A long-term fertilizer experiment investigating cotton-based cropping systems established in 1990 in central Asia was used to quantify the emissions of CO2, CH4 and N2O from April 2012 to April 2013 to better understand greenhouse gas (GHG) emissions and net global warming potential (GWP) in extremely arid croplands. The study involved five treatments: no fertilizer application as a control (CK), balanced fertilizer NPK (NPK), fertilizer NPK plus straw (NPKS), fertilizer NPK plus organic manure (NPKM), and high rates of fertilizer NPK and organic manure (NPKM+). The net ecosystem carbon balance was estimated by the changes in topsoil (0-20 cm) organic carbon (SOC) density over the 22-year period 1990-2012. Manure and fertilizer combination treatments (NPKM and NPKM+) significantly increased CO2 and slightly increased N2O emissions during and outside the cotton growing seasons. Neither NPK nor NPKS treatment increased SOC in spite of relatively low CO2, CH4 and N2O fluxes. Treatments involving manure application showed the lowest net annual GWP and GHG intensity (GHGI). However, overuse of manure and fertilizers (NPKM+) did not significantly increase cotton yield (5.3 t ha(-1)) but the net annual GWP (-4,535 kg CO2_eqv. ha(-1)) and GHGI (-0.86 kg CO2_eqv. kg(-1) grain yield of cotton) were significantly lower than in NPKM. NPKS and NPK slightly increased the net annual GWP compared with the control plots. Our study shows that a suitable rate of fertilizer NPK plus manure may be the optimum choice to increase soil carbon sequestration, maintain crop yields, and restrict net annual GWP and GHGI to relatively low levels in extremely arid regions.
