The objective of this project is to examine the possibilities and magnitudes of floods in the city of Smithfield, Utah as a result of precipitation runoff from the surrounding watersheds. Summit Creek is a small stream which meanders through Smithfield city that has had troubles with flooding in times past. There are two main watersheds which drain into Summit Creek that are both located to the east of Smithfield. Both Smithfield and Birch Creek canyons are steep, relatively small mountainous canyons that are located in the Cache National Forest. Figure 1 shows a Google Earth screen shot of the city of Smithfield with its drainage basins to the east in the mountains. Figure 2 shows a view from Smithfield looking east into the water shed of concern.

Figure 1: A layout of Smithfield with its drainage basins to the east.

Figure 2: View from Smithfield to its water shed.

This investigation will be based primarily on flood runoffs during precipitation events. To properly investigate the two watersheds of concern a base map, including a digital elevation map, of the region will be created using GIS. To accomplish this data will be acquired from the NHDplus data base and the USGS seamless server. With a base map in place the two watersheds will be able to be delineated to determine drainage areas, catchment slopes, and other critical hydrological information necessary to perform the flood analysis.

The flood analysis will be performed two ways. The first way will be to build a precipitation runoff model in GIS using the rational method to approximate discharge from the two canyons where the discharge Q = CiA. In this approximation Q is the discharge, C is a coefficient based on land surface attributes, i is the precipitation intensity, and A is the area of the drainage basin. The second way the runoff flood will be analyzed will be using a software package from the army corps of engineers called HEC-HMS. This software package will need values from the GIS base map to properly perform its analysis. Both methods will need precipitation intensities from the National Oceanic and Atmospheric Administration (NOAA) web site.

Objective

The main goal of this investigation will be to predict with reasonable certainty, the flow rate in the Summit Creek through the city of Smithfield resulting from a number of precipitation events acting within the two watersheds mentioned above. Then, if needed, a hydraulic analysis will be performed to determine how well Summit Creek will be able to route this flow rate through the city of Smithfield.

Introduction

This project was done in conjunction with research performed by the author for the city of Smithfield, Utah. Smithfield is concerned with the accuracy of current flood insurance rate maps (FIRM) within the city in an area proposed for new city offices. The water shed for Smithfield is a 23 square mile area of steep mountainous terrain known as Smithfield and Birch Creek Canyons. These Canyons both have a small base flow year round but discharges increase dramatically during spring runoff and in large precipitation events. Both Summit Creek (located in Smithfield canyon) and Birch Creek (located in Birch Creek canyon) have no gauging stations so base flow will be ignored for this analysis. Figure 3 and 4 show pictures, provided by Google Earth, of what some of the typical steep mountainous terrain in these canyons look like.

Figure 3: A picture in Smithfield Canyon.

Figure 4: A picture of Birch Creek canyon.

Data Acquisition and Base Map Construction

The information needed to construct a base map was acquired from the Utah GIS Portal (Utah 2008) and the NHDPlus data set (NHDPlus 2008), and the USGS seamless server (USGS, 2008). The Hydrologic region this project is located in is region 16B in the middle bear Idaho, Utah Huc-8 unit number 16010202. To build the base map for the proposed study the NHDPlus data was first acquired. This data consists mainly of NHD flow lines, catchments, and hydrologic region boundaries. With this data imported into arc map in GIS the next task was to determine where the proposed study was located in such a large region. Figure 5 shows a map layout of hydrologic region 16B with the middle bear and little bear-Logan Idaho, Utah hydrologic units highlighted in green. Figure 6 shows a layout of the middle bear and little bear-Logan Idaho, Utah hydrologic units associated with the Smithfield water shed, where the Smithfield watershed is highlighted in pink. Once the area of interest was separated from the rest of the data provided by NHDPlus the digital elevation map (DEM) for the area was downloaded from the Utah GIS Portal. Figure 7 shows the base map with the digital elevation map converted to a hill shade view.

Figure 5: Layout of hydrologic region 16B with the hydrologic units of the Smithfield water shed.

Figure 6: Layout of the hydrologic units the Smithfield water shed is in.

Figure 7: Smithfield watershed.

Next the precipitation values for each of the catchments in the Smithfield water shed was gathered from the NOAA web site (NOAA, 2008). These data are needed to build the precipitation runoff models in both GIS and HEC-HMS. Since this study was done in relation to city offices the precipitation storms with a 100 year return period are required for this analysis. This data was gathered from NOAA according to the latitude and longitude of the centroid of each catchment. Table 1 summarizes the precipitation intensities gathered from the NOAA web site and Figure 8 shows the numbered catchments in the Smithfield water shed along with the annual average precipitation depth in inches in each catchment. As is shown the precipitation values increase with higher elevations. Figure 9 shows the location of the rain gauges with the rain intensities associated with the 100 year 1 hour precipitation event.

Table 1: 100 yr precipitation values.

	Catchment Centroid		Precipitation Values
Catchment	Latitude	Longitude	100 yr 1 hr	100 yr 24 hr
(#)	(dd)	(dd)	(inches)	(inches)
1	41.911	-111.689	1.85	6.66
2	41.901	-111.678	1.87	6.82
3	41.891	-111.683	1.84	6.65
4	41.898	-111.694	1.85	6.62
5	41.89	-111.703	1.81	6.3
6	41.871	-111.699	1.76	6.02
7	41.854	-111.706	1.72	5.69
8	41.869	-111.758	1.55	3.96
9	41.853	-111.754	1.59	4.41

Figure 8: Smithfield water shed with annual average precipitation values

Figure 9: Rain gauges with 100 year 1 hour precipitation intensities shown.

Basin Characteristics

Two storm distributions were used in this analysis, namely a linear distribution and a more representative distribution. For the linear distribution the storm simply produces a uniform intensity over the entire duration of the storm. To produce a representative storm distribution in the state of Utah the rainfall was distributed using the 24-hour storm distribution obtained from the Utah State Engineer’s office. The time base of the 24-hour storm was modified to correspond to the duration of interest as this representative distribution was applied to the precipitation events modeled. The storm distribution from the State Engineer’s office is given in Table 3. The distribution is similar to the SCS 24-hour storm distribution. Indication from the State Climatologist was that this was a reasonable approach to distribute the rainfall for storms with durations shorter than 24 hours. Figure 10 is a plot illustrating the data in Table 3 along with a linear distribution.

Table 3. Storm distribution.

Ratio of Time to Storm Duration	Ratio of Rainfall Depth to Total Rainfall
0.0000	0.0000
0.0417	0.0094
0.0833	0.0219
0.1667	0.0500
0.2500	0.0875
0.3333	0.1500
0.3750	0.1969
0.3958	0.5031
0.4167	0.6344
0.4583	0.7125
0.5833	0.8063
0.7500	0.9125
0.8333	0.9500
0.9167	0.9781
0.9583	0.9906
1.0000	1.0000

Figure 10: Representative storm distribution

Based on the soil type layer in the base map and the land use in the catchment, the basin fell into Hydrologic Soil Group B which is defined as shallow loess or a sandy loam. According to McCuen, a soil classification of soil group B in an area of forest land cover with average slopes greater than 6% the runoff coefficient for the rational method should be between 0.17 and 0.18. Assuming antecedent moisture condition II, with the above given land cover and soil type the basin was assigned a Curve Number of 70. This number is used to compute the rainfall excess or amount of rainfall that contributes to runoff. The following equations are used for computing the initial abstraction (I_a), which is the amount of water that infiltrates into the ground before runoff begins.

where S can be found as shown below,

where CN is the curve number defined above. These equations produce an initial abstraction of 0.857 inches.

With this information gathered from the base map the two models can now be made.

Rational Flood Prediction Model

To develop a precipitation/runoff model in GIS; the rational method will be used. The rational method indicates that

where Q is the peak runoff due to precipitation, C is a runoff coefficient based on soil type, land cover, and land slope, i is the rainfall intensity in inches/hour, and A is the area of the catchment in acres. The traditional rational method rational method makes many assumptions (McCuen, 2004) including;

1) The rainfall intensity is constant over the entire storm duration

2) The rainfall is distributed uniformly over the water shed

3) The maximum runoff occurs when runoff from the entire water shed is contributing at the outlet

4) The peak runoff rate is some fraction of the rainfall intensity

5) The water shed system is linear

To develop a precipitation model in GIS the precipitation gauges were imported into GIS and converted to a raster. The raster calculator was then used to determine the runoff as the area multiplied the appropriate precipitation intensities and the runoff coefficient. The resulting flow rates (in cfs) in each catchment of the water shed are shown in Figures 11 and 12. Figure 11 shows the flow rates for the 100 year 1 hour storm and Figure 12 shows the flow rates for the 100 year 24 hour storm.

Figure 11: Catchment flow rates for the 100 yr 1 hr representative storm.

Figure 12: Catchment flow rates for the 100 yr 24 hr representative storm.

Once the flows in each catchment were calculated using the raster calculator the out flow hydrographs were calculated and routed through the channels to determine the maximum flow at the outlet of the water shed. The maximum flow will occur when all catchments are contributing to the runoff at the outlet. The results are shown below.

HECHMS model

All data used to construct and model the water shed in HEC-HMS was gathered from the GIS base map described above. The model was set up to represent the watershed to as close as possible as shown below in Figure 13. Note that catchments 1 and 4 were combined in the HMS model because catchment 4 is so small and their precipitation values are very similar.

Figure 13: HMS model screen shot

Because of lack of hydrologic data the watershed, the SCS Curve Number method was used to compute the runoff hydrograph in HEC-HMS. Other methods for computing the runoff are available in the HMS software; however the SCS method is well suited for both canyons. The method requires a representative Curve Number and basin lag time (time from the centroid of the rainfall to the peak of the runoff). The HMS software requires a lag time for each basin or sub-basin to be computed. The lag time in hours was computed for each contributing basin using the following equation that was obtained from the 1987 edition of the United States Bureau of Reclamations “Design of Small Dams” which provided typical lag times for mountain basins.

where L is the length of the main discharge channel in miles, L_c is the length from the outfall to a point perpendicular to the basin’s center of mass in miles, and S is the slope of the basin in feet per mile. Table 2 shows the characteristics of each drainage basin that were used in the HMS model. Notable is the fact that some of the basins are fairly steep and have very short lag times. Another important note is that the drainages are long and narrow which results in flood peaks arriving quickly, being much higher than anticipated, and having short durations.

Table 2: Basin characteristics.

Basin (#)	Area (sq. miles)	SCS Curve No. (#)	L (mi)	Lc (mi)	S (ft/mile)	Lag (hrs)
1	1.69	70	2.08	1.01	1173.7	0.72
2	0.56	70	0.92	0.74	2024.8	0.45
3	0.87	70	1.29	0.81	1456.3	0.55
4	2.76	70	1.90	1.17	348.5	0.89
5	1.91	70	2.42	1.65	1261.9	0.87
6	2.01	70	1.78	1.30	1295.6	0.73
7	8.21	70	6.12	4.12	257.5	2.09
8	5.03	70	5.29	3.78	437.1	1.77

Flood Routing

In addition to generating the runoff associated with each storm, the HMS software is capable of routing the runoff through reaches of each channel. As shown in Table 1, some of the drainage areas have significant channel lengths the flood must be routed through. The Muskingham method for routing the flood through the catchment channels was used to perform this task in HMS. The Muskingham parameters are given in Table 3 and were used in HMS to route the flood through the basins where significant reaches were present. Information for flow velocities in the channels was taken from “Applied Hydrology” by Chow, Maidment, and Mays.

Table 3: Muskingham flood routing parameters for each significant reach.

Reach - Basin	Muskingham X	Muskingham K
1 - 7	0.2	1.92
2 - 4	0.2	0.52
3 - 3	0.2	0.06
5 - 8	0.2	1.06

Once correct information including lag times, initial abstractions, rainfall hyetographs, and Muskingham routing parameters were entered into HEC-HMS, the model was ran and the results are presented below.

Discussion and Results

The results of the flow models are shown below in tables 4 and 5. As can be seen the flow rates for the 24 hour storm compare much better than the 1 hour storms. This is likely due to the routing of the one hour storms in the rational method model. It should also be noted that the linear storm distribution is not very representative for the 24 hour storm event.

Table 4: Flow rates calculated in the GIS model.

100-Year Storm Duration

(hr)

Peak Flow SCS Distribution

(cfs)

Peak Flow Linear Distribution

(cfs)

715

5,580

610

Table 5: Flow rates calculated in the HMS model.

100-Year Storm Duration

(hr)

Peak Flow SCS Distribution

(cfs)

Peak Flow Linear Distribution

(cfs)

479

483

5,813

2,372

Conclusions

A precipitation-runoff model was built using GIS and HEC-HMS to determine runoff resulting from 100 year precipitation events in the Smithfield watershed consisting of Smithfield and Birch Creek Canyons. In these models two precipitation events were modeled, including the 100 year 1 hour and 100 year 24 hour storms. Precipitation Hyetographs were created for both storms using a linear distribution and a more representative distribution obtained from the state engineer’s office. The maximum flow in this system happened during the 24 hour storm event where the calculated flows were greater than 5000 cfs. A small culvert analysis has shown that the maximum flow that can pass through the city of Smithfield in Summit Creek without flooding is 383 cfs. This implies that there will be flooding the case of any 100 year storm event and that Smithfield should look for a new site for their city offices.

References

McCuen, Richard H., (2004). “Hydrologic Analysis and Design”, Third Ed. Upper Saddle River, New Jersey.

National Hydrography Data Set Plus (NHDPlus) (accessed 2008, October 6) “Great Basin Hydrologic Region 16” http://www.horizon-systems.com/nhdplus/HSC-wth16.php

National Oceanic and Atmospheric Administration (NOAA) (accessed 2008, November 4) “Hydrometeorological Design Studies Center Precipitation Frequency Data Server”. http://hdsc.nws.noaa.gov/hdsc/pfds/sa/ut_pfds.html

USGS (United States Geological Survey) (Accessed 2008, December 1)a. The National Map Seamless Server. http://seamless.usgs.gov/website/seamless/viewer.htm