elasticity of 0.99) when lag times, based on curve number variations, are entered into a hydrologic model.

The discharge elasticity with respect to drainage area, calculated to be 1.06, is also elastic. Although discharge is more sensitive to drainage area than to curve number, errors associated with determining area are small. For instance, Area 3 has a CV for area of 1.52%. A 1% increase in area will result in a 1.61% increase in discharge. If curve number has a CV of 3%, the resulting discharge for a 1% increase in curve number will be 3%.

6.4

ADDITIONAL SOURCES OF ERROR

This analysis of parameter sensitivity in hydrologic modeling is general and meant to serve as a guide for TxDOT engineers. Several important elements were not considered in this report when calculating discharge.

First of all, the elasticity analysis does not account for curve number effects on excess rainfall. For each change in lag time within the HEC-1 model, curve numbers used to calculate excess rainfall were held constant.

In highly developed areas, parameter extraction methods used in this report may not apply. Without a thorough knowledge of the storm sewers and other urban structures it is difficult to judge flow paths. Including buildings and structures into the model does not account for the runoff flowing below the ground surface or under a bridge. Storm sewers may enter and leave watersheds, and water may flow along roads and enter into one watershed from another area.

134