Interferogram images derived from repeat-pass spaceborne synthetic aperture radar systems exhibit artifacts due to the time and space variations of atmospheric water vapor. Other tropospheric variations, such as pressure and temperature, also induce distortions, but the effects are smaller in magnitude and more evenly distributed throughout the interferogram than the wet troposphere term. Spatial and temporal changes of 20% in relative humidity lead to 10 cm errors in deformation products, and perhaps 100 m of error in derived topographic maps for those pass pairs with unfavorable baseline geometries. In wet regions such as Hawaii, these are by far the dominant errors in the Spaceborne Imaging Radar-C and X Band Synthetic Aperature Radar (SIR-C/X-SAR) interferometric products. The unknown time delay from tropospheric distortion is independent of frequency, and thus multiwavelength measurements, such as those commonly used to correct radar altimeter and Global Positioning System (GPS) ionospheric biases, cannot be used to rectify the error. In the topographic case, the errors may be mitigated by choosing interferometric pairs with relatively long baselines, as the error amplitude is inversely proportional to the perpendicular component of the interferometer baseline. For the SIR-C/X-SAR Hawaii data we found that the best (longest) baseline pair produced a map supporting 100 m contouring, whereas the poorest baseline choice yielded an extremely noisy topographic map even at this coarse contour interval. In the case of deformation map errors the result is either independent of baseline parameters or else very nearly so. Here the only solution is averaging of independent interferograms, so in order to create accurate deformation products in wet regions many multiple passes may be required. Rules for designing optimal data acquisition and processing sequence for interferometric analyses in nondesert parts of the world are (1) to use the longest radar wavelengths possible, within ionospheric scintillation and Faraday rotation limits, (2) for topography, maximize interferometer baseline within decorrelation limits, and (3) for surface deformation, use multiple observations and average the derived products. Following the above recipe yields accuracies of 10 m for digital elevation models and 1 cm for deformation maps even in very wet regions, such as Hawaii.¿ 1997 American Geophysical Union