Inconsistent recovery of nitrogen cycling after feral ungulate removal across three tropical island ecosystems

Abstract

AbstractNon-native ungulates (sheep, goats, and pigs) have significant negative impacts on ecosystem biodiversity, structure, and biogeochemical function throughout the Pacific Islands. Elevated nitrogen (N) availability associated with ungulate disturbance has been shown to promote the success of resource-exploitive invasive plants. While ungulate removal is a common restoration intervention, evaluations of its efficacy typically focus on vegetation responses, rather than underlying nutrient cycling. We used multiple chronosequences of ungulate exclusion (10–24 years duration) in three Hawaiian ecosystems (montane wet forest, dry forest, and dry shrubland) to determine N cycle recovery by characterizing gross mineralization and nitrification, soil inorganic N concentrations and leaching, N2O emissions, and plant tissue δ15N. Ungulate removal led to a 1–2 ‰ decline in foliar δ15N in most species, consistent with a long-term decrease in N fractionation via ecosystem N losses, or a shift in the relative turnover of N forms. This interpretation was supported by significant (dry forest) or trending (wet forest) increases in mineralization and decreases in nitrification, but conflicts with lack of observed change in inorganic N pool sizes or gaseous losses, and increased leaching in the dry forest. While results could indicate that ungulate invasions do not strongly impact N cycling in the first place (no uninvaded control sites exist in Hawai’i to test this hypothesis), this would be inconsistent with observations from other sites globally. Instead, impacts may be spatially patchy across the landscape, or ungulate invasions (possibly in combination with other disturbances) may have permanently shifted biogeochemical function or decoupled elemental cycles. We conclude that eliminating ungulate disturbance alone may not achieve restoration goals related to N cycling within the timeframe examined here.