On gauges: How do they do that?
by Mark Taratoot
I was just reading on the LCCC email list about a gauge on the Wilson that had become inaccurate. Well, sort of. The calculation of discharge (flow, cubic feet per second) had become inaccurate; the river stage (level in feet) was still just as accurate as ever. This can happen if floods or debris affects the river near the gauge. Sediment can be deposited or flushed away. Sediment isn’t always sand and silt; sometimes it’s boulders. Fluvial geomorphologists (scientists who study how rivers flow) call all this stuff sediment. It was really amazing to listen to giant bits of sediment probably weighing hundreds of pounds tumble under our rafts on the Alsek.
It’s actually pretty neat that we can even make good inference about how much water is flowing past a point with just a measure of the river stage. The relationship is far from linear. Geomorphologists develop “rating curves” to estimate discharge from state, and it takes a lot of work to develop a rating curve.
One thing you need is a way to accurately measure the river stage. There are a number of ways to do this. Sometimes it’s just a stick gauge that you use your eyes to read. Want to know the stage on the Calapooia? You’ll have to go look at the stick. Or you could trust Pat Welch’s most excellent estimate that’s based on adjacent watersheds. Other streams have automated gauges that send data back to us in town. All these gauges need to have a stilling well – a place connected to the river but away from the flow. This reduces the effect of waves and other nonlinear flow on the reading.
The most basic automated gauge used to just have a float not unlike the one in the back of your toilet. A sensor measured how high the float was floating. It was kind of a coarse measurement. A more accurate and precise way to measure river stage uses pressure transducers. The pressure transducer is connected to a bottle of inert gas (nitrogen). The transducer measures how much pressure it takes to create a bubble. This pressure is directly related to the depth of the water above the transducer. It’s a linear relationship, and it’s very accurate. Other modern gauges use radar to look down at the surface of the water in the stilling well to read the stage. This is also accurate and precise, and it doesn’t require that someone bring a new cylinder of gas up from time to time.
That’s all we actually measure – river stage. How high’s the water mama? Two feet high and rising. But what does it mean to be at two feet? This is where the rating curve comes in.
Geomorphologists and hydrologists need to develop this curve, and it takes time and data. The first thing that has to get measured is the river’s cross sectional area at the gauge. This measurement can be taken at any given flow level. Scientists measure the shape of the bed. They can measure what that cross sectional area is at different river stages. Streams tend to get wider as they move from the bed to the banks, and when the river gets out of the banks it spreads out even more. That cross sectional area changes at different stages. It’s still not hard to understand this mapping. When sediment is deposited or removed from the area near the gauge, this will clearly affect the accuracy of the equation that tells us discharge.
The other thing that is needed to refine a rating curve is what the actual flow velocity is at different places in the stream and at different river stages. Scientists can use current meters to take measurements at several depths at each of several places across the stream. This data gets more and more refined as measurements get made at different stages. The flow velocity in the same place might be significantly different at different water levels. When you see a cable with what looks like a boatswain's chair above a gauging station, that’s what they use to measure the current across the entire width of the river. For small streams, scientists can use hand-held equipment and wade across. For very small streams, scientists can run the entire flow through a weir and get very precise measurements of discharge at different stages.
There’s other methods too. Gordon Grant has been working on using remote sensing to estimate discharge just from looking at river waves. Pretty groovy!
Discharge, called Q in the hydrology world, is just the integration of cross sectional area and flow velocities. Having more measurements of flow velocities at different river stages improves the accuracy of the rating curve. When something happens to disturb the cross sectional area, a gauge can be recalibrated by just remapping the new area. If the flow velocities change, the rating curve won’t be as good as if new velocity measurements get added. The curve can always be refined.
We can also thank people like Pat Welch. Pat uses his understanding of physics to make estimates of discharge on streams that don’t have gauges. In a way it’s simple. In a way it’s not at all. Pat looks at a watershed that has an ungauged stream. He can look at adjacent watersheds where there are gauges and develop equations to estimate what the flow is in the ungauged stream. He can refine his models if real-world experience show that the calculations aren’t quite right. If you go paddle a stream that Pat has an estimate for, and if your experience says his numbers aren’t right, give him a buzz and let him know. He can tweak his calculations to make our inference better when we want to go run ungauged rivers or creeks. I think it’s great that Pat leaves up his estimate for the North Fork Willamette even though it has a gauge now. There is no rating curve for the gauge at Westfir, just river stage. For folks who’ve been paddling the North Fork for years, Pat’s estimate is a great way to decide if you want to go.
Over time, we might even get a good feel for what the river stage means to us as boaters.