Rhea presents an interesting target for the identification of cold traps due to its tenuous, seasonally varying exosphere comprised mainly of CO2 and O2 (Teolis et al., 2010, Teolis and Waite, 2016) and its very cold winter poles, revealed by thermal infrared spectra obtained by Cassini’s Composite InfraRed Spectrometer (CIRS) (Howett et al., 2016). Focal plane 1 (FP1) is the only CIRS detector that is able to detect temperatures seen on Rhea’s winter poles (≪60 K), but it has the lowest spatial resolution (3.9 mrad) of CIRS’ three detectors (Flasar et al., 2004). Thus, FP1 observations often cover multiple local times and terrains. Thermal models, which are able to predict the emission observed by FP1, can typically be categorized as ‘flat’ surface models that emulate bulk surface properties of large surface areas and are widely applicable, or highly sophisticated models that require very accurate surface topography, scattering and other physical processes that impact the surface thermal properties. We present results from an intermediate ‘rough’ surface model that derives topographic information from a digital elevation map (DEM) and uses slope and azimuth to improve the representation of surface insolation and is coupled to a 1-dimensional surface thermal model. While this model is tested on Rhea it is also applicable to similar airless bodies. We illustrate its impact compared to a flat surface model using a case study of 3 sets of CIRS polar observations of Rhea, in Southern Hemisphere winter. These were measured on 22nd December 2012, 9th March 2013 and 10th February 2015. We show that our thermal model agrees well with temperatures derived from CIRS observations of very cold (<40 K) polar scenes on Rhea. We demonstrate how using a realistic rough model that can represent topographic features can be used to probe anisothermality within scenes on sub-CIRS field-of-view (FoV) scales. We find temperature contrasts of up to +9 and -4 K within some CIRS FoV when compared to a ‘flat’ model, depending upon illumination conditions. Focusing on the coldest parts of the scenes, we see that for large areas annual temperatures don’t rise above 80 K all year. This is important because it would imply a number of volatiles should be quasi-permanently trapped there.