Soil moisture information is critical to the understanding of the global carbon cycle and its associated uncertainties. Over the last few years, availability of soil moisture data has grown considerably; for example, the NASA Oak Ridge National Laboratory Distributed Active Archive Center (ORNL DAAC, http://daac.ornl.gov) currently distributes about 0.7 TB (8085 files) of soil moisture products. This increase in data availability can be attributed mainly to the development of new soil moisture remote-sensing capabilities (e.g. AirMOSS), wider installation of in-situ soil moisture monitoring networks (e.g. SoilSCAPE, COSMOS), improved data assimilation (e.g. GRACE-DA-DM), and launch of the Soil Moisture Active-Passive (SMAP) satellite. One key challenge that may limit effective utilization of these data is their heterogeneity in terms of (1) spatial footprints (points vs. grid representation), (2) frequency (hourly to daily), (3) sub-surface measurement depth, and (4) method of measurements (in-situ vs remote-sensing). An integrated visualization and data distribution platform was developed for soil moisture datasets available for North America based on open source software libraries. Bringing these disparate datasets from multiple sources into a single system not only adds value to the existing data but also facilitates exploratory analysis and data discovery among different groups of stakeholders. This paper describes the challenges and lessons-learned in creating this data harmonization architecture, including three case studies that leverage the visualization system for real-world science applications.