Stream water quality and ecosystem health are controlled by a multitude of complex, interacting anthropogenic and natural processes. A generalized understanding of stream water quality dynamics under diverse hydro-climatic and biogeochemical conditions therefore still remains elusive, particularly in the context of coastal urban/natural environments. This dissertation focuses to investigate the major spatiotemporal controls of stream water quality, and leverages the knowledge to investigate stream biogeochemical-ecological similitude (parametric reduction), scaling, and emergent patterns across the East Coast of U.S.A. Total nitrogen (TN), total phosphorus (TP), chlorophyll-a (Chla), and dissolved oxygen (DO) were used to represent the stream water quality. Pearson correlation matrix, principal component analysis, and factor analysis were used to analyze the interrelations and groupings among the environmental drivers alongside tracking their individual linkages with the stream water quality. Partial least squares regression models were then developed to directly estimate the linkages by appropriately resolving multicollinearity among the predictor variables. Spatial analytics on 6 major streams in southeast Florida indicated the external sources (e.g., Everglades) and draining watershed's agricultural lands as the major drivers of in-stream TN, TP, and Chla. However, stream DO was most strongly influenced by the adjacent groundwater depth and watershed (agricultural and built-up) land uses. Spatio-temporal analytics on 50 stream sites across the East Coast showed strong controls of climatic (stream temperature) and biogeochemical (pH and salinity) drivers on stream DO. The dominant drivers were then involved in dimensional analysis to formulate mechanistically meaningful dimensionless numbers, which represented the contrasting as well as the collective influence of different hydro-climatic and biogeochemical drivers on stream water quality. A graphical exploration of the driver (predictor) numbers and stream DO number unraveled emergent patterns by collapsing observations from diverse environmental conditions into single dimensionless curves. Two environmental regimes ('climatic' versus 'metabolic' controls) were identified based on the critical threshold of the driver dimensionless number. The dimensionless numbers and emergent patterns led to the estimation of a scaling relationship, which was robust across the different environmental regimes. The scaling relationship was then utilized as a non-linear empirical model to successfully predict DO from numerous streams across the East Coast of U.S.A. (Nash-Sutcliffe Efficiency = 0.82). The research findings and insights are expected to guide effective water resources management for maintaining healthy stream water quality across the U.S. East Coast and similar regions around the world.