For the more information about water resources in the National Park Service, please visit

Hydrographic & Impairment Statistics Methods

Hydrographic and impairment statistics were derived using the 1:24,000 scale (1:63,360 in Alaska) National Hydrography Dataset (NHD) NHDinGeo data model. The methods explained below assume that the reader has some familiarity with the use of geographic information systems (GIS), particularly ESRI's ArcGIS. Each national park unit was processed within its own personal geodatabase, which included many feature classes such as hydrography, impairment data, and the park boundary. For parks that transverse more than one Watershed Boundary Dataset subbasin (the distribution unit for NHD data), the hydrography from each subbasin was combined into one geodatabase using the USGS "NHD Dataset Merge Utility."

Within each park geodatabase, the project focused specifically on five different NHD feature classes to derive hydrographic statistics:

Feature Class Description
NHDFlowline The core linear network of the surface-water drainage system primarily consisting of streams and artificial paths through polygons.
NHDWaterbody Polygons representing areal hydrographic waterbody features such as lakes, swamps, and reservoirs.
NHDArea Polygons representing areal hydrographic landmark features such as two-dimensional streams (left and right banks), canals, and oceans.
NHDPoint Points representing hydrographic landmark features such as springs, gaging stations, and waterfalls.
NHDLine Line features representing linear hydrographic landmark features such as reefs and nonearthern shores.

More detail on the NHD including technical references, tutorials, applications, and stewardship information can be obtained at U.S. Geological Survey's (USGS) NHD website.

Clipping the NHD

All NHD feature classes of interest were clipped to the park boundary contained in the NPS Land Resources Division's (LRD) administrative boundaries dataset. As of November 2009, only about half of the boundaries in the administrative boundaries dataset have been endorsed by the LRD. The other boundaries in the dataset originated from a variety of sources generally constituting the best available GIS boundary until updated by the LRD. Park boundaries are dynamic as Congress may expand, contract, or otherwise modify park boundaries. As better boundaries become available, WRD can reprocess a park to update its hydrographic and impairment statistics based on the improved boundary. More information on park boundaries can be found on LRD's website.

A clipper polygon was used to account for any adjacent water features. The clipper polygon is an auto-complete polygon that is adjacent to the park boundary and is digitized to include areas where adjacent water features occur. When a park has adjacent hydrography, the clipper polygon was used to clip the water features as opposed to the park boundary.

Attributing for Adjacencies

Adjacencies were defined as any water feature that shares a boundary with a park. It is important to note that identifying adjacent hydrographic features may include some level of subjectivity and require a certain degree of judgment by the processor. Other sources of information, such as reviewing land segment maps and contacting park staff, were utilized whenever possible to verify adjacent features. Below are some examples of adjacency and how they were dealt with.


This example of adjacency is rather clear cut. Since the Mancos River is contiguous with the Mesa Verde National Park boundary, the highlighted portion of the Mancos River was included in the adjacent perennial stream statistics.


The green line in this figure represents the park boundary of Glacier National Park. The North Fork Flathead River contains several braids, some of which are outside of the park boundary. Although no one channel clearly shares a border with the park, the entire river system was attributed as adjacent; however, only the highlighted portion of the river actually counts towards the adjacent perennial statistics. Refer to the centerline attribution section mentioned later on in the methods.


This figure depicts two adjacency issues in the southern parcel of Effigy Mounds National Monument. First, the linear portion of Johnsons Slough that is highlighted in cyan and the yellow areal portion do not appear to be adjacent to the park (since it is well within the park boundary). According to segment and topographical maps, however, Johnsons Slough is indeed adjacent and should not fall within the park. The other adjacency issue is that Johnsons Slough is a braided segment of the Mississippi River. A judgment call was made to exclude the Mississippi River to prevent overestimation of stream/river statistics. This decision was also heavily influenced by land ownership and the location of the Upper Mississippi River Wildlife and Fish Refuge. In this situation, only the highlighted linear segment of Johnsons Slough and the yellow areal polygon were attributed as adjacent and included in the statistics.


In Montezuma Castle National Monument, only a portion of Beaver Creek is adjacent to the park. The straddling section was parsed from the rest of Beaver Creek and identified as adjacent.


Sometimes adjacent water features are not apparent. In Poverty Point National Monument, the Macon Bayou and the park boundary barely even intersect. After reviewing the land segment maps and speaking with park personnel, however, the boundary of Poverty Point National Monument actually extends to the centerline of the Bayou. Therefore, all stream acres to the east of the centerline were tagged as adjacent. Additionally, the centerline was also tagged as adjacent because the boundary extends only to the center. Therefore, the hydrographic statistics for the Macon Bayou should include non-adjacent and adjacent stream acres, as well as adjacent stream miles.

Assessments of adjacent water features can be intricate. Decisions were typically made on a park-by-park basis. Although the situations above illustrate the most common ways the project dealt with adjacencies, each instance required some degree of research to make the best decision possible.

Attributing for Centerlines

Within the NHDinGeo data model, linear features within the NHDFlowline feature class not identified as centerlines weren't included in a park's hydrographic statistics. This was especially handy when dealing with "branches" and "braids," which were not considered part of the main river system.


In Colonial National Historical Park, the adjacent James River is represented as both a linear and polygonal feature. Several flowlines (in cyan) fall within the polygon to connect tributaries to the center pathway that is James River. These flowline "branches" exist in NHD to enforce network connectivity in order to model water flow. Therefore, they were not tagged as centerlines and, consequently, were excluded from the park's hydrographic statistics.


The Alagnak River in Alaska is a complex river system that contains several braids. Only the main channel (highlighted in cyan) of the river system is included in the statistics. The main channel is identified either by lowest NHD stream level or by USGS Geographic Names Information System attribution. Only the main channel is tagged as a centerline and included in the hydrographic statistics.


Parsing out centerlines for river systems located within swampy areas was often impractical. This is illustrated in Congaree National Park where several braids are distributed throughout the marsh. In these situations, all flowlines were tagged as centerlines and included in the stream statistics.

Similar to attributing for adjacent water features, tagging centerlines also requires some level of judgment from the processor. In situations where a braid extends for several miles, it is up to the processor to decide whether or not to include such a braid in the stream statistics. Consequently, not all instances encountered were as straightforward as the ones depicted above, especially for Alaskan parks where braids are ubiquitous. Parsing out centerlines in Alaskan parks required a great deal of time and effort and is prone to subjectivity and error.

Attributing "Artificial Paths" as Rivers and Canals

Artificial paths (NHD feature code 55800) exist in the NHDFlowline feature class and are assigned to features that overlap any polygonal features in the NHDArea and NHDWaterbody feature classes. The purpose of an artificial path is to delineate the direction of flow into and out of polygonal features. Artificial paths do not distinguish themselves as intermittent streams, canals, etc. Therefore, they must be reclassified to be accounted for properly in the hydrographic statistics.


Centerlines that overlap polygons classified as intermittent streams were reclassified from artificial paths to intermittent streams. This situation was especially prevalent in Intermountain Region parks such as Canyonlands National Park. Since Butler Wash was classified as an intermittent stream polygon by NHD, the centerline artificial path was reclassified as an intermittent stream.


Artificial path features that are centerlines and are located within canal polygons were reclassified to canal/ditch. For example, on the western portion of Biscayne National Park, several canals run into the park and empty into the Atlantic Ocean. The artificial path that represents Mowry Canal was reclassified as a canal to obtain the appropriate canal mileage for Biscayne National Park.

Attributing for Shared Water Features

The sum of hydrographic statistics from all national park units does not equal the overall service-wide statistics due to shared water features among parks with adjacent boundaries. Although a particular waterbody may be included within both individual parks' statistics, it is only counted once in the service-wide statistics.


Shared water features are included in each park's statistics, but they are not summed to calculate the service-wide statistics. In the case of Sequoia and Kings Canyon National Parks, each park gets credit for the 3.6 mile shared portion of the North Fork Kaweah River, but, to avoid double-counting, the service-wide summation only includes this segment once.

Parks with shared water features are listed in the table below.

Park Shares Water Features With
Alagnak Wild River Katmai National Park & Preserve
Appalachian National Scenic Trail Delaware Water Gap National Recreation Area & Middle Delaware National Scenic River
Harpers Ferry National Historical Park
Chesapeake and Ohio Canal National Historical Park
Great Smoky Mountains National Park
Shenandoah National Park
Blue Ridge Parkway
Big Hole National Battlefield Nez Perce National Historical Park
Blue Ridge Parkway Appalachian National Scenic Trail
Chesapeake and Ohio Canal National Historical Park Appalachian National Scenic Trail
Harpers Ferry National Historical Park
George Washington Memorial Parkway
Rock Creek Park
Delaware Water Gap National Recreation Area & Middle Delaware National Scenic River Appalachian National Scenic Trail
Fort Caroline National Memorial Timucuan Ecological and Historic Preserve
Fort Washington Park George Washington Memorial Parkway Piscataway Park
George Washington Memorial Parkway Chesapeake and Ohio Canal National Historical Park
Rock Creek Park
National Capitol Parks
National Capitol Parks Central
Fort Washington Park
Piscataway Park
Theodore Roosevelt Island National Memorial
Golden Gate National Recreation Area Point Reyes National Seashore
Great Smoky Mountains National Park Appalachian National Scenic Trail
Harpers Ferry National Historical Park Chesapeake and Ohio Canal National Historical Park
Appalachian National Scenic Trail
Katmai National Park & Preserve Alagnak Wild River
Kings Canyon National Park Sequoia National Park
National Capitol Parks George Washington Memorial Parkway
Theodore Roosevelt Island National Memorial
National Capitol Parks Central
National Capitol Parks- Central (National Mall & Memorial Parks) Theodore Roosevelt Island National Memorial
National Capitol Parks
George Washington Memorial Parkway
Nez Perce National Historical Park Big Hole National Battlefield
Piscataway Park George Washington Memorial Parkway
Fort Washington Park
Point Reyes National Seashore Golden Gate National Recreation Area
Rock Creek Park Chesapeake and Ohio Canal National Historical Park
Theodore Roosevelt Island National Memorial
George Washington Memorial Parkway
Sequoia National Park Kings Canyon National Park
Shenandoah National Park Appalachian National Scenic Trail
Theodore Roosevelt Island National Memorial George Washington Memorial Parkway
National Capitol Parks
Rock Creek Parks
Timucuan Ecological and Historic Preserve Fort Caroline National Memorial

Attributing for Park

Hydrography for each park was organized and processed within its own unique personal geodatabase. All clipped hydrography was tagged with the four letter park unit code.

Assessing Lake Counts in NHDWaterbody

A field called "LakeCountID" was added to the clipped NHDWaterbody feature class to quantify the number of lakes, ponds, and reservoirs within each park. In the field calculator of the attribute table, the "LakeCountID" field was set to equal the "ObjectID," which gave each feature a unique value in the "LakeCountID" field.

The clipped NHDWaterbody feature class was exported into KML (keyhole markup language) format and uploaded into Google Earth to more accurately quantify lakes using aerial photographs. Only features with an FCODE of 390xx, 36100, 43613-43619, and 43621 were assessed. Polygons that represent the same waterbody feature were given identical values in the "LakeCountID" field. Statistics for lake counts were derived using the number of unique values in the "LakeCountID" field.


In the instance of Whiskeytown National Recreation Area, Whiskeytown Lake exists in NHD as eleven separate polygons. Some of these polygons were created during processing to depict portions of the lake recognized by EPA as 303(d) listed (impaired) waters. In this instance, only the yellow portions of Whiskeytown Lake were recognized as impaired by EPA. The other polygons that comprise Whiskeytown Lake can be explained by satellite imagery.


As shown by this image, the tiny polygons that are nearly adjacent to the primary Whiskeytown Lake waterbody are also part of the lake, but separated by a road (or several bridge crossings) from the main waterbody. Consequently, all eleven polygons were assigned identical values in the "LakeCountID" field to account for there only being one Whiskeytown Lake. Failure to follow this step would result in counting Whiskeytown Lake eleven times.


Saint Croix River is an impounded river that is also partially a reservoir. When assessing lake counts, NHD shows adjoining waterbody polygons that have separate GNIS names. The yellow polygon is Lake Saint Croix and the green polygon is Lake Mallalieu. In situations where adjoining polygons have different GNIS names, they are each given a unique value in the "LakeCountID" field so they will be counted as separate lakes/reservoirs.


Quantifying lakes becomes difficult when they are located within marshy and/or estuarine areas. In the case of Cape Cod National Seashore, each red polygon represents a pond identified by NHD. Since the number of lakes within marshy areas are bound to vary (especially for coastal parks), each polygon was given a unique value in the "LakeCountID" field.


Coastal parks, such as Cape Krusenstern National Monument, may have varying numbers of lakes depending on the tide level. In this instance, the royal blue is the ocean, the red polygons are lakes, and the light blue area is marsh. In most coastal parks, unless lakes are inland and are not tidally influenced, then quantification of lake counts is strictly dependent on NHD delineation. Therefore, in Cape Krusenstern National Monument, each red polygon was given a unique value in the "LakeCountID" field.

Attributing for 303(d) Impairments

Clean Water Act (CWA) 303(d) impairments were georeferenced to the NHD. Georeferencing is the process of locating an entity in "real world" coordinates. As applied to this project, georeferencing consisted of adding 303(d) entity IDs to those NHD records which were 303(d) impaired. 303(d) entity IDs are unique IDs assigned to impaired waterbodies by either the State or the Environmental Protection Agency (EPA). To determine which waterbodies were impaired and the geospatial extent of the impairments, WRD primarily relied upon the EPA 303(d) shapefiles and TMDL Tracking System, State 303(d) lists, and State assessment shapefiles (if available).

All data sources used by the WRD have their limitations. Although the EPA impairment shapefiles were already georeferenced, they were typically dated and only available at 1:100,000 scale. State assesssment shapefiles were only available from a small number of States and were also typically at 1:100,000 scale. State 303(d) lists provided the most current listings, but offered poor textual descriptions of the geospatial extents. In some instances only sampling stations were identified as impaired with no indication of upstream or downstream extent. Similarly, in coastal areas States frequently made no effort to describe the geospatial extent of impaiments other than to locate the sampling stations. Since this project necessitated that all park impairments be georeferenced to NHD in order to calculate miles and acres of impairments, assumptions and judgment calls were often necessary.

Relate tables were created to house the impairment data (e.g. pollutants, sources, year listed, etc.) using the 303(d) entity ID as the relational key. Typically the impairment data was extracted from the EPA TMDL Tracking System. Impairment data was taken directly from State 303(d) lists when it was found to be more current than the EPA TMDL Tracking System.

Measuring the Clipped NHD

Prior to processing the hydrography for each park, the data frame within ArcGIS was projected on-the-fly to its appropriate Universal Transverse Mercator (UTM) zone. This was particularly challenging for some Alaskan parks, as their spatial extent may intersect multiple UTM zones. In those cases, the UTM zone that covered the majority of the park was selected. Parks within the U.S., Puerto Rico, and Virgin Islands were processed using North American Datum (NAD) 1983 UTM zones. However, parks in the Pacific Islands (i.e. Guam and American Samoa) were processed using World Geodetic System (WGS) 1984 UTM zones.

The length, perimeter, and area of the clipped NHD features were calculated using the "Calculate Geometry" tool for vector analysis. It is important to note that the "Calculate Geometry" tool generates planimetric measurements and does not produce surface length or area. Therefore, a certain degree of error is to be expected within the calculations, especially when processing hydrography for parks with high degrees of varying elevations and slopes.

Shoreline statistics were derived by converting lake/pond/reservoir/ocean polygons into polylines. The polylines that represent the perimeter of the clipped polygons, however, do not reflect the true shoreline mileage. As a result of clipping the polygons to the park boundary, any polygonal features that intersect the park boundary will contain "false" shorelines. "False" shorelines were manually deleted using the split tool. The "Calculate Geometry" tool was employed to calculate miles of all true shorelines within each park. The concept of "false" shorelines is illustrated below by overlaying the Isle Royale National Park boundary with the clipped lake/pond/reservoir polygons in NHDWaterbody.


Attributing Clipped Hydrography

Linear features within the NHDFlowline feature class were clipped to the park boundary and then exported as an event table using "Linear Referencing." The event table uses reach codes from NHDFlowline as its route reference. Several fields were then added to the event table to attribute for park, centerline, adjacency, stream, canal, CWA 303(d) entity ID, and length. Impaired flowlines that were assigned an entity ID were combined into a separate event table using the "Dissolve Route Events" tool.

Clipped polygonal features were assigned unique IDs for each record. Within individual park geodatabases, relate tables were created that link the unique IDs with various attributes such as adjacency, park unit, and CWA 303(d) entity IDs.

A separate relate table was created within each park geodatabase that contains impaired water features. This table contains CWA 303(d) impairment data and used the entity ID as the relational identifier. Information for impaired water features are maintained within this table and were derived from EPA TMDL Tracking System and the most current state 303(d), 305(b), and Integrated Reports.

Attribute and relate tables were imported into a master database where point, linear, and polygonal park and service-wide statistics were calculated. Statistical calculations were derived using queries and Visual Basic modules. Several QC procedures have been integrated to improve the accuracy of the statistics. The final product includes a spreadsheet of hydrographic and impairment statistics that can be viewed in both tabular and form views.

Data Chart


Last Updated: January 04, 2017