<p>Streamflow prediction in ungauged basins (PUB) is a process generating streamflow time series at ungauged reaches in a river network. PUB is essential for facilitating various engineering tasks such as managing stormwater, water resources, and water-related environmental impacts. Machine Learning (ML) has emerged as a powerful tool for PUB using its generalization process to capture the streamflow generation processes from hydrological datasets (observations). ML’s generalization process is impacted by two major components: data splitting process of observations and the architecture design. To unveil the potential limitations of ML’s generalization process, this dissertation explores its robustness and associated uncertainty. More precisely, this dissertation has three objectives: (1) analyzing the potential uncertainty caused by the data splitting process for ML modeling, (2) investigating the improvement of ML models’ performance by incorporating hydrological processes within their architectures, and (3) identifying the potential biases in ML’s generalization process regarding the trend and periodicity of streamflow simulations.</p><p>The first objective of this dissertation is to assess the sensitivity and uncertainty caused by the regular data splitting process for ML modeling. The regular data splitting process in ML was initially designed for homogeneous and stationary datasets, but it may not be suitable for hydrological datasets in the context of PUB studies. Hydrological datasets usually consist of data collected from diverse watersheds with distinct streamflow generation regimes influenced by varying meteorological forcing and watershed characteristics. To address the potential inconsistency in the data splitting process, multiple data splitting scenarios are generated using the Monte Carlo method. The scenario with random data splitting results accounts for frequent covariate shift and tends to add uncertainty and biases to ML’s generalization process. The findings in this objective suggest the importance of avoiding the covariate shift during the data splitting process when developing ML models for PUB to enhance the robustness and reliability of ML’s performance.</p><p>The second objective of this dissertation is to investigate the improvement of ML models’ performance brought by Physics-Guided Architecture (PGA), which incorporates ML with the rainfall abstraction process. PGA is a theory-guided machine learning framework integrating conceptual tutors (CTs) with ML models. In this study, CTs correspond to rainfall abstractions estimated by Green-Ampt (GA) and SCS-CN models. Integrating the GA model’s CTs, which involves information on dynamic soil properties, into PGA models leads to better performance than a regular ML model. On the contrary, PGA models integrating the SCS-CN model's CTs yield no significant improvement of ML model’s performance. The results of this objective demonstrate that the ML’s generalization process can be improved by incorporating CTs involving dynamic soil properties.</p><p>The third objective of this dissertation is to explore the limitations of ML’s generalization process in capturing trend and periodicity for streamflow simulations. Trend and periodicity are essential components of streamflow time series, representing the long-term correlations and periodic patterns, respectively. When the ML models generate streamflow simulations, they tend to have relatively strong long-term periodic components, such as yearly and multiyear periodic patterns. In addition, compared to the observed streamflow data, the ML models display relatively weak short-term periodic components, such as daily and weekly periodic patterns. As a result, the ML’s generalization process may struggle to capture the short-term periodic patterns in the streamflow simulations. The biases in ML’s generalization process emphasize the demands for external knowledge to improve the representation of the short-term periodic components in simulating streamflow.</p>

<p> The overall theme of my dissertation is applying frontier econometric models to interesting economic problems. The first chapter analyzes how individual consumption responds to permanent and transitory income shocks is limited by model misspecification and availability of data. The misspecification arises from ignoring unemployment risk while estimating income shocks. I employ the Heckman two step regression model to consistently estimate income shocks. Moreover, to deal with data sparsity, I propose identifying the partial consumption insurance and income and consumption volatility heterogeneities at the household level using Least Absolute Shrinkage and Selection Operator (LASSO). Using PSID data, I estimate partial consumption insurance against permanent shock of 63% and 49% for white and black household heads, respectively; the white and black household heads self-insure against 100% and 90% of the transitory income shocks, respectively. Moreover, I find income and consumption volatilities and partial consumption insurance parameters vary across time. In the second chapter I recast smooth structural break test proposed by Chen and Hong (2012), in a predictive regression setting. The regressors are characterized using the local to non-stationarity framework. I conduct a Monte Carlo experiment to evaluate the finite sample performance of the test statistic and examine an empirical example to demonstrate its practical application. The Monte Carlo simulations show that the test statistic has better power and size compared to the popular SupF and LM. Empirically, compared to SupF and LM, the test statistic rejects the null hypothesis of no structural break more frequently when there actually is a structural break present in the data. The third chapter is a collaboration with James Reeder III. We study the effects of using promotions to drive public policy diffusion in regions with polarized political beliefs. We estimate a model that allows for heterogeneous effects at the county-level based upon state-level promotional offerings to drive vaccine adoption during COVID-19. Central to our empirical application is accounting for the endogenous action of state-level agents in generating promotional schemes. To address this challenge, we synthesize various sources of data at the county-level and leverage advances in both the Bass Diffusion model and 10 machine learning. Studying the vaccine rates at the county-level within the United States, we find evidence that the use of promotions actually reduced the overall rates of adoption in obtaining vaccination, a stark difference from other studies examining more localized vaccine rates. The negative average effect is driven primarily by the large number of counties that are described as republican leaning based upon their voting record in the 2020 election. Even directly accounting for the population’s vaccine hesitancy, this result still stands. Thus, our analysis suggests that in the polarized setting of the United States electorate, more localized policies on contentious topics may yield better outcomes than broad, state-level dictates. </p>