Mixed method approach to assess atmospheric nitrogen deposition in arid and semi-arid ecosystems

Abstract

Arid and semi-arid ecosystems (aridlands) cover a third of Earth's terrestrial surface and contain organisms that are sensitive to low level atmospheric pollutants. Atmospheric nitrogen (N) inputs to aridlands are likely to cause changes in plant community composition, fire frequency, and carbon cycling and storage. However, few studies have documented long-term rates of atmospheric N inputs in aridlands because dry deposition is technically difficult to quantify, and extensive sampling is needed to capture fluxes with spatially and temporally heterogeneous rainfall patterns. Here, we quantified long-term spatial and temporal patterns of inorganic N deposition in protected aridland ecosystems across an extensive urban-rural gradient using multiple sampling methods. We compared long-term rates of N deposition from ion-exchange resin (IER) collectors (bulk and throughfall, 2006–2015), wet-dry bucket collectors (2006–2015), and dry deposition from the inferential method using passive samplers (2010–2012). From mixed approaches with IER collectors and inferential methods, we determined that 7.2 ± 0.4 kgNha−1y−1 is deposited to protected Sonoran Desert within metropolitan Phoenix, Arizona and 6.1 ± 0.3 kgNha−1y−1 in nearby desert ecosystems. Regional scale models overestimated deposition rates for our sampling period by 60% and misidentified hot spots of deposition across the airshed. By contrast, the easy-deployment IER throughfall collectors showed minimal spatial variation across the urban-rural gradient and underestimated deposition fluxes by 54%, largely because of underestimated dry deposition in throughfall. However, seasonal sampling of the IER collectors over 10 years allowed us to capture significant seasonal variation in N deposition and the importance of precipitation timing. These results, derived from the longest, spatially and temporally explicit dataset in drylands, highlight the need for long-term, mixed methods to estimate atmospheric nutrient enrichment to aridlands in a rapidly changing world.