THE EFFECTS OF URBANIZATION ON THE MINGO CREEK WATERSHEDby Timothy P. Marshall (1984)
Urbanization of flood plains presents additional challenges to the hydrometeorologist in flood forecasting and warning. One of the problems associated with urbanization is increased runoff rates from paved surfaces, storm sewers, and river channelization. A recent flood on the Mingo Creek watershed in Tulsa, Oklahoma, provided an opportunity to study the flood flow problems in this rapidly growing area.
Mingo Creek begins in southeast Tulsa, Oklahoma and flows northward to Bird Creek. The topography changes from rolling hills in the south to a broad, gently sloping flood plain in the north. Prior to 1950, the watershed remained virtually undeveloped. Land use was primarily agricultural. Since then, Tulsa has expanded rapidly eastward. Without regard to flood control, large residential areas have been constructed on the flood plain of Mingo Creek. Some homes, apartment complexes, and mobile home parks lie adjacent to the creek's channel.
On May 27, 1984, the worst flash flood in recorded history occurred along the Mingo Creek watershed. Rainfall varied from 5 to 15 inches, with the heaviest amounts reported in the center of the watershed. Most of the rain had fallen within a four hour period. Five fatalities occurred on swollen Mingo Creek tributaries. Although the meteorology situation was ideal for flash flooding, certain urbanization factors appeared to have contributed to the increased flood levels.
A ground survey was conducted within a few days after the flood disaster. The purpose of the survey was to document flood elevations and obtain channel dimensions. Subsequent study revealed channel flow problems that included highway berms, under-designed culverts, and other channel constrictions which restricted flood flow. In several areas, flood water accumulated behind obstructions causing extensive flooding beyond the 100 year flood plain. Results of the ground survey are presented along with comments on watershed planning and management.
2. URBANIZATION HISTORY
Mingo Creek has been a problem watershed ever since the first settlers arrived in the mid 1800's. Most early floods inundated farm land and were not life threatening. Recent floods occurred in 1943, 1959, 1961, 1968, 1970, 1974, 1976, and 1984. Generally, flooding has resulted when rainfall exceeded 5 inches. Flooding has been associated with slow moving thunderstorms which happen primarily in spring and summer months. Although flooding has been prevalent on the watershed for a long time, flood control was not implemented until 1975.
Urbanization has grown rapidly since the mid 1950's, when the city of Tulsa encompassed about 10 percent of the Mingo Creek watershed area. By 1984, approximately 90 percent of the 61 square mile watershed was within the city limits of Tulsa. Figure 1 shows the increase in watershed urbanization from 1955 to 1982. Rapid urbanization occurred initially west of the Mingo Creek mainstream. Large residential subdivisions, highways, shopping centers, office and industrial buildings were constructed. A recent study by HTB, Inc., (1981) mentioned that the riparian region of Mingo Creek was so heavily populated there was little open land left for construction of flood detention ponds. Flood control costs have escalated in recent years, since proposals have included buying back developed land for detention pond construction.
3. MINGO CREEK FLOOD CONTROL HISTORY
At the time of the May 27, 1984 flash flood, several flood control improvement projects were under construction and several were already completed on the Mingo Creek watershed. Improvement projects included channel widening, channel straightening, and the addition of detention basins to pond flood waters. Design criteria considered full urbanization of the watershed with a 100 year return period for flooding.
The U.S. Army Corps of Engineers (Corps) was originally asked to complete an interim flood control study of Mingo Creek in May, 1966. Although a flood in June, 1968 caused about 150,000 dollars in property loss, the study was discontinued in 1972 when flood control costs exceeded benefits. However, a flood in June, 1974 prompted a restudy of flood control plans, as property losses approached 20 million dollars. Even though a higher cost-benefit ratio was determined, lack of funding precluded further study.
Recognizing the serious nature of flooding along Mingo Creek, the City of Tulsa undertook its own flood control project in 1975 without federal assistance. Subsequently, ordinances were passed by the city council in 1977 restricting future flood plain development. Prior to the 1984 flood, three miles of channel had been widened, and one detention pond had been constructed. Twenty-two homes were purchased and removed to clear land for a second detention pond.
Another devastating flood on Mingo Creek happened in 1976, causing 26 million dollars in property loss (Corps, 1977). The Corps was again authorized to continue their flood control plan for Mingo Creek. A final project report was completed in 1980. According to the Federal Emergency Management Agency (1984), the Corps' 93.2 million dollar plan entails the construction of 23 detention ponds and channelization of 7.5 miles of the Mingo Creek mainstream and tributaries. After extensive review by the Corps, the flood control plan was submitted to Congress for approval in April 1984. The next flood disaster on Mingo Creek occurred the following month.
4. SURVEY RESULTS
A field survey was conducted after the flood to investigate channel flow restrictions. The survey entailed measuring channel characteristics such as depth and width. In addition, bridge dimensions and the extent of vegetation cover were recorded. Results showed flood water ponded upstream of bermed highways, channel constrictions, and channel obstructions (i.e. encroaching vegetation).
4.1 Highways Berms
There were five highway berms which traversed the Mingo Creek flood plain. These highways were: the Broken Arrow Expressway (Rt. 51), Skelly Drive (I-44), Admiral Street, Crosstown Expressway (I-244), and Pine Street. In each case, flow was directed through designed culverts which restricted discharge. Once the creek topped its banks, flood waters were essentially dammed by highway berms, and water proceeded to spread out horizontally.
Figure 2 shows the effect of the Broken Arrow Expressway berm on ponding flood waters along Bell and Fulton tributaries. The topography on both sides of the highway was similar, with a gradual slope downward to the north. North of the highway, flood flow remained close to the l00-year design flood. However, south of the highway, water pooled and spread out, inundating residential and business areas. Basin transfer occurred between creeks. The resulting lake stretched more than a mile wide. One method to minimize the damming effect of flood waters in the flood plain would have been to construct highways on columns rather than berms as shown in Figure 3. This would alleviate the ponding problem and allow flood waters to continue downstream. Although initial costs are much higher to elevate highways on columns, it is believed property losses would have been greatly reduced.
4.2 Channel Constriction
There were several areas along Mingo Creek where the channel width narrowed, and culvert sizes were smaller than those upstream. Water elevations upstream of flow constrictions were increased. Across the width of some culverts there was a two foot head difference. Generally, as the channel length increases, bridge culvert size should increase also.
Bridge and culvert data from the Corps (1970), show several locations where culvert sizes along the mainstream of Mingo Creek were smaller than adjacent culverts upstream. Widths of bridge openings are plotted in Figure 4. Of the twenty bridges traversing Mingo Creek, at least five culverts had widths 15 to 50 percent smaller than the adjacent ones upstream. Some of this difference can be explained by the ages of the bridges. In general, it was found that bridges constructed recently tend to have larger culvert openings than older bridges.
4.3 Channel Vegetation
The sloping sides of the mainstream of Mingo Creek had been excavated, surfaced with grasses, and appeared to be well maintained. On the other hand, tributaries of Mingo Creek, predominantly in undeveloped areas, had trees and brush growing along the creek. Figures 5 and 6 allow comparison of the amount of channel vegetation along Bell Creek near the Broken Arrow Expressway. Where the highway crosses the creek, a concrete box culvert was constructed, the channel sides excavated, and the slope was cleared of dense vegetation. Only a few hundred yards upstream in an undeveloped area, the situation was much different. The creek was clogged with trees and undergrowth. After the May, 1984 flood, debris marks revealed that the water exceeded the bank height by at least three to four feet.
Obstructions in the flow most likely caused a decrease in water velocity and an increase in water height. Investigation of land use upstream of the undeveloped area revealed a large residential subdivison with a concrete lined creek channel. High water velocities in the developed area appeared to contribute to the back flow problem downstream.
Rapid urbanization of the Mingo Creek watershed has increased the backwater problem. In the flood in May, 1984 pooling of water occurred upstream of highway berms, under-designed bridge culverts, and areas with dense channel vegetation. Mitigation of ponding flood waters will involve elevating bridges on columns instead of berms, widening Older bridges which have small culverts, and minimizing channel vegetation on the creek mainstream as well as tributaries. Urbanization has outpaced flood control on Mingo Creek. Although Mingo Creek has had a long history of flooding, flood control began only within the last decade. Without federal support initially, the City of Tulsa has had to fund local flood control projects. Urban problems were recognized more than twenty years ago but, flood control planning and implementation was a long, involved process. It is hoped that the problems presented herein are considered in planning and construction for future flood control projects.
I would like to thank the U.S. Army Corps of Engineers, Tulsa Division and Mr. Randy Zipser of the Oklahoma State Water Resources Board for providing background information on Mingo Creek. Mr. Ben Barker of the Tulsa National Weather Service Office provided detailed meteorology information on the May, 1984, flood. Mr. Richard Madison of Haag Engineering Company offered helpful comments and suggestions in review of this material.
Federal Emergency Management Agency, 1984: Interagency Flood Hazard Mitigation Report, FEMA-709-DP. 18pp.
HTB Inc., 1981: Mingo Creek Master Drainage Plan for Tributaries between 1-44 and Broken Arrow Expressway. 144 pp.
U.S. Army Corps of Engineers, 1970: Flood Plain Information on Mingo Creek, Southwestern Division, Tulsa, Division. 32 pp.
U. S. Army Corps of Engineers, 1977: Report on Flood of 30 May 1976, Southwestern Division, Tulsa, District. 30pp.