Soil core sampling in Tanzania was conducted during the dry season, in February 2011 and 2012. Here, one undisturbed soil core (12.5 cm inner diameter, 25 cm length) was collected at each of the 60 research plots. Cores were covered at the bottom and top with Parafilm M_ (Bemis Company, Inc. Oshkosh, WI, USA) lids to allow air exchange. They were shipped within a month to laboratory facilities at KIT, IMK-IFU, Germany, where they were kept air-dried at room temperature until further processing.Before being measured for GHG emissions (CO2, CH4, N2O), the soil cores were pre-incubated (Lovibond incubator ET 651-8, Tintometer GmbH, Germany) for 2 weeks at the site-specific mean annual temperature (MAT). They were watered with 4 mm of standard rain mixture equivalents after the first week and then again hr before the start of the measurement period (Breuer, Kiese, & Butterbach-Bahl, 2002). Each soil core was measured for GHG exchange once per week over a 5 week period (N = 5). During this time, moisture levels and temperature (_0.5°C) were kept stable. Measurements of CO2, CH4, and N2O concentration changes were monitored with a dual quantum cascade laser (Aerodyne Research Inc., Billerica, MA, USA, precision) at 1 Hz time intervals.To do so, soil cores were closed gas-tight with a lid (forming a headspace volume of 2 L). 50 ml/min of headspace air was then directed to the spectrometer and recirculated into the headspace volume over a 15 min time period. Mean ecosystem-specific GHG exchange rates were finally calculated from five replicated soil core measurements. As for the in situ GHG experiment, flux rates were calculated via the linear regression of concentration increases/decreases over time.
