What the finder searches
The tool is open to every Studio account. A country picker sets which national network you search, and the search modes change with it:
| Country | Network | Stations | Search modes | Records linked |
|---|---|---|---|---|
| South Africa | DWS flow, reservoir and meteorological stations | 2 896 | Browse map, point, polygon, watercourse, saved watershed | DWS station page, monthly peaks CSV, annual peaks CSV |
| United States | USGS NWIS stream gauges with peak-flow records | Nationwide | Point, upstream or downstream trace, polygon, watercourse, saved watershed | NWIS annual peak-flow file with qualification codes |
| Australia | Bureau of Meteorology National Flood Gauge Network | 3 836 | Point, polygon, watercourse, saved watershed | Annual maxima of the daily discharge record from Water Data Online, by July–June water year |
Not every DWS station measures river flow. The station list the tool carries breaks down like this:
| Station type | Count | Use for flood frequency |
|---|---|---|
| River | 1 297 | The primary target |
| Meteorological | 574 | No flow record |
| Canal | 313 | Regulated flow |
| Reservoir and Components | 261 | Stage; inflow only after a water balance |
| Closed Conduit | 214 | Pipe or tunnel flow |
| Downstream River Components | 150 | Below a dam; regulated |
| Eye (spring) | 70 | Groundwater emergence |
| Lake or Pan | 17 | Level monitoring |
Catchment area is recorded for 1 615 of the 2 896 stations, about 56 %. A catchment-area filter therefore hides every station without one.
Search modes and filters
- Point: the nearest gauges to a clicked point or typed coordinates, by great-circle distance, 1 to 50 results.
- Polygon: upload a catchment boundary. Gauges inside it are listed first, then the nearest outside it to fill the count, with distances from the polygon centroid.
- Watershed: the same search on a catchment saved from watershed delineation.
- Watercourse: gauges within a buffer of an uploaded river line. Use 5–10 km for a major river and 1–3 km for a tributary, so the buffer catches the main stem without reaching a parallel river.
- Trace (United States): follow the NHDPlus network upstream or downstream from a clicked river for a set distance.
Filters combine: station type, drainage region, catchment-area range, minimum record years, maximum distance, station name or number, and active only. A station counts as active when its last record is less than two years old. Record length is measured from the first to the last date, so it does not subtract gaps inside the record.
Worked example: a culvert site near Howick
A point search on a site at 29.47°S, 30.23°E in KwaZulu-Natal, run through the Studio search engine on the DWS station list (record lengths as at 8 October 2026), narrows like this:
| Filter applied (30 km radius) | Stations left |
|---|---|
| None | 25 |
| Station type: River | 15 |
| Minimum record: 30 years | 8 |
| Active only | 7 |
| Station | Place | Distance (km) | Catchment (km²) | Record (years) |
|---|---|---|---|---|
| U2H007 | Lions River (Mpofana River) @ Weltevreden | 8.5 | 358 | 71.5 |
| U2H006 | Karkloof River @ Shafton | 10.8 | 339 | 72.1 |
| U2H013 | Mgeni River @ Petrus Stroom | 13.9 | 299 | 65.6 |
| U2H011 | Msunduze River @ Henley Dam | 19.9 | 176 | 68.2 |
| U2H057 | Slang Spruit @ Pietermaritzburg | 21.5 | 48 | 30.6 |
| U2H058 | Msunduze River @ Masons Mill | 23.1 | 327 | 30.7 |
| U2H012 | Sterk River @ Groothoek | 25.6 | 438 | 65.4 |
The active filter removed one station: U2H001, Mgeni River @ Howick, 2.1 km from the site with 44 years of record that ended in 1993. For a catchment of a few hundred square kilometres, U2H006, U2H007 and U2H013 each offer more than 65 years from catchments of similar size, and U2H001 is worth reading too despite its closure. The next step is to send the annual peaks from three or more of them to flood frequency analysis, where the regional consistency check tests whether they behave as one group.
Checking a peak record before you fit it
A station’s record length is not its sample size. Before sending peaks to a frequency analysis, open the annual peaks file and check three things.
- Count the years that have a value. A 70-year station with a 15-year gap holds 55 annual maxima, and the frequency analysis should see 55.
- Read the quality flags. A year with missing data may have missed the flood, so its maximum can understate the true annual peak. Treat it as a lower bound or leave it out.
- Look for regulation. Compare the record against the construction dates of dams upstream. A record that spans a new dam is two different populations.
USGS peaks carry qualification codes that Studio reads on import and shows beside each value:
| NWIS code | Meaning |
|---|---|
| 1 | Maximum daily average, not an instantaneous peak |
| 2 | Estimated |
| 3 | Affected by dam failure |
| 4 | Less than the indicated value |
| 5 | Affected to an unknown degree by regulation or diversion |
| 6 | Affected by regulation or diversion |
| 7 | Historic peak, from outside the systematic record |
| 8 | Actually greater than the indicated value |
| 9 | Snowmelt, hurricane, ice-jam or debris-dam breakup |
| A, B | Year, or month and day, of occurrence uncertain |
| C | Affected by urbanisation, mining or channelisation |
| D | Base discharge changed this year |
| E | Only the annual maximum is available |
Codes 5, 6 and C mark a record that no longer describes a natural catchment. Code 7 matters most for Bulletin 17C, because a historic peak belongs in the historical-data input, not in the gauged series.
DWS station numbers and drainage regions
A DWS station number such as A2H019 reads as primary drainage region (A), secondary region (2), station type (H for flow, R for reservoir, M for meteorological) and sequence number (019). South Africa has 22 primary drainage regions, lettered A to X without I and O:
| Code | Region | Code | Region |
|---|---|---|---|
| A | Limpopo | N | Sondags |
| B | Olifants | P | Boesmans |
| C | Vaal | Q | Fish |
| D | Orange | R | Nahoon/Keiskamma |
| E | Olifants/Doorn | S | Great Kei |
| F | Buffels | T | Mzimvubu/Umbashe |
| G | Berg | U | Mvoti/Mgeni/Mkomazi |
| H | Breede | V | Tugela |
| J | Gouritz | W | Usutu/Phongolo/Mfolozi |
| K | Coastal Rivers | X | Sabie/Krokodil/Komati |
| L | Gamtoos | ||
| M | Swartkops |
Limitations
- Gaps. DWS records are not always continuous, so the record length shown can overstate the number of usable annual maxima. Open the annual peaks file and count.
- Rating-curve extrapolation. The largest recorded floods are often computed from stages above the highest gauged flow, which makes the peaks that matter most the least certain.
- Change over time. Dams, abstractions, transfers and land-use change can make an old record unrepresentative. A trend test in flood frequency analysis flags the obvious cases; local knowledge catches the rest.
- Coverage. The tool covers South Africa, the United States and Australia. Elsewhere, go to the national hydrological service directly.
From gauge to design flood
Each station row has Send to FFA, which opens a flood frequency analysis with the annual peaks loaded and their metadata kept: gauge height, the agency’s qualification codes and catchment area. USGS peaks coded as historic peaks are moved into the historical-data panel for a Bulletin 17C analysis. Select three or more stations and send them together for a regional consistency check. Results save to a project and export as a CSV station list or a Word or Excel report.
Where no usable gauge exists, the regional methods take over. The index flood method was calibrated on 383 DWS records of at least 20 years (Calitz and Smithers, 2020), the quantile regression technique gives an independent regional estimate, and daily rainfall statistics supplies the rainfall side for a rainfall-runoff cross-check.
Common questions
Where can I find stream gauge data in South Africa?
How do I read a DWS station number such as A2H019?
How many years of record do I need for flood frequency analysis?
Should I use a gauge that has closed?
Can the tool find USGS or Australian gauges?
search_stream_gauges. The agent passes the inputs, the engine does the calculation, and the record is saved to your project for you to review.
References
- Department of Water and Sanitation (DWS). Hydrology: Verified Data. Government of the Republic of South Africa, Pretoria.
- U.S. Geological Survey. National Water Information System (NWIS): peak streamflow data. USGS, Reston, Virginia.
- Bureau of Meteorology. Water Data Online and the National Flood Gauge Network. Commonwealth of Australia.
- World Meteorological Organization (2009) Guide to Hydrological Practices, Volume II: Management of Water Resources and Application of Hydrological Practices, 6th ed. WMO-No. 168, Geneva.
- Calitz, J.P. and Smithers, J.C. (2020) Development and Assessment of Regionalised Approaches to Design Flood Estimation in South Africa. WRC Report No. 2748/1/20, Water Research Commission, Pretoria.
- England, J.F., Cohn, T.A., Faber, B.A., Stedinger, J.R., Thomas, W.O., Veilleux, A.G., Kiang, J.E. and Mason, R.R. (2019) Guidelines for Determining Flood Flow Frequency, Bulletin 17C. U.S. Geological Survey Techniques and Methods, Book 4, Chapter B5.
Last reviewed by the Delta Hydro engineering team.