Forty-day photochemical model runs are compared with ground-based stratospheric ClO observations taken during the austral spring of 1993. Our purpose is to explore the range of required heterogeneous processing within which we can reproduce the duration and degree of chlorine activation within the Antarctic spring vortex. Heterogeneous processing on nitric acid trihydrate (NAT) polar stratospheric clouds (PSCs) or supercooled ternary solution (STS)-type particles is shown to be necessary to maintain chlorine in active forms during ozone hole formation in September, even for small HNO3 amounts, or chlorine deactivates sooner than observed. The lower limits for the surface areas required are quite small, however. Thus the record ozone losses observed during September of 1993 may be attributed to catalytic loss due to chlorine maintained in active forms by heterogeneous processing despite the sparse particle loading of the Antarctic lower stratosphere at that time. The ozone loss rates predicted by the model during the formation of the springtime Antarctic ozone hole indeed agree quite well with observations. The one-dimensional model is also able to reproduce both the observed timing and rate for subsequent deactivation of chlorine. Renitrification from PSC evaporation is not required for this deactivation, as HCl reformation is very rapid at low ozone values.¿ 1997 American Geophysical Union