Gas sampling

The study was conducted in riparian floodplains of the San Pedro River during the dry and monsoon seasons in summers of 2005 and 2007. The upper basin of the San Pedro (~7,600 km2) is drained by an unregulated, perennially flowing river near the Sonoran-Chihuahuan desert transition. Plots were established at two study sites, which were similar in hydrologic and biological features. Results did not differ significantly between sites and distinctions are therefore not made between the sites. The study sites lie at approximately 1300 m elevation and receive 300 to 750 mm of precipitation annually, largely confined to two distinct seasons: winter rains (December-March) and summer monsoon storms (July-September). A hot, dry summer (May-July) precedes the monsoon season. Median of daily stream discharge values during the dry season of both years was 0.05 m3 s-1 (USGS gage #09471000, located ~3 and10.5 km downstream from the study sites). Median discharge of the monsoon season was 1.42 m3 s-1 in 2005 and 0.58 m3 s-1 in 2007. Flash floods of instantaneous discharge >30 m3 s-1 inundated the floodplains during both study years.

Field sampling was conducted in wide riparian floodplains (>250 m) along gaining reaches (Pool and Coes 1999, Baillie et al. 2007) of the stream. The sites feature perennial stream flow, shallow groundwater, abundant hydric plant species, and productive vegetation. Such sites are typical of riparian ecosystems within the Upper San Pedro Basin (Brooks and Lemon 2007, Stromberg et al. 2007). Vegetative communities are dominated by cottonwood trees (Populus fremontii), seepwillow shrubs (Baccharis salicifolia and B. emoryi), native bunchgrass (Sporobolus wrightii) and two invasive grasses: Johnson grass (Sorghum halepense) and Bermuda grass (Cynodon dactylon). A layer of cottonwood leaf litter covers the soil surface throughout much of the study site. Soils are spatially heterogeneous and experimental plots were selected to encompass this variation. Soil texture of sampled plots ranged from sandy loam to clay loam.

Prior to wetting experiments, soils were collected from each plot for determination of nutrient and organic matter content, texture, and initial water content. Soils were collected manually using a copper ring pounded to a depth of 2 cm. Soils for nutrient determination were kept in a cooler until extraction, which was performed within 24 h of sample collection. Available nutrient content of soils was determined by extraction of soil in 2 M KCl by vigorous shaking for 1 h. Resulting extracts were filtered through pre-leached, ashless Whatman 42 filter papers and frozen until analysis for NH4+-N (phenol-hypochlorite method) and NO3--N (cadmium reduction method) on a Lachat flow-injection autoanalyzer. Soil moisture was determined gravimetrically by drying subsamples at 105oC for 48 hours. Organic matter content was determined as ash-free dry mass by combustion of subsamples at 550oC for 4 h. The density hydrometer method was used to determine soil texture following dispersion of clay in sodium hexametaphosphate (Robertson et al. 1998). Sand content was determined gravimetrically by rinsing subsamples through a 53-micro meter sieve. All analyses of soils were conducted on the less than 2-mm fraction.

Experimental wetting treatments were applied to each of 16 replicate plots with eight plots established at each of the two study sites. Plots were randomly stratified to account for spatial heterogeneity in soil characteristics of the sites and were located ~10 m to 200 m from the stream. In 2005, plots received only a precipitation treatment. A 1-cm precipitation event was simulated by applying water to 1-m2 plots using a watering can. In 2005, event water was synthesized to match regional precipitation chemistry and contained 1.5 mg N or C/L as NO3-, NH4+, and DOC (T.K. Harms, unpublished data and J. Koch, personal communication). In 2007, each (size?) plot was divided into two sub-plots, with one sub-plot receiving a precipitation treatment and the other a flood treatment. To facilitate application of simulated floods, water applied to plots in 2007 was collected from the stream. Water applied to plots in the dry season contained 0.06 mg/L NO3--N, 0.02 mg/L NH4+-N, and 1 mg/L DOC. In the monsoon season, water contained 0.05 mg/L NO3--N, 0.03 mg/L NH4+-N, and 2 mg/L DOC. Precipitation events (1 cm) were applied to 0.5-m2 plots and flood events (20 cm) were applied within the base of chambers used for gas sampling (0.14 m2; described below). C and N in experimental event water constituted a small fraction of the N and organic matter contained in the soils.

Gas and soil samples were collected from each sub-plot prior to application of treatments and minimally at 0.5 and 24 h following treatment (dry-season flood treatment). Other experiments were additionally monitored at 2, 4, and/or 48 h. Soil samples were collected from the treated area as previously described following the conclusion of gas sampling at each time interval using a 5-cm diameter core, stored in air-tight plastic bags, and analyzed for water content in the laboratory

In 2007, redox potential was measured in soils following gas collection. Probes were constructed of a platinum electrode and copper wire soldered to a brass rod. The Pt-brass junction was insulated with marine-grade epoxy and heat-shrink tubing. Electrodes were checked for accuracy and precision in the laboratory using Light solution and water, respectively (Wafer et al. 2004). Pt electrodes were inserted 2 cm into the soil at a 45o angle. Salt bridges constructed of perforated PVC filled with 2M KCl in agar were used to ensure electrical connection with soil. For each measurement, 2M KCl solution was added to the salt bridge and a Calomel reference electrode was placed in contact with the agar and the reference and Pt electrodes were connected to a current meter. Field measurements were corrected to the standard H2 electrode by adding 250 mV to each measurement

Emissions of CO2, CH4, and N2O were measured using static chambers. All chambers were attached to bases driven into the soil at least 2 h prior to sampling to minimize disturbance effects. Insulated stainless steel chambers were attached to stainless steel bases, with vacuum grease applied to weatherstripping at the chamber-base junction to provide a gas-tight seal. Total chamber volume was approximately 20 L. Chambers were incubated for 45 minutes, with gas samples withdrawn every 15 minutes. A 20 mL sample was collected into a nylon syringe via a septum in the chamber lid and air returned to the headspace passively via a vent. Headspace gas samples were immediately transferred to evacuated 10-mL vials that had been flushed with N2. Certified samples were similarly collected and transported to the field site to check for sample preservation. Air and soil temperature were obtained during chamber incubations. Trace gas concentrations were measured on a Varian CP-3800 gas chromatograph equipped with thermal conductivity, electron capture, and flame-ionization detectors for analysis of CO2, N2O, and CH4, respectively. Instrument precision was greater than 2%. Gas flux was determined as the linear change in headspace concentration over time.

A relatively long incubation time was necessary for accurate determination of N2O and CH4 as these gases occur in low concentrations. Such long incubations of static chambers result in underestimation of CO2 flux. Although all incubations produced linear changes in CO2 concentration within the chamber headspace, a decreasing concentration gradient between the soil air and chamber headspace results in reduced flux of CO2 across the soil-air interface (Norman et al. 1997, Pumpanen et al. 2004). With the incubation time used here, CO2 flux is underestimated by approximately 4 times compared to dynamic methods (T.K. Harms, unpublished data). Caution should be applied toward comparison of these values with those collected via alternate methods. Patterns in CO2 flux following experimental wetting reported here were qualitatively similar to patterns measured using infrared gas analyzer to measure fluxes following experimental precipitation in 2005.

grid sampling

The study was conducted within the upper basin (catchment area ~7,600 km2) of the San Pedro River in southeastern Arizona, USA (Fig. 1). The upper basin is drained by an unregulated stream located near the Sonoran-Chihuahuan desert transition. The study sites lie at approximately 1200 m elevation and average annual precipitation in the basin ranges 300 to 750 mm, largely confined to two distinct seasons: winter rains (December-March) and summer monsoon storms (July-September). Weather preceding the monsoon storms is hot and dry (May-July). Two reaches were selected that are bordered by wide (>100-300 m) floodplains and contrast in hydrologic characteristics. The stream at the xeric site is characterized by losing hydrology and at the mesic site by gaining hydrology (Baillie et al. 2007; Pool and Coes 1999). Because stream channels of losing reaches are higher in elevation than the regional groundwater table, water discharges from the stream to the aquifer. The xeric site is thus characterized by ephemeral discharge and dry periods occur in summer prior to onset of floods during the monsoon season, whereas the stream at the mesic site is perennial. Recurrence interval of discharges capable of inundating the entire floodplain at the xeric site is on the order of years, whereas multiple floods inundate the mesic floodplain each year. During the period of this study, overbank floods did not inundate the entire floodplain at the xeric site. At the xeric site, median discharge was 0 m3/s during the dry season and 1.95 m3/s in the monsoon season (USGS gage # 09471550, ~7 km downstream of the study site). At the mesic site, median discharge was 0.04 m3/s during the dry season and 2.46 m3/s in the monsoon season (USGS gage # 09471000, ~10.5 km downstream from the study site).

Both floodplains are bounded by a mesquite (Prosopis velutina)-dominated terrace situated 3-5 m above the riparian zone. Vegetative communities of the floodplains are dominated by cottonwood (Populus fremontii) and mesquite trees, seepwillow shrubs (Baccharis salicifolia and B. emoryi), native bunchgrass (Sporobolus wrightii), and Johnson grass (Sorghum halepense), a non-native grass. The xeric site also contains rabbitbrush (Ericameria nauseosa), a perennial shrub. Compared to the mesic site, the xeric site supports sparse vegetation cover with fewer cottonwood trees, and obligate wetland species (e.g., seepwillow) are confined to nearstream locations. Surface elevation of the xeric floodplain is undulating whereas elevation increases gradually away from the stream at the mesic site (Fig. 2). However, absolute change in topography is similar between the two sites.

We established one 20 m x 100 m study plot along each reach. Plots were located adjacent and perpendicular to the active channel, extending 100 m into the riparian floodplain (Fig. 1). To facilitate geostatistical analysis, a regular grid was overlain with random points to yield 175 sampling locations per reach. The regular grid was spaced at 5 m intervals on the axis parallel to the stream, and random points were established at a density of 1 point/25 m2. Spacing of the regular grid was chosen to reflect average spacing of perennial trees and shrubs. Soils were sampled twice at each point, once during the dry season and once during the monsoon season.

Analysis of soils included available inorganic N, moisture, organic matter, texture, and potential rate of denitrification. Soil inorganic N was estimated as the ion-exchangeable fraction. Resin bags were constructed of 4 g mixed-bed ion-exchange resin enclosed in a nylon bag. In May 2006 we buried two replicate resin bags to 5 cm depth at each sampling point. Following 30 days of incubation, we retrieved the resin bags to individual plastic bags and collected one soil sample from each point. Soils were collected manually from within a copper ring (8 cm diameter) inserted to 8 cm depth. Soils and resin bags were transported on ice to the laboratory. Resin bags were then frozen until extraction. Soils were analyzed for moisture and organic matter content, texture, and potential rate of denitrification. Except for texture, these analyses were completed within 3 days of collection. Resin bags were again deployed during the monsoon season. Bags were deployed at the mesic site in August and two weeks later at the xeric site. Floods prevented access to the xeric site until this date. Floods again prevented access to both sites until mid-September when resin bags and soil samples were collected. Deposition of sediment at the mesic site resulted in significant loss of resin bags during the monsoon season (~60%).

Mass of nutrients adsorbed to resins was measured by extraction of resins in 2 M KCl with vigorous shaking for 1 h. Resulting extracts were filtered through pre-leached, ashless Whatman 42 filter papers and frozen until analysis for NH4+-N (phenol-hypochlorite method) and NO3--N (cadmium reduction method) on a Lachat flow-injection autoanalyzer. Analyses of bulk soil were conducted on the less than 2mm fraction. Soil moisture was determined gravimetrically by drying subsamples at 105oC for 48 hours. Organic matter content was determined as ash-free dry mass by combustion of subsamples at 550oC for 4 h. The density hydrometer method was used to determine soil texture following dispersion of clay in sodium hexametaphosphate (Robertson et al. 1999)and sand content was determined gravimetrically by rinsing subsamples through a 53-μm sieve. Soil texture was analyzed only during the dry season.

Potential rate of denitrification was assayed using the acetylene-block technique (Yoshinari et al. 1977). Soils were amended with a solution of dextrose (100 mg dextrose-C L-1) and nitrate (100 mg NO3--N L-1). Chloramphenicol (10 mg L-1) was added to prevent new protein synthesis. Resulting slurries were made anoxic by addition of N2 and acetylene was added to approximately 10% v/v headspace. Bottles were then vented to bring headspace equal to atmospheric pressure, shaken manually until all particles were suspended, and initial headspace gas samples were collected into evacuated vials. A final headspace sample was collected after 4 hours of incubation, again following vigorous shaking. Resulting gas samples were analyzed for N2O on a gas chromatograph equipped with an electron-capture detector and Porapak-Q columns. Temperature-specific Bunsen coefficients were used to account for N2O in the aqueous portion of slurries.

Soil and water solute timeseries sampling

Field sampling was conducted in the upper basin of the San Pedro River, AZ, USA (~7,600 km2), an unregulated, perennially flowing river near the Sonoran-Chihuahuan desert transition. The study site lies at approximately 1300 m elevation and receives 300 to 750 mm of precipitation annually, largely confined to two distinct seasons: winter rains (December-March) and summer monsoon storms (July-September). Streamwater samples were collected from a hydrologically gaining reach (Pool and Coes 1999, Baillie et al. 2007) and soil and shallow groundwater were collected from the adjacent floodplain. The floodplain as well as spatial and temporal patterns in riparian soil characteristics are described in detail in Harms and Grimm (2008).

Surface water samples were collected March 2001 to September 2007 at various time intervals (Fig. 1). Shallow groundwater was collected on a subset of these dates. The dataset adequately captures seasonal patterns and although it contains some observations following storms, event-based sampling was not the focus of data collection. Surface water samples were collected in triplicate at each of two locations on the stream bordering the floodplain. Means of values for the two locations for each date were used in statistical analyses. Shallow groundwater was collected from 6 wells March 2001 through February 2003, and from an additional 3-5 wells March 2003 through September 2006. Median values across all wells sampled on each date were used in statistical analyses. Wells were constructed of 5 cm (inner diameter) polyvinyl chloride (PVC) with 1-m screened sections in the zone of saturation and unscreened sections in the vadose zone. The interspace between the PVC and soil was filled with sand, followed by native substrate, and a bentonite seal at the surface. Triplicate groundwater samples were collected by pumping for 1-2 minutes using a peristaltic pump, enough time to purge standing water within the well before sample bottles were rinsed and samples collected. Dissolved oxygen (DO) was measured using an YSI-85 portable field probe in water that was continuously pumped from the well.

Water samples were transported to the laboratory on ice and, within 24 h of collection, samples were centrifuged or vacuum-filtered through ashed, stacked Whatman® GF/D and GF/F filters (nominal pore sizes 1 µm and 0.7 µm, respectively). Centrifuged samples were analyzed within 3 d of collection for nitrate (NO3--N; cadmium reduction method), ammonium (NH4+-N; phenolate method), chloride (Cl-; mercuric thiocyanate method), and soluble reactive phosphate (SRP; ascorbic acid method) on a Lachat flow-injection autoanalyzer. Filtered samples were analyzed for dissolved organic carbon (DOC) on a Shimadzu TOC-5000 or a Shimadzu TOC-VC/TN total organic carbon/nitrogen analyzer within 7 d of collection and total dissolved N on a Lachat autoanalyzer (inline persulfate-UV digestion followed by cadmium reduction) or on a Shimadzu TOC-VC/TN (chemiluminescence method) within 30 d of collection. Dissolved organic N (DON) was calculated as the difference between total dissolved N and the sum of NO3--N and NH4+-N. In our analyses, we also considered molar ratios of NO3-: NH4+, NO3-:SRP, DOC:DON, and DON:TN.

Soil samples were collected February 2003 to September 2005 on a subset of the same dates as groundwater sampling. Soils were collected +1 m from each well. We analyzed extractable and microbial pools of C and N, and rates of microbial N transformations. We collected soils from the surface (after removing unconsolidated litter) and from the region of seasonal saturation (RoSS; sensu Baker et al. 2000) using an auger (17 cm x 8 cm diameter). The location of the RoSS varied from 60 to 180 cm depth among sampling locations and was estimated from a previously collected year-long dataset of seasonal water-table elevations for a subset of the groundwater wells. These data showed that water table elevation varied on average 40 cm annually, creating a seasonally inundated region during the summer monsoon and winter rains (J.D. Schade and D.B. Lewis, unpublished data).

Soils were transported on ice to the laboratory and were processed within 24 h of collection. Samples were passed through a 2-mm sieve to remove the coarse fraction and large roots. Soil water content was determined as mass loss after drying subsamples at 105oC for 48 h and soil organic matter (SOM) content was determined as mass loss on ignition at 550oC for 4 h. Soil NO3- and NH4+ concentrations were determined by extracting a subsample with 2M KCl by shaking for 1 h. Soil particles were removed from the extract by gravity filtration through pre-leached, ashless Whatman® 42 paper filters. Samples were frozen until analysis on a Braunn+Luebbe TrAAcs autoanalyzer or a Lachat QuickChem 8000 autoanalyzer using the cadmium-reduction method for NO3- and the phenolate method for NH4+. We determined net N-mineralization and nitrification rates by incubating a subsample at room temperature in the dark for ~28 d, with soil water adjusted to fresh content weekly. Rates were determined by the difference of pre- and post- incubation extractable NH4+ and NO3-, respectively (Robertson et al. 1999). We estimated microbial biomass C and N as the difference in extractable DOC and TN between a pre-fumigation and chloroform-fumigated subsample (Robertson et al. 1999). Soils were fumigated in the dark under a chloroform atmosphere for 3 d followed by extraction in 0.5M K2SO4. Samples were frozen until analysis for dissolved organic carbon (DOC) and total dissolved N as described for water samples. During periods of low soil water content, this method can yield negative biomass values. We set all negative microbial biomass values to 0 for analysis.