Field studies in a hot, humid climate were conducted to investigate the thermal and hygrothermal performance of ventilatedattics and nonventilated semi-conditioned attics sealed with open-cell and with closed-cell spray polyurethane foam insulation.In the ventilated attics the relative humidity drops as the attic air warms; however, the opposite was observed in the sealed attics.Peaks in measured relative humidity in excess of 80%–90% and occasionally near saturation (i.e., 100%) were observed fromroughly solar noon till about 8 p.m. on hot, humid days. The conditioned space of the test facility is heated and cooled by an airto-air heat pump. The space was not occupied and had no interior moisture load due to occupancy. Moisture pin measurementsmade in the sheathing and absolute humidity sensor data from inside the foam and from the attic air show that moisture is beingstored in the foam and the roof sheathing. The moisture in the foam diffuses to and from the sheathing dependent on the pressuregradient at the foam-sheathing interface, which is driven by the irradiance and night-sky radiation. Ventilated attics in the samehot, humid climate showed less moisture movement in the sheathing than those sealed with either open- or closed-cell spray foam.The temperature measured on the underside of the sheathing was 10°C cooler (north-facing roof deck) and 20°C cooler (southfacingroof deck) for the ventilated attic as compared to the sheathing temperatures in the sealed attics. Foam was physicallyremoved and moisture was observed around the rafter and sheathing interface. Observations, sheathing temperature, and partialpressure measurements suggest two-dimensional heat and moisture flow that is cyclic moving in and out of the depth of the roofand that also moves along the plane of the sheathing towards the rafters. The use of permeable spray foam in a hot, humid climateinadvertently allows moisture to be held against the roof deck. The moisture transfers back to the attic air as solar irradiancebears down on the roof.