Abstract. We compare geocoded heights derived from the interferometric mode (SARIn) of CryoSat to surface heights from calibration–validation sites on Devon Ice Cap and western Greenland. Comparisons are included for both the heights derived from the first return (the point-of-closest-approach or POCA) and heights derived from delayed waveform returns (swath processing). While swath-processed heights are normally less precise than edited POCA heights, e.g. standard deviations of ∼ 3 and ∼ 1.5 m respectively for the western Greenland site, the increased coverage possible with swath data complements the POCA data and provides useful information for both system calibration and improving digital elevation models (DEMs). We show that the pre-launch interferometric baseline coupled with an additional roll correction ( ∼ 0.0075° ± 0.0025°), or equivalent phase correction ( ∼ 0.0435 ± 0.0145 radians), provides an improved calibration of the interferometric SARIn mode. We extend the potential use of SARIn data by showing the influence of surface conditions, especially melt, on the return waveforms and that it is possible to detect and measure the height of summer supraglacial lakes in western Greenland. A supraglacial lake can provide a strong radar target in the waveform, stronger than the initial POCA return, if viewed at near-normal incidence. This provides an ideal situation for swath processing and we demonstrate a height precision of ∼ 0.5 m for two lake sites, one in the accumulation zone and one in the ablation zone, which were measured every year from 2010 or 2011 to 2016. Each year the lake in the ablation zone was viewed in June by ascending passes and then 5.5 days later by descending passes, which allows an approximate estimate of the filling rate. The results suggest that CryoSat waveform data and measurements of supraglacial lake height change could complement the use of optical satellite imagery and be helpful as proxy indicators for surface melt around Greenland.