soil oxygen and nitrogen gas emissions data data

<strong>Methods</strong> <em>Study Location</em> This study was conducted in the Luquillo Experimental Forest (LEF), a U.S. National Science Foundation funded Critical Zone Observatory and Long Term Ecological Research site in the USDA Forest Service managed El Yunque National Forest in northeastern Puerto Rico (Fig. 1a). The LEF is a premontane tropical forest landscape that receives 3 to 4 m of rain annually with a mean annual temperature of 23°C (Scatena, 1989; Garcia-Montino et al., 1996). Rainfall at the time of sampling was typical of this location and preceded the 2015 drought, which began in May of 2015 (O’Connel et al., 2018). We sampled in forests dominated by Tabonuco (<em>Dacryodes excelsa </em>Vahl) and underlain by volcaniclastic parent material in the Bisley 3 watershed (400 m asl; Silver et al., 1999) and near the El Verde field station (500 m asl, Crow 1980). <em>Experimental design</em> We collected soils from two sites to examine how topographic controls on soil O₂ influence N gas emissions. First, we measured how N emissions varied across a macrotopographic gradient by collecting soil cores from ridge, slope and valley positions (4 replicates per topographic position) from sites near the El Verde field station (Fig. 1b) that are described in O’Connell et al. (2018). Macrotopography soils were collected and analyzed for N gas emissions at 0%, 10% and 20% headspace O₂ in April of 2015. Second, we measured how N gas emissions varied across a microtopographic gradient, which consisted of 16 plots (1 m²) that differed in soil moisture and soil O₂ dynamics, as described by Hall et al. (2013). Because we did not measure O₂ availability at the time of this soil collection, the microtopography sites were binned by historic O₂ status into three categories: low, medium and high O₂ based on Hall et al. (2013) (see <em>Soil oxygen and soil moisture </em>below). Microtopography soils were collected from Bisley watershed 3 (Fig. 1c) and were analyzed for N gas emissions at 0%, 5%, 10% and 20% headspace O₂ in January of 2015. <em>Soil sampling</em> Mineral soils (0-10 cm) were collected using a split core sampler (3.4 cm width) and hammer auger in order to avoid compaction and to keep soil core structure intact. Soils were collected in plastic sleeves, stored at room temperature in Ziploc bags and transported to the Cary Institute of Ecosystem Studies in Millbrook, NY USA the same day they were sampled. All laboratory analyses occurred within 3 days of soil sampling. Soil chemical analyses and N₂, N₂O and CO₂emissions were measured at the Cary Institute, while soil O₂ and soil moisture data were collected in the field (see <em>Soil oxygen and soil moisture </em>below). <em>N₂, N₂O and CO₂ emissions </em> The Nitrogen Free Air Recirculation Method (N-FARM) laboratory-based gas flow incubation system was used to directly measure N₂, N₂O, and CO₂ emissions from intact soil cores (Burgin et al., 2010). Soil cores were sealed in jars and cyclically flushed and evacuated with a given helium-oxygen mixture (0, 5, 10 or 20% O₂) for ~16 hrs to completely replace the N₂ atmosphere. After this flush, headspace samples were directly transferred via sample lines to Shimadzu GC 14 gas chromatographs equipped with thermal conductivity (for quantification of N₂ and CO₂) and electron capture (for N₂O) detectors (Burgin et al., 2010). Soil cores were incubated at multiple headspace O₂ concentrations, and each incubation took place on a separate day using fresh soil cores in order to avoid soil nitrate depletion. As a quality control measure, empty jars and autoclaved soil cores were run to ensure that the system was leak-free and that N₂ was not leaking from soil pores that were not adequately flushed prior to measurement. The method detection limit for the system was 3.98 ppm N₂, 0.0678 ppm N₂O, and 10.5 ppm CO₂, based on the standard deviation of six replicate standards (MDL = <em>T</em>_(df,0.01) × SD; Morse et al., 2015). <em>Soil oxygen and soil moisture</em> We relied on previous year-long continuous soil O₂ measurements at the microtopography plots (Supplemental Figure 1; Hall et al., 2013) to classify these sites into three categories: low, medium and high O₂. At the macrotopography plots, we measured soil moisture and O₂ for ~1 year using time-domain reflectometry (Campbell Scientific, Logan, Utah) and galvanic cell sensors (Apogee Instruments, Logan, Utah) housed in polyvinyl-chloride tubes installed at a depth of 10 cm. Soil O₂ levels and precipitation during the time of soil O₂ data collection (2010) and soil core collection (2015) were typical of this forest under non-drought conditions (Supplemental Figure 1; Silver et al. 1999, Silver et al., 2013; O’Connell et al., 2018). Soil O₂ concentrations measured in the field are hereby referred to as “<em>in situ</em>” in order to differentiate them from incubation headspace O₂ manipulations. Gravimetric soil moisture was also determined on a subsample of all soil cores at the time of sampling by calculating mass loss after oven drying soils at 105°C for 48 hrs. Soil cores were stored in Ziploc bags at the time of collection until the time of analysis to maintain field moisture. <em>In situ</em> soil O₂ data collected across the microtopography gradient several years prior to N measurements were used to bin the microtopography plots into groups that were comparable with headspace manipulated O₂ levels. Plots with mean <em>in situ </em>soil O₂ concentrations from 5-10% are referred to as “Low”, 10-15% are referred to as “Medium”, and 15-20% are referred to as “High”. Since no plot had a mean soil O₂ concentration &lt; 5%, a fourth O₂ category was omitted from our analyses (Fig. 3).