Larkin, John2020-02-262020-02-262019-12https://hdl.handle.net/11299/211699University of Minnesota M.S. thesis. December 2019. Major: Civil Engineering. Advisor: Paul Capel. 1 computer file (PDF); viii, 93 pages + 2 supplementary files.The majority of the 20th century saw a marked increase in environmental degradation, and it was not until creation of the Clean Water Act (CWA) in 1972 that the United States took its first major step towards active improvement of the environment. The CWA dictates that all public waters must be fishable and swimmable. To be fishable the water must be clean enough to sustain fish and other aquatic organisms. To be swimmable means the water must be suitable for recreational purposes and not cause illness when humans or animals come into contact with it. Water pollution can come from many sources and is not just limited to acutely toxic chemicals. Rather, anything that makes the water not fishable or swimmable can be considered a pollutant. A source of water pollution not often considered is agricultural operations as there is a heavy chemical reliance, which can often result in runoff ladened with pesticides (insecticides, herbicides, and fungicides) and fertilizers. Pesticides are used to control unwanted organisms that compete with, or are detrimental to, crops. Fertilizers, however, are used to increase the nutrient content of the soil and so pollution from fertilizer use is often referred to as nutrient pollution (USGS Surface Runoff, 2018). To achieve the goal of fishable and swimmable, the Federal government, through the Environmental Protection Agency (EPA), works with states to reduce or eliminate pollution to waters via implementation of Total Maximum Daily Loads (TMDLs). A TMDL for nutrients and sediment was implemented in the Chesapeake Bay watershed in the late 1990s across Pennsylvania, Maryland, Virginia, District of Columbia, New York, Delaware, and West Virginia (EPA, 2015) The annual limits set in this TMDL were 185.9 million pound of nitrogen, 12.5 million pounds of phosphorus, and 6.45 billion pounds of sediment – reductions of 25%, 24%, and 20%, respectively (EPA, 2015). Since implementation there has been interest in determining whether the goals laid out in the TMDL have been achieved, which has been difficult to definitively determine. One complicating factor in the analysis of real-world observed data has been the year-to-year variation in streamflow and chemical loads created by the annual variation in weather. Other difficulties exist with constantly changing landscape as agricultural and urban land use spread. This study focuses on the comparison of various factors affecting the loads of a conservative chemical and nitrogen and the relative impact those factors have on the nitrogen load. Precipitation falling on the surface of land can either be recycled to the atmosphere through evapotranspiration (ET), be stored in the upper soil layers, exit the surface as runoff, or exit the upper soil as recharge to groundwater. The groundwater then moves to a discharge point, a iii process that may take years. Precipitation can introduce some nitrogen to the land surface, known as atmospheric deposition, and is the main source of pre-anthropogenic nitrogen. Additional nitrogen is introduced to the environment primarily through urban and agricultural activities. Water movement then carries nitrogen through, and out, of the landscape to either groundwater or surface water. The first chapter examines how variation in precipitation leads to variation in observed streamflow and water quality. As there may be numerous factors affecting water quality, it was determined that the model would be run with a simple, generic chemical C that was conservative, non-sorptive, and non-reactive chemical. The goal was to better understand how the magnitude of groundwater flow and C load to the stream, the variation in the flow and load of C, and the time it may take the groundwater to reach the stream are all affected by the variation in weather and land use. A standardized field, referred to as a unit field, was used for comparisons between scenarios. A square field was created with sides of 400 m by 400 m, for a total area of 16 hectares (39.54 acres). Annual precipitation was simulated using a random normal distribution with mean of 1.0 ± 0.2 m. From this it was assumed that 0.5 m left the system as ET and 0.25 m as groundwater. The remaining precipitation left the system as runoff. A lumped-parameter model (LPM) was used to model the groundwater. The LPM followed an exponential distribution and the total groundwater discharge to the stream was added to the runoff to create the streamflow. The total load of nitrogen was also accounted for using the LPM and runoff concentrations. Using the model, four basic land use scenarios were examined: natural vegetation, agricultural land use, and agricultural land use with a Best Management Practice (BMPs) reducing the nitrogen levels by 10% and 30%. Results showed that, as in real-world observed data, identifying trends year-to-year was nearly impossible and that the groundwater component of the model greatly increased uncertainty. The model showed that, due to the lag resulting from the groundwater, the stream required four times the groundwater transit time to approach steady state – e.g. a groundwater transit time of 20 years meant the stream required approximately 80-100 years to reach steady state. It also demonstrated a change in load of 25% was not easily identifiable without a large dataset and may take decades to observe (if all the change occurred in a short period). In the final chapter a second model was used to examine how land use, land management, and climate might affect streamflow and nitrogen load. APEX, Agricultural Policy/Environmental eXtender model, was used to model various land uses (agriculture, urban, perennial vegetation), land managements (BMPs, nitrogen application rates, percent urban), and different climate projections to identify relative importance when it came to streamflow and nitrogen load. A new iv stand-alone model in R was then created that combined the APEX output, along with the groundwater model used in chapter 1, and developed a theoretical watershed. The purpose of creating this watershed was to analyze how a change in part of the watershed may impact the streamflow and nitrogen load. This information is useful to water resource scientist and policy decision makes as it helps to emphasize what changes may be most important from a water quality perspective. A by-product of this analysis is that this tool in R is now available for anyone to use as it is a simple model based off of APEX simulations. First ArcGIS and APEX were used to delineate a field near Bridgeville, Delaware. This would be the unit field for the second chapter and was 13.21 ha (32.64 acres) in size and used for all scenarios. The model outputs for the hydrologic and nitrogen cycles were captured and annual water and nitrogen budgets developed. A standard land use and management was chosen as the basis for comparison: corn with conventional tillage, recommended nitrogen application rates, no BMPs, and historically observed weather. All other scenarios were compared to this. It was identified that reducing nitrogen application rates (where applicable) and implementing buffer strips were most effective at reducing total nitrogen in the stream for all crops. No till increased the amount of water and nitrogen to the groundwater. The timing of rainfall was critical, as uniform increases in rainfall intensity resulted in large increases in nitrogen export, but a single large event did not substantially increase nitrogen loads. Building the theoretical watershed in R, the watershed can be any size (by number of unit fields) and composition (land use and land management). For purposes of this study, a watershed comprised of 1,000 unit fields was modeled from pre-settlement into the future. While it shows that it is not possible to achieve pre-settlement levels of nitrogen, implementation of BMPs can help reduce nitrogen loads by 25%, the level required in the Chesapeake Bay TMDLs. This information thus allows water resource scientist and policy makers, when they identify regions for improvement, to quickly recommend options and provides background on how they might see their watershed responded to those changes.enThesis or Dissertation