What is a weir and how does it work?
Each of the numbered watersheds at the Hubbard Brook Experimental Forest encompasses the upper section of a first order stream and has a weir at the bottom. The purpose of the weir is to intercept the water flowing down the stream (in theory, all of the water coming out of the watershed) in order to determine flow rates and amounts over time and to track seasonal and annual patterns.
1. Weir structure
2. Chart recorder
3. Gathering and interpreting data
1. Weir structure
The structure consists of a ponding basin, a stilling well, a V-notch "weir", a flume, and concrete "wings", which reach slightly outwards uphill to catch and direct the stream and any nearby ground water into the flume and basin. The weir is anchored in the bedrock so that no water can flow under it. The amount of water flowing through the weir is measured continuously as is passes through one of two components--a V-notch at the downstream end of the basin for most flows, or the flume for very high flows. When we use the term "weir", sometimes we are referring to the whole of the complex and sometimes just to the V-notch.
The V-notch functions to control the flow of the water out of the basin. Stream water passes between the concrete "wings", through the flume and ponds in the concrete basin. Water flows out through a steel V-shaped notch situated in the basin's downstream wall. The V-notch controls the flow of water out of the ponding basin and the more water there is coming down the stream, the higher it will rise in the basin and V-notch as it spills over. Therefore, the amount of water flowing out of the watershed can be measured by recording the height of the water in the basin, and translating this height to a flow rate in cubic feet per second or liters per second. Flow rates for any given height in the V-notch were measured empirically at the time the weir was installed and are checked periodically. Mathematical functions are prepared which relate the height of the water in the basin to the flow rate.
The height of water in the basin is measured continuously by a chart recorder housed in an outhouse sized building. This building sits over a square concrete "stilling well" which is connected to the ponding basin by two pipes through which water flows freely (see diagram above). In this arrangement, as the water rises or falls in the basin, the water equally rises or falls in the stilling well. A float on the well water surface is connected to the chart recorder, and its rise and fall is recorded as changes in water height. In the closed and protected well, the water surface is unaffected by wind and waves and remains very still, so that irregular height recordings are minimized.
In winter, ice formation interferes with the normal functioning of the weir. If the ice builds up in the V-notch, the height of the water will artificially increase and the calibration is thrown off. To avoid this problem, a wooden cover is installed over the ponding basin and a gas fired heater beneath this cover keeps the the V-notch from freezing. Similar heaters are placed in the chart recorder house and by the flume recorder.
Sometimes stream flow is so great that water overtops the V-notch and flow can no longer be accurately estimated by water height in the ponding basin. When this happens, the San Dimas flume is used to record flow rate. The flume lies just upstream of the ponding basin and beside the chart recorder house. It is a long rectangular steel "chute" that stream water flows through before reaching the ponding basin. It is the black metal structure in the upper left of the photo. There are slits in the sides of the flume, which allow water to flow into a small box beside the flume where another height recorder is housed. Normally water through the flume is not much more than a trickle, but at high flows water can be several inches to a foot or more deep. The flume cannot be calibrated as the V-notch was to develop an empirical height-flow relationship, but hydrologists and engineers have developed theoretical rating curves for these types of devices. Recorded height data from the flume is translated to flow rate just as it is for the V-notch, but with less precision.
2. Chart recorder
A clock driven chart recorder is housed in a small structure next to the weir. The large wheel to the right of the recorder is driven by a counter-weighted cable attached to the float on the water surface in the stilling well. The counter weight is visible in the photo just under the table. The steel band which passes from the counter weight over the wheel and down to the float below is visible just to the right of the counter weight under the shelf. As the water level in the well rises and falls, the wheel is turned back and forth. The wheel is in turn connected by a gear system to an ink pen which draws a continuous line on the chart paper. A clock slowly feeds a roll of chart paper through the flat drawing bed, to ultimately produce a graph of water height over time.
Click here to see a hydrograph chart
The height of the line on the chart paper is a record of the height changes of the water in the stilling well which is a reflection of the height of the water in the ponding basin which is controlled by the V-notch weir. The flow rate of the water over the V-notch is calibrated against the height of the line on the chart. This calibration is called a "rating curve" and is developed by actually measuring the flow of water over the V-notch over a range of water heights in the basin (which is a function of stream flow rate). This calibration is done using a large tank of known volume. It is placed below the V-notch and the flow is directed into the tank and timed with a stop watch until it is full. Then the flow rate is determined and the position of the pen on the graph paper is noted. This process is repeated at a number of basin heights so that x-y coordinates of the pen on the graph paper can be interpreted as actual flow rates of the stream.
3. Gathering and interpreting data
Once a week the paper on the chart recorder is replaced and the clock is rewound. The chart is returned to the lab and the flow curve is digitized using a device, which when moved along the line can record the x-y coordinates for various points along the curve. The x-y points are recorded every 30 minutes during periods of minimally changing flow and every 6 minutes when the flow begins to increase during a rain event as the stream rises. From these digitized time-flow data it is possible to calculate the amount of water leaving the watershed for any period of time.
The continuous record of hour by hour flow rates are on the Hubbard Brook web site. We have extracted a short segment of data as illustrated in the table below. Lets translate these data. The data is from Watershed 6 on April 12, 1963. Let's take data from 11:30 AM and 12:30 PM, which we have highlighted. The gage height is recorded in feet and is the height of the water in the V-notch. The hydrologists, using their computer programs and the rating function described above, have calculated that the flow rate in cubic feet per second at 11:30 AM was 0.2309 and at 12:30 PM was 0.3134. If we average these two numbers we can get the average flow rate for this one hour interval in tens of cubic feet per second--0.27215. Since there are 3600 seconds in an hour, there were 979.74 cubic feet of water that left the watershed in the one hour period.
Rain is measured in inches that fall on a given point and it is useful to know what 979.74 cubic feet of water would amount to in inches spread out evenly over the watershed. In this example there were 0.00822 inches of water which left the system in this hour. This number is listed under "Interv. Inch" (interval inches), which means inches of water which left the watershed in the one hour period. We can calculate that number ourselves from the flow rate and by knowing the area of the watershed, which is 13.23 hectares. Since we think in hectares and hydrologists think in feet and inches and acres, we need to make some conversions. There are 2.471 acres in a hectare, thus Watershed 6 is 32.6913 acres in area. However it would be convenient to get that into square feet. There are 46500 square feet in an acre so that comes to 1,424,034 square feet. Now if we divide the 979.74 cubic feet by the 1,424,034 square feet we get the feet of water on each square foot of area. This comes out to 0.0007. But that is in feet and we need inches, so multiply by 12 inches per foot and now we get 0.0082 inches of water, the value recorded in the table below. The other column headings are for summaries which will appear at the end of each day and at the end of each month.
WS MO DA YR TIME GAGE DISCH. INTERV. ACCUM. MEAN DAILY ACCUM. ACCUM.
FT. C.F.S. INCH INCH DAILY MM. INCH MM.
(DAILY) C.S.M. (MONTH) (MONTH)
6 4 12 63 930 .357 .1979 .05945 .05945
6 4 12 63 1030 .360 .2021 .00607 .06552
6 4 12 63 1130 .380 .2309 .00656 .07208
6 4 12 63 1230 .430 .3134 .00822 .08030
6 4 12 63 1330 .494 .4415 .01138 .09168
6 4 12 63 1430 .540 .5501 .01500 .10669
6 4 12 63 1530 .561 .6044 .01750 .12419
6 4 12 63 1615 .571 .6314 .01406 .13825
6 4 12 63 1730 .571 .6314 .02394 .16219
6 4 12 63 1830 .566 .6178 .01895 .18114
6 4 12 63 2130 .530 .5252 .05195 .23309
6 4 12 63 2400 .509 .4753 .03793 .27101
So, if you were born after 1956, you can see what the flow rate was on the day you were born and if you know the time you were born you can see the flow rate at the time you were born. For fun look up the flow rate for 2 to 3 in the afternoon of July 10, 1973!!
By these calculations the amounts of water flowing out of W6 per day, month or year can be calculated, thus providing the "output" part of the water budget for the watershed. Approximately 6 million liters of water (about 1.6 million gallons) flow out annually, but that can vary quite a lot.
A couple of further notes:
At Hubbard Brook, the use of "old fashioned" clock driven recorders rather than modern electronic devices is due to the proven record of reliability of the mechanical recorders, and probably to some extent, to the cost of a change-over at the 9 weirs in the valley which would require recalibration. Also there is no electricity at the weirs. The "modern" recorders would leave no physical record as a chart recorder does so that if problems arise during a weekly period there is a physical record to work with.
Web page created January 2001
by Thomas Siccama and Ellen Denny