Digital water level recorders on small, flashy streams are frequently set at 5‐minute readout intervals to define fluctuations during critical storm flow periods. Consequently, during nonstorm periods far more data points are punched on the recorder tapes than are needed to define the long‐term hydrograph. A two‐step reduction process eliminates more than 95 of the original 105,000 data points per year of record. The first reduction is obtained by translating to IBM cards only every twelfth (hourly) head value punched on the tape during nonstorm periods. Storm periods are translated at 5‐minute intervals. A computer reduction step on both hourly and 5‐minute data then systematically rejects head values not essential to hydrograph definition. Rejection is based on differences in elevation and slope between successive head values. Computer output consists of cards and a listing of the red
展开▼
