U.S. streamflow map: every gauge, against its own normal.
Every reporting USGS stream gauge in the country, coloured by how today's flow ranks against its own history for this calendar day. Search any river, isolate the extremes, and see where the water actually is.
Furthest below normal.
Recorded September 9, 2026 from the USGS daily-values feed, ranked by how far each gauge sits below its own median for this date — not by raw volume, which would only ever list the smallest creeks.
| Gauge | Flow | Normal | Of normal |
|---|---|---|---|
| Rio Grande Below Elephant Butte Dam, NM | 0 ft³/s | 874 ft³/s | 0% |
| Rio Grande AT Albuquerque, NM | 0 ft³/s | 394 ft³/s | 0% |
| Withlacoochee R AT Inglis Dam Near Dunnellon, FL | 0 ft³/s | 237 ft³/s | 0% |
| Ward Lake Near Bradenton FL | 0 ft³/s | 201 ft³/s | 0% |
| Canadian River Near Sanchez, NM | 0 ft³/s | 48.0 ft³/s | 0% |
| Pecos River Near Anton Chico, NM | 0 ft³/s | 39.6 ft³/s | 0% |
| Dirty Devil R AB Poison SP Wsh NR Hanksville UT | 0 ft³/s | 30.0 ft³/s | 0% |
| Little Creek AT Green Swamp Road Near Clermont, FL | 0 ft³/s | 29.1 ft³/s | 0% |
What you're looking at.
Every USGS stream gauge in the country, ranked against its own history for today's date.
Daily mean discharge
Each dot is a U.S. Geological Survey stream gauge reporting discharge — the volume of water moving past it, in cubic feet per second. The figure shown is the daily mean, the statistic USGS publishes once a day's record closes, and the same statistic the historical comparison is computed from. Comparing a daily mean against a daily baseline is what makes the two numbers on this map directly comparable.
Percentile of the historical record
A raw discharge figure means nothing without context: 200 cubic feet per second is a drought on one river and a flood on another. So every gauge is ranked against its own record for this calendar day — all the August 13ths USGS has ever measured there, sorted, and today's reading placed among them. A gauge in the 5th percentile is running lower than 95% of the days on record for this date.
The class boundaries are USGS's own — the 10th, 25th, 75th, and 90th percentiles, the same cuts the agency's streamflow maps have used for two decades. Inventing our own thresholds would mean publishing a national water signal that disagrees with the agency's, so only the colours here are ours.
The five classes
The scale diverges from the middle: normal conditions are drawn small and dim, and both directions brighten as they get further from normal. Roughly two in five gauges sit in the normal band on any given day, so dimming them is what lets the dry and wet anomalies read at a national glance.
- Much below normalBelow the 10th percentile
- Below normal10th to 25th percentile
- Normal25th to 75th percentile
- Above normal75th to 90th percentile
- Much above normalAbove the 90th percentile
- No baselineToo short a record to rank
Why some gauges are hollow rings
A percentile is only as good as the record behind it. A gauge installed three years ago can technically be ranked — against three numbers — and calling the result "much below normal" would dress up noise as a drought signal. Any gauge with fewer than ten years of record for this date is drawn as an empty ring instead: present on the map, visibly not classified.
Folding those into the normal band would have been easier and would have looked cleaner. It would also have meant the map quietly asserting that conditions are fine at hundreds of sites where nobody actually knows.
Updated daily, and dated on the page
This map is rebuilt once a day rather than queried from your browser on every visit. That is a deliberate choice: assembling the national picture takes roughly sixty-five requests against a public federal API that charges nothing and rate-limits nobody, and pushing that cost onto USGS once per visitor would be an abuse of a service everyone depends on. Harvesting once and serving the result from our own CDN means USGS sees a single pass a day no matter how many people open the page.
Little is lost by it. The map compares a daily mean against a daily percentile, so the classification genuinely only moves once a day. The date the data was gathered is stamped on the map itself — and for the live hydrograph, provisional-data caveats, and sub-daily readings at any single gauge, the detail panel links straight through to that gauge's page at USGS.
What streamflow decides.
The same number drives access, habitat, and risk decisions across very different programs.
Whether a river is runnable
Paddle trails and water trails live and die on flow. Below a reach's minimum, boats drag and portages multiply; above its safe maximum, the same reach becomes a hazard for the recreational paddlers it was designed for. Programs that manage public access are making a go/no-go call against gauge readings constantly, and usually by hand.
What fish and restoration work can survive
Sustained low flow concentrates fish into shrinking pools, raises water temperature, and drops dissolved oxygen — the combination behind most warm-season fish kills. It also strands newly planted riparian work and shifts the timing of spawning runs. High-flow extremes scour redds and move sediment through reaches that restoration crews have just stabilised.
Where a watershed is trending
A single gauge below its 10th percentile is a fact. Twenty adjacent gauges below their 10th percentiles for three weeks is a watershed heading somewhere, and it is visible on a national map well before it shows up in a drought declaration.
The same feed, wired into the daily work
Driftwise watches these same USGS gauges as one of more than a dozen federal inputs on every monitored river reach, alongside NOAA flow and flood forecasts, EPA water-quality data, wildfire proximity, and air-quality readings. A reach drifting toward its low-flow tail crosses an edge, and that routes recommended actions to the crews responsible for it — which is the difference between a map you check and a system that tells you when to act.
About streamflow data.
What the federal record behind this map does and doesn't tell you.
Where does this streamflow map get its data?
Every reading comes from the U.S. Geological Survey's Water Data service, the authoritative federal source for streamflow in the United States. Current values are USGS daily mean discharge; the historical comparison is USGS's own day-of-year percentile statistics, computed from each gauge's full period of record. Nothing on this map is modeled or estimated by Driftwise — the readings, the baselines, and the percentile thresholds are all the agency's.
What does a streamflow percentile actually mean?
It is today's flow ranked against every historical measurement at that same gauge for that same calendar day. If a gauge sits at the 15th percentile on August 13th, it means 85% of the August 13ths in its record ran higher than today. Ranking by calendar day is what makes the comparison fair: rivers have a natural annual cycle, so low summer flow is normal on most rivers, and the percentile asks whether today is low for this time of year rather than low in absolute terms.
What do the five colours mean?
They are the standard USGS classes. Much below normal is under the 10th percentile; below normal is the 10th to 25th; normal is the 25th to 75th, which spans the middle half of each gauge's own historical range and is where roughly two in five reporting gauges sit on a typical day; above normal is the 75th to 90th; and much above normal is over the 90th. The scale diverges from the middle — normal is drawn small and dim, and both dry and wet extremes brighten away from it — so anomalies are what you see first at a national zoom.
Why are some gauges drawn as empty rings?
Those gauges are reporting flow but have too short a historical record to rank honestly — fewer than ten years of measurements for this calendar day. A percentile computed from three or four years would be noise presented as a drought or flood signal, so those sites are drawn as hollow rings: plotted, but explicitly not classified. They are not the same as gauges reading zero, and they are deliberately not folded into the normal band.
What is cfs?
Cubic feet per second, the standard U.S. unit for river discharge. It is the volume of water passing a fixed point each second: one cubic foot per second is about 7.5 gallons per second, or roughly 646,000 gallons a day. Small headwater creeks typically run in the single or double digits, mid-sized rivers in the hundreds to low thousands, and the largest rivers in the hundreds of thousands.
How often does this map update?
Once a day. The national dataset is assembled at build time and served as a single file from our own CDN rather than queried from your browser, because building the national picture takes roughly sixty-five requests against a free public federal API — a cost that should not be multiplied by every visitor. The date the data was gathered is stamped on the map. Since the map compares a daily mean against a daily percentile, the classification only changes once a day anyway.
Is this data provisional?
Recent USGS readings are published as provisional and are subject to revision as the agency reviews them — this is normal for any near-real-time hydrologic data and applies to every source that carries it, including this map. The historical percentiles are computed from approved record. For the full hydrograph, provisional-data caveats, and sub-daily readings at any single gauge, the detail panel links straight through to that gauge's own page at USGS.
Can I use this for flood or safety decisions?
No. This is a situational-awareness map built from daily federal data, not a flood warning system. It is refreshed once a day and shows a daily average, so it cannot represent a fast-rising river. Our flood outlook map carries NOAA's own flood categories and is the better starting point for that question — it is forward-looking and refreshed hourly — but it is not a warning system either. For flood forecasts and warnings, use NOAA's National Water Prediction Service and your local National Weather Service office; for life-safety information, follow your local emergency management agency.
Why did WaterWatch stop working?
USGS retired WaterWatch and the legacy waterservices.usgs.gov endpoints, which for two decades were how most people got national streamflow percentiles. Their replacement is the api.waterdata.usgs.gov family — a modern OGC-standard service for current values plus a separate statistics service for percentile normals. This map is built on the replacement APIs, which is why it can still show the classification WaterWatch made familiar.
Does Driftwise use streamflow data in its monitoring platform?
Yes. Driftwise watches USGS gauges as one of more than a dozen federal inputs on every monitored river reach, alongside NOAA flow and flood forecasts, EPA water-quality data, wildfire proximity, and air-quality readings. Sustained low flow on a monitored reach crosses that input's edge and routes recommended actions — access closures, fish-kill watches, restoration timing — to the crews responsible for it.
River levels by state.
Every state, with the number of USGS gauges reporting there.
- Alabama140 gauges
- Alaska90 gauges
- Arizona186 gauges
- Arkansas111 gauges
- California437 gauges
- Colorado320 gauges
- Connecticut69 gauges
- Delaware17 gauges
- District of Columbia3 gauges
- Florida374 gauges
- Georgia283 gauges
- Hawaii79 gauges
- Idaho193 gauges
- Illinois185 gauges
- Indiana193 gauges
- Iowa135 gauges
- Kansas178 gauges
- Kentucky148 gauges
- Louisiana54 gauges
- Maine64 gauges
- Maryland128 gauges
- Massachusetts108 gauges
- Michigan199 gauges
- Minnesota110 gauges
- Mississippi71 gauges
- Missouri226 gauges
- Montana223 gauges
- Nebraska113 gauges
- Nevada145 gauges
- New Hampshire53 gauges
- New Jersey117 gauges
- New Mexico155 gauges
- New York297 gauges
- North Carolina233 gauges
- North Dakota102 gauges
- Ohio215 gauges
- Oklahoma134 gauges
- Oregon239 gauges
- Pennsylvania296 gauges
- Puerto Rico99 gauges
- Rhode Island34 gauges
- South Carolina122 gauges
- South Dakota104 gauges
- Tennessee106 gauges
- Texas515 gauges
- Utah174 gauges
- Vermont51 gauges
- Virginia198 gauges
- Washington234 gauges
- West Virginia102 gauges
- Wisconsin181 gauges
- Wyoming104 gauges
The other public maps.
Free, no-login maps built from the same federal data the Driftwise engines read.
Watershed Explorer
Click any point in the country and read what six federal datasets say about it — flow, drought, fire, flood forecast, satellite cyanobacteria screening and Clean Water Act impairment — each one labelled with what it is actually a fact about.
Wildfire & Smoke Map
Official NIFC fire perimeters and incident points under NOAA’s HRRR-Smoke near-surface forecast, filterable by size, state, and containment.
U.S. Drought Map
The U.S. Drought Monitor — the country’s official weekly drought assessment — drawn across five severity categories from D0 to D4.
Flood Outlook Map
NOAA’s river forecast centres call the next few days at every forecast gauge in the country, on the National Weather Service’s own flood categories — action, minor, moderate and major.
Impaired Waters Map
Every subwatershed in the country against the states’ own Clean Water Act assessments — what is impaired, what is still on the 303(d) list, what is causing it, and how much of the country nobody has assessed at all.
Cyanobacteria Bloom Map
NASA and EPA’s satellites read a pigment index over roughly 2,300 U.S. lakes large enough to resolve at 300 m. A pre-screening indicator, not a health assessment, and never a substitute for your state’s advisory.
Critical Habitat & Protected Waters Map
Every river and stream the federal government has designated as ESA critical habitat, the listed species each unit is designated for, and who manages the protected land around it.
One platform for three monitoring programs.
Run a mark-recapture program, coordinate a multi-partner invasive species response, or manage the waterway program for a state park or conservation nonprofit. Bring your own program to the demo and we'll run Driftwise on it, live.
Request a demoA free 30-minute working session with your data on screen.
Scope a pilot for your program.
Tell us about your program — a watershed, an infestation, a trail system. We'll walk through how a field record becomes the report your agency and funders sign off on.
A working call, not a pitch
We spend the time on your actual program — the data you have now and the report you owe.
Guided onboarding
We import your existing records and configure the setup alongside your team. Role-based in-product tours carry the rest of the crew after the call.
Start with one season
Run a single pilot first and expand when it earns the next watershed.
Send us your corridor, watershed, or program and we'll build the working demo around your records. We reply within one business day.
Click to reveal — we obscure the address to cut spam.