Anthropogenic nitrogen (N) deposition has been shown to have wide-reaching consequences for terrestrial ecosystem properties and functions, yet we know little about deposition’s effects on drylands. Studies suggest critical N loads may be relatively low in drylands due to limited N stocks and extremes in water availability and temperature. However, direct evidence that increasing N inputs have sustained effects in drylands has been limited and conflicting. Here, we used a long-term fertilization experiment to assess the fate of N inputs in semiarid grasslands. In 2011, we established 3 study sites along a soil texture gradient in Arches NP, with fertilization plots receiving 0, 2, 5, or 8 kg N/ha/y for 7 years (n=5 per treatment per site). We used a two-pronged approach to determine if and how N inputs affected coupled carbon (C) and nutrient cycling. First, we measured multiple biogeochemical properties prior to each fertilization event to assess the longevity of fertilization effects and whether effects compounded through time. Second, to better understand the immediate fate of N, we tracked a suite of N cycling metrics (e.g., soil inorganic N, gaseous N losses) just before and at multiple timepoints for 90 days following the spring addition of N fertilizer in 2013.
Results/Conclusions
After 7 years of N treatments, we found that total soil N and foliar N were unaffected by N inputs, regardless of N input amounts and regardless of soil texture. Instead, soil biogeochemical cycling was more strongly correlated with fine scale variability in edaphic conditions across our 3 study sites. Over the short-term, soil nitrate and ammonium concentrations increased significantly following fertilization (P<0.01 and P< 0.05, respectively), but returned to pre-fertilization levels after just 3 weeks. Similarly, N oxide fluxes increased in the 8 kg N/ha/y plots relative to control plots immediately after fertilization (P<0.05), and quickly returned to pre-fertilization levels. Variability in soil moisture across the study sites was the strongest control of gas flux rates in our assessment of added N’s fate in the experimental plots. Taken together, our results suggest that N is lost more rapidly and completely from dryland soils than has been predicted based on their low background concentrations of N and their relatively limited times of biological activity. Our results also highlight the importance of fine scale differences in dryland soil texture as a driver of coupled biogeochemical cycling over short and long time scales.