Analysis of Stormwater Infiltration Ponds on the North Carolina Outer Banks (Report No. 254) September 1990

The State of North Carolina adopted the current Stormwater Runoff Disposal Rules (15A NCAC 2H.1000) in 1988 requiring stormwater management plans for new development in 20 coastal counties. Stormwater infiltration pond systems are approved by the State as an option for retaining storm water on the developed site; however, the long-term performance of these systems has not been measured or determined. Likewise, computer models or other predictive techniques to describe the performance of stormwater infiltration ponds over long periods of time have not been developed and tested. Such models and techniques would provide valuable assistance in evaluating and designing these infiltration systems.

The overall objective of this project was to develop a model that continuously simulates the performance of stormwater infiltration ponds on the North Carolina barrier islands. The hydrology of two operating infiltration ponds systems was studied in an 18-month field study. A three-dimensional numerical simulation model for combined unsaturated and saturated groundwater flow was developed and used to analyze pond seepage under geometries characteristic Of the North Carolina barrier islands. The numerical solutions were used to modify and test approximate methods for quantifying pond seepage. The approximate methods were incorporated into a reservoir model for monitoring pond stage and overflow. The reservoir model was included in a hydrologic simulation to predict water table elevation, runoff, and pond performance over a long period of time. The reliability of the model was tested using data collected from the field experiments.

Two stormwater infiltration ponds in operation at Surf City, N.C. and Bald Head Island, N.C. were used for field study. The water table elevations at selected locations along the transect of the island were monitored continuously. Water table elevations at additional locations were monitored on a biweekly basis. The water surface elevations in the infiltration ponds were also monitored on a continuous basis as was the rainfall at the site. Soil hydraulic conductivities and soil water characteristic relationships were determined at both field sites. The subsurface geology was described at the Surf City site and an aquifer pump test was performed to determine aquifer transmissivity and specific yield.

The hydrology of the infiltration ponds at the two research sites was very different. The runoff water stood in the pond at Bald Head island for much shorter periods than at Surf City. This was due to greater pond drawdown rates, less impervious area in the watershed, and a more shallow pond depth at Bald Head Island. The higher pond drawdown rate at Bald Head Island resulted from the shorter distance between the pond and the river and the greater elevation of the pond bottom above the water table and mean sea level.

Both of the infiltration ponds in the field studies effectively served their primary purpose of retaining on site the stormwater runoff from the first 38 mm (2.5 in) of rainfall. During the field study, there was no evidence that any stormwater runoff from either developed site flowed overland to the sounds. In nearly every case, the pond seepage rate was sufficient to completely draw down the pond in 5 days. The single case during the 18-month study when drawdown was greater than 5 days occurred during a period of high tides. Even with sufficient drawdown rates, water did stand in the pond for as much as 18 consecutive days due to successive rainfall events. No adverse effects to property, environment, or pond performance were observed during these periods.

A numerical solution to the Richards equation for combined saturated and unsaturated flow in three dimensions was developed to determine seepage rates from infiltration ponds. The solution used the Gauss-Seidel finite difference method with successive over-relaxation (GS-SOR). To increase the rate of convergence, the numerical solution was modified for the multigrid method (MG3D). Details of the multigrid method were presented for a two-level description of linear and non-linear systems of equations. Solutions for three sample problems were obtained with the GS-SOR and MG3D algorithms. The rnultigrid solutions resulted in a reduction of total CPU time ranging from 71% to 82% over GS-SOR.

The three-dimensional numerical simulation model for combined unsaturated and saturated groundwater flow was used to analyze pond seepage under geometries characteristic of the North Carolina barrier islands. Ponds of constant size spaced from 10 to 190 m apart were simulated using varying distances from the ponds to the sinks and to the restricting layer. For the geometries and soil properties considered, pond seepage was dominated by flow in the saturated zone. The spacing between adjacent ponds affected seepage rates, having a greater effect for large island widths.

The numerical solutions were used to develop an approximate analytic solution for calculating three-dimensional pond seepage rates. The solution is for ponds located at the island midpoint and is based on radial flow theory in the vicinity of the pond and an equation for one-dimensional lateral flow in the remainder of the flow domain. For deep profiles, a method is presented for approximating the increase in the length of the flow path due to flow out of the bottom of the pond. The approximate solution predicted nearly the same pond seepage rates as determined from the numerical solutions for the geometries considered.

A computer model, DMPOND, was developed to simulate the performance of stormwater infiltration systems over long periods of climatological record. DMPOND was composed of three components: a seepage component using the approximate method for calculating three dimensional pond seepage, a hydrologic component for calculating stormwater runoff and water table elevation in the pond watershed, and a reservoir component for predicting pond stage, storage, and overflow. DMPOND also evaluated pond performance by predicting the number of pond overflow events and drawdown delays (water standing in the pond more than 5 days) occurring during the simulation period.

DMPOND was used to simulate the hydrology of the Surf City infiltration pond system for 1988. DMPOND predicted values for pond stage and drawdown time that compared well to measured values except during an extended period of unusually high tides.

DMPOND simulations were conducted to evaluate the effects of the following design variables on pond performance: island width, island length, pond bottom elevation, pond length to width ratio, pond volume, distance from pond to sink, and impervious area S,The performance of a pond was evaluated by serveral parameters: the number of drawdown delays, the maximum time period that water stands in the pond, the number of overflow events, and the percent to total runoff that is retained on site. The simulations indicated that higher values for for pond length to width ratio, island length, and elevation to pond bottom produced better pond performance. Lower values for island widths distance between pond and sink, and impervious area also improved pond performance. Increasing pond volume by increasing the surface area reduced pond over but increased the time that water is in the pond.

The sensitivities of the pond performance parameters to the design variables were different for the set of conditions used in this study. The number of drawdown delays was most sensitive to pond bottom elevation and pond length to width ratio. The number of pond overflow events was most sensitive to volume and impervious area. Only three variables (elevation of pond bottom, island width, and impervious area) significantly affected the maximum time period that water was in the pond. Only two variables (pond volume and impervious area) significantly affected the percent of total runoff retained and the total and maximum overflow volumes. The sensitivities reported here may not apply for all conditions encountered on the barrier islands; therefore, infiltration pond systems should be designed using design variables encountered on the specific site.

Recommendations

Properly designed stormwater infiltration ponds on the North Carolina barrier islands will effectively retain on site the stormwater runoff from the first 38 mm (1.59) of rainfall. Both infiltration ponds analyzed in this field study and most of the hypothetical ponds simulated for a 30-year period using a long-term simulation model performed within the limits of North Carolina regulations. We recommend the use of stormwater infiltration ponds for reducing the volume of runoff from developed areas to coastal waters. This recommendation is based on the hydrology of the pond system; however, the effects of infiltration ponds on the water quality of the groundwater has not been considered and should be the topic of future research.

Stormwater infiltration ponds must be properly designed in order to perform effectively. The island width (distance from ocean to sound) and island length (perpendicular to width ie. distance between property boundaries or hydraulic sources) are specific for each situation and cannot be easily changed by design. other variables that affect pond performance can be controlled in the design of a pond system. These variables are the distance of the pond from the ocean or sound, the distance between adjacent ponds, the elevation of the pond bottom above mean sea level, the length-to-width ratio of the pond, and the amount of impervious area. When designing a pond the engineer should consider that pond performance improves by

1. increasing the elevation of the pond bottom above mean sea level,

2. increasing the ratio of pond length to pond width,

3. increasing the distance along the island length between the pond and adjacent ponds or other hydraulic sources,

4. decreasing the distance between the pond and the sound or the ocean, and

5. decreasing the area of impervious surfaces.

The models (DMPOND and MG3D) and equations developed in this project will be very useful to engineers for designing effective stormwater infiltration pond systems. Work is continuing toward making these design tools more user-oriented and toward developing simple design methods.