Different activities utilize water resources at different rates, rely on different water sources, and thus have a different “footprint” in regard to their water demand. However, the water footprint of these activities differs depending on the region and the time when the activity occurs. In popular literature, the footprint metric has seen some success in the ability to represent the large magnitude of human resource consumption, to more relatable per-capita quantities. However, these depictions are often based on global surveys that mask regional variability in water use. This thesis presents a framework for agricultural water footprint assessment, implemented using free and open source software. It explores the variability of the water footprint of select agricultural activities in the State of California from 2008 through 2015. The results describe a diverse landscape of water use variability driven primarily by crop choice. The study reveals that crop specific water footprints can have significant, inter-annual variability, independent of climatic conditions. Supported by water distribution infrastructure, agricultural activities can be found across the state. Water footprint assessments can be used by growers and resource managers, who wish to maximize utility per unit of water allocated, planners who wish to understand national risks and strengths, or the informed citizen who wishes to evaluate the sustainability of their consumptive activities.

In the present day, we are collecting more information about our built andnatural environments than ever before, to enable an unprecedented level of global prosperity. Additionally, we are relying on these insights to help us manage our natural and human systems so that we can maintain them sustainably for the foreseeable future. There still remains the motivation to push further in efficiently managing our built and natural systems, by achieving the gold-standard of intelligence --- that is, systems that are able to reproduce and/or exceed human-level ability in surveying and environmental modeling. Such an ambitious task will either require a lot of labor or the strategic use of intelligent autonomous systems.

For autonomous robots tasked with surveying the spatial and temporal dynamics ofa changing environment, there are a variety of (sometimes conflicting) criteria that must be satisfied over the course of a surveying mission: Where should the robot travel to? Which areas are worth observing? Does it make sense to revisit a previous area? How should the task be divided if there are multiple robots in the team? When should the mission stop? We can view this objective through an information-theoretic lens, where the robot is tasked with finding a path through a domain that maximizes the information that it collects along the way. In the robotics literature, this objective is known as /informative path planning/ (IPP), which is NP-hard.

This dissertation explores a few dimensions of the IPP problem and presents afew information-theoretic approaches of generating solutions that satisfy various surveying criteria. The methods used in this dissertation are grounded in geostatistics and explore different applications of a core motivation: from surveying a scalar field at fixed monitoring locations with a team of robots, to reconstructing a time-varying phenomena in a continuous planning space. By specifying an information-theoretic utility function, it is possible to reconcile prior knowledge about a system with new knowledge obtained by the robotic surveyors. The reader will gain insights into the nuances embodied the practical applications of this method in single-robot and multi-robot systems.

Finally, the reader will be presented with a collection of topics that serve aspoints of departure from this dissertation into separate lines of further inquiry. The methods explored in this dissertation can be applied to a variety of environmental monitoring tasks, enabling a host of analyses from national security objectives, to the analysis of trade flows, and other agricultural and ecosystem management objectives.