The stable isotopic composition of water (δ¹⁸O) preserved in terrestrial sediments has been used to reconstruct surface elevations. The method is based on the observed decrease in δ¹⁸O with elevation, attributed to rainout during air mass ascent. We use a climate model to test the δ¹⁸O–elevation relationship during Tibetan–Himalayan uplift. We show that δ¹⁸O is a poor indicator of past elevation over most of the region, as processes other than rainout are important when elevations are lower.