Interannual variability of wheat yield in the Argentine Pampas during the 20th century

• Published in: Agriculture, ecosystems & environment ecosystems & environment, 2004-06, Vol.103 (1), p.177-190
• Main Authors: Verón, Santiago R, Paruelo, José M, Slafer, Gustavo A
• Format: Article
• Language: English
• Subjects: Agricultural statistics; Agronomy. Soil science and plant productions; Argentina; Biological and medical sciences; climatic factors; environmental control systems; Environmental controls; Fundamental and applied biological sciences. Psychology; General agronomy. Plant production; Generalities. Agricultural and farming systems. Agricultural development; Generalities. Production, biomass, yield. Quality; Grain yield; linear models; regression analysis; temporal variation; Triticum aestivum; Wheat; Yield variability
• ISSN: 0167-8809; 1873-2305
• DOI: 10.1016/j.agee.2003.10.001

National long-term agricultural records offer an unique opportunity to improve our understanding of the patterns of variability of crop yield. In this paper, our objectives were: (i) to describe yield and yield stability of wheat crops ( Triticum aestivum L.) and (ii) to identify some controls for their spatial and temporal variability. Their environmental controls (precipitation (PPT), photothermal coefficient and drainage type) were explored using a database covering the period between 1923 and 2000 and the main subregions of the Pampas. Linear or bi-linear regression models were fitted to the relationship between yield and year for each of the 97 counties analyzed. Two traits derived from the temporal dynamics of wheat yield were used to characterize the spatial variability in yield trends: the inflection point (IP, the year at which the rate of increase in yield changed) and the difference between yield at the beginning and at the end of time-period considered (Δyield). Wheat variability trends were analyzed by plotting the absolute residuals and relative residuals of the regression models (absolute residual expressed as percentage of predicted yield) against year. Forty counties displayed linear relationships between yield and time, while in the other 57 counties the relationship was bi-linear. On average, the IP occurred in 1970. Only a small portion of the overall variance in the IP, 11%, was explained by environmental variables (photothermal quotient (PTQ), as the ratio between mean monthly incident short-wave radiation and mean monthly temperature) suggesting that the modernization of agriculture was driven mainly by local factors (such as risk aversion or land tenure). The difference in yield between 1923 and 2000 was positively associated with the average precipitation during the crop cycle and the photothermal quotient between September and November. These variables accounted for 34% of the total spatial variability. Ninety-six counties showed higher yield variability at the end than at the beginning of last century. However, in 92 counties the increase in yield variability was lower than the increase in grain yield. This finding suggests that during the last century wheat production systems of the Pampas have been successful in increasing yield while maintaining or increasing relative yield stability. Absolute variability was associated with the photothermal quotient (13%), and precipitation (7%). The proportion of the spatial variance in the