Irrigated Wheat Production Systems: Too Much Nitrogen Not Enough Water

Irrigated wheat production systems (spring, facultative and winter wheat) comprise nearly 55% of the wheat area and roughly 65% of wheat production in the developing countries. Somewhere between 35 – 45 % of these production systems involve wheat in rotation with flooded paddy rice and the rest with wheat in rotation with a large number of potential upland crops like maize, soybean and cotton. The vast majority of this area is characterized by: 1) use of intensive tillage systems, often with crop residue removal or burning; 2) largely inefficient irrigation water delivery by gravity delivery systems (mainly by flooding) and 3) use of comparatively high levels of N fertilizers.

Excessive tillage, especially when residues are removed or burned, is clearly contributing to a “wearing down” of the foundation for sustainable production for these crop intensive situations through degradation of soil productivity and/or through creation of conditions leading to diminishing input use efficiencies. However, the problems associated with marked reductions in tillage combined with high levels of surface retained crop residues for surface/gravity irrigation water delivery systems (especially flood irrigation systems) have discouraged most irrigation researchers and farmers from trying to reduce tillage and retain residues. CIMMYT agronomists nonetheless, in collaboration with NARS scientists and farmers, have developed new technologies and machinery to allow zero/reduced till planting with crop residue retention which are being extended in south Asia, including to small scale farmers. Furrow irrigated, bed-planting systems have greatly facilitated the scope to manage the crop residues as well as dramatically reduce tillage.

There is a continuing need to improve the efficiency of irrigation water use in wheat production because water does present a major production cost to most farmers. Yet more importantly, there is a world wide, accelerating competition for scarce water resources and agriculture will undoubtedly lose the battle to maintain event its current share, especially since most irrigation systems and farmer irrigation practices are notoriously inefficient, wasting excessive amounts of water. It is a forgone conclusion that marked increases in the efficacy of irrigation water use must be achieved if production levels are to be maintained let alone be increased since we will need to produce more from less.

Similarly, N fertilizer use efficiency in irrigated wheat must be improved, not only in view of its increasing contribution to the cost of wheat production but also because of the detrimental environmental effects that are associated with improper N management and its excessive use.

This presentation, then, attempts to illustrate how breeders and agronomists can work together to develop needed management strategies to enhance water and nitrogen use efficiency and then identify suitable genotypes to fit these new management strategies. To do this, management by genotype interactions must be understood and then utilized to identify the right genotypes.

A key part of crop management strategies which CIMMYT agronomists are using to improve both water and N use efficiency entails the use of furrow irrigated bed planting systems. Farmer trials/observations as well as station trials have indicated up to a 30% saving of irrigation water as compared to typical flood irrigation systems in Mexico as well as in China, India, Pakistan and Iran. This planting system allows new wheat management opportunities for planting orientation on the beds as well as for N timing and placement because of opportune field access to accomplish management operations by tracking in the furrows between the beds.

The results reported below originate from trials conducted at CIANO, Cd. Obregon, Sonora, during the 1998/99 and 1999/00 crop seasons.

Irrigation Strategies

Figure 1a presents the two-year (1998/99 and 1999/00) averaged yield results for seven bread and seven durum wheat genotypes which were grown with either five irrigations (554 mm h2o applied) or four irrigations (392mm h2o applied). It is clear that the performance of the durum wheat lines over the two irrigation treatments was decidedly different from the bread wheat lines. The average yield for the durum was not effected by reducing the irrigation whereas a small but significant yield reduction occurred for the bread wheat. However, there were significant irrigation*crop and irrigation*genotype within crop interactions indicating differential performance patterns which can offer positive selection opportunities for breeders. Only small year alone or interactions of the other treatment factors with year were noted.

Figure 1b presents results from a similar trial where two durum and two bread wheat genotypes were produced in the 1998/99 and 1999/00 crop cycles with either five irrigations (508mm h2o applied) or 4 irrigations (392mm h2o applied). The genotypes were planted using either 3 rows/bed (14 cm between rows) or 2 rows/bed (28 cm between rows) on 80cm beds (width from furrow center to furrow center). As can be observed, average grain yield was higher with 2 rows/bed using five irrigations whereas yield for four irrigations were higher for 3 rows/bed. Again, however, there were highly significant irrigation*genotype and row #/bed*genotype interactions indicating that differential genotypic performance patterns must be carefully considered in order to be utilized in developing new lines that will provide higher water use efficiencies under the most feasible planting methodology.

Fig. 1a. Effect of irrigation frequency on the average grain yields of seven bread wheat and seven durum wheat genotypes common to both the 1998/99 and 1999/00 cycles at CIANO/Obregon^*
^*Both irrigation x crop and irrigation x genotype within crop interactions were significant at the 0.01 level.

Fig. 1b. Effect of irrigation frequency and row number per bed on the average grain yield of bread wheat and durum wheat genotypes (2 each) planted on beds with conventional tillage during the 1998/99 and 1999/00 crop cycles at CIANO/Obregon^*
^* Irrigation x row #/bed, irrigation x genotype, row #/bed X genotype and the three way interaction were all significant at the 0.01 level.

N Management Strategies

N rates that many farmers use for irrigated wheat tend to be markedly higher that those used by most rain fed wheat producers because of higher yield potential/expectations. However, this can be apparently exorbitant as is observed in the Yaqui Valley of Sonora where the current average N application to wheat by farmers is over 275 kg/ha. As in most irrigated wheat situations, farmers in the Yaqui Valley tend to apply a large part of the N pre-plant or at planting (commonly between 50-80% of the total N applied). Our research has consistently demonstrated that when there is a marked reduction in the amount of fertilizer N applied at or before planting combined with the bulk applied at near the 1^stnode growth stage, yield is normally enhanced and remarkable grain quality improvement occurs.

Figure 2a presents the yields for four durum wheat varieties grown for two years (1998/99 and 1999/00) at CIANO where 225 kg N/ha were applied using three different timing patterns. Altar 84 currently is the most widely grown durum wheat in the Yaqui Valley and can be considered as a check. It is obvious that applying all N at planting (similar to farmer practice) was grossly inferior to the other two application strategies using split applications. A small year*genotype interaction for yield was observed but no other interactions were significant.

Figure 2b presents the % flour protein for the same varieties and N management treatments and serves as a quality indicator. The figure clearly indicates the exceptional advantages for the split applications in quality expression.. As can be observed there were large yield and quality differences between the varieties. Concerning the split application treatments, applying 1/3 N at planting and 2/3 at 1^st node provided highest yields and obvious increases in % flour protein compared to applying all N at planting. Applying 2/3 at 1^st node and 1/3 at boot stage provided an intermediate yield increase but a greater increase in % flour protein. There was a significant N management*variety interaction while all other interactions were not significant.

Fig. 2a. Effect of nitrogen management on grain yield of four durum wheat varieties averaged over the 1998/99 and 1999/00 crop cycles at CIANO, Cd. Obregon^*
^* Year x variety interaction was significant at 0.01 level. No other interactions were significant.

Fig. 2b. Effect of nitrogen management on percent flour protein of tour durum wheat varieties averaged over the 1998/99 and 1999/00 crop cycles at CIANO, Cd. Obregon^*
^*The year x N management and year x variety interactions were not significant. The N management x variety interaction was significant at the 0.01 level.

Figures 3a and b present similar information for a series of bread wheat genotypes grown during the 1999/00 crop cycle at CIANO and for which 225 kg N/ha was applied with different timing. Rayon 89 is currently the most widely grown bread wheat in the Yaqui Valley and is the check. Also included are the mean yields for four genotypes obtained from the rust resistance program and five from the bread wheat program.

The yield performance illustrated in Fig. 3a also indicates the inferiority of applying all N at planting. Yields were higher for the two split application treatments, which were at par. The splits were 1/3 at planting and 2/3 at 1^st node versus 1/3 at planting, 1/3 at 1^st node and 1/3 at boot stage. Large genotypic differences occurred but there was a significant N management*genotype interaction.

Fig. 3b presents the % grain protein values for the same N management and genotypes. As observed above for durum wheat, the split applications not only increased bread wheat yield but markedly increased grain protein content as compared to applying all N at planting. All genotypes responded in a similar manner for protein content although there were large genotypic differences. The 3 way split was better for both yield and protein for all genotypes except yield of Rayon 89. The N management*genotype interaction for protein was not significant.

Fig. 3a. Effect of nitrogen management on grain yield of Rayon 89, five advanced lines from the bread wheat program, and four advanced lines from the rust resistance program at CIANO, Cd. Obregon, during the 1999/00 crop cycle*
*N management by genotype interaction was significant at the 0.05 level.

Fig. 3b. Effect of nitrogen management on % grain protein of Rayon 89, five advanced lines from the bread program, and four advanced lines from the rust resistance program at CIANO, Cd. Obregon, during the 1999/00 crop cycle*
*N management x genotype interaction not significant

The three examples given above attempt to indicate the sharp differences in crop and genotype performances that can be obtained with different crop management alternative strategies. Furthermore, they illustrate the differential management and genotype interactions that can occur and can be exploited in the variety development process. It seems clear that breeders and agronomist have not traditionally worked closely enough to truly exploit these kinds of elements to more efficiently develop the varieties farmers need. This tends to be especially true when faced with new technologies like reduced/zero till planting systems with residue retention or bed planting systems for wheat or the inevitable constraints imposed by less available irrigation water or ever more costly sources of fertilizers. The breeders and agronomists in the CIMMYT wheat program are trying to improve how we manage these circumstances to develop better germplasm and to provide a purposeful example for our NARS colleagues.