15.3.3 Phenology, drought and yield

Printer-friendly version

Agricultural produce (cereals, grain legumes, fruit) comes from reproductive development, and because most crops experience water deficit at some stage in their life cycle the impact of water deficit on reproductive development is of particular importance. Furthermore, reproductive performance is the ultimate determinant of evolutionary fitness in both natural and agricultural ecosystems, so responses to water deficits are under intense natural and human selection. Responses to drought observed in both natural and managed communities will often have considerable adaptive significance.

In Australian agriculture (and especially in areas with predominantly winter rain), water deficits most commonly occur towards the end of the life cycle of annual crops and are very predictable. There is little chance of these deficits being relieved and, in any case, the closer a plant is to maturity, the less flexibility it has to compensate after relief of the stress. Consequently most research on the effects of water deficits on annual crop growth and reproduction has concentrated on deficits occurring in the latter half of the crop’s life cycle. However, in Mediterranean climate zones such as south-western Western Australia rainfall is erratic at both ends of the growing season. Advances in agricultural technology that have allowed earlier sowing also mean that water deficit very early in a crop’s life cycle is becoming increasingly important in those areas. This deficit has a good chance of being relieved, and plants have a much greater capacity to compensate following relief, so responses to early water deficit can be quite different from those that occur later in a given season.

A generalised sequence for vegetative growth and reproductive development in a cereal plant (Figure 15.16) can be used to highlight key physiological phases (i.e. crop phenol-ogy) and their comparative sensitivity to water deficit. Overall, the earlier a water deficit occurs, the less effect it has on final yield. However, cereals are particularly sensitive to water deficit at or just before anthesis (Fischer 1973).


Figure 15.16 A schematic representation of phenology for a cereal plant, showing major developmental phases, from germination to grain maturity. Band widths represent relative sensitivity to water deficit. (Notional values compiled by D.G. Abrecht and R.J. French)

Water deficit at anthesis affects yield by reducing the number of grains per ear rather than ear number or grain size. This suggests that water deficit interferes with fertilisation. In wheat, male fertility is more strongly affected than female fertility since droughted plants fertilised with pollen taken from well-watered plants produce normal numbers of grain, but well-watered plants fertilised with pollen from droughted plants produce few grains (Saini and Aspinall 1981). Saini and Aspinall also found that the period when grain set was most sensitive to water deficit coincided with pollen meiosis and microspore release from the tetrads.

By contrast, water deficit impedes embryo development after fertilisation in maize rather than pollen fertility per se (Westgate and Boyer 1986). Drought stress can also delay silking so that pollen is shed before plants have receptive stigmas (Herrero and Johnson 1981).

Water deficit during the vegetative phase of a cereal crop affects grain yield via vegetative growth and tillering. Water deficit reduces leaf expansion rates (Barlow 1986; Section 6.2). In wheat and maize, drought stress can also reduce the rate of leaf appearance (Abrecht and Carberry 1993; Armstrong et al. 1996). Therefore stressed crops have lower leaf area indices and intercept less radiation. This is the major cause of reduced crop growth caused by water deficit during vegetative growth. Robertson and Giunta (1994) have shown that water deficit prior to anthesis did not affect radiation use efficiency in wheat.

Water deficit experienced during vegetative growth also reduces the rate of tiller appearance, but compensatory responses do occur. For example, Aspinall et al. (1964) found that barley stressed during vegetative growth grew more tiller buds after rewatering, so plants stressed early produced more tillers although many of these tillers bore no heads. There is no compensation for water-deficit-induced reductions in tiller number in wheat following rewatering, and wheat sub-jected to water deficit during vegetative growth consequently carries fewer fertile tillers per plant (Armstrong et al. 1996).

Water deficit after anthesis influences grain size, although grain number per ear is also reduced if the deficit occurs in the first two weeks after anthesis (Fischer 1973). Later deficits reduce grain size by shortening the duration of grain filling (Brookes et al. 1982).

Phenology and partitioning

The balance between vegetative growth and reproductive development is crucial to final yield. Vegetative growth establishes a carbon source in leaves, roots and support struc-tures which is used by reproductive sinks. This balance is largely achieved by developmental partitioning, especially in determinate plants such as wheat. Growth and developmental phases correspond to phenological events, and consequently the timing of photoassimilate partitioning is largely determined by phenology.

Indeterminate plants, including most grain legumes, express a complicated phenology because flowering is followed by a phase of mixed vegetative and reproductive growth. This is usually followed by a period of purely reproductive growth. The previous mixed phase enables grain legumes to set enough reproductive sinks to yield satisfactorily. In species such as field pea, lentil, chickpea and faba bean new pods are set sequentially. Moreover, each pod or pair of pods is formed together with supporting stem and a subtending source leaf (French 1990). Other grain legumes, such as narrow-leafed lupin, carry additional groups of pods on late-formed racemes.

This pattern of prolonged reproductive development in grain legumes becomes compressed in response to water deficit (Figure 15.17), resembling drought stress effects on node formation and pod set in field peas (French and Turner 1991).

Reproductive growth is less sensitive to water deficit than vegetative growth, and increased partitioning to reproductive structures is one consequence. Ong (1984) provided a very clear example of this in groundnut. Once seed is set, growth rates are often maintained under water deficit. Soybean and wheat behave similarly. In general, mild water deficit can even accelerate seed growth (French and Turner 1991) because vegetative growth is severely constrained and photoassimilate is diverted to reproductive structures. If drought stress intensifies, carbon assimilation diminishes and stored reserves are then mobilised for seed growth. Seed growth rate is thus sustained, but duration of filling can be shortened, resulting in smaller seeds.

In groundnut, mild water deficit prior to or just after flowering increases pod set in a selective way. Development of small pods is reduced and fewer of them become mature pods (Chapman et al. 1993). Water deficit in cotton also reduces the proportion of flowers setting bolls (Turner et al. 1986).

Such contrasting effects of water deficit on reproductive structures are commonplace. By and large, small structures suffer in favour of large structures. In the tropical pasture legume Macroptilium atropurpureum water deficit retards small bud development prior to floral initiation. Later develop-mental stages are not affected but survival of floral buds is reduced (Kowithayakorn and Humphreys 1987). Again, in soybean, water deficit at anthesis reduces pod set at high floral positions while improving basal pod set (Westgate and Peterson 1993). In cotton, small fruiting buds and bolls are shed in response to water deficit, while pods formed later are also shed in groundnut in response to water deficit. A decrease in pod numbers in field peas (Figure 15.17) again results from a loss of young pods.


Figure 15.17 In field pea, water deficit curtains duration of flowering which results in fewer pods per plant. Reproductive nodes formed on each plant are not as severely reduced, but their effectiveness in forming pods is diminished by drought stress. (Based on unpublished data, D.G. Abrecht and R.J. French)

Such patterns of crisis reaction to drought stress where small structures are shed preferentially ensure that repro-ductive structures nearing maturity are retained and thus enhance prospects for successful completion of a life cycle, a feature of plant development with obvious selection advantage.