BIORETENTION/CISTERN/IRRIGATION TO ELIMINATE STORMWATER RUNOFF AT THE UNIVERSITY OF MARYLAND

Water quality of parking lot (~1,858 m2) stormwater runoff and its treated effluent flow were analyzed for total phosphorus (TP), total nitrogen (TN), total suspended solids (TSS), electrical conductivity (EC), copper, lead and zinc. The novel system under investigation, located at the University of Maryland, College Park, Maryland, includes a standard bioretention facility, underdrained to a cistern to store treated stormwater, and pumped to a vegetable garden for irrigation. The site abstraction, the average bioretention abstraction, and bowl volumes were estimated to be 8500, 4378, and 895 L, respectively; this indicates that rain events of more than 0.45 cm are necessary to produce runoff and more than 0.75 cm will produce system overflow. The cistern water quality indicates good-to-excellent treatment by the system. Compared to local tap water, cistern water has lower concentrations of TP, TN, EC (non-winter), copper, and zinc, indicating a good water source for irrigation.

SEQUENTIAL DECISIONS AND PREDICTIONS IN NATURAL LANGUAGE PROCESSING

Natural language processing has achieved great success in a wide range of ap- plications, producing both commercial language services and open-source language tools. However, most methods take a static or batch approach, assuming that the model has all information it needs and makes a one-time prediction. In this disser- tation, we study dynamic problems where the input comes in a sequence instead of all at once, and the output must be produced while the input is arriving. In these problems, predictions are often made based only on partial information. We see this dynamic setting in many real-time, interactive applications. These problems usually involve a trade-off between the amount of input received (cost) and the quality of the output prediction (accuracy). Therefore, the evaluation considers both objectives (e.g., plotting a Pareto curve). Our goal is to develop a formal understanding of sequential prediction and decision-making problems in natural language processing and to propose efficient solutions. Toward this end, we present meta-algorithms that take an existent batch model and produce a dynamic model to handle sequential inputs and outputs. Webuild our framework upon theories of Markov Decision Process (MDP), which allows learning to trade off competing objectives in a principled way. The main machine learning techniques we use are from imitation learning and reinforcement learning, and we advance current techniques to tackle problems arising in our settings. We evaluate our algorithm on a variety of applications, including dependency parsing, machine translation, and question answering. We show that our approach achieves a better cost-accuracy trade-off than the batch approach and heuristic-based decision- making approaches. We first propose a general framework for cost-sensitive prediction, where dif- ferent parts of the input come at different costs. We formulate a decision-making process that selects pieces of the input sequentially, and the selection is adaptive to each instance. Our approach is evaluated on both standard classification tasks and a structured prediction task (dependency parsing). We show that it achieves similar prediction quality to methods that use all input, while inducing a much smaller cost. Next, we extend the framework to problems where the input is revealed incremen- tally in a fixed order. We study two applications: simultaneous machine translation and quiz bowl (incremental text classification). We discuss challenges in this set- ting and show that adding domain knowledge eases the decision-making problem. A central theme throughout the chapters is an MDP formulation of a challenging problem with sequential input/output and trade-off decisions, accompanied by a learning algorithm that solves the MDP.

Understanding Urban Stormwater Denitrification in Bioretention Internal Water Storage Zones

Free-draining bioretention systems commonly demonstrate poor nitrate removal. In this study, column tests verified the necessity of a permanently saturated zone to target nitrate removal via denitrification. Experiments determined a first-order denitrification rate constant of 0.0011 min-1 specific to Willow Oak woodchip media. A 2.6-day retention time reduced 3.0 mgN/L to below 0.05 mg-N/L. During simulated storm events, hydraulic retention time may be used as a predictive measurement of nitrate fate and removal. A minimum 4.0 hour retention time was necessary for in-storm denitrification defined by a minimum 20% nitrate removal. Additional environmental parameters, e.g., pH, temperature, oxidation-reduction potential, and dissolved oxygen, affect denitrification rate and response, but macroscale measurements may not be an accurate depiction of denitrifying biofilm conditions. A simple model was developed to predict annual bioretention nitrate performance. Novel bioretention design should incorporate bowl storage and large subsurface denitrifying zones to maximize treatment volume and contact time.