Impacts of an African Green Revolution on greenhouse gases and pollution precursors: Nonlinear trace nitrogen gas emission responses to incremental increases in fertilizer inputs in a western Kenyan maize field

Background/Question/Methods

Over the last several decades, agricultural soils in many parts of sub-Saharan Africa have become depleted of nitrogen (N) and other nutrients, creating challenges to achieving food security in many countries.  At only 8 kg N ha^-1 yr^-1, average fertilizer application rates in the region are an order of magnitude lower than typical rates in the United States, and well below optimal levels.  Increased use of nutrient inputs is a centerpiece of most African Green Revolution strategies, making it important to quantify the impacts of this change in practices as farmers begin moving towards 50-80 kg N ha^-1 yr^-1.

Increased N inputs are invariably accompanied by losses of trace N gases to the atmosphere, including the greenhouse gas nitrous oxide (N₂O), and nitric oxide (NO), a precursor to tropospheric ozone pollution.  Several investigations of N₂O emissions and one investigation of NO emissions from sub-Saharan agricultural systems have been conducted over the last 20 years, but they typically include only two levels of fertilizer additions, and so are unable to identify potentially important thresholds in the response of trace gas emissions to fertilization rate.  Here we examine the response function of NO and N₂O emissions to 5 different levels of inorganic fertilizer additions in a maize field in Yala, Kenya during the 2011 long rainy season.  We used a randomized complete block design incorporating 5 levels of inorganic fertilizer additions (0, 50, 75, 100, and 200 kg N ha^-1) in 4 blocks.  After each of 2 fertilizer applications, we measured trace gas fluxes daily, and conducted weekly measurements until trace gas emissions subsided to control levels.   We fit the data to linear and exponential models relating N gas emissions to N input levels, and conducted a model comparison using AIC.

Results/Conclusions

Preliminary data suggest that emissions do respond in a non-linear fashion, with emission rates increasing more rapidly at high fertilization levels, as has been observed in several commercial agroecosystems. The identification of emission response thresholds can be combined with information on crop yield responses to provide insight into how to manage fertilizer use to optimize crop production per unit greenhouse gas emitted.