Understanding Streamflow Data

I am new to kayaking in rivers. I don’t understand how to read the USGS Real-Time Data on Streamflow. All I know is that I was on the Brazos River in Texas over the Memorial weekend. A couple of places were pretty low and I had to port my kayak at one point and ended up with some pretty bad scrapes on some rocks in shallow water. Depth height/feet I am guessing is just the depth of the river. I don’t understand discharge ft3/s and long-term median flow. Can someone explain those terms.

Some Rough Info

– Last Updated: May-28-09 11:51 PM EST –

Gauge height, or "gage" height as the USGS spells it, is usually arbitrary. Years ago, someone nailed a big ruler to the side of a bridge pier to keep track of river elevations, and ever since then, the gauge readings have been based on that original measurement method. Okay, it may not ALWAYS be that simple, but usually it's something like that. Therefore, the gauge reading is only meaningful if you are familiar with that particular section of that particular stream.

Discharge in cubic feet per second is also meaningful only in local context. A flow volume of 3000 cubic feet per second on one stream might be suicidal paddling conditions during a major flood, while the same volume of water moving down another, much larger river, might be associated with a very slow current during summertime drought conditions where you end up needing to drag your boat half the time. Again, you need to be familiar with what the various flow rates mean on a particular stretch of a particular stream.

Long-term median flow is simply the median value for a particular day, calculated for the entire history of that particular gauge. Again, that isn't terribly meaningful, as flood conditions tend to have flow rates of many times the normal rate, while low-flow conditions are usually just some "reasonable fraction" of the normal flow. Thus, flood conditions can skew the average, and what's "normal" often appears to be "less than normal" when compared to long-term median flow. This may not always be the case, but it certainly can be, and I provide this as another example of how you really need to know the area and have some familiarity with what kind of paddling conditions occur at certain gauge readings.

Locals who paddle the river on a regular basis are likely to know this stuff, so newcomers can find out a lot from them, IF you know who to ask.

As GBG said, the numbers themselves are arbitrary, but they gain meaning when you combine them with people’s prior experience on the river at similar gage conditions. It’s a good idea for river paddlers to keep a record of the gage levels every time they visit a given segment.

Gage height and water discharge are related in that both go up together. If one moves higher, the other moves higher also. Because streams get wider as they get higher, the discharge rate moves up faster than the gage height.

Generally, gage height is the most commonly used statistic to describe river conditions. On some rivers, by local convention, most people use discharge rates. I’m not sure why, except that it seems to be mostly whitewater rivers and often the smaller ones where discharge rates are used.

The median streamflow is intended to be a guide to normal river levels. Actually, GBG, I don’t think it will usually be distorted by floods, because it is a median average (midpoint of all observations) rather than a mean average (sum up all observations and divide by number). However, it has other problems. The main problem is that it is an average for the entire year, whereas (for most of the country) you can expect normal fluctuations by season.

A better guide, in my experience, is to create a graph on the USGS site of daily readings going back a year or several years. In particular, this lets you see a baseline of how low the river gets in different seasons. I checked a few spots on the Brazos, and Memorial Day weekend did indeed have levels very close to the low points on the history graph, so it’s not surprising that you were dragging bottom. (Which gage were you near?)

A few other helpful numbers you can get are several statistics about flood levels (each gage station has numbers for Action Stage, Flood Stage, Moderate Flood and Major Flood – Action is “banks full”) and record levels (they usually list the 4 or 5 highest and lowest readings on the gage in its history). Of these, the most helpful to me are Action Stage and the highest of the low water records. Those two numbers create a range of where the river usually is, and can be compared to the current reading.

One final point in regard to the your experience – there are many dams on the Brazos, and these of course have a greater impact on water levels than rainfall. When planning a trip, it is often helpful to try to find out whether the dams upstream of you are likely to be releasing water. Most major dam projects have their own websites or basin websites with this information.

gauge in feet versus CFS

– Last Updated: May-29-09 9:42 AM EST –

Some stick gauges are not meaningfully placed, being situated in relatively wide pools where an enormous increase in river volume may result in only a small rise in river inches. Downstream of such a pool, if the river channelizes, an increase in gauge height of an inch might result in a doubling of flow through the narrow section.

Stick gauges that read feet often seem to be geared more to judging where the river level is relative to flood stage. Whitewater boaters often prefer to go by CFS or "cubes" since it at least provides some non-arbitrary measure of volume, although as GBG said, 1500 CFS on a small creek may result in Class V flow and 1500 CFS on a mature river may result in an unfloatable level. Whitewater boaters will, however, use what they can get, and stick gauges, at least, provide an instantly visible measure of river volume. Boaters experienced with a certain river may be able to fairly accurately estimate flow in CFS, but you can't read CFS off a stick gauge or gauge painted on a bridge piling.

Realize also that the same river often has stick gauges at multiple locations and the readings at the same moment in time sometimes seem illogical due to the completely arbitrary nature of stick gauges. For example, an upstream gauge may be reading 2' 6" and another gauge 5 miles downstream may read 8" below zero. A stick gauge at a pool may be nearly unchanged when a second in a narrowed channel shows an increase of a foot or more. So make sure you are reading the right gauge when trying to judge river level based on verbal or printed recommendations.

Does anyone know of any info online about the actual gauges and how they work? I’m trying to solve a local mystery (well, of interest to only me:) ) on how/why the relationship between the height and the CFS recently changed on a local river. I’m interested in finding out more about the hardware itself and how it measures the flow.

Answered my own question… here’s a good page from the USGS:


What I was thinking about Median Flow
The “median flow” that I was thinking about is the median value for a particular day, over the entire history of the gauge. The mean values are platted on the daily charts, and every day on the chart has a different value for long-term median flow. On the lower Wisconsin River, “normal” summer and autumn flows (flow rates that anyone who sees the river on a regular basis would consider typical) are always less than the long-term median flow values shown on the “real-time” charts. During summer and fall, if the water level is into the realm of what we would call “high”, say 1.5 to 2.0 feet higher than “normal”, the current flow rate at or slightly greater than the posted long-term median flow rate for that day.

This is very strange. I figured it had something to do with flow rates on days during occassional big floods more than just cancelling-out the days when flow was very low, since “very low” flows occur at gauge readings not more than half a foot lower than what you see on an “average” day, in short, because floods go a lot higher than the degree to which low water is “lower”. It might also be true that this particular river simply has a whole lot less water in it than it did for most of it’s recorded history, but I find that prospect a lot harder to believe. It’s all speculation on my part. I was just guessing at a possibility that seemed to make some sense.

River info

Check the gauge record for the area you paddled on the day you paddled it. There’s your first reference point – you know what that number means in real life. It sounds like you’d want a bit more water before you ran that stretch again.

The American Whitewater site has some information on interpreting the gauges – the Brazos is currently shown as “low runnable”, but it’s up a bit from where it was last weekend.

what mystery?
Tell us about your local mystery on the gage, I’m curious.

One systemic change I know of in the last few years is that they’ve gone from manual readings to automatic readings. This required them to install new gages at all locations. I noticed several of the local gages change charater somewhat when that change took place - like, the previous typical minimum flow went up or down by a few tenths of inches, or new record highs were set several times a year for the first few years.

Also, changes to the riverbed can change gage readings, like if they channelize a segment upsream of the gage.

How we do it
We use the gage height.

We have kept a record for about the past ten years of the height of a river when we run it.

Then we can tell by the gage height if the river is higher or lower than when we ran it.

For instance for the past week I have been tracking the New River in NC since we have a race coming up ther a week from now.

Last year the gage height was 2.06. right now it is 4.03 which translates to kevlar boats will be good if it stays about like it is.

It does not translate to flow speed though since just recently we ran a river that was so high that it had overlowed the banks and was into the woods creating a swamp on both sides, yet the river was not any faster than a year ago when it was well witin its banks.



good points
Good points about reasons to use CFS.

One thing, it actually would be possible to make a CFS gage and nail it up on a bridge. As long as the river bed stays the same, 3 feet on the height gage will always produce the same number on the flow gage, whatever it is. You can get the history for the two numbers off the USGS web site, calculate the equivalent numbers and then put them up next to the height gage.

People should do this for those rivers where CFS is the standard talking number. It wouldn’t be hard, although you’d want to be sure it was right so you aren’t spreading misinformation. Probably a good idea to coordinate with a local paddling club so they could check it over time.

Also, if you’re not sure about river conditions over time, you can call the USGS and talk to the individual analyst who covers that river. Because the CFS numbers are calculations based on estimates and the estimates are critical, usually a trained geologist on the spot has made the estimates. They keep notes on this, so the current analyst should usually be able to quickly check his notes file and tell you the last time there was a change in the riverbed that affected flow.

BTW, the analysts are usually paddlers too (much of their work is on the river itself, so they’ve at least been on it in jon boats and waded it in places), and they are cool companions to have along on trips. I’ve met and canoed with several who cover various southern rivers.

what river?
Hey, Jack, what river was that, and what are the dates of the flood race and the regular race? Also, what’s the location of the race, and what is your evidence that it was no faster? I’d like to look at those cases and see if I can explain it for myself.

I’ve been studying this topic a bit over the past year, mostly in regard to the Mississippi River, but the principles should be the same on any river. As I understand it, you should expect the surface speed to be faster in flood UNLESS counter currents are created by the shape of the drainage in flood. Thus, I should be able to look at topos and aerial photos of the river you mention and find some cause of counter currents.

The general rule, in absense of counter currents, is that the speed of the surface flow is a function of the distance from that spot to all sources of friction in a 180-degree arc (mostly the sides and bottom of the stream, but also any standing obstacles in it). Thus, any increase in the surface level is going to increase all of those distances.

This relationship is just as steady and just as dependent upon gage height as is flow. In other words, 20 feet on the height gage corresponds to a given number on the CFS meter and a given speed for a designated spot on the surface.

For example, near the mouth of the Red River, the Mississippi has the following relationship of gage height to surface speed in the fastest part of the channel:

Height MPH

10 2.3

20 2.8

30 3.5

40 4.6

50 6.2

60 8.1

So, to extend my comments about posting a gage showing CFS, you could really have a 3-column river stage gage for a given set of conditions, showing river Level, Flow in CFS and Current Speed. However, Current Speed would probably take too much explaining, so I guess that kind of river stage info might not be useful.

Black Water River Race on the
Edistoe in SC a few weeks ago vs last year at the same time.

I don’t have technical evidence. Just basing the overall times of the paddlers each year.

It also seems to be that way in the annual Suwannee River Race, and the annual Lumber River race in NC.

It only seems to be on the rivers where the water is way over the banks and into the woods for a half a mile or so creating one big flowing swamp.