X hits on this document

PDF document

Parameter Sensitivity in Hydrologic Modeling - page 102 / 163





102 / 163

the standard deviation of each parameter divided by its mean, is used in addition to the standard deviation to evaluate each of the watershed parameters. Assuming that errors are normally distributed, in 68% of the cases the observed variation will fall within one standard deviation above or below the mean, or ± CV expressed as a percent of the mean.

The standard deviation for area and slope is small for all three areas. For both Area 1 and Area 2, the drainage area can be determined within approximately ± 0.01 square miles (2.07% for Area 1 and 0.81% for Area 2), and ± 0.13 square miles (1.52%) for Area 3. For Area 3, using traditional delineation techniques on paper maps produces an area smaller than any of the other techniques (0.29 square miles below the mean). This is most likely attributed to area measurement using the planimeter.

Similar to area, the standard deviation (as well as the CV) for slope is small for all three areas. The standard deviation for slope is highest for Area 1, at a value of 0.05%. The highest coefficient of variation for slope, as calculated for Area 1, is 3.65%.

Longest flow path and perimeter show different results from area and slope, in that there is much more parameter variation among extraction methods. Longest flow path was determined within ± 0.13 miles for Area 1, ± 0.12 miles for Area 2, and ± 0.22 miles for Area 3. Precision in determining the longest flow path increases as the area becomes larger, with a CV for Area 1 of 8.42% and a CV for Area 3 of 3.22%. This is most likely because the flow path for Area 1 is small, and hence there is more error in this measurement. Secondly, for Area 3 a


Document info
Document views507
Page views507
Page last viewedTue Jan 17 11:20:12 UTC 2017