Stormwater Remediation Strategies for Rural Nicaragua A case study based in the Dulce Nombre region of Jinotepe, Nicaragua

Transcription

1 Engineers Without Borders Portland State University Stormwater Remediation Strategies A case study based in the Dulce Nombre region of Jinotepe, Nicaragua Austin Peters, Cole Presthus, PE, Juanita Platz, Seth S. Moody, William Templeton, Hamid Moradkhani, PHD, PE. March 1, 2009

2 ABSTRACT Annual stormwater flooding occurs in the Dulce Nombre District of the municipality of Jinotepe, Nicaragua. In this study we discuss the stormwater remediation project centered at the Dulce Nombre district. It includes a detailed analysis of the problem, hydrologic data, modeling and analysis of stormwater and recommendations for remediation. Development of the model has been difficult due to incomplete or lack of requisite data, inconsistent communication with local government and community leaders, and the sensitive variability of stormwater calculations. In order to minimize miscalculations due to erroneous data, several methodologies have been used to determine storm cycle volume and peak flow rates. Developing a workable remediation strategy presents its own difficulties. Typical stormwater engineering materials such as large pipes and culverts are expensive to procure. Inconsistent electricity and risk of theft makes pumping or more elaborate remediation strategies unfavorable. Simple solutions may require community involvement, additional property acquisition and a high level of diplomacy. Public health risks are also a factor due to local waste disposal methods. In order to determine the best approach to flood remediation several strategies are evaluated. The aforementioned difficulties will make up the framework of the evaluation schema and provide a recommendation for implementation to the community and municipal government. The paper also discusses not only the particular methodology and remediation strategies used in this case study but also constructs an outline for developing stormwater solutions in Nicaragua to aid in future public works projects there. March Stormwater Remediation Strategies

3 SECTION 1 Introduction and Background Jinotepe, Nicaragua March Stormwater Remediation Strategies

4 INTRODUCTION The Portland State University (PSU) student chapter of Engineers Without Borders (EWB) has been working in Nicaragua for three years within the Municipality of Jinotepe. During the summers of 2006 and 2007, PSU-EWB participated in a joint project in collaboration with PSU Institute on Aging to renovate an elder care facility. During the summer of 2007, participating students and professional mentors met with Jinotepe government officials from the mayor s office to review additional engineering projects in the Jinotepe area that would benefit from future PSU-EWB efforts. Of the several projects that were reviewed two of these projects were taken on by the student chapter and assessed during the summer of the following year. The projects chosen included a stormwater remediation project located in the Dulce Nombre district of Jinotepe and a drinking water design project for the city of Huehuete located on the Pacific coast approximately 30km west of Jinotepe. In June, 2008 three students and two professional mentors spent two weeks in Nicaragua assessing the nature of these projects, gathering government records, and collecting field data. Partnership with the Jinotepe Municipal Mayor s Office as well as working directly with municipal employees, local university engineering students, and community leaders accelerated data mining of municipal records. These included collecting pertinent engineering reports, rainfall IDF graphs, and geologic data. This data was analyzed using both standard and more advanced stormwater methodologies to determine the best remediation solution. BACKGROUND Annual stormwater flooding occurs in the Dulce Nombre District of the municipality of Jinotepe Nicaragua. Though several stormwater engineering projects have already occurred in the area, recurring flooding of the district s elementary school continues during the rainy season. The school is mid slope in a wide depression that collects runoff from two neighborhood public streets and surrounding drainage basins. During the rainy season, in October and November, government officials and school administrators estimate the water depth in the school yard to be 30 to 40-cm. This flooding greatly affects the educational process for the approximately 150 elementary school students. During the 2007 implementation trip for the elder care facility renovation the PSU-EWB chapter was asked by Jinotepe Municipal Mayor Alvaro Portocarrero Silva to conduct a full assessment and assist with implementation for remediation of this stormwater problem. Assessment Assessment of the stormwater issues surrounding the Dulce Nombre School occurred June 15 June 20, This assessment included conducting a thorough elevation survey of the drainage basins surrounding the school, talking with community residents about observed stormwater flows, and gathering information on rainfall and geologic soil types from the local government. Although the main area of interest is the flooding of the school itself, additional information was gathered on surrounding neighborhood stormwater problems and is partially included in this report. March Stormwater Remediation Strategies

5 Existing Conditions A detailed survey of the schoolyard is shown in figure 1-1. This figure shows the shape of the school grounds as a collection basin for stormwater. This basin collects stormwater from its own surface area as well as two surrounding natural drainage basins and two street right-of-ways which at high water drain into the schoolyard. Curbing installed on the streets funnel water from surrounding drainage basins into the roadways to a local flat point at the street intersection of Camino al Cementerio and Camino al Rosario. At this intersection flood waters are said to breach the curbing on the north side of Camino al Rosario and contribute additional flood water to the school basin. Additional flood areas were identified to the west of the school where houses west of the intersection of Camino al Cementerio and Camino al Rosario sit in a local depression that floods up to 45-cm in depth during the rainy season. Water from both directions along Camino al Rosario drains around and through the church via constructed stormwater control structures. This water floods an auxiliary street to the south of the church which shows significant erosion damage from concentrated storm flows. Of key concern is the flooding of the school buildings themselves. Table 1-1 provides both the elevations of the school building and height above the schoolyard low point for the finished floors of each school building. Table 1-1: Elevations of School Buildings Finished Schoolyard Lowpoint Height Above Building Floor Elevation Low-point Number Elevation [m] [m] [m] Water depth estimates of 0.3 to 0.4-meters from community members indicate that at least two of the three school buildings floors are beneath the water surface during the rainy season floods. This is not only a poor teaching environment but also presents health risks. Waste management in this community consists of garbage and latrine pits. Major rainstorms bring not only flooding concerns but water quality concerns as well. The elementary school itself has three pit latrines located on the north end of the school property. Based on this information, stormwater management alternatives in the area must consider human health risks. Provided Data The Municipality of Jinotepe provided limited survey and stormwater report data from a previous stormwater project completed by a local engineering firm. This project was done in an area adjacent to the Dulce Nombre School and included IDF curves, local methods for calculating stormwater runoff, CAD drawings of nearby streets, and a basic overview of standard stormwater remediation practices. March Stormwater Remediation Strategies

6 Figure 1-1: Area Overview Map Including survey of schoolyard

7 SECTION 2 Computer Model Development and Calibration March Stormwater Remediation Strategies

8 COMPUTER MODEL DEVELOPMENT AND CALIBRATION In order to develop initial conditions for use in a stormwater model and utilize more advanced methods to analyze and evaluate available alternatives a digital terrain model was built and populated with accumulated flow channels. Development of the model was made possible using tools from ESRI ArcGIS (ESRI) and the freely available ArcMap extensions ArcHydro, (Maidment) and HEC-GeoRAS (Army Corps of Engineers). Actual stormwater modeling was performed using the freely available HEC-RAS software package (Army Corps of Engineers). GIS was used to assemble the various data sources into one location. This included collected GPS data from the site visit, survey data, rudimentary maps acquired from the municipality, and associated observational data. The GPS data was converted into a CAD readable format using the export tools provided by the proprietary Garmin MapSource software (Garmin). All data was assigned to the 1983 North American Datum, Zone 16, and projected alignments were calibrated to maintain consistency. Digital Elevation Model In order to enable the use of more advanced terrain analysis tools, creation of a Digital Elevation Model (DEM) was required. Since detailed elevation information is not readily available in Nicaragua, the DEM was created by hand correlating data from 20m contour 1:250,000 scale maps, EWB survey data, collected GPS points, and photographs of the area. These interpolated elevation points were drawn on a gridded overview map at ten by ten meter intervals, keyed into the Microsoft Office spreadsheet application Excel (Microsoft, Excel), and then imported into the GIS software ArcMap as a raster file using the ASCII to Raster conversion tool in ArcToolbox. The DEM raster was further refined into a Triangulated Irregular Network (TIN) using the conversion feature provided in the Spatial Analyst extension. 0.2m contour lines were constructed from the TIN and used in subsequent representations to visually represent terrain data. Figure 2-1 shows the steps used to generate the TIN and contour lines graphically. In order to provide context and geographic alignment to the newly created TIN, the AutoCAD linework provided by the city with additional linework added by the Engineers Without Borders team was brought into the GIS map. This became the first, of many, computer/compatibility related issues. Traditionally regular AutoCAD software does not provide information interchange tools for use between GIS platforms. However with the advent of AutoCAD Map (Autodesk, MAP) this feature is provided with capability to export CAD layers to popular GIS formats built in. Careful attention was paid to the units and data projection settings before exporting to ArcMap to avoid display/rendering issues. Even still, some re-projection of the data was necessary for correct alignment of all the digital data. For this project AutoCAD Map proved very useful for its GIS compatibility features, however it is not a very powerful design tool and AutoCAD Civil3D was needed for designing stormwater remediation alternatives for reasons that will be discussed in the design section. ArcHydro The tools in the ArcHydro extension were used to delineate watersheds and create stream networks although no streams actually exist in the study area during dry periods. Due to the high intensity of rainfall events and the low permeability of soil layers at this location, however March Stormwater Remediation Strategies

9 precipitation quickly forms into shallow channelized flow justifying the creating of a stream network. Site observations made by the assessment team during storm events proved this assumption showing clear areas where stormwater collected into well defined channels. Using the terrain preprocessing tool in the ArcHydro extension the developed DEM was augmented to make it ready to accept watershed and stream network delineation algorithms. This included filling sinks and defining existing streams. As noted above, there were no permanent streams in the area. However, stormwater channelization did occur rapidly during storm events. This happened most significantly on the roadways where curbing on either side of the road prevented runoff. In order to include these features in the model the cm curb elevations were added to the DEM during the terrain preprocessing step in ArcHydro. The ArcHydro tools provided automated steps for calculating flow direction, flow accumulation, stream definition and catchment grid delineation. The main area of user input during this process lay in the adjustment of the stream threshold parameter. This parameter defined the number of adjoining cells required to define a stream. When set to 4 cells it provided a very detailed number of streams but also way too many watersheds. By modifying the stream threshold to 20 cells, larger watersheds were created but very few streams were delineated. The solution to this was to combine the two sets using the stream delineation from the 4-cell threshold but the watersheds from the 20-cell threshold. To further simplify the watershed calculations the watersheds were remapped once more to collect sub-watersheds together to form identifiable regions that could be used during the design process. The final results of this manipulation with ArcHydro is shown in Figure 2-3. The associated directional stream network created in this step is shown in figure 2-2. ArcHydro defines stream flowlines in a network as HydroEdges. At the beginning or end of a HydroEdge or the intersection of two HydroJuction Flow Arrows HydroEdges ArcHydro creates a HydroEdge Figure 2-2: ArcHydro Network including Hydrojunctions, Hydroedges, and directional flow arrows node in the network called a HydroJunction. Because these stream networks are given a digital direction based on topography, arrows can be displayed on the display map using the network analyst tool in ArcMap. This stream network was used to generate flowlines later in the stormwater process. model March Stormwater Remediation Strategies

10 a. The DEM was originally assembled with all relevant data sources by hand in a x grid format with relevant data overlayed using AutoCAD. b. The hand written numbers were brought into Excel and then exported to an ASCII text file. c. The ASCII text file was converted in ArcMap into a DEM. The assembled linework from AutoCAD was also brought into ArcMap at this time. d. The DEM was converted to a TIN and then re-colored for clarity. Contour lines were added at 0.2m intervals to delineate terrain changes. FIGURE 2-1: The steps required to create a TIN from raw data. March Stormwater Remediation Strategies

11 Figure 2-3: Dulce Nombre School.4 Arc Hydro - HydroEdges and Catchments Meters DrainagePoint_fine HydroEdge_fine DrainagePoint 98 DrainageLine HydroJunction_fine Legend Engineers Without Borders - Portland State University : Catchment contours GPS Track_line

12 Figure 2-4: Basin Overview Map The tools available from ArcGIS and ArcHydro also allowed further analysis of the terrain and its relationship to the projected flowlines. In order to calculate flowrates using the rational method discussed in section 3, the basin area and slope of streamlines became important. ArcGIS tools automatically calculated the area of polygons like those that defined each of the basins. Customizing the attribute tables using Microsoft Access database (Microsoft, Access) software allowed direct input of the basin labeling system into the basin s GIS database file. Updating labeling criteria in layer properties allowed the display of multiple attributes in the ArcMap window to display the basins as shown in figure 2-4. Using the 3D Analyst extension for ArcGIS, the longest, most significant flowline in each basin was traced with a 3D line. A surface elevation profile was generated from the 3D line s intersection with the TIN so that the average slope could be discerned from the resulting graph. Additionally, the data used to generate the surface elevation profile was also exported to Excel spreadsheet software so that the average slope could be calculated directly from the high and low elevation points and the distances between them. The results of these calculations is given in section 3. A sample surface elevation profile from basin 8 is shown in figure 2-5. Figure 2-5: Surface elevation profile for the main drainage in basin 8 March Stormwater Remediation Strategies

13 HEC-GeoRAS After processing with ArcHydro, HEC-GeoRAS was employed to export the GIS data from ArcMap into a data format readable by the hydraulic modeling software HEC-RAS. Incidentally, the required version of the APFramework application, which is an information interchange protocol, is not consistent in ArcHydro and HEC-GeoRAS. Furthermore neither version is cross compatible with both extensions. Therefore ArcHydro tools had to be completely uninstalled and the APFramework application removed before HEC-GeoRAS could be successfully installed and vice-versa. This process was quite problematic when performing calibration that required adjustments in both paradigms. Preparation of the ArcGIS data for export into HEC-RAS included defining flowlines, stream banks, cross-sections and land use polygons. The flowlines were derived from the ArcHydro model stream delineation layer. Because no permanent streams existed, stream banks were not necessarily relevant but were still required to enable the export process to proceed. The land use polygons were used to define pervious and impervious areas as well as the frictional flow coefficient, Manning s n, tied to area subsections. The extension toolbar highlighting some of the available automated features is shown in figure 2-6. Digitization of the required features was done using ArcEditor. Though this tool only offers simple drawing tools it was sufficient for use in this particular model. If more complicated linework had been required it would have been simpler to do this in AutoCAD and then imported back into GIS using the export features in AutoCAD Map. Figure 2-6: The HEC-GeoRAS extension toobar including some of the required layers that must be created in order for the export file to function correctly in HEC-RAS. Creation of a HEC-RAS readable export file was a lesson in trial and error. Because the data was gathered from multiple sources and in many cases input by hand there was much that was either incomplete or inconsistent. Each attempt to export the model would yield various error messages that were subsequently researched, remedied, and reapplied before the next error message would arise. A similar process was used to run the model after it was imported into the HEC- RAS software. The working geometry as imported into HEC-RAS is shown in figure 2-7. March Stormwater Remediation Strategies

14 Once the model was successfully running in HEC-RAS several steadystate flow conditions were applied to the modeled stream network to calibrate flowrate to modeled flood depths at the school site. Because community interviews and observation had predicted annual flood depths between 30 and 40- cm the flowrates were adjusted to yield those depths in the school yard. To determine the height of water above land surface, the results from the HEC-RAS Figure 2-7: The geometric data window showing the cross sections and the central flowlines for the working model imported into HEC-RAS. model were exported from HEC-RAS into a GIS readable format and imported into ArcMap using the HEC-GeoRAS tools. A water surface raster was created from the HEC-RAS export file and this raster was subtracted from the land surface DEM to determine the flood water depth for a given steady state flow rate. The calibrated model with a flowrate of.-m 3 /s and 0.15-m 3 /s is shown in figure 2-9 and 2- respectively. Based on the flood depths predicted in the model the overland flow during a typical annual storm that floods the school is approximately 0.15-m 3 /s Figure 2-8: For clarification of the modeling process used in this project a schematic of the software utilized to set up the data and run the model is given in figure 2-8. March Stormwater Remediation Strategies

15 Dulce Nombre School Flooding Model 0.-m3/s Figure 2-9: Legend : GPS Track_line 0.cms Flowrate Flood Deph [m] Meters

16 Dulce Nombre School Flooding Model 0.15-m3/s Figure 2-: Legend GPS Track_line : 0.15cms Flowrate Flood Depth [m] Meters

17 SECTION 3 Stormwater Runoff Flow Methods and Calculations March Stormwater Remediation Strategies

18 STORMWATER RUNOFF FLOW CALCULATION METHODS Collection of accurate rainfall data proved difficult. Though rainfall gauging stations do exist at locations nearby Jinotepe, acquiring this rainfall data was not possible despite several attempts. Past projects and experience in Nicaragua has shown over and over again that non-personal communication ( and telephone) with local government officials and in-country contacts is inconsistent and unreliable. Current political events in Nicaragua have paralyzed both national and local government leaving many questions unanswered until political accord is reached. Without the possibility for another personal visit given the project timeline the original IDF curves provided by the local Jinotepe government were used as the primary source of rainfall data. Though methods do exist for converting IDF curves back to 24-hour storm depth totals it was shown to be inaccurate to do this based on the computer modeling. Typically, the area underneath a chosen IDF curve can be determined using rectangular approximation or integration of a fit curve. However, since the high intensity storms shown in the IDF curves do not typically last for 24 hours the results from the aforementioned method gave extremely high rainfall depths. The Rational Method Since more accurate rainfall data was not available the IDF curves were used directly with the rational method. The rational method is based on the following formula: (Gupta, 674) Q T C i A <3-1> T Where: Q T = Estimate of the peak runoff rate (m 3 /s) for some return period, T. C = Runoff coefficient; fraction of runoff, expressed as a dimensionless decimal fraction, that appears as surface runoff from the contributing drainage area. i T = Intensity of precipitation for a duration equal to time of concentration, T C, at a chosen return period, T. A = The contributing tributary drainage area to the point of design in square meters which produces the maximum peak rate of runoff. T C = Time of concentration. The time required for runoff from the hydraulically most remote part of the drainage area to reach the point of reference. Though this method is generally straightforward it is open to some analytical interpretation in determining the runoff coefficient, C and the time of concentration, T C. Previous hydrology reports from neighboring areas were provided by the local government to show regional methods for determining stormwater parameters and stormwater control methods. The acquired rainfall data is based on rainfall recording stations located in Masatepe, Nicaragua a city with similar characteristics located approximately -km from Jinotepe. The maximum rainfall intensities over a 28-year period for given time intervals are shown in table 3-1. The IDF curves associated with this rainfall data are shown in figure 3-1. This data was used in conjunction with the developed survey to determine stormwater flows for the -year return period storm using the rational method. March Stormwater Remediation Strategies

19 Year - Intensities [mm/hr] Table 3-1: Maximum Annual Intensities from the Masatepe Weather Station Records from: Time [min] Average: Maximum: Minimum: March Stormwater Remediation Strategies

20 Figure 3-1: Intensity Duration frequency curves provided from rainfall data from the Masatepe, Nicaragua rainfall station. 250 Intensity, Duration, Frequency Curves Intensity [mm/hr] years 2 years 5 years years 15 years 25 years 50 years 0 years Time [min] Runoff Coefficient The runoff coefficient is generally determined from published tabled values. The coefficient is based on the average soil type in the basin, type of vegetation coverage, and land slope. Table 3-2 provides a sampling of generally accepted runoff coefficients. (Haan, 84-85) Table 3-2: Runoff Coefficients Urban Areas: the use of average coefficients for various surface types, which are assumed not to vary through the duration of the storm, is common. The range of coefficients, classified with respect to the general character of the tributary reported in use is: Description of Area Runoff Coefficients Downtown Areas 0.7 to 0.95 Neighborhood areas 0.5 to 0.7 Single-Family Areas 0.3 to 0.5 Parks, cemeteries 0.1 to 0.25 Playgrounds 0.2 to 0.35 Unimproved Areas 0.1 to 0.3 March Stormwater Remediation Strategies

21 Note: It is often desirable to develop a composite runoff coefficient based on the percentage of different types of surface in the drainage area. This procedure is often applied to typical 'sample' blocks as a guide to selection of reasonable values of the coefficient for an entire area. Coefficients with respect to surface type currently in use are: Character of Surface Runoff Coefficients Streets 0.7 to 0.95 Roofs 0.75 to 0.95 Lawns, sandy soil Flat, 2% 0.05 to 0.1 Average, 2 to 7% 0.1 to 0.15 Steep, 7% 0.15 to 0.2 Lawns, heavy soil Flat, 2% 0.13 to 0.17 Average, 2 to 7% 0.18 to 0.22 Steep, 7% 0.25 to 0.35 Note: The coefficients in these two tabulations are applicable for storms of 5-year to -year frequencies. Less frequent higher intensity storms will require the use of higher coefficients because infiltration and losses have a proportionally smaller effect on runoff. Table 3 Continued Open Clay and Tight Sandy Silt Loam Clay Topography and Gegetation Loam Woodland Flat 0-5% slope Rolling 5-% slope Hilly -30% slope Pasture Flat 0-5% slope Rolling 5-% slope Hilly -30% slope Cultivated Flat 0-5% slope Rolling 5-% slope Hilly -30% slope Reference: Haan, 84 Time of Concentration In order to determine the correct intensity to use, the time of concentration had to be determined for the appropriate basins. Time of concentration (T C ) is the time it takes for a theoretical drop of March Stormwater Remediation Strategies

22 water to make it from the most remote area of the basin to the outfall point; several methods have been developed to determine this. For the purposes of this study it was difficult to know which method would provide the most accurate results. For that reason, each method was used to calculate the time of concentration and the results were evaluated for comparability to the modeled flowrates acquired in the previous section. Time of concentration depends on several factors including ground cover, slope, flow length, and rainfall volume. Flow type is also extremely important in determining T C as water moves much faster in concentrated flows. In the beginning droplets of water act independently and travel in a thin layer through variable flow-paths. This is defined as sheet flow. Once stormwater has accumulated over space and time it usually becomes channelized in shallow concentrated flow. The following empirical equations are commonly used in the U.S. to determine t C. (U.S. Dept. of Ag., TR-55) For sheet flow, for a surface with a maximum length of 300-ft: ( nl) T t 1 <3-2> ( P ) S Where: T t1 = Sheet flow travel time, min n = Manning s roughness coefficient, typical values are given in Table 3-3 L = Flow length, ft P 2 = 2-year 24-hour rainfall, in S = land slope, unitless 2 Table 3-3: Overland Flow Roughness Coefficients Surface type Manning's n Concrete, asphalt, bare soil Gravel, clay-loam eroded Sparse vegetation, cultivated soil Short Grass Dense grass, bluegrass, bermuda grass Woods Adapted from Gupta, 681 For shallow concentrated flow: Where: T t2 = Concentrated flow travel time, min L = Concentrated flow length, ft L 1 T t 2 <3-3> V 60 March Stormwater Remediation Strategies

23 V = Flow velocity, ft/s (Figure 3) P 2 = 2-year 24-hour rainfall, in S = land slope, unitless Because these equations are empirical they rely on english units and though conversion is possible it is not convenient for use in areas where metric units are commonly used. Figure 3-2: Typical velocities for various ground covers at differing slopes. (Reproduced from Gupta, 682) Municipal engineers in Jinotepe identified several methods commonly used in Central America to determine time of concentration. The empirical formula used by the Colombian Engineering School: Where: T c min = mimimum time of concentration for the basin, min A = area of the drainage basin, hectares 8.68 T min A C <3-4> March Stormwater Remediation Strategies

24 Another method used by the Central American Hydro/Meteorological Project: L T C 1 <3-5> 2 S Where: T C = The time of concentration for the basin, min L = Length of the channel from the end of the basin to the point of departure, m S = Slope of the channel, unitless Finally, the empirical formula used by the Ministry of Public Works of Venezuela: L.0195 T C <3-6> H Where: T C = The time of concentration for the basin, min L = Length of the channel from the end of the basin to the point of departure, m H = the difference in elevation from end to end of the channel, m These methods generated wildly different time of concentration values which are discussed later. Return Periods The data provided by the municipality included IDF curves for many different return periods from 1.5-years to 0-years. Typical stormwater systems are designed for the or 25-year storm so as not to over-design the system and cause unfavorable hydrologic flow conditions during average storms. Local Nicaraguan engineers recommend using the -year return period for design purposes using the rational method. The flowrates acquired during the modeling stage described in the previous section were considered to be for a typical major storm events. This was decided to be analogous to a 2-year return period event. The rational method was calibrated using the 2-year return period IDF curve and the stormwater design utilized a -year return period storm event. March Stormwater Remediation Strategies

25 HYDROLOGICAL STUDY DATA AND CALCULATIONS Using the rational method techniques described in the previous section, information was compiled from several sources as discussed in section 1 to generate design values for stormwater control alternatives. Basin Area The calculated areas of each basin are shown in table 3-4 The overall basin area was acquired through the use of the ArcHydro extension for delineating catchments as described in section 2. The impervious areas were derived from personal knowledge from site visits, hand drawn maps, and aerial photographs. These areas were also input into the GIS database as polygons such that their area could be measured accurately. Table 3-4 Basin Areas, Pervious and Impervious Area, and dominant use. Impervious No. Total Area Pervious Area Area Description [m] [m] [ha] [m] [ha] Land Use Basin Med. Residential Basin Light Residential Basin Light Residential Basin Med. Residential Basin School Basin Med. Residential Basin Light Residential Basin Undisturbed Basin Undisturbed Basin Light Residential Basin Med. Residential Runoff Coefficient Calculation of the runoff coefficient, C required knowledge of the land use, vegetation type and general slope of each basin. Because of the historical volcanic activity in the area the general soil type is primarily silty clay containing weathered volcanic minerals. This soil type is described as silty clay loam topsoil with an increase in silt and clay content below the sandy loam layer. (Vogel, 6) The upper 23-cm contains high concentrations of organic carbon. (Vogel, 6) This information was used to choose tabled values and combine them using weighted average techniques to achieve a single runoff coefficient for each basin. These runoff coefficients are shown in Table 3-5. March Stormwater Remediation Strategies

26 Basin No. Table 3-5: Calculation of the Runoff Coefficient, C Description of Runoff Coefficient, C - Pervious Runoff Coefficient, C - Impervious Land Use Tabled Value* % of Area Tabled Value* % of Area Weighted Average Basin 1 Med. Residential % % 0.42 Basin 2 Light Residential % % 0.21 Basin 3 Light Residential % % 0.23 Basin 4 Med. Residential % % 0.45 Basin 5 School % % 0.82 Basin 6 Med. Residential % % 0.38 Basin 7 Light Residential % % 0.23 Basin 8 Undisturbed 0. 90% 0.98 % 0.19 Basin 9 Undisturbed 0. 94% % 0.15 Basin Light Residential % % 0.26 Basin 11 Med. Residential % % 0.63 * Referenced from table 3-3 Slope Slope was calculated on an average basis for each basin by using the digital elevation model and graphical elevation profiles described in section 2. The data from the profiles was exported to an Excel spreadsheet such that average slope could be calculated directly. These slopes are shown in Table 3-6. March Stormwater Remediation Strategies

27 Table 3-6: Average Slope Length Elevations Slope Basin No. Upper Lower [m] [m] [m] [m/m] Basin % Basin % Basin % Basin % Basin % Basin % Basin % Basin % Basin % Basin % Basin % Time of Concentration Several methods to determine the time of concentration were discussed previously. These formulas yielded dramatically different results ranging from only 3 minutes to nearly 60 minutes for the same basin. Several calculations and determinations were used to acquire the necessary data for input into the T C equations. This included determining Manning s coefficient for each basin and calculating flow velocity and 2-year rainfall depth. These values are given for each basin in table 3-7. Since velocity was acquired from figure 3-2 in English units it is presented in this table in the inconsistent units of ft/s. The units were converted in later calculations depending on the empirical methodology requiring use of either metric units or English units. March Stormwater Remediation Strategies

28 Table 3-7: Necessary data for calculating T C Manning s n Est. Velocity 2-year Rainfall Basin No. [unitless] [ft/s] [mm] Basin Basin Basin Basin Basin Basin Basin Basin Basin Basin Basin In order to determine the 2-year Rainfall depth, further analysis of the IDF curves was necessary. The 2-year IDF curve was modeled using an exponential regression (R 2 =.97). The curve was integrated to determine the area beneath the curve and thus the total rainfall depth for a 6-hour design storm. Both the original equation and the integrated one are shown below: 2-year IDF regression: 2-year IDF regression integrated: C i T <3-8> D TC <3-9> Where: i = intensity, mm/hr T C = time of concentration, hours D = rainfall depth, mm The integral was evaluated between 0.08 and 6 hours to achieve the value of 151-mm. Finally, Table 3-8 gives the calculation of the time of concentration using the four different methods. March Stormwater Remediation Strategies

29 Table 3-8: Calculation of the Time of Concentration Basin Tc (min) No. Calculated (1) Calculated (2) Calculated (3) Calculated (4) Calculated (5) Tc Average [min] Basin Basin Basin Basin Basin Basin Basin Basin Basin Basin Basin Formula for Sheet Flow (TR-55) 2 Formula for Shallow Concentrated Flow (TR-55) 3 Formula from the Columbian Engineering School 4 Formula from the Central American Hydro/Meteorological Project 5 Formula from the Ministry of Public Works of Venezuela In order to determine the best value to use for T C the flowrates resulting from using the rational method and the 2-year return period intensity curve were compared with the results from hydraulic modeling previously discussed in section 2. Typical flood depths that matched relevant observational data were found during modeling a steady state flowrate of 0.15-m 3 /s. These flows would be a combination of flows from basins 5 and 6. In order to determine the flowrates for these basins the design storm intensity needed to be calculated first. March Stormwater Remediation Strategies

30 Intensity The intensity was calculated using each of the Time of Concentration methods discussed previously. Using the 2-year IDF regression given in equation 3-8 storm flows for the 2-year return period storm could be calculated. Table 3-9 shows the resulting rational method flowrates in basins 5 and 6 from using each of the T C values given in table 3-8 and the 2-year return period IDF curve. The combined peak flow from these two basins was compared to the calibrated model steady state flow of 0.15m 3 /s. The most accurate results from a single method are achieved using the Sheet Flow TR-55 equation. However, the most accurate flows are acquired by averaging all the T C results. Table 3-9: Flowrates resulting from Time of Concentration Values (2-year IDF) Flowrate (m 3 /s) % Difference Basin Basin From Model T C Calculation Method 5 6 Total Flow (0.15m 3 /s) Sheet Flow TR % Shallow Flow TR % Columbian Eng. School % Central American Hydro Project % Public Works Venezuela % Average of All % Based on the results shown in table 3-9 the average time of concentration from all five methods was used to calculate the peak flow design values for the -year return period storm. The - year return period IDF curve was modeled using an exponential regression (R 2 =.98) as shown below: -year IDF regression: C i T <> Where: i = intensity, mm/hr T C = time of concentration, hours This equation, along with the average T C value given in table 3-8 was used to determine the - year, peak flow, design storm intensities. Using the intensity, basin areas, and associated weighted runoff coefficients, the peak flowrate in each basin could be calculated using the rational method (Equation 3-1). The final results of this analysis are given in table 3-. The total combined flow from Basins 5 and 6 for a -year storm event by this analysis is 0.188m 3 /s. March Stormwater Remediation Strategies

31 Table 3-: Summary of C, I, and A values and associated peak flowrates. Intensity, I Flowrate Basin, Weighted Runoff Total Basin Basin No. Coefficient, C (mm / h) Area, A (m 2 ) Q (m 3 / s) Basin Basin Basin Basin Basin Basin Basin Basin Basin Basin Basin March Stormwater Remediation Strategies

32 SECTION 4 Available Alternatives March Stormwater Remediation Strategies

33 AVAILABLE ALTERNATIVES Several alternatives were proposed to remedy the flooding problems surrounding the Dulce Nombre School. These included the use of stormwater swales, grading and related earthwork in the schoolyard, elaborate conveyance systems, holding ponds, and dry wells for added infiltration. These alternatives were explored to determine their effectiveness, cost, community impact, and health/sanitation impact. Though additional flood remediation infrastructure is required for other areas in the vicinity a solution for the school was determined to be the priority at the request of the municipality. Future expansion projects of the chosen remediation strategy may be designed and implemented at a later date to provide stormwater solutions to the other areas. Based on the modeling and mathematical analysis performed in the previous two sections the selected remediation systems explored were designed to accommodate flows from a -year storm event. Manning s equation was used to compute the steady state hydraulic calculations required to size swales, piping, culverts, and related stormwater control structures. The general Manning s equation is: (Gupta, 541) Where Q = Flowrate, m 3 /s K = 1 in SI units, in English units n = Manning s roughness coefficient, dimensionless A = Area, m 2 R = Hydraulic radius, m S = Bottom slope, unitless K Q A R S <4-1> n The computer software Flow Pro 2.1 (ProSoft Apps) was used to quickly and easily solve Manning s equation for the variety of different solutions presented. March Stormwater Remediation Strategies

34 Stormwater Swales Stormwater swales collect stormwater from the surrounding area in designed surface water channels that aid in both direct ground infiltration at the sight of the swale as well as surface conveyance to other locations. A typical constructed stormwater swale is shown in figure 4-1. Figure 4-1: Cross section of a vegetated swale. 6 Based on area topography and established drainage areas the swales should be constructed on the north and east side of the school yard. This will collect runoff from the contributing drainage basins before it enters the school yard. The north 6 Adapted from the City of Eugene Stormwater Manual. swale would collect run off from Basin 5 and the East Swale would collect flows from the North Swale and runoff from Basin 6. In order to capture the -year storm the North Swale must be designed to handle m 3 /s and the East Swale must handle m 3 /s. A summary of the input parameters and the resulting water depth in each of the swales are shown in tables 4-1a and 4-1b. The Manning s n for these calculations was chosen from tabled values for rocky vegetated channels. (Gupta, 541) Table 4-1a: Depth of flow in North Swale for -year storm event Table 4-1b: Depth of flow in East Swale for -year storm event Flowrate, m 3 /s Flowrate, m 3 /s Width, m 1.20 Width, m 1.20 Manning's n Manning's n Bottom Slope, V:H 0.01 Bottom Slope, V/H 0.01 Side Slope, H:V 3 Side Slope, H/V 3 Water Depth, m Water Depth, m Velocity, m/s Velocity, m/s March Stormwater Remediation Strategies

35 Construction of these swales would be achieved with manual labor and hand tools, both available in Nicaragua. Swale depths are shallow and would not require extensive excavation to create a channel deep enough to carry design storm flows. Though constructed swales typically make use of proprietary permeable filter fabric to protect groundwater from chemical infiltration these measures are not required. Natural rocky soils and vegetation can be used to anchor the excavated soil and prevent sediment transport. Grading and Earthwork Current grading of the schoolyard creates a perfect settling basin for stormwater after a rain event. Because the earth is compacted in this area, infiltration is not significant and stormwater can remain in the pooled area for several days after rainfall. Though these conditions are ideal for playing football during dry weather they do not help in preventing stormwater accumulation. To do this the school yard should be re-graded to eliminate the areas where water pools. To achieve reasonable flow velocities across the school grounds a minimum 1% slope should be used. This will require a minimum grade drop of 0.06-m over the 60-m school basin. Given the fixed finished floor elevations of the school buildings grading must accommodate this slope without undermining building foundations or causing local low points at building locations. Shallow trenches should be constructed around each school building to eliminate risk of localized flooding. Grading of the schoolyard must be combined with other remediation techniques to convey stormwater runoff from the graded low-points to other locations. Conveyance System Several options exist for routing the stormwater to outfall locations in the area. Because of community impact concerns, city officials had recommended piping the water from the school east along Camino al Rosario to an uninhabited discharge area as shown in figure 4-2. The pipe would be designed to carry runoff flows from both Basin 5 and Basin 6, totaling m 3 /s. Based on computer calculations for flows of this Table 4-2: Depth of flow in Conveyance Pipe for -year storm event Flowrate, m 3 /s Diameter, m 0.4 Manning's n Bottom Slope, V:H 0.01 Water Depth, m Velocity, m/s 1.75 magnitude, this option would require approximately 350-m of 40-cm diameter pipe. In order to achieve adequate slopes to minimize settling in the pipe and also effectively drain the school yard, the pipe would be buried at an average depth of 1.61-m and a maximum depth of 3.21-m. A summary of the input data and resulting water depths and velocity is given in Table 4-2. This option requires a significant amount of materials and excavation at some great expense. Long excavations at depths in excess of 3-meters will require intensive labor and poses worker safety concerns due to the lack of adequate excavation shoring and personal protective equipment. Large diameter pipe is expensive and the required equipment to handle heavy piping March Stormwater Remediation Strategies

36 would be difficult to acquire. Maintenance of the system would incur additional expense for both excavation and repair. As an alternative to a long pipe run, a culvert could be constructed to cross the Camino al Rosario south of the school yard to drain with the natural run of the land into adjacent private properties using simpler stormwater management techniques. The sizing of the culvert would need to be such that it could not only convey the required flowrate but also deliver the water across the street without drastic increase in flow velocities that could cause significant scour and erosion at the outfall. Tables 4-3a and 4-3b show two possible concrete culvert design parameters and the resulting flowrates and velocities. The smaller culvert is more easily constructible, requiring less concrete and excavation. There is, however, significant concern of clogging with foreign materials (eg plastic bottles) for a culvert of this size. The larger culvert requires more excavation and concrete but would decrease the chances of small objects clogging the culvert flow and causing failure of the stormwater system. Furthermore, a larger culvert will create minimal entrance losses and decrease the likelihood that water will back up into the east swale. In the long run this approach is preferable. Table 4-3a: Depth of flow and Velocity in a 0.40-m x 0.35-m Culvert for - year storm event Table 4-3b: Depth of flow and Velocity in a 1.2-m x 0.35-m Culvert for -year storm event Flowrate, m 3 /s Flowrate, m 3 /s Width, m.4 Width, m 1.3 Manning's n Manning's n Bottom Slope, V:H 0.01 Bottom Slope, V:H 0.01 Side Slope, H:V 0 (vertical) Side Slope, H:V 0 (vertical) Water Depth, m Water Depth, m 0.12 Velocity, m/s 1.39 Velocity, m/s 1.21 Though the use of a culvert to allow water to follow the natural existing land drainage has a greater impact on the community, its expense and feasibility is much more reasonable than the more complex large pipe run discussed previously. However, additional stormwater control structures may be required after the culvert is installed to aid in dispersion of stormwater flows. Holding Ponds Due to the large volume of water produced during a given storm cycle holding ponds are not recommended. These ponds would require not only large areas of land to be acquired but also significant excavation. March Stormwater Remediation Strategies