|
<br />increase in flood peaks. Urbanization does not always
<br />increase floods, however. Some aspects of urbanization
<br />can decrease an area's flood potential. For instance, if
<br />the lower part of a basin is urbanized and the upper part
<br />left in its natural condition, rapid removal of floodwaters
<br />from the lower part may occur before the upper part can
<br />contribute significant runoff. Some cities reduce flooding
<br />by storing the water in designated areas (detention ponds)
<br />and releasing it slowly. As discussed above (Malcolm,
<br />1980), culverts, bridges, storm sewers, and roadway
<br />embankments may inhibit flooding and cause temporary
<br />storage behind them, thus reducing peak-flow rates.
<br />Obviously, evaluating the effects of urbanization on
<br />flood potential involves many factors. The data accum-
<br />ulated for this study show that for some basins the urban
<br />flood-frequency curve is below an equivalent rural curve.
<br />Also, there are several instances in which the two flood-
<br />frequency curves cross, with low-order floods increased
<br />by urbanization and high-order floods decreased.
<br />
<br />DATA BASE
<br />
<br />The second phase of this study was the compilation
<br />of a comprehensive data base for drainage basins af-
<br />fected by urbanization. Contact with district offices of
<br />the Geological Survey revealed that at least 3 years of
<br />runoff data from almost 600 urbanized sites were avail-
<br />able nationwide. Data collected by- other agencies were
<br />also sought, but these data did not meet the following
<br />selection criteria established for the study:
<br />I. A watershed selected must have at least 15 percent of
<br />the drainage area covered with commercial, indus-
<br />trial, or residential development.
<br />2. Reliable flood-frequency data must be available for
<br />the watershed. These could be based on actual peak
<br />flow records if records were available for 10 or more
<br />years, or from synthesized data if such data were
<br />based on a rainfall-runoff model specifically cali-
<br />brated from actual flood and rainfall data for that
<br />basin.
<br />3. The period of actual flood data, or the period of cali-
<br />bration for synthesized data, must have been one of
<br />relatively constant urbanization. This was the most
<br />difficult criterion to meet, and in some cases only
<br />part of a long record could be used. As a general
<br />guideline, "relatively constant urbanization" was
<br />defined as a change in development of less than
<br />50 percent during the period of record. If a basin was
<br />30 percent urbanized at the beginning of the period
<br />of record, it could be no more than 45 percent urban-
<br />ized at the end of the period.
<br />An appraisal of all available sites resulted in a
<br />final list of 269 sites that met the selection criteria. These
<br />sites represent a broad spectrum of watershed conditions
<br />
<br />and metropolitan areas, ranging from the East Coast to
<br />tbe West Coast and Hawaii. A few States, such as Illi-
<br />nois, Texas, and Missouri. have had extensive urban
<br />data-collection programs, as reflected by the large num-
<br />ber of sites for which records are available in those
<br />States. Many other States, however, also are well repre-
<br />sented. Gaging sites are included for 31 States and 56
<br />cities or metropolitan areas. Table I lists cities or metro-
<br />politan areas and the number of sites used in this study.
<br />Table I also includes a city skew value and the source of
<br />equivalent rural discharges, which will be discussed
<br />later. Figure I illustrates the geographical distribution
<br />of sites.
<br />The data compiled for each urban site includes a
<br />comprehensive list of topographic and climatic variables,
<br />land-use variables, indices of urbanization, and f1ood-
<br />frequency estimates. The main sources of information
<br />were as follows:
<br />I. Department of the Interior, U.S. Geological Survey,
<br />Water Resources Division, District Offices
<br />a. Peak-discharge data
<br />b. Basin characteristics
<br />c. Indices of urbanization
<br />2. Department of the Interior, U.S. Geological Survey,
<br />Topographic Division
<br />a. Topographic maps
<br />b. Land-use maps
<br />3. Department of Agriculture, Soil Conservation Service
<br />a. Land-use data
<br />b. Soils data
<br />c. Basin characteristics
<br />4'. Department of Commerce, Bureau of the Census
<br />a. Population data, 1970 census reports
<br />A complete listing of the data base cannot be
<br />included in this report because of its size. The complete
<br />data base is stored on the Geological Survey computer
<br />in a "Statistical Analysis System" (SAS) data set (SAS
<br />Institute, Inc., 1979), to whicb access can be obtained
<br />from the Chief, Data Management Section, U.S. Geo-
<br />logical Survey, Mail Stop 437, National Center, Reston,
<br />Va. 22092. A brief description of all variables, as well as
<br />a detailed description of tbe significant variables, is pro-
<br />vided in the following paragraphs and the glossary.
<br />Appendix I contains a listing of selected data for all gag-
<br />ing stations used in this study. Data descriptions are
<br />subdivided into four groups: (I) topographic and cli-
<br />matic variables, (2) land-use variables, (3) indices of
<br />urbanization, and (4) flood-frequency estimates. Some
<br />parameters could justifiably fit in more than one of
<br />these groups but were assigned to only one group for
<br />convenience. Not all data items are available for aU gag-
<br />ing sites, mostly because base maps were not universally
<br />available.
<br />Most of the basin parameters, or variables, were
<br />compiled for the entire drainage basin and represent a
<br />
<br />Data Base 3
<br />
|