The maxim ‘it’s not what you’ve got, it’s what you do with it’ applies particularly to farming, and water use efficiency benchmarks are the best and simplest way yet of assessing how well farmers are using one of their key resources (water). There are other important resources, however, and ultimately crop simulation models may be the best way of integrating them.
The use of water use efficiency benchmarks has been twisted somewhat in the many interpretions of earlier work (eg. French and Schultz 1984). The original idea was that potential yield could be estimated from rainfall, and if available stored soil water, using some of the equations in French and Schultz. The equation is:
Potential Yield = (Growing Season Rainfall + Stored Water – Soil Evaporation) x Transpiration Efficiency
Ideally GSR would be the rainfall between sowing and physiological maturity of the crop. Stored water would be the difference between soil water at sowing and at physiological maturity, to the depth of the root zone, less drainage and runoff. Soil evaporation was estimated to be 110mm, but found to range between 50 and 150mm. They suggested an estimate – that 60% of growing season rainfall be used, up to growing season rainfall of 150mm; 180mm is neater. Transpiration efficiency was estimated at around 20kg/ha/mm for the standard height wheats of the time. This is approximately a 35% harvest index of crop biomass produced with a transpiration efficiency of 55kg/ha/mm (a fairly robust number, which varies with the dryness of the atmosphere). In this project we have measured a wide range of harvest indices (Figure 32), but we would also expect a wide range in soil evaporation.
The difference between potential yield and actual yield would help to diagnose crops in which some factor (eg. nutrition, disease, weeds) was limiting yield, over and above rainfall. Crops could be compared from year to year and paddock to paddock on the basis of how close they were to achieving the yield potential set by rainfall – allowing for other environmental factors that might also affect yield potential.
The current practice of using French and Schultz’s equations is to rearrange them, often ignoring stored water, and approximating growing season rainfall with April-October rainfall:
Transpiration Efficiency = Potential Yield / (Growing Season Rainfall – Soil Evaporation)
and to further assume that soil evaporation is fixed, at 110mm or 60mm. This means that the estimate of transpiration efficiency (mistakenly called water use efficiency) becomes the benchmark describing not only whether potential yield was achieved, but in addition any errors caused by approximating GSR with April-October rainfall, ignoring stored water, drainage and runoff, and in the assumption of soil evaporation.
Here we have to estimate GSR with April-October rainfall (because daily rainfall measurements aren’t available, and soil water wasn’t measured exactly at pre-sowing or physiological maturity). We also have to estimate soil evaporation, but rather than try to discuss an erroneous estimate of transpiration efficiency as though it described only water use efficiency, we will estimate maximum transpiration efficiency as being 28 kg/ha/mm - a 50% harvest index of 56 kg/ha/mm transpiration efficiency for biomass, and talk instead about potential yield.
Rainfall
Unfortunately rainfall was not always measured on each paddock. Where rainfall data was not available for one paddock, we have used data from the other paddock in the pair. When no rainfall data was available, data from the rainfall interpolation exercise has been used.
Growing season rainfalls were low especially in 2006, and only a little better in 2008 (Table 16). The year 2007 was best for growing season rainfall, but much of this came in winter , with August-September being much drier than usual and below-average growing season rainfall recorded as a result. Summer fallow (November-March) rainfalls were almost always at least 60mm, apart from the Donald paddocks in 2008. Most of the pairs had similar rainfall recorded (often because the same gauge was used). Patchewollock was the most uneven pair, with the Till gauge receiving an additional 10mm in the growing season in 2007 and 2008.
Table 16. Monthly rainfall, total fallow (November – March) rainfall, and growing season (April – October) rainfall for each paddock, 2005-2008.
The other part of water use (assuming drainage and runoff are negligible) is use of water from the soil. In the focus paddock data set, pre-sowing soil measurements were made at a useful time for measuring stored water before the growing season (March-April), but post-harvest measurements often occurred after post-growing season rain. Rather than attempt to adjust (which is difficult with monthly rain data), it was assumed that because of the very dry springs in all years, crops would draw water down to the wilting point. Wilting point was estimated as the driest measurement of the 0-100cm soil water between 2006 and pre-sowing 2009; this was usually harvest 2006.
Potential yield
Water determined yield potential was often much higher than crops were able to achieve (Table 17). Crops were closest to approaching potential in 2008, where Patchewollock and Yaapeet No Till and Sea Lake Till crops achieved (or exceeded) water-limited yield potential. The year 2006 was most difficult, with only Donald No Till crop bettering 50% of potential yield. In general where both Till and No Till paddocks had cereal crops, the No Till paddock was closer to water-limited potential yield, the exception being the Sea Lake paddocks in 2008 and possibly 2006, where yield in the sample area of the Till paddock was much reduced by frost. This generalisation does not include Till hay crops though.
Table 17. Potential yield and water use efficiency for focus paddock cereal crops. GSR = Growing Season Rainfall, WU = Water Use (GSR + change in soil water), WUE = Yield/GSR, SE = Soil Evaporation estimate; 0.6*GSR < 180mm GSR after French and Schultz (1984), TE = Transpiration Efficiency estimate; Yield/(WU-SE), “WUE” = Transpiration Efficiency estimate without accounting for change in soil water; Yield/(GSR-SE).
The difference between actual and potential yield encompasses a number of factors, which can be partitioned using other biomass measurements. Here we have broken potential water use (soil water + growing season rainfall) into components, by assuming that the transpiration efficiency for biomass was 56 kg/ha/mm (this would vary with vapour pressure deficit, but should have been similar for both paddocks in a pair). This can be used, with maximum biomass (usually at GS65 or GS99) to estimate the amount of water that went through the plant. The balance estimates evaporation, runoff, drainage or soil water that was not used. In some crops biomass drops between GS65 and GS99, and this drop has been converted to transpiration and labelled as ‘Biomass loss’. The amount of transpiration represented by the GS99 biomass has been partitioned into ‘grain’ and ‘biomass/hay’ according to the harvest index.
In 2006 the lentil crops on the Minyip paddocks used little of the available water and most available water was evaporated or unused (Figure 37a). Slightly more water was available on the Till paddock and this corresponded to higher measured lentil biomass. At Donald there was much more water available in the Till paddock, but only a fraction of this was converted to additional biomass, and less still to yield. This may be related to a difference in sowing date. At Culgoa later sowing in the Till paddock in 2006 reduced biomass accumulation and increased the balance of evaporation, although the available water was also less. At Patchewollock there was slightly more available water in the Till paddock, but biomass and yield were less and the balance may be accounted for by grass weeds. At Sea Lake available water was higher in the Till paddock (due to fallow) and good use was made in accumulating biomass; this was not converted into yield (due in part to frost). At Yaapeet available water was similar in the Till and No Till paddocks, but the earlier sown barley in the Till paddock was able to convert more to biomass before being cut for hay.
In 2007 potential water use was higher in all paddocks (Figure 37b), but evaporation was also a relatively large proportion of potential water use, probably because most rain fell over winter. Note that estimates of evaporation by these methods are higher than the typical French and Schultz estimates. In 2007 the Culgoa Till paddock had some water retained from 2006, and was able to convert this to biomass but not grain. Donald? The cereal Till paddock at Minyip made some use of again greater available water than the No Till paddock, although the chickpea final biomass and yield were not measured. The No Till crop at Patchewollock used relatively more water than the oaten hay on the ‘Till’ paddock. Potential water use at Sea Lake was relatively low; similar to the Culgoa paddocks, but owing to the sandier soil type there much less was used in evaporation and biomass and yield were both higher. The early sown barley hay crop at Yaapeet Till had a greater water supply and was able to convert more to biomass than chickpeas in the No Till paddock, which had other problems.
In 2008 the very dry spring, following a relatively dry winter, led to large amounts of dry matter loss between anthesis and maturity in many crops (Figure 37c), particularly Till crops at Donald, Minyip and Patchewollock (albeit sown No Till). Evaporation was also very low in 2008. No Till cereal crops at Minyip and Patchewollock were able to convert more transpiration into grain, despite having less potential water use. At Sea Lake, the Till crop on fallow did the reverse, having relatively less evaporation and converting more transpiration into grain. The vetch hay crop at Yaapeet Till had relatively high evaporation/low water use, whereas the No Till crop had very low evaporation. This is probably because crops on the No Till paddock never completely dried that paddock out and the estimate of wilting point used in these estimates was too high as a result.
Figure 37. Breakdown of water use by crops on focus paddocks into evaporation or not used, and transpiration related to biomass loss, grain, and biomass remaining at harvest.