Chemical and biological composition of surface materials and physical structure and arrangement of those materials determine the intrinsic spectral reflectance of Earth’s land surface at the plot scale. As measured by a spaceborne or airborne sensor, the apparent reflectance depends on the intrinsic reflectance, the surface texture, the contribution and attenuation by the atmosphere, and the topography. Compensation or correction for the topographic effect requires information in digital elevation models (DEMs). Available DEMs with global coverage at ~30 m spatial resolution are derived from interferometric radar and stereo-photogrammetry. Locally or regionally, airborne lidar altimetry, airborne interferometric radar, or stereo-photogrammetry from airborne or fine-resolution satellite imagery produces DEMs with finer spatial resolutions. Characterization of the quality of DEMs typically expresses the root-mean-square (RMS) error of the elevation, but the accuracy of remote sensing retrievals is acutely sensitive to uncertainties in the topographic properties that affect the illumination geometry. The essential variables are the cosine of the local illumination angle and the shadows cast by neighboring terrain. We show that calculations with globally available DEMs underrepresent shadows and consistently underestimate the values of the cosine of illumination angle; the RMS error increases with solar zenith angle and in more rugged terrain. Analyzing imagery of Earth’s mountains from current and future missions requires addressing the uncertainty introduced by errors in DEMs on algorithms that estimate surface properties from retrievals of the apparent spectral reflectance. Intriguing potential improvements lie in novel methods to gain information about topography from the imagery itself.