Reducing bias on soil surface CO2 flux emission measurements: Case study on a mature oil palm (Elaeis guineensis) plantation on tropical peatland in Southeast Asia

Large-scale conversion of tropical peat swamp forests to agricultural plantations has resulted in substantial carbon dioxide (CO2) emissions. Despite consensus on the importance of these emissions, the cause of the large range in the magnitudes of reported values remains uncertain. Differences in reported fluxes might result from site specific factors and/or potential limitations of the manual flux chambers commonly used. It is important that any biases at the site level are explored as they ultimately affect regional and global emission estimates. Therefore, the aim of this study is to determine if measurement timing of commonly used infrequent manual chamber measurements leads to biased emission estimates. In this study we make use of six months of automated chamber data to provide a semi-continuous timeseries. This timeseries is used to explore the potential for time-of-day sampling biases in infrequent, monthly manual chambers measurements in a peatland oil palm plantation in Malaysian Borneo. Fluxes from Palm Base, Harvest Path, Frond Pile, Drain and Inter row microforms were recorded hourly using automatic chambers. From these hourly data, mean diurnal patterns of fluxes were produced. These diurnal patterns were used to characterize the biases in a larger, monthly flux manual chamber dataset. This monthly manual dataset was collected over six years at the same site and microforms, with individual measurements made in the daytime. Bias range was widest for Harvest Path (-18 to 24 %), followed by Palm Base (-13 to 11 %), Drain (-10 to 9 %) and Frond Pile (-5 to 3 %). Estimates of annual plantation scale emission over six years, corrected for sampling bias ranged from 36 – 53 Mg CO2 ha−1 yr−1. We recommend careful consideration of artefacts sample timing might introduce in any sampling design, and where possible fluxes should be corrected with measured diurnals for each microform considered.