USGS - science for a changing world

Water Science for Maryland, Delaware and the District of Columbia

Home >> Publications >> Online Publication - WRIR-99-4240 - Online Report
Loch Raven

Sediment Accumulation and Water Volume in Loch Raven Reservoir, Baltimore County, Maryland

By William S. L. Banks and Andrew E. LaMotte

Introduction

Baltimore City and its metropolitan area are supplied with water from three reservoirs, Liberty Reservoir, Prettyboy Reservoir, and Loch Raven Reservoir. Prettyboy and Loch Raven Reservoirs are located on the Gunpowder Falls (figure 1). The many uses of the reservoir system necessitate coordination and communication among resource managers. The 1996 Amendment to the Safe Drinking Water Act require States to complete source-water assessments for public drinking-water supplies. As part of an ongoing effort to provide safe drinking water and as a direct result of these laws, the City of Baltimore and the Maryland Department of the Environment (MDE), in cooperation with other State and local agencies, are studying the Gunpowder Falls Basin and its role as a source of water supply to the Baltimore area. As a part of this study, the U.S. Geological Survey (USGS), in cooperation with the Maryland Geological Survey (MGS), with funding provided by the City of Baltimore and MDE, is examining sediment accumulation in Loch Raven Reservoir.

The Baltimore City Department of Public Works periodically determines the amount of water that can be stored in its reservoirs. To make this determination, field crews measure the water depth along predetermined transects or ranges. These transects provide consistent locations where water depth, or bathymetric, measurements can be made. Range surveys are repeated to provide a record of the change in storage capacity due to sediment accumulation over time. Previous bathymetric surveys of Loch Raven Reservoir were performed in 1943, 1961, 1972, and 1985. Errors in data-collection and analysis methods have been assessed and documented (Baltimore City Department of Public Works, 1989). Few comparisons can be made among survey results because of changing data-collection techniques and analysis methods.

Purpose and Scope

This report presents maps of sediment thickness and water depth, and quantifies water volume in Loch Raven Reservoir. Maps of sediment thickness are based on bathymetric measurements made in 1998 compared to information available from 1913 topographic maps. Water depths are reported in ft (feet) below pool level. Pool level represents the maximum water-surface elevation the reservoir was designed to maintain and is the elevation of the spillway of the dam -- 240 ft m.s.l. (above mean sea level). Water volume was determined on the basis of 1998 bathymetry and reported in gallons.

Background

Residents of Baltimore City and Baltimore County, Md., have used the Gunpowder Falls as their primary water supply since 1873. For the past 117 years, since the construction of the first of three dams, this resource has been developed to provide water to residents of Baltimore City and adjacent parts of Baltimore, Howard, Harford, and Anne Arundel Counties.

Looking East Across Gunpowder Falls at Merrymans Mill Road Bridge During Construction of Concrete Gravity Dam, 1912.

Loch Raven Reservoir was created in 1881 with the construction of a masonry dam. This original structure had a pool level of 180 ft m.s.l., but by the turn of the century sediment accumulation and a growing population compelled the City of Baltimore to increase its water-supply holdings. Freeman and Stearns (1910) investigated impoundments in several local watersheds but concluded that further development of the Gunpowder Basin would meet the City's long-term water needs and would be cost effective. Volumetric calculations made in 1910 showed that flooding the valley of the Gunpowder River to 240 ft m.s.l. would impound about 23.7 Ggal, or billion gallons, of water (Freeman and Stearns, 1910). In 1914, a concrete gravity dam was constructed about 2,100 ft upstream from the original masonry dam, which greatly enlarged the reservoir (figure 2). This new dam was raised to its current height of 240 ft m.s.l. in 1923.

Description of Gunpowder Falls Basin

Figure 1. Location of Study Area and Reservoir Drainage Basin, Baltimore and Surrounding Counties, Maryland.

The Gunpowder Falls Basin drains parts of the Piedmont and Coastal Plain Physiographic Provinces in north-central Maryland and a small part of southern York County, Pennsylvania (figure 1). The basin has an area of approximately 350 square miles and trends from the northwest to the southeast (O'Bryan and McAvoy, 1966). The 30-year average precipitation for Baltimore, Md. (located about 5 miles south of Loch Raven Reservoir), is 40.76 inches per year (National Oceanic and Atmospheric Administration, 1997). Activities in the basin have included paper manufacturing, logging, and agriculture (O'Bryan and McAvoy, 1966). Today, most of the basin is lightly wooded or has been cleared of vegetation for agricultural activities and to support housing for the suburbs of Baltimore City. The city-owned property surrounding the reservoir is protected from development and serves as a buffer between a major metropolitan area and its water supply. Loch Raven Reservoir is 7 miles west of the Chesapeake Bay and lies entirely within the Piedmont Physiographic Province. The Piedmont is an area underlain by crystalline rocks and located east of the Appalachian Mountains and west of the Coastal Plain.

Acknowledgments

William Stack and Eugene Scarpola of the City of Baltimore were instrumental in providing historical documents, property access, and logistical support. Randall Kerhin and Richard Ortt, Jr. of MGS collected all bathymetric data and provided logistical and technical support. Editorial, layout, and graphics assistance by Valerie Gaine, Jean Hyatt, and Lonnie Lanham of USGS.

Study Approach

Aerial photographs provided by the Maryland Department of Natural Resources (MDDNR) were used to define the shoreline of the reservoir. Topographic maps produced in 1913 by the City of Baltimore showing parts of the Gunpowder Falls Basin were used to create a reference surface for sediment accumulation. Bathymetric data collected in 1998 by MGS were used to create a representation of the current reservoir bottom surface. The shoreline and the 1913 topography were digitized and registered to the Universal Transverse Mercator (UTM) coordinate system, zone 18, North American Datum of 1983 (NAD83) through the use of a geographic information system (GIS). The 1998 bathymetric data were imported directly into the GIS. The GIS was then used to create surfaces representing the 1913 topography and the 1998 bathymetry, bounded by the shoreline. These surfaces consist of an array of regularly spaced elevations referred to as a grid. The grids of the 1913 and 1998 surfaces were then compared to determine areas of sediment accumulation and the change in storage capacity. Change in storage capacity is derived from the difference between the potential storage capacity calculated from the 1913 topography and the current storage capacity based on the 1998 bathymetric data.

Methods of Data Collection and Compilation

Comparison of the different data sets required that they be registered to the same coordinate system. The data sets were projected to UTM coordinate system, zone 18, NAD83.

1994 Shoreline

Four aerial photographs taken on April 4 and 8, 1994 were used to represent the reservoir shoreline. These photographs, called digital orthophoto quarter quadrangles, or digital orthophoto quadrangles (DOQs), are computer-generated images of aerial photographs in which displacements caused by terrain relief and camera tilts have been removed (U.S. Geological Survey, 1997a). The DOQs meet the National Map Accuracy Standards at 1:12,000 scale for 3.75-minute quarter quadrangles (MDDNR, 1991). The shoreline was traced from the DOQs by digitizing the edge of the water. Based on records kept by the City of Baltimore, the altitude of the reservoir surface as measured at the dam on the dates the photographs were taken was 240.98 and 241.0 ft m.s.l., respectively. For purposes of shoreline delineation, the altitude of the reservoir surface was taken to e 240 ft m.s.l. This corresponds to the elevation of the dam spillway.

1913 Topography

Topographic maps were prepared by Baltimore City water engineers in 1913 for the construction of a second dam on the Gunpowder Falls to enlarge Loch Raven Reservoir. These maps display the elevation of the land surface prior to construction of the second dam, and were used to assess the potential storage capacity (figure 3). The maps were hand-drawn on 13 linen sheets at a scale of 1:3,600 with a 10-ft contour interval. The survey was based on a local coordinate system with an origin located about 80 ft south and west of the original (1881) dam. The maps show registration marks every 2,000 ft and identify many physical and political landmarks still in existence today.

Figure 3. Topographic map mosaic of Loch Raven Reservoir, 1913.

No permanent monument marking the origin of the coordinate system was found after thorough inspection of the site. Vertical control of the 1913 survey was checked at two locations -- the southwest corner of the dam and the northwest corner of the bridge at Sparks Station, Md. (figure 1). These locations represent the two most distal points on the 1913 maps and could be easily verified. The elevation of the dam spillway corresponds with the 240 ft m.s.l. noted for the proposed structure referenced by the City of Baltimore (Baltimore City Water Department, 1913) and on the USGS 7.5-minute Towson quadrangle (U.S. Geological Survey, 1974c). The elevation of the Sparks Station Bridge on the 1913 maps is 255.9 ft m.s.l. A second-order benchmark identified as JV1550 by the U.S. Coast and Geodetic Survey was used to determine that the 1998 elevation of the bridge was 256.6 ft m.s.l. Since the bridge deck had been dislodged or removed at least once since 1913, the 0.7-ft elevation difference between the 1913 measurement and the 1998 measurement is not excessive.

To create a surface that would represent the original volume of the reservoir, contour lines from the 1913 maps were digitized and attributed according to their elevation. The data were registered using 36 control points that could be matched from the 1913 linen sheets to point locations found on four USGS 7.5-minute topographic quadrangles (U.S. Geological Survey, 1974a, 1974b, 1974c, 1986), and four DOQs (MDDNR 1994a, 1994b, 1994c, 1994d).

The digital contour lines were converted into a grid of regularly spaced points that were based on hydrography and the shape and spacing of the contour lines for contours between 180 ft and 240 ft m.s.l. The 180-ft contour represents the spillway elevation of the original masonry dam built in 1881; therefore, no data were collected below that elevation in 1913. All interpolated data below 180 ft m.s.l. were converted to 180 ft. No assumptions were made as to the depth or shape of the stream channel below 180 ft m.s.l.

1998 Bathymetry

Figure 4. Bathymatric Map of Loch Raven Reservoir, 1998.

In the spring and summer of 1998, MGS conducted a hydrographic survey of Loch Raven Reservoir. Approximately 32,000 data points were collected using a fathometer operating at a frequency of 200 kilohertz. Point locations were recorded by a differentially corrected global positioning system with a horizontal accuracy of plus or minus 5 ft. The accuracy of depth measurements was 0.1 ft plus or minus 1 percent of the height of the water column. In areas where water depth was 5 ft, this represents a potential error of less than 0.2 ft. In areas of the reservoir where water depths approach 70 ft, the potential error would be slightly less than 1 ft (0.8 ft). Shallow areas, where the depth of the water was less than about 2 ft, were not measured. These areas were generally near shore and therefore considered insignificant to water volume (Ortt and others, 2000).

A data file containing latitude, longitude, and depth of water measured and adjusted to pool level to the nearest tenth of a ft was imported into a GIS. The resulting collection of points and their interpolated bathymetry were converted into a grid that was representative of the 1998 bottom surface of Loch Raven Reservoir (figure 4).

Spatial Representation of Data

This study mapped Loch Raven Reservoir from its dam to about 0.3 miles east of Merrymans Mill Road bridge (figure 1). Previous surveys of the reservoir considered the Northern Central railroad bridge (about 1 mile north of the Merrymans Mill Road bridge) as the upper extent of the reservoir. Sediment accumulation has rendered this area unnavigable; therefore, it was not surveyed.

The digitally defined shoreline provided a consistent boundary for the 1913 and 1998 surfaces. All data from the MGS 1998 survey fell within this boundary. The 240-ft contour line from the 1913 data set was also compared to the shoreline. In this case, the two were not consistent everywhere. The potential surface area of the reservoir based on the 1913 data is 2,096 acres, and the surface area of the reservoir taken from the DOQs is 2,090 acres. The discrepancy between the two represents a 0.1-percent difference in surface area and can be attributed to the use of different map scales.

Grid Generation

Grid representations of the reservoir bottom-surface elevations were created by use of a GIS. To determine the net change in sediment thickness, the 1913 interpolated surface was subtracted from the 1998 interpolated surface. Changes in bottom-surface elevation indicate a change in the quantity of sediment at a given location due to scour, displacement, or deposition. Final representation of the 1913 and 1998 bottom surfaces was achieved by using a locally adaptive approach to an iterative finite-difference interpolation gridding technique (Hutchinson, 1996). This technique allows for the use of terrain characteristics such as stream channels and the shape and spacing of contour lines to calculate surface characteristics for each grid cell. The process calculates values for grid cells at increasingly finer resolutions, starting with a coarse initial grid and successively halving the grid spacing until the final user-specified resolution is reached (Golub and Van Loan, 1983). Cell values are generated by linear interpolation from the preceding grid. At each resolution, the elevation value at a given cell represents an estimate of the average of the actual surface elevation across that particular grid cell. Eventually, the resolution is sufficiently fine for there to be little or no data-point averaging. At this point, the grid is considered to be stable and all information has been extracted from the source data (Hutchinson, 1988). The spacing of the final grid is established prior to interpolation and is representative of the source data. For this study, a cell resolution of 10 ft was selected on the basis of National Map Accuracy Standards for maps at scales of 1:20,000 and larger (U.S. Geological Survey, 1997b).

Comparison of the interpolated grids with the measured data was quantified by a root-mean-square (rms) error analysis. A rms error analysis shows the spread or scatter of the interpolated data relative to the measured data. Both the 1913 and 1998 interpolated data sets were subtracted from their respective coincident measured values. The differences were squared and the square root of their average yielded the rms error. The 1913 interpolated data set had a rms error of 0.98 ft. The 1998 interpolated values resulted in a rms error of 0.16 ft.

Sediment Accumulation and Water Volume

Figure 5. Sediment Accumulation in Loch Raven Reservoir, 1913 - 1998.

From 1913 to 1998, about 11.23 million yd (cubic yards) of sediment were deposited in Loch Raven Reservoir (figure 5). Water volume in the reservoir decreased from 21.445 Ggal to 19.175 Ggal from 1913 to 1998. This represents about an 11-percent decrease in overall water volume over the past 85 years (table 1). Sediment accumulated at an average-annual rate of about 132,000 yd /yr (cubic yards per year) -- about 168,000 yd /yr lower than the rate predicted during the design phase of the reservoir (Freeman and Stearns, 1910).

Sediment accumulation (which causes a loss of water volume) has varied spatially over the reservoir. To quantify this variability, the reservoir was divided into five distinct sections (table 1). Analysis of sediment and water volume by section determined which sections were gaining or losing capacity relative to other sections. Water volume comparisons and sediment accumulation calculations in section A were made on the basis of data collected between 180 ft and 240 ft m.s.l. No data were collected below 180 ft m.s.l. in 1913. As part of their calculations for sedimentation rates, Freeman and Stearns (1910) reported that in 1908, 178 Mgal (million gallons) of water were stored behind the 1881 masonry structure. Calculations made from the 1998 grid show that 205 Mgal of water are now stored below 180 m.s.l.--the elevation of the 1881 dam. By including water volume measured below 180 m.s.l. in 1998, but not accounted for from the 1913 topography, the total water volume in Loch Raven Reservoir increased from 19.175 to 19.38 Ggal.

Summary

A cooperative study between the City of Baltimore, the Maryland Department of the Environment, the Maryland Geological Survey, and the U.S. Geological Survey was undertaken to determine the accumulated volume of sediment and present storage capacity for Loch Raven Reservoir, Baltimore County, Maryland. The reservoir represents a major part of Metropolitan Baltimore's water supply and is periodically measured to quantify its total storage capacity. Topographic maps drafted in 1913 by Baltimore City water engineers were used to create a reference surface that provided a base for sediment accumulation measurements. Bathymetric measurements made in 1998 were collected by the Maryland Geological Survey. A geographic information system was used to generate surfaces representing the 1913 topography and the 1998 bathymetry. Digital orthophotos were used to provide a consistent boundary for volume calculations and the two surfaces were subtracted to determine volumes of sediment accumulation. Change in water volume is based on the difference between the potential water volume as calculated from the 1913 topography and the current (1998) storage capacity based on the 1998 bathymetric data.

Between 1913 and 1998, about 11.23 million cubic yards of sediment were deposited in Loch Raven Reservoir. Water volume in the reservoir has decreased from 21.445 billion gallons as calculated from the 1913 grid to 19.175 billion gallons as calculated from the 1998 grid. This represents an 11-percent decrease in water volume over the past 85 years. By including the volume of water below the elevation of the existing impoundment in 1913, water storage increases to 19.38 billion gallons. Sediment has accumulated at an average-annual rate of about 132,000 cubic yards per year. This rate is lower than the rate predicted during the design phase of the reservoir. The reservoir has accumulated the most sediment in its upper reaches and the least sediment near its dam.

References Cited

For further information contact:

District Chief
U.S. Geological Survey
5522 Research Park Drive
Baltimore, MD 21228

Visit the USGS Chesapeake Bay Homepage:
http://chesapeake.usgs.gov/

Water-Resources Investigations Report 99-4240


Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://md.water.usgs.gov/publications/wrir-99-4240/html.htm
Page Contact Information: webmaster@md.water.usgs.gov
Page Last Modified: Thursday, September 14, 2017