The Nitrogen Legacy: Understanding Time Lags in Catchment Response as a Function of Hydrologic and Biogeochemical Controls
MetadataShow full item record
Global population has seen a more than threefold increase over the last 100 years, accompanied by rapid changes in land use and a dramatic intensification of agriculture. Such changes have been driven by a great acceleration of the global nitrogen (N) cycle, with N fertilizer use now estimated to be 100 Tg/year globally. Excess N commonly finds its way into both groundwater and surface water, leading to long-term problems of hypoxia, aquatic toxicity and drinking water contamination. Despite ongoing efforts to improve water quality in agroecosystems, results have often been disappointing, with significant lag times between adoption of accepted best management practices (BMPs) and measurable improvements in water quality. It has been hypothesized that such time lags are a result of the buildup of legacy N within the landscape over decades of fertilizer application and agricultural intensification. The central theme of my research has been an exploration of this N legacy, including (1) an investigation of the form, locations and magnitudes of legacy N stores within intensively managed catchments; (2) development of a parsimonious, process-based modeling framework for quantifying catchment-scale time lags based on both soil nutrient accumulations (biogeochemical legacy) and groundwater travel time distributions (hydrologic legacy); and (3) use of a statistical approach to both quantifying N-related time lags at the watershed scale, and identifying the primary physical and management controls on these lags. As a result of these explorations I am able to provide the first direct, large-scale evidence of N accumulation in the root zones of agricultural soils, accumulation that may account for much of the ‘missing N’ identified in mass balance studies of heavily impacted watersheds. My analysis of long-term soil data (1957-2010) from 206 sites throughout the Mississippi River Basin (MRB) revealed N accumulation in cropland of 25-70 kg ha-1 y-1, a total of 3.8 ± 1.8 Mt y-1 at the watershed scale. A simple modeling framework was then used to show that the observed accumulation of soil organic N (SON) in the MRB over a 30-year period (142 Tg N) would lead to a biogeochemical lag time of 35 years for 99% of legacy SON, even with a complete cessation of fertilizer application. A parsimonious, process-based model, ELEMeNT (Exploration of Long-tErM Nutrient Trajectories), was then developed to quantify catchment-scale time lags based on both soil N accumulation (biogeochemical legacy) and groundwater travel time distributions (hydrologic legacy). The model allowed me to predict the time lags observed in a 10 km2 Iowa watershed that had undergone a 41% conversion of area from row crop to native prairie. The model results showed that concentration reduction benefits are a function of the spatial pattern of implementation of conservation measures, with preferential conversion of land parcels having the shortest catchment-scale travel times providing greater concentration reductions as well as faster response times. This modeling framework allows for the quantification of tradeoffs between costs associated with implementation of conservation measures and the time needed to see the desired concentration reductions, making it of great value to decision makers regarding optimal implementation of watershed conservation measures. To better our understanding of long-term N dynamics, I expanded the ELEMeNT modeling framework described above to accommodate long-term N input trajectories and their impact on N loading at the catchment scale. In this work, I synthesized data from a range of sources to develop a comprehensive, 214-year (1800-2104) trajectory of N inputs to the land surface of the continental United States. The ELEMeNT model was used to reconstruct historic nutrient yields at the outlets of two major U.S. watersheds, the Mississippi River and Susquehanna River Basins, which are the sources of significant nutrient contamination to the Gulf of Mexico and Chesapeake Bay, respectively. My results show significant N loading above baseline levels in both watersheds before the widespread use of commercial N fertilizers, largely due to 19th-century conversion of natural forest and grassland areas to row-crop agriculture. The model results also allowed me to quantify the magnitudes of legacy N in soil and groundwater pools, thus highlighting the dominance of soil N legacies in the MRB and groundwater legacies in the SRB. It was found that approximately 85% of the annual N load in the MRB can be linked to inputs from previous years, while only 47% of SRB N loading is associated with “older” N. In addition, it was found that the dominant sources of current N load in the MRB are fertilizer, atmospheric deposition, and biological N fixation, while manure and atmospheric deposition account for approximately 64% of the current loads in the SRB. Finally, long-term N surplus trajectories were paired with long-term flow-averaged nitrate concentration data to as means of quantifying N-related lag times across an intensively managed watershed in Southern Ontario. In this analysis, we found a significant linear relationship between current flow-averaged concentrations and current N surplus values across the study watersheds. Temporal analysis, however, showed significant nonlinearity between N inputs and outputs, with a strong hysteresis effect indicative of decadal-scale lag times between changes in N surplus values and subsequent changes in flow-averaged nitrate concentrations. Annual lag times across the study watersheds ranged from 15-33 years, with a mean lag of 24.5 years. A seasonal analysis showed a distribution of lag times across the year, with fall lags being the shortest and summer lags the longest, likely due to differences in N delivery pathways. Multiple linear regression analysis of dominant controls showed tile drainage to be a strong determinant of differences in lag times across watersheds in both fall and spring, with a watershed’s fractional area under tile drainage being significantly linked to shorter lag times. In summer, tile drainage was found to be an insignificant factor in driving lag times, while a significant relationship was found between the percent soil organic matter and longer N-related lag times. By moving beyond the traditional focus on nutrient concentrations and fluxes, and instead working towards quantification of the spatio-temporal dynamics of non-point source nutrient legacies and their current and future impacts on water quality, we make a significant contribution to the science of managing human impacted landscapes. Due to the strong impacts of nutrient legacies on the time scales for recovery in at-risk landscapes, my work will enable a more accurate assessment of the outcomes of alternative management approaches in terms of both short- and long-term costs and benefits, and the evaluation of temporal uncertainties associated with different intervention strategies.
Cite this version of the work
Kimberly Van Meter (2016). The Nitrogen Legacy: Understanding Time Lags in Catchment Response as a Function of Hydrologic and Biogeochemical Controls. UWSpace. http://hdl.handle.net/10012/10896