[en] Early on March 14, 1989, a thermal plasma probe on the Defense Meteorological Satellite Program (DMSP) F9 spacecraft detected extensive and dramatic decreases in the ion density at 840 km, near 2130 LT, during two consecutive transequatorial passes over South America. The order of magnitude decreases in the ion density extended more than 4,000 km along the satellite track. The depletions were accompanied by upward and westward plasma drifts, both in excess of 100 m/s. Their onsets and terminations were marked by extremely sharp density gradients. A partial depletion was detected over the eastern Pacific during the following orbit. The DMSP F9 ground track passed slightly west of a Brazilian total electron content (TEC) station and two Brazilian ionosondes during the first depletion encounter. The TEC fell far below normal during the night of March 13-14. The ionosonde measurements indicate that, in the hour after sunset, before DMSP passed through the depletions, the F₂ layer rose rapidly and disappeared, but at the time of the first depletion encounter, h_mF₂ was decreasing over one of the stations. The authors develop a phenomenological model reconciling DMSP F8, F9 and ground-based measurements. The calculations show that rapid upward drifts sustained for several hours can produce depletions in the equatorial ion density with sharp gradients at their high-latitude boundaries, consistent with the data. They discuss possible contributing mechanisms for generating these upward drifts. These include direct penetration of the magnetospheric electric field to low latitudes, the electric fields generated by the disturbance dynamo, and the effects of conductivity gradients near the dusk terminator and the South Atlantic anomaly