Ecosystem recovery from chronic fertilization: biotic mechanisms underpinning soil nitrogen legacies in burned and unburned grasslands

Abstract

Nitrogen (N) is an essential element for life. Though most forms of N cannot be used by biota, humans have transformed the global N cycle and enriched terrestrial ecosystems with bioavailable N. Since soil microorganisms, plants, and soil organic matter all have incredible capacity to retain N from exogenous inputs, N accrued from the past can persist as a N source to support the functions of future biota once exogenous input ceases. This N legacy may cause ecosystem recovery from higher N conditions to lag, or it may lead to asynchronous recovery of ecosystem N pools and fluxes. Moreover, fire is a key driver of N dynamics in mesic grasslands, and it may shape responses to past and current input of N by combusting exogenous N in aboveground litter. Therefore, the overarching goal of this dissertation is to assess how fire affects plant and microbial mechanisms of ecosystem N cycling and loss after the cessation of decades of chronic N input. To address this goal, I measured N availability, soil microbial community structure and function, and plant and soil N and carbon (C) stocks in a long-term N fertilization and cessation experiment crossed with different fire treatments at the Konza Prairie Biological Station, KS, USA. Before cessation of fertilization, N-addition and fire suppression both independently decreased soil microbial enzymatic investment for N relative to C, indicating reduced microbial N-limitation, and this response was connected to substrate availability feedbacks, rather than life history tradeoffs or pH-related suppression of activity. In the first year after ceasing N- fertilization, soil N availability, measured as resin-sorbed N, dropped by 86%, nitrification potential recovered more than denitrification potential, and key microbial N-cycling populations were differentially sensitive to fertilization, together suggesting that microbial N removal and fire could be equally important mechanisms of ecosystem recovery in unburned and burned prairies, respectively. Three years after cessation of N-fertilization however, nitrification potential recovered in burned prairies only, while denitrification potential recovered by the fourth year in both fire treatments. Five years after ceasing N-fertilization, while soil N stocks in previously fertilized prairies were recovering in both fire treatments, soil microbial community structure did not recover in either fire treatment, and aboveground net primary productivity remained elevated in burned prairies. Plant shoot C:N ratios recovered in both fire treatments, and root C:N ratios recovered in burned prairies, but soil C:N ratios did not recover, suggesting different mechanisms support plant, soil, and soil microbial recovery. Collectively, while ecosystem attributes in burned and unburned prairies are not recovering similarly because of contrasting plant and soil mechanisms, loss of the soil N accumulated over decades of fertilization is happening at similar rates in both fire treatments, with fire and nitrate leaching the likely respective loss pathways in burned and unburned prairies. Overall, this dissertation shows that ecosystem N pool and flux responses to cessation of N eutrophication are caused by different feedbacks driven by fire history, which have large-scale consequences to N loss during ecosystem recovery; and that N-fertilization legacies in plant and microbial community structure can persist even as ecosystem function recovers.