Multivariate linear modeling for the application in the field of hydrological engineering

The representation and measurement of change is both an interest and a fundamental concern to hydrology. Traditional approaches to studying change have used multiple regression. Also, linear regressions are designed to measure a specific type of relationship between variables, which take a linear form. However, the variability of hydrological processes is increasing, and consequently, events with values and trends far from the historical mean behavior are more frequent. In order to face this new hydrological reality and with the aim of anticipating and forecasting it, new approaches, methodologies, and techniques have emerged that incorporate this changing behavior. This chapter aims to provide a conceptual overview of univariate and multivariate linear model statistical techniques for hydrological data analysis. In this sense, some of the most powerful modeling approaches are examined, and issues related to data structures and model building are discussed. Developed statistical methods such as the H_C-MLM can be embedded as a tool/module for a wider Decision Support System aimed to reach a sustainable and efficient integrated management of water systems.