Historical Runoff Potential in Eastern Cache Valley

Pre-Development and Post-Development Runoff Comparisons

Term Report

Eric Major

GIS in Water Resources 2010

Utah State University

Created For Dr. David G. Tarboton,

Dr. David R. Maidment, and Dr. Ayse Irmak

3 Dec 2010

Objective

The Eastern portions of Cache Valley have experienced both periods of tremendous growth and steady development over the last forty years. As the cities along the benches expand, there comes a need to manage stormwater and implement design standards to meet city specifications and future stormwater system requirements. With strict requirements on the doorstep by way of water quality in Cutler Reservoir and Bear River tributaries, it is important for cities in the Cache Valley to have standards and master plans in place for stormwater (Shaw 2010). Often, a development is required to find the historical or pre-development runoff from a site to mitigate stormwater impacts with new impervious areas. Requirements vary by city, but knowing the historical land cover, runoff ratios, and post-developed runoff is valuable for planning and mitigation measures. This project explores several methods to establish the current pre-development runoff potential using ArcGIS for the developed areas of Eastern Cache Valley, from Paradise to Cove as shown in Figure 1. It is the hope to develop a dataset that will be useful for the cities under study for an aide in stormwater master planning, implementing development requirements, and reviewing submitted plans for stormwater utility design and approval.

Figure 1: Study Area of Project in Eastern Cache Valley

Terrain Analysis and Stream Delineation

The datasets that were utilized in this project include the 30-meter National Elevation Dataset (NED) and the National Land Cover Datasets from USGS Seamless Server. A shapefile dataset obtained from the National Resource Conservation Service (NRCS) of the irrigated and agricultural areas, and stream flowlines, waterbodies, and watershed boundaries from NHD Plus data shown below in Figure 2. The two watersheds included in the study were the Middle Bear to the North and the Little Bear on the South, shown with outlines in red below.

Figure 2: DEM and Flowlines in Watersheds in Cache Valley (NHDPlus 2010)

Before adding any land cover or soil type data and methodology, the DEM and NHD Flowlines were clipped by the Extract by Mask tool in GIS using both watersheds as the boundary. Once these data were clipped to a smaller area, the DEM reconditioning could take place with an ultimate goal of defining streamlines and catchments. The NHD Flowlines were inspected and compared to base maps and were found to have gaps and assumptions that did not work well with this analysis. For example, several of the canal systems were shown as stream lines that had missing reaches, dead ends, and disconnected streams with their actual destinations. Therefore, the DEM was used to define stream lines using the following steps illustrated by layouts showing each layer and feature class produced in ArcMap 10.

DEM reconditioning began by the Fill function to remove pits and then followed by the Flow Direction tool to assign a flow direction to the grid. From this Flow Direction raster, the Flow Accumulation tool in the Spatial Analyst Tool set was used to calculate cells and areas flowing to each other. Figure 3 shows the Flow Accumulation raster with lighter colored cells having a higher value than the dark or no flow cells.

Figure 3: Flow Accumulation Raster Example

Once Flow Accumulation was completed, Streams were defined using the Raster Calculator tool and a threshold of 1000 by the map algebra expression (“FlowAccumulation”>1000) creating a raster named Stream1000. The delineated streams were then divided into segments by the Stream Link tool with an unique value for each link that would later be used to define catchmenst. Stream networks and stream links in the study area are sampled below in Figure 4 and unique catchments found by the Watershed tool are shown in Figure 5.

Figure 5: Catchment Rasters

Figure 4: Stream Networks and Links

It was observed that in the flatter areas were the majority of the development was taking place had catchments that were very long and not very wide. This is obviously not exactly correct as canals and storm water systems bisect these catchments. More discussion on this observation will follow later in the report. The catchment raster was used to create polygons of each catchment with a unique identifying number or ID. The Raster to Polygon tool created catchments shown below in orange in Figure 6. The both watersheds held part of the study area, so only a selection of Catchment Polygons inside of the project area were used in further runoff calculations.

Figure 6: Study Area Catchment Polygons

Soil Data

Now that a catchment data set was selected for study and identifying ID numbers and areas were available in attribute tables, more data about the specific land uses, impervious areas, and soil types could be added to the map. The soil data shown in Figure 7 was obtained through the NRCS Soil Data Mart and was interpreted using a Web Soil Survey also from the NRCS. It was useful to know the soil descriptions and abbreviations to link data populated in attribute tables in GIS with Hydrologic Soil Group classifications given in the Custom Soil Report (NRCS 2009). These classifications were necessary for SCS runoff calculations and Curve Number development. Table 1 shows a few samples of soil types and descriptions with the assigned Hydrologic Soil Group used in the SCS Curve Number calculation.

Figure 7: Study Area Soil Data Raster (NRCS 2010).

Table 1: Soil Type Abbreviations, Descriptions, and Hydrologic Soil Groups

Land Cover Data

Using data obtained from the USDA – NRCS Geo-Spatial Data Gateway, a raster from the National Land Cover Data set was imported into the map with a 30 by 30 meter cell size (Homer 2004). Figures 8 below shows the raster and accompanying legend for land cover in the study area.

Figure 8: 2001 National Land Cover Dataset for Utah (Homer 2004).

Once land cover classification was available, a determined hydrologic condition was matched to the different cover types. Table 2 contains the SCS Curve Numbers assigned to each land cover type, with the corresponding ID found in each grid cell attribute (NRCS 1986). As discussed in the previous section, the Hydrologic Soil Group (HSG) give a varied Curve Number for each cover type and is important in accurately calculating the Curve Numbers for the SCS runoff calculations.

Table 2: Runoff Curve Numbers for NLCD Land Types and Hydrologic Soil Group

GIS Methods & Curve Number Calculations

After catchments were defined with soil data and land cover data imported and classified, GIS tools were utilized to assign a Composite Curve Number (CCN), or weighted average Curve Number for each catchment. For the SCS Method calculations, a CCN is necessary to find runoff potential for a given precipitation event. The first GIS tool useful for summing all land cover and soil attributes per catchment was the Tabulate Area tool, with Figure 9 shown below as an example of input rasters, fields, and class fields. Using the Catchment ID as the identifier, the output table given by this tool gave values for the area of each land cover type (or soil type) in the catchment. Figures 9 and 10 are examples of land cover tables and Figures 11 is similar for the soil data.

Figure 9: Tabulate Area Tool Inputs for Land Cover Areas

Figure 10: Output Table for Land Cover Values per Catchment

Figure 11: Output Table for Soil Type Values per Catchment

Composite Curve Number Calculation

Using the output tables for both land cover and soil values in each catchment, the data was sorted in Excel. For this report, an example calculation in Catchment 2597 is used to illustrate the methodology of the CCN calculations for the SCS method (NRCS 1986). Following the direction of the NRCS TR-55 publication, a HSG was assigned by summing the areas of soils falling in each group. Table 3 shows the values given in the output table for this catchment with the total area for each group at the bottom. Using the maximum value as the determining factor, an HSG was assigned to the catchment. In this case, HSG B was the dominant group and is then used in Table 4 to multiply the correct row of Curve Numbers with the areas of each land cover group. These products are then summed and divided by the total area of the catchment to find the weighted CCN. In this case, Catchment 2597 is assigned a CCN of 73.

Table 3: Example Hydrologic Soil Group Areas for Catchment 2597

Table 4: Example CCN Calculation by Land Cover for Catchment 2597

SCS Runoff Calculations

The Soil Conservation Service (SCS) method for runoff calculation in English units is portrayed below in Figure 12 with the given parameters from the TR-55 report. So far, all the inputs are defined for this study area except precipitation.

Figure 12: SCS Equation for Runoff Calculation (NRCS 1986)

To conform to the design standards of the study area, the Northern Cache Valley Storm Water Design Standards (LCPW 2009) document was used to find required storm durations and return periods. For these specifications adopted by several Cache Valley cities, a return period of 100 years and duration of 24 hours is standard. Table 5 is a reproduction of the depth-duration summary table from the KVNU Logan Station for NOAA Atlas 14 precipitation. The depth of 3.02 inches was used for runoff depth computations.

Table 5: Precipitation Depth per Northern Cache Valley Design Standards (LCPW 2009)

In the Excel spreadsheet, the calculations for CCN, SCS runoff, and volumes generated from 10, 25, 50, and 100 year storms were arranged in a simple table for export back into ArcGIS as shown in Figure 13 below.

Figure 13: Excel Table for Export into ArcGIS

GIS Calculations & Tabular Joins

With valuable runoff data linked to catchments in the study area, the data in Excel was imported into the GIS project and a tabular join with the attribute table of the Catchment features. Figure 14 shows the Field Calculator used to populate the values in the attribute table fields. Figure 15 shows the final product, an attribute table with values for CCN, runoff for each return period, and a field with runoff volume (depth multiplied by catchment area) for the 100 year event.

Figure 14: Field Calculator for Join Data Population

Figure 15: Complete Attribute Table for Catchments

Results

The computed data for runoff potential can be easily shown in GIS by creating attribute labels for displaying CCNs (Figure 16), runoff depths, and volumes. Rasters were created for runoff depths in each catchment for 10, 25, 50, and 100 year events, with a sampling shown in Figures 17 and 18. The volume of the 100 year event was also reproduced as a raster by the Polygon to Raster tool and is given in Figure 19. The depths ranged from zero to 3 inches of runoff, which is expected as impervious areas have little infiltration and high CCNs.

Figure 16: CCN Values for Study Area Catchments

Figure 17: 10 Year Event Runoff Raster

Figure 18: 100 Year Event Runoff Raster

Figure 19: 100 Year Event Runoff Volume Raster

Observations & Conclusions

The runoff potential for the current land cover and developed areas has been calculated using the SCS Method for runoff and CCNs have been designated for each catchment area in the eastern areas in Cache Valley. This project has brought out some useful applications of ArcGIS tools for a runoff model study. With several points discovered about the model, it is the hope to refine the catchments by possibly eliminating the long and thin catchments using canals to reflect actual boundaries in the field. It was also discovered that catchments vary greatly in land cover when stretching East and West. Future work would eliminate catchment areas and use real land parcels to designate CCN and runoff from one area. This would suit developments more accurately and allow regulatory officials the ability to check the existing or historical runoff potential by parcel. Another variable that could be defined better is the CCN calculation and using a more accurate method to define HSG and Curve Number values for the areas. In research for this method, an application called HEC-GeoHMS was discovered that allows one to calculate a CN grid by inputs of land and soil types (Merwade 2010). It is also of great interest to find actual streamflow data to calibrate the model by runoff generated by actual storms. This would refine the CCNs to reflect actual runoff seen in a measured event.

The current results reflect a trend from runoff depths increasing in developed areas, which can be expected. The land cover data from 2001 is in need of an update and will soon be available through the National Land Cover Dataset. Comparisons between growth in these past 10 years would help illustrate the increase in runoff due to development and more impervious areas. The project is definitely a work in progress and has exciting possibilities for those involved. The process and tools for the runoff calculations in ArcGIS will allow for manipulation of this model to become a valuable asset for stormwater planning and design.

References

Homer, C. C. Huang, L. Yang, B. Wylie and M. Coan. 2004. Development of a 2001 National

Landcover Database for the United States. Photogrammetric Engineering and Remote Sensing, Vol. 70, No. 7, July 2004, pp. 829-840.

Logan City Public Works (LCPW). 2009. “Northern Cache Valley Storm Water Design

Standards”. November 18, 2010. http://loganutah.org/public_works/Engineering/stdsspecsdesign.cfm

Merwade, Venkatesh. 2010. Creating SCS Curve Number Grid using HEC-GeoHMS. School of

Civil Engineering, Purdue University. November 18, 2010. http://web.ics.purdue.edu/

~vmerwade/education/cngrid.pdf

National Hydrography Dataset Plus (NHDPlus). 2010. Great Basin (Hydrologic Region 16).

October 2010. http://www.horizon-systems.com/nhdplus/HSC-wth16.php

Natural Resources Conservation Service (NRCS). 2010. Soil Data Mart. Soil Types for UT603:

Cache Valley and Box Elder County. November 18, 2010. http://soildatamart.nrcs.usda.gov/Download.aspx?Survey=UT603&UseState=UT

Natural Resources Conservation Service (NRCS). 2009. Custom Soil Resource Report for Cache

Valley Area, Parts of Cache and Box Elder Counties, Utah. November 18, 2010. http://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx

Natural Resources Conservation Service (NRCS). 1986. Urban Hydrology for Small Watersheds.

Technical Release 55. June 1986. pp. 1-164.

Shaw, Chuck. Logan City GIS Division Head. Personal Communication. October 2010