Turbulence Induced Errors On Surface Temperature Measurements Over Heterogeneous Landscapes

Surface temperature (Ts) is an important parameter for the investigation of the surface energy balance. Kilometre resolution satellites are already in service (MODIS, VIRSS) and are extensively used. At the same time studies are underway to develop a new thermal infrared (TIR) mission combining high spatial resolution and high revisit capacities (projects Mistigri at CNES, Hyspiri and ECOSTRESS at NASA, THIRSTY CNES-NASA).  In this context, recent reports (Lagouarde et al., 2015) indicate that the precision of surface temperature measurements at high resolution (50m) may be limited by surface atmospheric turbulence induced temporal fluctuations. As for kilometric resolution such errors -which can originate in planetary boundary layer turbulence- have not been explored yet.  In this context an experiment was performed using a TIR camera looking over a cliff face towards i) a typical agricultural landscape with small cultured crop parcels ii) a peri-urban landscape in the east of France near Grenoble. Data were recorded at 12.5Hz on three sunny cloudless days over 20 minute periods between 10 and 14h TU at different times in the year during 2015. The objective of this experiment is to investigate the temporal and spatial variability of Ts at different image resolutions, to define the error related to the instant nature of a satellite image. as this error has both consequences for the definition of specifications of future satellites and defines the limits for which existing and future image data can be used to study surface energy balance and evaluate the error on derived products. This paper will present the experiment in detail as it represents a unique dataset of high resolution Ts measurements and initial results indicating the error associated with IRT satellite measurements at different resolutions.