Browsing by Subject "Mulit-level Linear Regression"

Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    Modeling N2O Emissions from Agricultural Soils Using a Multi-level Linear Regression
    (2007-08-29T23:43:14Z) Carey, Kyle
    Nitrous oxide (N2O) is a potent greenhouse gas emitted from soils through the microbial processes of denitrification/nitrification. In agricultural soils these natural processes and thus N2O emission are significantly enhanced by the use of nitrogen fertilizers. Use of fertilizer has grown rapidly since the 2nd half of the 20th century and will likely continue to grow as more countries develop advanced agricultural management strategies and demand for food increases. Though fertilizer enhanced N2O emissions from agriculture contribute only a small amount to the current anthropogenic global climate warming, it is expected that these emissions will play a larger role in the future. This study consists of a statistical meta-analysis of an agricultural N2O emission database made up of data taken from peer-reviewed literature. A multi-level linear regression is used to investigate the relationship between N fertilizer input and N2O emissions; (1) independent of variation between each study, (2) as a categorical function of crop type and (3) as a categorical function of fertilizer type. An understanding of these relationships could help to establish management strategies to more efficiently use N fertilizers, reducing N2O emissions and lowering expenses for agricultural producers. The results of the multi-level linear analysis of the dataset indicate that the relationship between N input and N2O emissions is not independent of the conditional variation between studies. The categorical analysis of differences in crop type also did not have a significant influence on N2O emissions. The categorical analysis of N fertilizer forms did show a significant influence on emissions. Differences in the slopes of the fertilizer type models provide relative comparability of expected N2O emissions of different chemical forms of N fertilizer for a given N input. The analysis performed in this study yielded important insight into the factors influencing N2O emissions from agricultural soils. These findings can be used to guide both management strategies and further research dealing with an increasingly important topic.