During the autumn of 2012, we collected litter from individuals of triangle-leaf bursage (Ambrosia deltoidea), a dominant, drought-deciduous herbaceous shrub that has been shown to accumulate excess N and P in these plots. A. deltoidea litter was collected from twenty or more individuals at each of the 5 urban sites inside the city (totaling 100+ shrubs), from both the N+P fertilization plots and unfertilized (control) areas, and pooled by treatment to create one homogenized bag of each treatment. Similarly, litter was collected and homogenized within each of the N+P fertilized and control treatments in the 5 non-urban, outlying desert sites. Collected litter from each region was dried, weighed, and placed in litterbags (see below) and then set out on the soil surface in one location within the region from which the litter originated: inside the city (SME) and outside the city (SRR). In other words, we decomposed plant litter in the region in which it was grown (inside or outside the city), allowing us to understand the influence of the urban location on both litter production and decomposition environment while avoiding artifacts that may result from decomposer communities working on non-native litter. The field sites SME and SRR were selected because they have comparable 10-year average annual rainfall, though rainfall during 2013 when the experiment was conducted differed between the two locations (SME received 130 mm while SRR received 185 mm).

We decomposed A. deltoidea litter in litterbags composed of UV transparent plastic film (ACLAR 33C Film) and UV opaque plastic film (CON-TROL-CURE UV Blocking Films, UV Process Supply Inc.) in each of the two desert locations (inside and outside the city). Plastic film was shaped into 10 cm × 10 cm pouches, sealed with UV-transparent tape on all sides, and ~2 mm holes were drilled through the film on both sides to allow water movement. Three grams of A. deltoidea litter from both the N+P and control treatments were placed into 16 replicate UV-transparent and UV-opaque litterbags (2 locations x 2 fertilization treatments x 2 bag types = 8 treatments x 16 replicate bags each = 128 litterbags). All litterbags were placed in the field in late December 2012 in non-fertilized areas on level ground, in direct sunlight away from plants. Litterbags were arranged in four replicate blocks, with backs laid out randomly in grids with approximately 30 cm of spacing in between them to avoid cross-contamination. Four replicates of each of the 8 treatments were retrieved after 10 weeks in early February, 20 weeks in late April, 30 weeks in early July, and 40 weeks in late September 2013.

Immediately following litter bag collection from the field, approximately five leaves from each litter sample were put into 10 ml of a 4% formalin solution to preserve the bacteria present on the leaves. The remaining litter samples were air-dried for at least a week. Once dry, accumulated sediment was scraped off of the litter, and the sediment and litter (now mostly sediment-free) were weighed separately. Five leaves were weighed to estimate the mass of leaves preserved in formalin, then the entire litter sample was ground using a ball mill (Spex-Certiprep 8000D, Metuchen, New Jersey, USA). Ash-free dry mass (AFDM) was determined by incinerating a subsample of the ground litter at 600°C in a muffle oven for 3 hours, and mass loss calculated as proportion of initial AFDM remaining. The difference in mass loss between the UV opaque and transparent bags for any given Location×Fertilization treatment is assumed to be the result of photodegradation, which can be used to calculate what proportion of total mass loss in the UV-transparent bags is attributed to photodegradation. Litter %C and %N were measured from a 3 mg subsample of ground litter on a CHN elemental analyzer (Perkin Elmer PE2400, Waltham, Massachusetts, USA). Litter %P was measured using nitric acid digestion. Briefly, 100 mg of ground sample was inserted into a combustion safe tube with 10 ml of 70% nitric acid. Samples sat in the acid for 15 min followed by microwaving at 175°C for 20 min using a microwave digestion system (MARS 5 Microwave Reaction System, Matthews, North Carolina, USA). Once digested, the sample tubes were slowly opened in the fume hood, then poured into containment vials and analyzed for P content on an Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES, Thermo iCAP6300, Hudson, New Hampshire, USA). Nutrient ratios (C:N, C:P, N:P) were calculated using these data. Litter lignin was measured using the sequential acid digestion method (Van Soest, 1994). Briefly, 0.5 g of ground sample was placed into fiber filter bags and then digested in a fiber analyzer (ANKOM A200 Fiber Analyzer, Macedon, New York, USA) using neutral detergent, followed by acid detergent, then 72% sulfuric acid to identify cellulose, hemicellulose, and lignin, respectively. Filter bags were then ashed at 500°C for 5.5 hrs to correct the lignin content for non-organic recalcitrant compounds. Litter bacterial cell counts were conducted by sonicating the formalin-preserved leaves for 30 minutes in an ice water bath. After sonication, 2 ml of the formalin solution was added to an ethanol-rinsed filtration manifold, stained with 500 µl 25× SYBR Green for 15 minutes, then vacuum filtered through a 0.2 µm black polycarbonate filter (supported by a 0.45 µm backing filter). The column was rinsed with 1 ml sterilized deionized H2O, again vacuum filtered, then the filter slide mounted and stored in the refrigerator in the dark until counted. Cells were enumerated using epifluorescent microscopy (1000×) by counting ten random fields and categorizing cells into shape (coccoid or rod) and size class (small and large). Cell biovolumes were calculated using equations for the geometric shape size classes based on diameter measurements using a stage micrometer and expressed as biovolume per g AFDM. Size and shape are necessary to calculate biovolume, and are then summed into a total biovolume for each sample. Since previous conversion factors from biovolume to biomass have not been established for this site, data are presented as biovolume.