Drought is a complex phenomenon with varying degrees of impacts and monitoring methods. No drought is alike, creating a challenge for both water managers and communities. No area is immune to a drought. Due to the cyclical nature of drought events, clear information to those impacted is necessary to reduce risk and move towards proactive responses, as opposed to reactive responses. To better provide communication and mitigation tools, Drought Early Warning Information Systems (DEWIS) have been developed in various regions and contexts. To improve early warning, an understanding of the end user’s perceptions of risk, and the applicability of data and methods is valuable. This dissertation discusses findings from three related studies exploring the use of multiple methods and streams of data for drought risk in the United States. Chapter 1 provides a brief overview and outline of this dissertation. Chapter 2 presents findings from a study looking at the perceptions and usage of the term ‘flash drought’ among water managers in the United States. Chapter 3 is focused on modelling methods of using crop insurance occurrence data as a proxy of drought impacts for specialty crops. Chapter 4 is a systematic literature review of the use of mixed methods in drought science. Chapter 5 is the final study that investigates the benefits and validation of mixed methods in understanding drought risk to dry beans in North Dakota and Minnesota. Finally, in Chapter 6, I provide a brief conclusion of the findings and areas for future research. This dissertation demonstrates the use of various data types in understanding drought risk for improving drought early warning systems.