Nitrogen addition has minimal effect in Wyoming big sagebrush (Artemisia tridentata) communities

Nitrogen additions are known to elicit variable responses in semi-arid ecosystems, with responses increasing with precipitation. The response of semi-arid ecosystems to nitrogen are important to understand due to their large spatial extent worldwide and the global trend of increasingly available nitrogen. In this study, we evaluated the impact of nitrogen additions on semi-arid big sagebrush (Artemisia tridentata) ecosystems. This is important given that sagebrush ecosystems are poorly understood, despite their prevalence in the western US. In addition, large-scale nitrogen additions have begun on sagebrush landscapes in Wyoming in order to mitigate population declines in mule deer (Odocoileus hemionus). The study objectives were (1) to evaluate the effectiveness of nitrogen fertilization in increasing sagebrush biomass and forage quality, and (2) to assess effects of nitrogen addition on soil biogeochemistry and vegetation community structure. We fertilized 15 plots across 5 locations in western Wyoming using nitrogen (5.47g N m-2), in the form of urea. In addition, we immobilized available nitrogen through surface hay treatments (254g hay/m2). Nitrogen additions failed to increase growth of sagebrush, alter nitrogen content of sagebrush leaders, or alter greenhouse gas efflux from soils. However, the vegetation community shifted; nitrogen-fertilized plots were only 72% similar to the controls (Bray-Curtis). Over the two years of this study, we did not find indications of nitrogen limitation of ecosystem processes, despite a wet growing season in 2014. Thus, we have found a general lack of response to nitrogen in sagebrush ecosystems, with some shifts in the plant community composition.

In a meta-analysis of research studies on the effects of N fertilization studies in water-47 limited drylands worldwide, Yahdjian et al. (8) found that in almost all cases the addition of N 48 increased net primary productivity (NPP), showing that nitrogen limitation -even if minor -is a 49 global phenomenon. The ability to predict the effects of increased N on ecosystem process and 50 structure is important because of the global trend of increasing N deposition (9). Given that 51 drylands cover 40% of the Earth's terrestrial surface (10,11) and 20% of the US (12), it is critical 52 to investigate the effects of increasing N availability in dryland systems, due to their co-53 limitation by both water and N. In addition, dryland systems are home to 44% of all cultivated 54 land (13,14), making understanding the effects of increasing N particularly important due to the 55 high use N-based fertilizers. 56 In the western US, big sagebrush (Artemisia tridentata spp.) ecosystems are the dominant 57 dryland system (15,16), covering 48 million hectares (ha) and 13 states (17). Despite its current 58 extent in the US, up to 60% of the historic, pre-settlement big sagebrush range has been partially 59 or completely converted to exotic annual grasslands because of increased disturbance to these 60 systems and conversion to grazing and agricultural lands (18). In addition, significant portions of 61 the big sagebrush range continue to be affected by energy development, agriculture, and 62 expansion of urban areas (19)(20)(21)(22). These drivers of the loss of big sagebrush -unsustainable 63 use of natural resources, population expansion, globalization, as well as climate change -are 64 the same as those exacerbating desertification in drylands globally (13). Big sagebrush 65 ecosystems are expected to continue to decline in coverage over the next decade with 2.3-5.5 66 million hectares projected to be impacted due to oil and gas development (23). While agricultural 67 additions of N account for a majority of the increase in N availability globally, energy 68 development also affects N deposition locally, through the production and release of NO x 69 compounds (9,24,25).

70
The effects of nitrogen additions on big sagebrush ecosystems are poorly understood due 71 to both the small number of studies and the complexities associated with potential co-limitation 72 of ecosystem processes by water and N in dryland systems (26

216
We analyzed big sagebrush growth and forage quality data by site to examine treatment 217 effects and by individual subplots (n=45) to analyze correlation between soil N and big 218 sagebrush growth. Tukey's HSD test was used to evaluate the site differences between 219 treatments for four metrics of growth and three metrics of forage quality: leader length, 220 ephemeral branch length, leader dry mass, and leader dry density, along with percent C, percent 221 N, and C:N ratio of the big sagebrush leaders. We examined the effect of soil N availability for 222 its ability to predict leader length with multiple models, using the 'gls' function in the 'nlme' 223 package. Models were evaluated using Bayesian Information Criteria (BIC) and the model with 224 the lowest BIC was selected (Table 1). 225 We averaged plant community data by site (n=5) for total plant cover and species 226 richness. Total plant cover was assessed by summing the cover class midpoint for all species in 227 each quadrat and then averaging across all quadrate to determined mean total percent cover. We 228 used Tukey's HSD test to evaluate differences in cover and richness between treatments. We 229 also calculated Bray-Curtis dissimilarity using the 'distance' function within the `ecodist` We found no treatment effect on total plant cover or species richness. However, plant 277 community composition shifted slightly between treatments; a Bray-Curtis similarity analysis 278 showed that the largest differences were found when N was added, with a 72% similarity to both 279 the control and N removal treatments. The N removal and control treatments were 79% similar 280 with respect to plant community composition.

281
Soil trace gas emission displayed no effect from treatments for any gas species.
282 Differences in flux among treatments types were not significant using a Tukey HSD. Fluxes 283 were relatively consistent with other studies of gas fluxes in big sagebrush plant communities 284 (Table 2).

299
The lack of significant increase in soil N availability in the second year in fertilized plots 300 is likely due to 'normal' N losses through N mineralization, gaseous loses (volatilization and 301 microbially-mediated), and leaching. Additions of N were conducted at 4.5 g N m -2 (or 45 kg N 302 ha -1 ), which is 150-900% of annual N mineralization (38). Trace gas emissions of N-containing 303 gases (N 2 O and NO x ) accounted for an additional 10-20 kg N ha -1 yr -1 loss from the soil, which 304 could contribute to the lack of increase observed in inorganic soil N among treatments in year 305 two (5,26). While we found lower fluxes of N 2 O compared to these studies one year after 306 fertilization, it may be that fluxes were greater immediately after fertilization or during the spring 307 when water was more available, therefore not captured in our sampling period.

308
Increased N availability had minimal effect on the length of big sagebrush leaders over 309 the two years of the study. Additional N did not result in an increase in forage quality; C:N ratio 310 and mass of N were the same between the control and N fertilized plots. The minimal difference 311 between the N fertilized and the control subplots indicates that while N fertilization in the big 312 sagebrush communities of the Pinedale Anticline increased the amount of N available to plants 313 for one growing season, it did not increase forage or forage quality such that mule deer and other 314 herbivores would benefit. Our prediction that there would be no differences between treatments 315 was generally supported, despite the minor increases in leader length correlated with increasingly 316 available soil N estimated in the first year.

317
While the plant community composition changed slightly with N addition as predicted -318 becoming more dissimilar with the addition of N, there was no change to either total plant cover 319 or species richness by treatment. This is not surprising as soil N availability was only increased 320 for the first year after treatment. In addition, drylands -and big sagebrush ecosystems -are 321 primarily water limited, making short-lived increases of N unlikely to affect cover or richness.
322 Interestingly, the removal of N resulted in greater differences than N addition. This is likely due 323 to increasing the degree of N limitation within the hay treatment.

324
There was no effect of N additions on GHG efflux (i.e., CO 2 , CH 4 , and N 2 O) from soil.