Annual freeze-thaw cycles are one of the major problems that impact the performance of granular-surfaced roadways in Iowa. They can cause severe damage to granular roads, especially in areas where the timing of heavy agricultural traffic coincides with that of spring thawing. This is a critical problem for Iowa since granular roads play a significant part of the road networks that enable the transport of goods to rural areas. There are several ways to control the damage caused by these annual freeze-thaw cycles. One of these methods is the seasonal load restrictions (SLR), widely used by agencies in cold regions. Seasonal load restrictions are applied during the thawing period to control excessive road damage caused by freeze-thaw cycles. The application periods are based on local regulations or according to subgrade conditions. The timing of such restrictions is generally determined based on engineering experience, field testing, subsurface instrumentation, and modeling.Subsurface monitoring is a reliable method for agencies to determine the status and freeze-thaw behavior of the subgrade in real-time. It can also help local road agencies plan better for annual budgets and frost embargos. In order to obtain accurate field data, an appropriate sensor network and data acquisition system must be carefully planned and installed. A comprehensive subgrade sensors network was installed below a granular-surfaced roadway in Hamilton County, Iowa, to capture spatial and temporal variations across a roadway. The network continuously measures temperature, volumetric water content and matric potential to a depth of 213 cm below the aggregate-subgrade interface, and atmospheric conditions by a weather station installed along with the subsurface sensors. Development and installation procedures for the system and lessons learned throughout the process were documented. Several significant issues were highlighted regarding selecting the sensor and data acquisition system, laboratory and field checks, borehole sensor installation tools, and post-installation troubleshooting. Laboratory and field trials were performed to ensure a successful installation. Data obtained from this sensor network for a freeze-thaw period in the winter of 2019-2020 were analyzed using contour mapping. The results were presented for the road cross-section and showed some unsymmetrical behavior under the road surface. The freezing front was found to be variable throughout the road. It was deeper towards the west midpoint and shallower at the shoulders. Volumetric water content and matric potential measurements were also found to be consistent with the temperature data during freezing and thawing periods. Monitoring data were also used alongside numerical modeling to develop reliable freeze-thaw predictions for the soil profile beneath a granular-surfaced roadway. The Simultaneous Heat and Water (SHAW) model was used to simulate temperature, volumetric water content and matric potential with some adjustments for granular road conditions. Simulation results were compared to the measured in-situ measurements by temporal and spatial analysis, and the differences were quantified with statistical analysis. The road center exhibited significantly lower differences between measured and simulated temperature than the shoulder, whereas the center and the shoulder had similar differences for volumetric water content. In addition, the SHAW model provided more accurate temperature and water content results at the road center than at the shoulder in general due in part to factors not considered in the simulations, such as snow cover, nearby vegetation, or the effect of the drainage ditch. In conclusion, this study revealed the unsymmetrical freeze-thaw behavior from the evaluations of in-situ measurement and computational modeling, which helps develop more accurate and reliable freeze-thaw prediction models to be used by the local transportation agencies with more efficient SLR organization and resource planning.