Radiation transfer from turbulent nonpremixed jet flames and plumes is important in many applications such as energy-efficient combustion systems, temperature sensitive pollutant control, and detection, control, and suppression of accidental fires. Combined spatial and temporal correlations of scalar values such as temperature and species concentrations affect the emitted radiation intensity. Spatiotemporal correlations and radiation intensity measurements downstream of the reacting parts of flames (plumes) have received limited attention. Motivated by this, planar time-dependent narrowband radiation intensity measurements are acquired of a turbulent nonpremixed flame and its plume using an infrared camera. Temporally and spatially correlated instantaneous realizations of local scalars and path integrated intensity values are calculated using a stochastic time and space series analysis, a narrowband radiation model, and the radiative transfer equation. The time-dependent infrared images reveal intermittent, low intensity regions in the plume characteristic of buoyancy-dominated transport. High radiation intensity structures are observed in the flame characteristic of momentum dominated flow and vorticity driven mixing. Normalized intensity fluctuations are nearly constant in the flame region, but increase by up to a factor of three in the plume. Normalized temporal correlations, power spectral density functions, and spatial correlations of the intensity are independent of the spatial location throughout both the flame and the plume. Spatial correlations of the radiation intensity exhibit approximately linear decay to half an integral length scale followed by an exponential decrement. The radiation intensity fluctuations remain spatially correlated up to separation distances two times larger than the integral length scale. Space–time cross correlations of the intensity fluctuations are measured for the first time and are shown to be more isotropic in comparison to the product of the spatial and temporal correlations. This suggests that a correction factor should be applied to the space–time correlation model in future stochastic calculations to account for the anisotropy. The infrared imaging technique, illustrated in this paper, is promising to be a useful qualitative and quantitative nonintrusive technique for studying both reacting and nonreacting flows.