Block averaging can hide the inlet pulses that determine surface ponding in bioretention systems. We asked what information must accompany the averages to keep the storage requirement within known limits. We derived bounds for a constant-outflow store and evaluated them in 776 new hydraulic model runs, including controlled pulse experiments; 22 monitored six-minute inlet records provided the main forcing. With an outflow equivalent to 30 mm/h over 1,800 m², four-hour averaging underestimated required storage by a median 89.2% among 13 records requiring more than 1 m³. Retaining two values—the largest positive and negative cumulative departures from the block averages—reduced the median conservative storage estimate from 6.50 to 1.75 times the six-minute requirement. These values bound averaging error, but they cannot determine storage: in a controlled layered-model experiment, two inflows with identical block volumes, peaks and reported extrema produced maximum surface depths of 130.25 and 82.74 mm. Maximum inflow volumes over nine selected durations provided a lower bound that complemented the upper error bound. Together, the bounds enclosed required storage for every tested record under constant outflow. With one-hour averages at 30 mm/h, the median ratio of the upper to lower bound was 1.29. The resulting interval identifies when summarized inflow supports sizing and when the available information leaves the requirement unresolved.