to calculate crop water use (0-1 m) according to the water balance equation:

ETc = crop evapotranspiration; P = rainfall; Ia = depth of irrigation water; CR = capillary rise; ΔS = change in profile soil water content; RO = runoff; and DP = deep percolation. Potential crop evapotranspiration (ETo) was determined from Deniliquin class A pan data (Epan) according to

ETo = 0.88 (Epan)0.91 (R2 = 0.98, P < 0.001, n = 1721).


Relative responses to spring irrigation

The total depth of water applied to each plot was calculated (WA = Ia + R – ΔSRO). WA varied from 300 to 600 mm and, in response, Y varied from 4.8 to 7.3 t/ha in the wheat and from 2.9 to 3.6 t/ha in the canola. Regression analysis showed there was a strong linear relationship between grain yield (Y) and WA within each crop type (R2 = 0.97; se obs. = 0.3; P < 0.001). Y increased by 9 and 3 kg/ha/mm of WA in the wheat and canola respectively. Septoria reduced wheat yields by an estimated 0.75-1 t/ha in the IFull treatment in 1998.

Farmers are interested in the yield response per megalitre (ML) of irrigation water. To assess this, the yield response to water applied in spring (i.e. WS = I+R-RO) was determined for each crop. Step-wise multiple linear regression showed there was a strong correlation between Y and WS within each crop type (R2 = 0.97; se obs. = 0.25; P < 0.001), with yields increasing by 4.6 and 1.9 kg/ha/mm of WS for the wheat and canola respectively. The response of canola yield to WS was not significantly different from zero (t = 0.73).

Local crop factors and Kc curves

Crop factors (Kc) were determined for consecutive 7 to 10 day periods for the crops grown in 1998:

This data was used to derive Kc curves for each crop (Figure 1), with growth stage lengths defined by Allen, et al. (1998). Neither the wheat nor the canola were water stressed between 20th Aug and 8th Sept and a zero flux plane existed at 100 cm. Therefore, ΔS (0–1.0 m) during this period reflected maximum crop water use in the mid-season stage and Kc mid was estimated to be 1.2 and 1.15 for the wheat and canola respectively.

Production functions for flood irrigated wheat and canola

Quadratic yield (Y) and applied water (WA) production functions were obtained for flood irrigated wheat and canola crops growing on red-brown earth soils in the southern Murray-Darling Basin (MDB) using data from this and other experiments (Figure 2). The production functions were determined to have the form:

Ycanola = -3.40 + 2.5910-2.WA – 2.4410-5.WA2 (R2 = 0.92)
= -2.25 + 2.3010-2.WA – 1.2510-5.WA2 (R2 = 0.94).

The potential transpiration efficiency for grain (Te) of canola and wheat crops in the southern MDB was estimated from this data to be 15 and 20 kg/ha/mm respectively. Direct soil evaporation (i.e. x-axis intercept) was estimated to be 130 and 110 mm for canola and wheat respectively (Figure 2).


Following English & Raja’s (1996) method, revenue and cost functions were used to determine WA when (1) yields are maximised (Wm), (2) when profits are maximised if land is limiting (Wl) and (3) when profits are maximised if water is limiting (Ww). It was assumed that: crop price equals the long term average ($180/t for wheat and $350/t for canola); crops receive median growing season rainfall (250 mm); crops are pre-irrigated in autumn (100-150 mm); and spring irrigations apply 60 to 75 mm. The results showed that:

  • irrigating to maximise yield is not a profit maximising strategy. The fact that yields at Wl are not much lower than at Wm indicates that water is cheap ($41.90/ML; NSW DPI 2008) compared to crop price.
  • the profit maximising strategy when land is limiting is to irrigate wheat with 4.9 ML/ha to achieve 8 t/ha (i.e. max net return/ha; Yaron & Bresler 1983). However, if water is plentiful, CMV farmers will irrigate rice rather than wheat as rice has a net return of $1,500 to $2,000/ha (NSW DPI 2008) compared to optimal net returns of $700/ha and $620/ha for wheat and canola respectively estimated using he production function analysis.
  • the profit maximising strategy when water is limiting is to maximise average net returns per ML (Yaron & Bresler 1983) and this is achieved by spreading available supplies over the largest feasible area by pre-irrigating canola in autumn and, if necessary, irrigating once in spring.

Figure 1. Average Kc of consecutive 7-10 day periods (●) for the canola (top) and wheat (bottom) at Site1 in 1998. The solid line (––) shows the Kc curve for the IFull treatment, with growth stages lengths determined according to Allen et al’s (1998) criteria and crop observations. The dashed line (– –) shows Kc ini from Allen et al. (1989) which is unadjusted for wetting frequency and ETo and then Kc end showing the effect of deficit irrigation.

Figure 2. Grain yield (Y) as a function of seasonal applied water (WA) for canola (left) and wheat (right) in the southern MDB. A regression line was fitted to the data from this experiment (○) as well as data from dryland crops at Deniliquin (●, Ian Lea, pers. comm.) and from canola experiments at Tatura (∇, Taylor et al. 1991) and from wheat experiments at Yanco (Δ, Cooper 1980) and Griffith (▲, Steiner et al. 1985). Vertical lines show median April-Oct rainfall at Deniliquin; horizontal lines show average district yields; dashed lines show transpiration efficiency (Te) for grain. Tasmanian data (♦, Mendham et al. 1984) helped define canola Te.

Phenological development in the wheat and canola was well predicted by the models of Stapper & Fischer (1990) and Robertson et al. (2002). This allowed crop stage lengths and thence Kc curves to be predicted for wheat and canola in the CMV for a range of sowing dates. Together with Deniliquin weather data (1960 to 2005), these were used in the single Kc model (Allen, et al. 1998) to estimate wheat and canola irrigation requirements. The model was run using two irrigation intervals based on: (1) a readily available water capacity (RAWC) of 55 mm; and (2) a RAWC of 75 mm. The model output showed that, in the CMV:

  • canola requires one less irrigation on average than wheat if fully irrigated.
  • wheat will nearly always experience at least one period of water stress in spring if not irrigated.
  • spring irrigation of canola was not needed in 6% and 11% of years when RAWC was 55 and 75 mm respectively. This is attributed to a shorter mid-season stage and lower Kc mid .
  • increasing RAWC from 55 to 75 mm reduced the median number of irrigations required in spring from 3 to 2 and reduced deep percolation, and hence WA, by approximately 10 mm on average over all years.
  • early sown crops require fewer spring irrigations than later sown crops.

The lack of response to spring irrigation by the canola is, in part, attributed to its indeterminate nature. This permitted leaf area and flowering to recover in I0 and IDef when rain fell during the vegetative and flowering stages. Such a recovery was not possible in the (determinate) wheat. It is postulated that the inability of the canola to achieve yields as high as Mendham et al. (1984) and Taylor et al. (1991) (i.e. 4-5 t/ha) is due to:

  • high temperatures during pod/seed development. It is estimated that yields are reduced by 0.7 t/ha for every 1oC that average daily mean temperature during pod/seed development exceeds 13oC
  • a root system which is less able to penetrate dense/massive clay sub-soils. This, combined with very low soil hydraulic conductivities and high evaporative demand in spring, leads to greater levels of crop water stress in canola compared to wheat when grown on soils similar to the Birganbigil loam.


With limited irrigation supplies in the cMV, the recommended profit maximising irrigation strategy for wheat is to pre-irrigate in autumn and apply 3-4 spring irrigations to attain a target yield of 6-7 t ha-1. Applying more water to attain a higher yield is not recommended because of diminishing marginal returns. A single irrigation timed for flowering is not recommended as it does not ensure the greatest average return per ML and because there is a high risk of crop failure if spring rainfall is insufficient to support a large biomass. The profit maximising irrigation strategy for canola is to pre-irrigate and aim for a target yield of 3 t/ha by irrigating once in spring to ensure 1) good moisture until 3 weeks past full flower and 2) that biomass at flowering is maximised. Canola’s relative insensitivity to water stress during pod fill and grain ripening means that deficit irrigation can be practiced after flowering. When water is limiting, profits will be maximised if available supplies are allocated to canola. Its lower water requirement allows water to be spread over a larger area, maximising average net returns per ML. Its poor response to irrigation, however, makes this a risky strategy. To mitigate this risk, it is recommended to:

  • sow early - to avoid high temperatures later in spring and to reduce the number of irrigations required.
  • only irrigate canola on better soil types – hard soils restrict root growth and water supply to the crop and this combines with high evaporative demand and high temperatures in spring to limit yield potential.


Allen RG, Pereira LS, Raes D and Smith M (1998). Crop evapotranspiration: guidelines for computing crop water requirements. Irrigation and Drainage Paper 56. FAO, Rome.

Cooper JL (1980). The effect of N fertiliser and irrigation frequency on a semi-dwarf wheat in SE Australia. II. Water use. Australian Journal of Experimental Agriculture and Animal Husbandry. 20, 365-369.

English M and Raja SN (1996). Perspectives on deficit irrigation. Agricultural Water Management. 32, 1-14.

Isbell RF (1996). The Australian Soil Classification. CSIRO: Collingwood, Australia

Mendham NJ, Russell J and Buzza GC (1984). The contribution of seed survival to yield in new Australian cultivars of oil-seed rape (Brassica napus). Journal of Agricultural Science, Cambridge. 103, 303-316.

NSW Department of Primary Industries (2008).Farm budgets and costs. Accessed April 2008.

Robertson MJ, Watkinson AR, Kirkegaard JA, Holland JF, Potter TD, Burton W, Walton GH, Moot DJ, Wratten N, Farre I and Asseng S (2002). Environmental and genotypic control of time to flowering in canola and Indian mustard. Australian Journal of Agricultural Research. 53, 793-809.

Smith R (1945). Soils of the Berriquin Irrigation District. CSIR, Melbourne.

Stace HCT, Hubble GD, Brewer R, Northcote KH, Sleeman JR, Mulcahy MJ and Hallsworth EG (1968). A Handbook of Australian Soils. Rellim. Tech. Pubs.: Glenside, S.A.

Stapper M and Fischer RA (1990). Genotype, sowing date and plant spacing influence on high-yielding irrigated wheat in southern NSW. Australian Journal of Agricultural Research. 41, 997-1019.

Steiner JL, Smith RCG, Meyer WS and Adeney JA (1985). Water use, foliage temperature and yield of irrigated wheat in south-eastern Australia. Australian Journal of Agricultural Research. 36, 1-11.

Taylor AJ, Smith CJ and Wilson IB (1991). Effect of irrigation and nitrogen fertilizer on yield, oil content, nitrogen accumulation and water use of canola (Brassica napus L.). Fertilizer Research. 29, 249-260.

Yaron D and Bresler E (1983). Economic analysis of on-farm irrigation using response functions of crops. In Advances in Irrigation. Volume 2. Ed. Hillel D. pp. 223-255. Academic Press: New York.

Top Of PageNext Page