| Weather has a strong impact on crop growth and nitrogen uptake and loss in soil-crop systems, also soil and field managements are especially important for crop and nitrogen. Increasing nitrogen fertilizer response and improving nitrogen management have long been recognized as a most challenge of agricultural development, because worldwide nitrogen use efficiency (NUE) for cereal production is only estimated at 33%. As a consequence, current agricultural and environmental concerns demand improved NUE and reduced N losses to protect environment, while maintaining an optimal yield.Meta-analysis is a methodology used to intergrate a set of statistical techniques developed to provide an objective comparison of research results from independent studies. A meta-analysis combines the results of several studies that address a set of related research hypotheses. In this study, a meta-analysis was performed on the relationship between the application rates of nitrogen fertilizer and the corn yield response in order to understand the effects of weather (air temperature, rain) and other factors (including soil pH, soil surface textures, plant density, previous crop, nitrogen rates and application methods). The database is constituted by two parts: a database of nitrogen response trials covering the years 1997 to 2008 in 60 locations in the grain corn growing area of Quebec, and a database is built up by collecting publications from Africa, Europe and America published between 2000 and 2009, inclusively. Those data are submitted to meta-analyses procedures in conjunction with a corresponding weather database. The response ratio (RR) used to measure effect size indicates significantly positive response of corn yield.A better understanding of the effects of weather, soil and managements would allow for adjustments of the nitrogen rate targeted for optimal N supply. In this dissertation, meta-analysis is used to determine: (1) Which weather variables are the most determinant for corn response to nitrogen rates; (2) Which soil and managerial factors are the most predominant in term of corn response to nitrogen rates; (3) The relationship between corn response to N rates and specific ranges of weather variables and other factors. Also, the main results obtained are as follow: (1) Traditional statistics are not appropriate when the data come from numerous studies, while meta-analysis, which uses effect size and weighted statistical models, should therefore be applied in the analysis of a large collection of results originating from different studies. It can be used to reflect the magnitude of the treatment effect or the strength of a relationship between two variables. (2) High-yield years do not necessarily mean high-RR years, and high-RR years are characterized by cooler and wetter growth season, together with the appropriate rainfall evenness. (3) The results indicate that RRs are not influened by previous crop, and the highest RRs were found by the percent of clay (28~40%), soil pH (6~7) and plant density (60, 000~70, 000 plants/ha-1). The starter N and top-dressing N should be recommended for high RR. (4) A major result of this analysis is that rainfall was more important than corn heat units for RR, and the amount and application method of nitrogen are more important than other soil and factors relating to field management. (5) The results indicate that weather condition before side-dressing is more important than after side-dressing for RR. (6) The meta-analysis shows that response ratios to weather factors tend to be higher in High-N and Med-N than in Low-N, but there is no significant difference between High-N and Med-N. Therefore total nitrogen range (100~150 kg ha-1) and starter N range (0~50 kg ha-1) rate application under the different weather conditions can be considered the optimal selection to maintain cron productivity and limit environmental losses. RRs have no significant differences from one another within three starter N rates.
|
PDF Full Text Request