Farmington Canyon Sediment Production

Iosefa Matagi

December 4, 2003

Introduction

The question of sediment production is fairly new to me and my limited knowledge. I have found this question to be both intriguing and a good subject to research a total annual and volumetric yield for Farmington Canyon. I have some background in the range of sedimentation engineering and plan to use that background and newly accrued knowledge of the Geographic Information System to help in the calculations of yield. This drainage has a history of producing high levels of sediment yield to the point where this yield becomes a debris flow. The location of this drainage is shown below. This research project has been directed toward a group of four and each member has been given a topic that will significantly or insignificantly affect the yield.

Figure 1: Location of Farmington Canyon.

Objective

The objective of this research is to roughly find out the quantitative sediment yield that may be discharged out of the Farmington Canyon drainage. With this knowledge we can apply engineering techniques that may help in the sediment load of the river itself. There will be different parameters that may have to be taken into consideration and will be discussed later in this paper.

Problems

There are a few problems that are related to sediment production that have occurred in this drainage. In July 2003, there was a small fire covering around 1935 acres located fairly close to the mouth of the canyon. This fire will decrease the amount of cover and vegetation that is available for erosion control. There may be a small amount of hydrophobicity involved yet judging by the photos available below, the severity was not high enough for soil armoring to occur.

Figure 2: The Farmington fire of 2003 (Utahfireinfo 2003)

Other problems in this drainage that may contribute to the sediment production can be the presence of mountain roads that are ranked by the forest service to be level four mountain roads. These roads may produce a considerable amount of sediment when a rainstorm or even snow melt runoff is introduced to it. The pictures depict the location of the roads are within the riparian zone and also in open areas crosscutting the hillslope. Roads are the number one pollutant of mountain streams.

Figure 3: Roads in Farmington Canyon

Assumptions and Parameters

The consideration of different parameters includes:

· The calculations differing from a snow pack runoff to an assumed rainfall event.

· The type of yield calculation technique used to calculate yield.

· Differences between mudslides and debris flow with sediment yield.

The assumptions were very numerous yet may be understandable later in the paper. The assumptions are:

· We will be assuming that we will apply the yield techniques to a rainfall event that will be delivered over a 6-hr period with a depth of 1 inch.

· We will assume that a snowmelt flood will be in concordance with the largest peak flow in the last 50 years which will be the peak flow from the flood of 1983.

· We will also assume that there will be a peak flow of 150% of the normal runoff.

· The assumption will be made that the soil ratio that we collected by hand will represent the entire watershed as a whole.

· We assume that slope data will be the average slope over an entire calculated area.

· We will also assume that the Soil Delivery Ratio will be 1 for steep slopes.

· Finally, I will assume that the majority of the soil cover in a calculated area will represent the entire calculated area.

Yield Method

We conducted a site visit where we could perform one of the two yield techniques on this drainage. This technique is called the PSIAC method of 1968. This is named after the group who established it called the Pacific Southwest Inter Agency Committee. This method is very subjective and can be described with the following equation:

Annual Yield = .0833 e ^ (.0359 * FR)

FR = Sediment Rating Factor = sum of (9) different factors.

These factors are completely up to the surveyor who will rate different parameters of the watershed. These parameters are as follows:

· Surface Geology (ranked between 0 and 10, 0 being hard rock)

· Soil (0-10, 0 being rocky)

· Climate/Precipitation (0-10, 0 for snowmelt)

· Runoff (0-10, 0 for low peak)

· Topography (0-20, 0 for mild slopes)

· Ground Cover (-10-10, -10 for 100% protection)

· Land Use (-10-10, -10 for unused)

· Upland Erosion (0-25, 0 for no evidence of slope erosion)

· Channel Erosion (0-25, 0 for bedrock)

Five of us did the survey and after, I performed the calculations and our results are depicted in the following table.

Table 1: PSIAC results for the Farmington Canyon.

The final yield we came up with is 44,350.39 tons/year. This means that there will be that amount of sediment produced each year regardless of the event.

In the notes for Sedimentation Engineering written by Dr. William Rahmeyer, he quotes the Pacific Southwest Inter-Agency Committee Report saying that the method is intended as an aid to the estimation of sediment yield for the variety of conditions encountered in the Pacific Southwest. The classifications are intended for broad planning purposes only, rather than for specific projects where more intensive investigations of sediment yield would be required (Rahmeyer 2003).

As a result of the PSIAC method being somewhat general, to refine our research of this watershed, we have decided to institute the MUSLE equation to calculate the sediment yield pertaining to our assumed rainfall event and snow melt event. This equation is called the Modified Universal Soil Loss Equation and is stated as follows:

Ys=a*[Q*qp] ^β*K*LS*CP*SDR

Ys – Total Tons per Event

Q – Storm Runoff (acre-ft)

qp – peak runoff (cfs)

α, β – storm factors

Typical Rain Storm Snow Melt Flood

a=95 a=120

β=.3 β=.56

K – Soil Erodibility Factor

LS – Slope Factor (length and steepness)

CP – Cover and Management Practice Factor

SDR – Sediment Delivery Ratio

Process

I had to perform the majority of the GIS steps in order to obtain data to perform the MUSLE technique. Everyone in the group would obtain the data and calculate their factors and apply them to the data processing that I performed. In order to do this I had to download the necessary Digital Elevation Model and National Hydrologic Data for stream reaches. I did this using the websites shown.

Figure 4: Steps and source of data processing.

Once I obtained the data of the general area, I performed the necessary steps to preprocess the terrain. In these steps, the raster data was processed to fill the sinks, calculate the flow accumulation, and calculate the flow direction and other processing of the terrain and raster data. The step I was interested was the processing of different catchments in my Digital Elevation Model (DEM). Once I pointed out the watershed that I wished to work on (Farmington Canyon), I was able to separate it out of the rest of the data. When things were separated and I obtained a single drainage that I can work on, I used Spatial Analyst to create contours and find relative slopes around the DEM.

Figure 5: Steps of singling out the Farmington Canyon drainage.

Now that I have a singled out drainage, I can run the ArcHydro software package once again to divide that singled out drainage into separate catchments and refine the boundary of the watershed. I also used the TauDEM software package to delineate the streams into different orders to more define the flow accumulation in the drainage.

Figure 5: Steps to delineate and separate the singled watershed into catchments.

This is the final product that I planned to use for splitting up the watershed into catchments and applying the MUSLE equation to each one. I delivered this product to the rest of the group so that they can apply their different ratios to each of the separate catchments.

Equation Ratios

The different ratios that this equation entails were found by each of the groups members. The Q and qp values were found with the assumption of our rainfall event and knowledge of hydrology. Knowing this, we can incorporate the alpha and beta values to come up with the value known in other equations as R.

The K value was found after visiting the site and collecting soil samples, sieving the samples, and analyzing the data while interpolating or projecting the data throughout the entire watershed. This was explained as one of our assumptions.

The LS value was found using the spatial analyst tools to find the slope of each raster cell and then using zonal statistics to find the average slope in each of my calculated catchments. This value was assumed to represent each one of the catchments.

The SDR factor was already assumed to be 1. The reason for this was that the SDR factor had a calculation that was only useful if the slopes were somewhat mild. Our average slopes in the watershed exceeded our perception of a mild slope and therefore the SDR can be assumed to take on the value of 1.

The CP Ratio

Michigan State University says Surface cover is material in contact with the soil surface that intercepts raindrops and slows surface runoff. The total percent of the surface covered is the characteristic used by MUSLE to compute how surface cover affects erosion. Surface cover includes all cover that is present, including rock fragments, live vegetation, cryptograms, and plant residue. The only minimum size requirement for material to be counted as surface residue is that it either be of sufficient size or attached to the surface such that is not removed by runoff (MSU 2003).

MUSLE accounts for surface roughness in the C value calculation. The surface roughness ponds water in various depressions while reducing erosivity of raindrop impact and water flow. If the depositions are sufficiently deep, much deposition occurs in them. Over time, roughness disappears as the depressions fill with sediment and the soil subsides after the tillage operations that formed the depressions (MSU 2003).

The cover control factor is applied in the equation as a single unit. It accounts for the effects of all erosion control measures that may be implemented on any particular site, including vegetation, mechanical manipulation of the soil surface, chemical treatments, etc. It does not include structures such as berms and ditches. These are part of the topographic factor, LS (Rahmeyer 2003).

Standing vegetation exerts its influence on the CP factor in proportion to its aerial density and type of root system. Apparently all grasses that are suitable for erosion control and adapted to the site can be grouped together as can all forbs such as legumes, weeds, etc (Rahmeyer 2003).

Structures such as sediment traps and settling basins do not decrease erosion but serve only to catch the sediment after it has left the source area. The amount of sediment captured in such structures can be measured or calculated and subtracted from the total soil loss, calculated by the equation, to determine actual loss (Rahmeyer 2003).

On our site visit, I took pictures of the vegetation and the differences between the south and north facing slopes. On the south facing slope, there was minimal canopy cover and large amounts of area that included grasses and forbs. On the north facing slope there was considerable canopy cover and little grass or forbs.

Figure 6: Vegetation in Farmington Canyon

Figure 7: National Land Cover Data

To manipulate the data that I acquired from the seamless data site with the USGS, I first projected that data on to the catchments that I processed earlier. This data had too many options with regard to land cover to manipulate. As a result of this, I had to reclassify the data using the spatial analyst tools. I broke this up in a manner that the major land cover uses that existed in this watershed were visible. Doing this made the manipulation more convenient and user friendly. With this data, I now knew the value of the land cover but I did not know what that value meant. The USGS site had values and names for these values.

Figure 8: Value of the NLCD and its’ catchment.

Table 2: Names for the Values of NLCD.

11 Water

12 Perennial Ice Snow

21 Low Intensity Residential

22 Hi Intensity Residential

23 Commercial/Industrial/Transportation

31 Bare Rock

32 Quarries/ Mines

33 Transitional

41 Deciduous Forest

42 Evergreen Forest

43 Mixed Forest

51 Shrub land

61 Orchards/ Vineyard

71 Grasslands/Herbaceous

81 Pasture/Hay

82 Row Crops

83 Small Grains

84 Fallow

85 Urban/Recreational Grasses

91 Woody Wetlands

92 Emergent/Herbaceous Wetlands

Now that I know the names for the values in the ArcGIS, I can now use the zonal statistics to find out the value of the majority land cover existing in each catchment. Running this program separated the data and I could then export that data into a spreadsheet where a list of the catchments and a list of their majority land cover were presented.

Now I know the majority value of the land cover in each catchment. I also now know what that value represents. What I do not know is the CP value for the representations of the land cover values.

In order to find what the CP value was corresponding to, its value in ArcGIS, and the name given that value, I consulted a previous paper that discussed these values. The following table gives the CP value for each type of land cover depicted in the watershed.

Table 3: CP values for different types of land cover existing in the Farmington drainage (Grams 2003).

Now inputting these known CP values for each catchment into a spreadsheet facilitated the calculation of the sediment yield. The final result of collecting the CP values for each catchment is given in the following table.

Table 4: Results of the CP value data collection through manipulation in GIS.

GRIDID	Area (acres)	MAJORITY	Type of land cover	Cp Value
17	540.583454	51	Shrubland	0.06
18	451.6267805	51	Shrubland	0.06
34	196.2631126	51	Shrubland	0.06
35	237.7949223	51	Shrubland	0.06
36	139.1078576	51	Shrubland	0.06
39	97.79733175	51	Shrubland	0.06
40	101.5779975	51	Shrubland	0.06
41	149.5601863	42	Evergreen Forest	0.004
43	42.47696007	42	Evergreen Forest	0.004
44	229.3993074	21	Low Intensity Residential	0.24
45	120.7600854	43	Mixed Forest	0.004
47	21.79410096	51	Shrubland	0.06
49	211.7745336	42	Evergreen Forest	0.004
50	165.1272175	42	Evergreen Forest	0.004
51	170.3538159	43	Mixed Forest	0.004
52	65.0506986	42	Evergreen Forest	0.004
53	101.2444642	41	Deciduous Forest	0.004
55	248.1918335	51	Shrubland	0.06
56	258.088477	42	Evergreen Forest	0.004
57	128.0434939	51	Shrubland	0.06
61	109.4738591	51	Shrubland	0.06
65	97.68643268	42	Evergreen Forest	0.004
66	53.81915028	43	Mixed Forest	0.004
67	56.54344508	41	Deciduous Forest	0.004
69	149.4488692	42	Evergreen Forest	0.004
70	66.66236903	41	Deciduous Forest	0.004
74	188.3675083	51	Shrubland	0.06
80	69.8319965	42	Evergreen Forest	0.004
81	132.3244275	42	Evergreen Forest	0.004
84	126.6533002	42	Evergreen Forest	0.004
85	203.1565095	41	Deciduous Forest	0.004
90	96.40765252	42	Evergreen Forest	0.004
91	38.08496657	43	Mixed Forest	0.004
100	179.8612836	42	Evergreen Forest	0.004
101	231.5677561	42	Evergreen Forest	0.004
106	215.444043	42	Evergreen Forest	0.004
107	101.0221515	41	Deciduous Forest	0.004
117	200.2101271	51	Shrubland	0.06
118	196.3183371	41	Deciduous Forest	0.004
127	10.3413331	41	Deciduous Forest	0.004
128	170.6318675	51	Shrubland	0.06
129	234.0137421	41	Deciduous Forest	0.004
131	107.6392651	41	Deciduous Forest	0.004
135	208.7161589	41	Deciduous Forest	0.004

Conclusion

The final result of this project was to input all of the factors that we have all calculated for each of the catchments that I separated and run them through the MUSLE equation. This was done using a spreadsheet that is shown below in Table 5. Inputting all of the factors in the MUSLE equation and producing a volumetric yield for each catchment allowed us to sum all of those values to produce a total volumetric yield for our assumed rainfall event.

We found that for the assumed rainfall event, we would find that the sediment yield would be 8,539.48 tons of sediment. We also found that for the snowmelt flood, the sediment yield would be 10,372.57 tons of sediment.

It was also found that the amount of sediment yielded would change between catchments. The south facing, less canopy cover side of the watershed would yield far more sediment than the established, north facing side. This is somewhat of an expected find.

These numbers represent the amount of sediment that could be removed from the drainage under the assumed events and does not represent an annual yield such as the PSIAC method does. Such results do not take into account of the streams bed load capacity or suspended load capacities. Further research can take these parameters into account and then differentiate between a debris flow and a sediment yield.

For an in-depth look at the other MUSLE factored ratios please click on the links below.

· Q, qp, (R); Ryan McBride

· LS, SDR; Jeff Jensen

· K; Brad Taylor

Table 5: Results of the MUSLE equation with all factors of each catchment.

				Snowmelt Flood			Rainfall Event
Catchment	LS	K	CP	Q	q	Total Sediment	Q	q	Total Sediment
GridID				(acre-ft)	(cfs)	(tons/event)	(acre-ft)	(cfs)	(tons/event)
17	124.5	0.09	0.06	46.06	15.23	576.19	13.51	40.88	2192.95
18	74.3	0.09	0.06	38.48	12.72	308.68	11.29	34.15	1069.99
34	143.5	0.08	0.06	16.72	5.53	321.25	4.91	14.84	722.01
35	116.0	0.09	0.06	20.26	6.70	327.94	5.94	17.98	814.38
36	110.4	0.08	0.06	11.85	3.92	201.16	3.48	10.52	378.04
39	142.4	0.08	0.06	8.33	2.75	209.89	2.44	7.40	328.44
40	126.0	0.08	0.06	8.65	2.86	190.06	2.54	7.68	303.34
41	152.6	0.08	0.004	488.70	161.55	172.65	2.12	6.41	20.00
43	84.2	0.08	0.004	561.80	185.72	103.57	0.60	1.82	2.69
44	29.2	0.08	0.24	590.00	195.04	2216.98	3.25	9.83	370.25
45	120.6	0.09	0.004	429.89	142.11	142.15	1.71	5.18	13.99
47	91.5	0.09	0.06	405.53	134.06	1562.28	0.31	0.93	23.40
49	133.3	0.08	0.004	546.32	180.60	161.18	3.00	9.08	25.78
50	104.9	0.07	0.004	14.07	4.65	12.35	2.34	7.08	13.43
51	123.0	0.08	0.004	519.94	171.88	144.37	2.41	7.30	18.64
52	123.0	0.07	0.004	383.41	126.75	105.28	0.92	2.79	5.55
53	111.4	0.09	0.004	339.38	112.19	113.87	1.43	4.34	10.60
55	116.9	0.08	0.06	308.76	102.07	1505.72	6.20	18.77	765.23
56	130.8	0.07	0.004	21.99	7.27	20.14	3.66	11.06	27.62
57	98.9	0.08	0.06	10.91	3.61	171.45	3.20	9.68	308.63
61	49.5	0.09	0.06	9.32	3.08	87.85	2.74	8.28	145.78
65	135.7	0.08	0.004	8.32	2.75	13.33	1.38	4.19	11.04
66	81.3	0.08	0.004	279.28	92.32	65.78	0.76	2.31	3.39
67	84.2	0.08	0.004	263.78	87.20	65.80	1.41	4.28	7.01
69	127.4	0.08	0.004	12.73	4.21	16.15	2.12	6.40	16.67
70	80.5	0.08	0.004	246.23	81.40	60.36	0.94	2.86	4.27
74	85.1	0.08	0.06	182.75	60.41	800.37	4.71	14.25	409.11
80	185.7	0.09	0.004	157.37	52.02	119.74	1.75	5.28	22.03
81	85.7	0.09	0.004	57.79	19.10	30.30	1.87	5.67	11.01
84	47.1	0.08	0.004	134.11	44.33	24.52	1.79	5.43	5.12
85	31.4	0.09	0.004	17.31	5.72	5.39	2.88	8.71	6.53
90	118.7	0.08	0.004	8.21	2.71	11.57	1.37	4.13	9.51
91	139.4	0.08	0.004	38.30	12.66	34.23	0.54	1.63	3.95
100	86.7	0.08	0.004	15.32	5.07	12.29	2.55	7.71	13.97
101	91.1	0.09	0.004	19.73	6.52	16.90	3.28	9.92	21.91
106	55.8	0.07	0.004	18.35	6.07	7.71	3.05	9.23	9.62
107	112.7	0.07	0.004	104.96	34.70	44.33	1.43	4.33	8.33
117	31.8	0.06	0.06	17.06	5.64	53.98	2.84	8.58	64.89
118	36.7	0.15	0.004	79.29	26.21	26.17	2.78	8.41	12.24
127	13.7	0.15	0.004	27.83	9.20	5.22	0.15	0.44	0.17
128	105.4	0.1	0.06	14.54	4.81	271.21	2.42	7.31	300.01
129	44.0	0.12	0.004	19.94	6.59	10.96	3.32	10.03	14.29
131	34.7	0.15	0.004	9.17	3.03	6.78	1.52	4.61	5.90
135	49.8	0.15	0.004	17.78	5.88	14.46	2.96	8.94	17.76
				Snowmelt			Rainfall
				Total Yield	10372.57	tons	Total Yield=	8539.48	tons

Sources

Michigan State University; http://www.iwr.msu.edu/rusle/cfactor.htm Nov. 19, 2003

Grams, Paul; http://moose.cee.usu.edu/giswr/archive99/termp/grams/tp.html Nov. 19,

2003

UtahFireInfo; http://www.utahfireinfo.gov/previousfires/farmingtonfire.htm Nov. 14,

2003

UtahFireInfo; http://www.utahfireinfo.gov/firephotos/index.htm Nov. 14, 2003

Dr. Rahmeyer, William; Sedimentation Engineering Volume 2, spring 2003

Dr. Tarboton, David; Personal Contact, Oct. 1- Dec. 2, 2003