Optimum sampling time and frequency for measuring N2O emissions from a rain-fed cereal cropping system

Abstract

Annual cumulative nitrous oxide (N2O) emissions from soil have historically been calculated from intermittent data measured manually via the static chamber method. The temporal variability in emissions, both diurnally and between days, introduces uncertainty into the up-scaling of static chamber data. This study assessed the most appropriate time of the day to sample and the best sampling frequency to ensure reliable estimates of annual cumulative emissions. Sub-daily N2O emissions were measured using automatic gas sampling chambers over three years in a sub-tropical cereal crop system. The sub-daily dataset was divided into eight time periods per day to assess the best sampling time of the day. Daily mean N2O emissions were subsampled from the dataset to simulate different sampling frequencies, including pre-set and rainfall-based scenarios. Annual cumulative N2O emissions were calculated for these scenarios and compared to the ‘actual’ annual cumulative emissions. The results demonstrated that manual sampling between mid-morning (09:00) and midday (12:00), and late evening (21:00) and midnight (24:00) best approximated the daily mean N2O emission. Factoring in the need to sample during daylight hours, gas sampling from mid-morning to midday was the most appropriate sampling time. Overall, triweekly sampling provided the most accurate estimate (± 4% error) of annual cumulative N2O emissions, but was undesirable due to its labour intensive high sampling frequency. Weekly sampling with triweekly sampling in the two weeks following rainfall events was the most efficient sampling schedule, as it had similar accuracy (± 5% error) to the triweekly sampling, the smallest variability in outcomes and approximately half the sampling times of triweekly sampling. Inter-annual rainfall variability affected the accuracy and variability of estimations of annual cumulative emissions, but did not affect the overall trends in sampling frequency accuracy. This study demonstrated that intermittent samplings are capable of estimating the annual cumulative N2O emissions satisfactorily when timed appropriately.