Site-specific agriculture aims at optimising inputs on field and farm level. The approach should benefit the farmer in terms of net return and the environment through lower emission levels. Variations in grain yields are often significantly related to changes in available water capacity within one field but the temporal variation often overrules the spatial variation. The objective of the present study was to evaluate the impact of varying plant available soil water and precipitation (or irrigation) conditions on yield and various plant characteristics of winter wheat. Knowledge of these variables and of the effect that they have on variations in the response to fertilizer is necessary when planning a fertilizer program.
The two-year trial was conducted in the tertiary foothills of the Bavarian Alps. A completely randomised two-factorial experiment was set up with two different N fertilizer treatments (180 N kg ha-1 and 120 N kg ha-1) and three different water supply treatments (stress by rainsheltering, irrigation, and control, i.e. rain-fed only). The experiment was conducted simultaneously on two sites of different plant available soil water within one field in each experimental year. Soil water content and soil water matric suction were regularly measured in different depths by portable capacitance probes and tensiometers, respectively.
A speedy and cheap method to calibrate Diviner capacitance sensors in heterogeneous fields was developed that highly increased the relation between gravimetrically obtained soil moisture and instrument readings in the trial fields.
At both sites, depletion of plant available soil water was slowed down by irrigation or accelerated by the rain-shelter treatment compared to the control treatment during the application of the water supply treatment. After the water supply treatment was terminated in the beginning of June, the level of depletion reached quickly a similar level on all treatments of each site. A conspicuous difference between the two sites could be recognized in the pattern of depletion of plant available soil water.
The depletion of plant available soil water was similar for both N treatments except for the topsoil at the site of high plant available soil water where a stronger depletion was found on the low N treatment than on the high N treatment.
The important yield increase from 2000 to 2001 on all treatments can readily be explained by the unfavourable distribution of precipitation in 2000. The high N treatment generally yielded more than the low N treatment but the response of grain yield to higher N application appeared to be dependent on site and water supply.
At the sites of high plant available soil water, the differences in grain yield between rainshelter and irrigation treatment were of a similar order than the grain yield differences due to the application of additional N. At the sites of low plant available soil water, the differences in grain yield between rain-shelter and irrigation treatment were far more important than the average increase caused by a higher N application. The maximum efficiency for each kg of N applied at the sites of high plant available soil water was found on the rain-shelter treatments whereas at sites of low plant available soil water the maximum grain yield increase for each kg of the N increment applied was on the irrigation treatment. The importance of sufficient water supply during vegetative growth at sites of low plant available soil water to maintain a potential for high yield was also demonstrated in this study.
All plant characteristics examined in this study except for N and K concentration in grain were affected by site. Many were also affected by the water supply treatment but only seed weight and N concentration in grain were affected by N fertilization. N removed by the plants exceeded in most cases the amount of applied N fertilizer.
Water use efficiency was higher in 2001 than in 2000. Water was more efficiently used at sites of high plant available soil water than at the sites of low plant available soil water. Within the water supply treatment, a more efficient use of soil water was observed on the high N treatment every time in which a yield increase occurred compared to the low N treatment. The results of the present study also complements findings in other works insofar as they demonstrated that the effect of water supply on water use efficiency is also dependent on the water holding capacity of the soil.
Differences in soil mineral N at the various sampling times could primarily be attributed to site and to the water supply treatment, whereas the N fertilization treatment did not produce a significant effect on soil mineral N. A reduced water supply during stem elongation and booting seems to increase the risk of N leaching at the sites of low plant available soil water. Soil water matric suction is predominantly affected by site. Variables computed on the basis of average monthly soil water matric suction generally perform better in identifying yield variability than variables based on the average matric suction of the entire spring-summer growing period. The quality and character of correlations between soil water matric suction and yield varied considerably between years which can in part be attributed to yearly rainfall patterns.
In conclusion, between the three factors, i.e. site, water supply (precipitation and irrigation) and N fertilization, site is the primary effect that accounts for variability of grain yield while water supply, and in particular its distribution during the growing season, affects the annual yield level in a given year. Increased N fertilization within the water supply treatment generally increases yield, but in particular on coarse-structured sites, N efficiency can come to nothing if climatic conditions are unfavourable. Stress during tillering and jointing limits yield potential that is not regained when stress is relieved. However, if stress is prevented until heading by a sufficient precipitation, there is still potential for maximum yields even on sites of low plant available soil water if sufficient precipitation occurs during heading and anthesis. A below average water supply during tillering and jointing on sites of high plant available soil water may eventually result in higher yield expectation and thus, may suggest a higher N fertilization. On coarse-structured sites, only if rainfall during tillering and jointing is considerably above average, an increase in the amount of applied N to obtain higher yield and grain quality may be indicated, but generally decreased N uptake, a higher soil mineral N content during and after the trial, and a reduced rooting capacity advocate a low-level N fertilization strategy on sites of low plant available soil water. «