Stormwater Outfalls, Impervious Cover, and Stream Monitoring in Washington, DC
A stormwater source exploration map
This report is an independent portfolio project prepared from public data. It was not produced for, and has not been reviewed or endorsed by, DOEE, DC Water, USGS, EPA or any other agency. The report maps stormwater outfalls and impervious cover in Washington, DC and places recent stream monitoring results next to them for context. Results are not compared with water quality standards, and the report is not a regulatory assessment.
Features on the map can be selected for details. The search box at upper left finds outfalls by receiving waterbody or outfall ID. The 1 km outfall density grid and the turbidity sites are off by default and can be turned on in the layer list at upper right. Outfalls are drawn as circles, and monitoring sites as squares (E. coli) and diamonds (turbidity) with darker fills for higher medians. Combined sewer areas cover 51.3 km² (39% of the sewered area). They are shown in grey on the map and are not analyzed.
Summary
The District has 587 mapped outfalls from its separate storm sewer system (MS4). They drain 556 sewersheds covering 81.3 km², of which 44% is impervious. The sewersheds are small: 529 have a single outfall, the median area is 3.3 ha, and 90% are smaller than 31 ha. The outfalls discharge to 35 named waterbodies. Rock Creek receives 128 outfalls, the Potomac River 105 and the Anacostia River 91.
A field survey recorded flow at 147 of the 369 outfalls with a flow record (40%). The survey consisted of a single visit to each outfall, mostly between 2004 and 2006, and describes conditions on those days only.
Outfalls by receiving waterbody
| Receiving waterbody | Outfalls | Median pipe diameter (in) | Flow observed (% of those recorded) | Median sewershed area (ha) | Median sewershed impervious cover (%) |
|---|---|---|---|---|---|
| Rock Creek | 128 | 24.0 | 42 | 2.9 | 39 |
| Potomac River | 105 | 24.0 | 28 | 2.3 | 41 |
| Anacostia River | 91 | 24.0 | 36 | 4.7 | 50 |
| Oxon Run | 62 | 27.0 | 47 | 4.9 | 51 |
| Watts Branch | 58 | 30.0 | 50 | 4.3 | 48 |
| Pope Branch | 24 | 24.0 | 36 | 3.6 | 36 |
| C&O Canal | 18 | - | - | 0.8 | 25 |
| Pinehurst Branch | 15 | 19.5 | 36 | 3.2 | 33 |
| Piney Branch | 8 | 21.0 | - | 1.2 | 37 |
| Fort Davis Tributary | 6 | 15.0 | - | 0.6 | 49 |
| Bladgen Run | 5 | 15.0 | - | 1.2 | 8 |
| Luzon Branch | 5 | 15.0 | - | 0.7 | 51 |
| All other named waterbodies | 53 | 22.5 | 60 | 5.6 | 44 |
| Not recorded | 9 | 24.0 | 25 | 3.3 | 33 |
Largest sewersheds by impervious area
| Sewershed ID | Watershed | Area (ha) | Impervious cover (%) | Impervious area (ha) | Outfalls | Receiving waterbody | Pipe diameter (in) | Land use |
|---|---|---|---|---|---|---|---|---|
| 998 | Anacostia River | 450.4 | 43 | 192.1 | 1 | Anacostia River | 18 | - |
| 175 | Anacostia River | 341.1 | 49 | 167.1 | 1 | Anacostia River | - | Residential |
| 218 | Anacostia River | 266.2 | 53 | 139.8 | 1 | Hickey Run | - | - |
| 999 | Anacostia River | 294.5 | 43 | 128.1 | 1 | Anacostia River | - | - |
| 1016 | Potomac River | 292.1 | 43 | 126.4 | 1 | Potomac River | - | Res/Comm/Ind |
| 85 | Potomac River | 358.3 | 34 | 122.8 | 1 | Potomac River | 109 | Res/Comm |
| 147 | Anacostia River | 251.7 | 43 | 107.2 | 1 | Anacostia River | - | Res/Ind |
| 878 | Rock Creek | 235.9 | 45 | 105.9 | 1 | Rock Creek | 72 | Res/Comm |
| 2004 | Rock Creek | 166.3 | 51 | 85.6 | 1 | Luzon Branch | - | - |
| 851 | Rock Creek | 127.2 | 49 | 62.7 | 1 | Rock Creek | 66 | Res/Comm/Ind |
A dash indicates that the survey left the field blank.
Stream monitoring (2018 and later)
Water Quality Portal results collected at DC stream sites since 2018 were summarized by site. A site is included only when it has at least 20 results for the parameter, and site-level comparisons also exclude sites where half or more of the results are below detection. Under these minimums, 32 sites are available for E. coli, 32 for turbidity and 29 for total suspended solids (TSS). TSS is not charted because 15 of its 29 sites are at least half non-detect.
The chart compares each site’s median result with the percentage of impervious cover within 1 km of the site. Sites are shown when at least 12.5% of the buffer lies inside MS4 sewersheds (30 sites for E. coli and 30 for turbidity). For E. coli, the Spearman rank correlation is ρ = 0.28 (-0.14, 0.62) (95% bootstrap interval in parentheses), and for turbidity ρ = 0.25 (-0.12, 0.58). The intervals include zero in all 9 combinations of parameter and buffer radius tested (500 m, 1 km and 2 km), so no relationship between impervious cover and site medians is demonstrated.
Replacing the circular buffer with the upstream catchment of each E. coli site changes the result. Among the 29 sites with enough DC sewershed area in the catchment, impervious cover in the DC part of the catchment is positively correlated with E. coli (ρ = 0.54 (0.17, 0.81)). This is treated as a hypothesis only, because a median of 8% of each catchment lies inside DC. Please refer to the limitations and caveats below for how these results should be read.
Data cleaning
| Field | Distinct values, raw | Distinct values, cleaned | Records filled, cleaned | Note |
|---|---|---|---|---|
| Receiving waterbody | 64 | 35 | 578 | Bank/side suffixes and case variants merged; unmatched names kept as written |
| Pipe size | 80 | 6 | 370 | Parsed to inches where a single diameter is given (see records filled); box and multi-barrel sizes stay as written in the popup |
| Pipe shape | 24 | 6 | 408 | Case and spelling variants merged |
| Pipe material | 30 | 6 | 393 | Case and spelling variants merged |
| Land use | 5 | 5 | 322 | Case standardised |
| Dry-weather flow | 2 | 2 | 369 | 1 = flow observed, 0 = none observed at the single survey visit |
| Cleaning step | Result rows |
|---|---|
| raw rows (bbox, stream sites) | 120,537 |
| sites coded DC (StateCode 11) | 77,832 |
| drop QC samples (blanks, replicates) | 73,260 |
| drop groundwater | 73,260 |
| valid date | 73,260 |
| keep harmonizable units / fractions | 66,184 |
| numeric non-negative value present | 65,348 |
| drop exact duplicates within the same site | 59,246 |
| drop cross-organization duplicates (keep original source) | 59,112 |
| drop all-non-detect organization x parameter series | 58,061 |
Water quality sites were limited to those coded as DC in the Water Quality Portal. Blanks, replicates and duplicates were removed, including Chesapeake Bay Program copies of USGS results. E. coli results reported as MPN or CFU were pooled. Non-detects were retained as censored values. Many non-detects have a blank result unit, and the detection-limit unit was used for these because removing them would bias summaries upward. One E. coli series (1,260 results, all non-detects) was excluded.
Outfalls were linked to sewersheds by OUTFALL_ID_TEXT because the numeric ID is duplicated for some sewersheds. Thirty-one outfalls had no sewershed under the published ID. These were assigned to the nearest sewershed within 200 m and are flagged in the data.
Limitations and caveats
- Outfall survey: attributes come from a single field visit per outfall, mostly 2004 to 2006, and many fields are incomplete (land use is recorded for 322 of 587 outfalls). Flow observations describe the day of the visit only. Outfall locations were GPS-captured in the same period.
- Combined sewers: 39% of the sewered area drains to combined sewers. These areas and their overflow outfalls are not analyzed, so monitoring sites downstream of them are not explained by the MS4 data shown here. No public layer of combined sewer overflow outfall points was found.
- Outfall density: outfalls per km² describes how the sewer network is divided, not the amount of stormwater reaching a stream. Around monitoring sites it is strongly related to median sewershed area (ρ = -0.76). Across the 30 sites meeting the data minimums it is negatively related to E. coli (ρ = -0.31), but the pattern is not consistent across watersheds: ρ = -0.60 within Rock Creek (n = 14) and ρ = 0.27 within the Anacostia River (n = 13). Outfall density is therefore not used to compare monitoring results.
- Number and clustering of sites: most monitoring sites are on Rock Creek or the Anacostia River, and nearby sites resemble each other (Moran’s I of model residuals ranges from 0.14 to 0.46). When sites within 1.5 km are resampled together, the correlation intervals include zero in 4 of 4 E. coli and turbidity models.
- Monitoring programs: most sites are routine DOEE monitoring. The 4 USGS sites include storm-targeted sampling and have a much higher median E. coli (4,050 compared with 214 for DOEE sites). Turbidity is 95% FNU and the remainder NTU, and these units are treated as equivalent.
- Weather: most monitoring results carry no weather information, so wet and dry conditions cannot be separated. Flow and tidal influence on the lower Anacostia River are not modeled.
- Multiple comparisons and causation: 18 correlations were screened and 3 remain after a false-discovery-rate adjustment, all involving outfall density. The adjustment assumes independent sites. The results describe association only, and combined sewers, sewershed size, parkland and sampling design vary together with the stormwater measures.
- Impervious cover: values come from DC’s 2025 impervious surface polygons. The DOEE sewershed impervious field agrees with them almost exactly but is probably derived from the same data, so it is not an independent check.
- Older data: conductance, nutrient and chloride results back to the 1980s were cleaned but not analyzed because recent coverage is thin.
Data sources and licenses
| Data | Source | License |
|---|---|---|
| MS4 outfalls, MS4 sewersheds, combined-sewer sewersheds, impervious surface 2025 | Open Data DC (Government of the District of Columbia, DOEE / OCTO) | CC BY 4.0 |
| Stream water quality | EPA/USGS Water Quality Portal (DOEE, USGS, Chesapeake Bay Program, others) | Public data; cite the originating organizations |
DC layers were retrieved on 2026-10-30 and water quality data on 2026-10-31. Suggested citation for the DC layers: “Contains data from Open Data DC, Government of the District of Columbia, licensed CC BY 4.0.” All results on this page are regenerated from code in the project repository. The analysis CRS is EPSG:6487 (NAD83(2011) / Maryland, metres) and the web map uses EPSG:4326.