Yield, technological quality and water footprints of wheat under Mediterranean climate conditions: A field experiment to evaluate the effects of irrigation and nitrogen fertilization strategies

The evaluation of the role of different agronomic strategies in achieving sustainable wheat yields under variable Mediterranean climate conditions may involve the use of resource-use indicators that combine productivity and environmental impact. A two-seasons field experiment was conducted in South Portugal to study the effect of water regimes and nitrogen fertilization on wheat yield, grain quality and water use evaluated with water productivity and water footprint indicators. The water regime treatments were full irrigation, supplemental irrigation, and rainfed. Nitrogen fertilizer treatments, including conventional and enhanced efficiency fertilizers (EEF) were distinguished by N splitting and timing over the crop cycle. Contrasting meteorological variables in the two years caused distinct wheat productive responses. Although leading to lower grain yields, supplemental irrigation guaranteed a water productivity similar to full irrigation. The use of EEFs in which 50% of the total nitrogen was applied at the booting phase had a positive significant effect on grain protein content and on dough rheologic properties, indicating that late nitrogen applications benefit the technological quality of wheat. The average total water footprints estimated for the two seasons showed no relevant differences but the partition of the green, blue and grey components in irrigated wheat varied, with an increased importance of blue water consumptive use in the second year of the experiment. In fact, the ratio blue water footprint/green water footprint increased from 0.40 to 2.00 due to higher irrigation requirements. High grey water footprint in rainfed wheat was mostly influenced by lower yields in 2018–2019, and by an advantageous rainfall distribution during the 2017–2018 season. No significant reduction in grey water footprint was observed when using EEFs. A multivariate statistical approach through factor analysis (FA) and multivariate linear regression (MLR) was used to examine the data structure and correlation. FA resulted in three-factor models of yield and water use, yield components and wheat quality, in the first season. In the second, drier, season, variables most related with irrigation water use were clustered in one detached factor. The stepwise MLR pointed to a good prediction capability of water footprints from NDVI measured with proximal sensors at booting, anthesis, maturation and/or tillering.