The increasing demand for renewable energy may lead to the conversion of millions of hectares into bioenergy plantations with a possible substantial transitory carbon (C) loss. In this study we report on the greenhouse gas fluxes (CO2, CH4, and N2O) measured using eddy covariance of a short-rotation bioenergy poplar plantation converted from agricultural fields. During the first six months after the establishment of the plantation (June-December 2010) there were substantial CO2, CH4, and N2O emissions (a total of 5.36 +/- 0.52 MgCO2eq ha(-1) in terms of CO2 equivalents). Nitrous oxide loss mostly occurred during a week-long peak emission after an unusually large rainfall. This week-long N2O emission represented 52% of the entire N2O loss during one and an half years of measurements. As most of the N2O loss occurred in just this week-long period, accurately capturing these emission events are critical to accurate estimates of the GHG balance of bioenergy. While initial establishment (June-December 2010) of the plantation resulted in a net CO2 loss into the atmosphere (2.76 +/- 0.16 Mg CO2eq ha(-1)), in the second year (2011) there was substantial net CO2 uptake (-3.51 +/- 0.56 Mg CO2eq ha(-1)). During the entire measurement period, CH4 was a source to the atmosphere (0.63 +/- 0.05 Mg CO2eq ha(-1) in 2010, and 0.49 +/- 0.05 Mg CO2eq ha(-1) in 2011), and was controlled by water table depth. Importantly, over the entire measurement period, the sum of the CH4 and N2O losses was much higher (3.51 +/- 0.52 Mg CO2eq ha(-1)) than the net CO2 uptake (-0.76 +/- 0.58 Mg CO2eq ha(-1)). As water availability was an important control on the GHG emission of the plantation, expected climate change and altered rainfall pattern could increase the negative environmental impacts of bioenergy. (C) 2012 Elsevier B.V. All rights reserved.