How’d you make that chart I was looking at?

NOTE: This page has been updated 08 November 2014 to reflect revisions to the NAVD88 elevation of the NGS bench mark F-45 used to control this survey.  The Maine Department of Transportation has recently published a revised elevation for F-45 based on a recently completed GPS control survey. The new elevation of the bench mark is 178.746′ (NAVD88) or 0.746′ higher than the original NGS value. As of November 5, 2014 I have revised my GPS control point elevations higher by this amount. Here’s the new F-45 Maine DOT Survey Control Points Report. All elevations except those in the narrative have been adjusted accordingly.

A number of people want to know more about the water level measurements I’ve been making- where did I get the elevations from, how do I know what the elevation of the dam is, how do I measure the distance below the top of the dam, etc. In other words, how did I come up with the chart you were looking at? Well to start, it helps working for a Land Surveyor who’d let me borrow tens of thousands of dollars worth of surveying gear. I borrowed 4 of my boss’s Locus(tm) GPS receivers and on the afternoon of 18 December 2011 I ran a differential GPS level loop around 4 points, 3 in Whitefield and 1 in Jefferson:

• For control, a USGS bench mark designated F-45 with a published (3rd Order) NAVD88 elevation of 178.00 feet. (Note: this data sheet has been superseded, see new MDOT data sheet with new elevation of 178.746′);
• a drill hole on the bridge abutment on Route 218 at the dam designated “STATION D” (STAD);
• a drill hole in ledge on the shoreline at my house designated “STATION G” (STAG);
• a spike in the edge of the pavement in David Hodsdon’s driveway in Jefferson designated “STATION H” (STAH);

The whole GPS survey took about 4 hours. Crunching the data was finished the next day. Positions are in feet (Maine State Plane Coordinate System, West Zone). Elevations are based on the North American Vertical Datum 1988 (NAVD88). The accuracy of the survey was superb. The network adjustment resulted in the following site positions:

8 November 2014: The elevations in the following table have been adjusted upwards by 0.746′ to reflect the revised elevation of control point F-45:

Site                                  95%   Fix    Position
ID      Site Descriptor Position      Error Status Status

1 STAG  DH on shore East. 3111535.922 0.009        Adjusted
                    Nrth. 502329.591  0.009
                    Elev. 150.476     0.015

2 STAD DH at Bridge East. 3105721.273 0.007        Adjusted
                    Nrth. 508931.243  0.007
                    Elev. 152.033     0.012

3 STAH Nail/Hodsdon East. 3114759.139 0.010        Adjusted
                    Nrth. 502762.628  0.008
                    Elev. 162.274     0.016

4 F-45 USGS Bench   East. 3100425.726 0.000        Adjusted
                    Nrth. 507111.659  0.000
                    Elev. 178.746     0.000        Fixed

The narrative from here on includes elevation data that has not been updated and is for illustrative purposes only.

With the bench mark elevations accurately established, it was a simple matter to pick a point on the top of the dam and accurately locate it in 3 dimensions via traditional remote resection (i.e, locating a point in space by measuring distances to it from 3 or more known locations). This gave me the elevation at the top of the dam: 151.64′ NAVD88. Finding the elevation of the lake surface is then a simple matter of making a level observation of the bench mark on the shore at my house and measuring the amount the lake surface is below the bench mark. Subtracting the elevation of the lake from the elevation of the top of the dam gives the distance below the top of the dam in feet. Multiplying that figure by 12 gives you inches. 

For example, if I determine by level observation that the surface of the lake at my house is 3.50′ below my bench mark (elevation 149.73′)then the lake surface is at 149.73′ – 3.50′ = 146.23′ The distance below the top of the dam is then 150.64′ – 146.23′ = 4.41′ or 52.9″ (4.41 x 12). A note about these elevations: the level of the lake below the top of the dam I am talking about is not to be confused with the actual elevation of the water as measured at the dam. This is because when the dam is open and water is flowing out, there is a gradient (i.e., a difference in elevation) between the lake surface and the water level at the dam. Think about it: water only flows down hill so if there is no gradient, there is no flow. By observation this difference has been determined on average to be approximately -0.17′ (i.e., -2″) though during times of high runoff this difference can be much greater, as much as half a foot or more. In other words, if the dam is open and rainfall or snow melt runoff at the time  is minimal or non-existent, the water level at the dam will be ~2″ lower than the level of the lake. Clearly when referring to water levels, elevations should measured at the lake and not at the dam. To make the actual water level measurements, I use a simple vernier reading non-optical clinometer (pictured above) with an adjustable bubble vial that mounted on a cheap camera tripod and a folding 6′ ruler graduated in feet, tenths, and hundredths. The vernier scale reads directly to 10’of arc (about the angle subtended by pack of cigarettes at 100′) which for these purposes is way more accurate than it needs to be as a simple hand level would suffice. In any event this unit adds precision to the measurements, which is a good thing. Normally a clinometer is used to measure vertical angles but with the scale set to 0° and the bubble kept level, it can be used as a level. The clinometer is set up level over a hole chopped in the ice and adjusted up or down by adjusting the tripod legs until, siting through the eye piece, it is level with the bench mark located 15′ away on the shore. The height of the clinometer above the water surface is then measured with the ruler. Simple! Today’s (11 March 2012) reading was -3.47′ (i.e., the water surface is 3.47′ below the elevation of the bench mark). This translates into a water surface elevation of 146.26′ (149.73 – 3.47) which is 52.56″ below the elevation of the top of the dam. The rest of the story is pretty tedious. Every day, or every few days, after making the measurement I plug the data into a spread sheet I designed to keep track of the data which does all the simple calculations. I also periodically download precipitation data from the Weather Underground web site and plug that into the spreadsheet. I’ve told the the spreadsheet which columns of data to graph, and the software does the rest. All in all it really doesn’t take long on a daily basis to collect the data- about 5 minutes.

[Spring Update 10 April 2012]

Since writing the above, the spring has arrived and the ice has melted. Making water level measurements now is a simple matter of going down to the lake with a 10′ straight edge, a carpenter’s level, and a folding ruler. No chopping holes in the ice to find the water!

Interpreting the Chart

There is a lot of information to be gleaned from the chart. First a note about precipitation: inches of precipitation (the scale on the left side) is a measure of how much actual water fell whether it was in the form of rain, or snow. Inches below the top of the dam is shown on the right hand scale and is based upon the measurements I have made as described above. The correlation between precipitation and lake level is clear but is complicated slightly by the fact that snow doesn’t immediately translate into a corresponding rise in the lake level as if the snowfall had been rain. Obviously, when it rains or when snow melts, the lake level rises (the weight of snow on the ice also serves to raise the level of the water some, but I’m not sure how much). How much it rises can be surprising. If the lake were on the top of a mountain and the only rain that made it to the lake was what actually fell in the lake, then when it rained an inch the lake would only rise and inch. However, the area that drains into Clary Lake is more like 9 square miles, roughly 8 times the size of the lake so when it rains, much more water than actually lands in the lake ends up in the lake. This would suggest that when the ground is frozen pretty much all the precipitation should run off and an inch of rain would bring the lake up as much as 8 times as much, or 9″ total (don’t forget the inch that fell in the lake directly!), a rather large multiplier! This winter we’ve seen multipliers in the range of 5 to 7 for certain rain/snow events which is pretty close to the theoretical 8x multiplier. At other times of the year when the ground isn’t frozen or completely saturated and more or less water is retained in the ground, less water runs off and the multiplier is more like 3x or 4x i.e., and inch of rain would bring the lake up 3″ to 4″. From this information we can conclude that since the lake is currently down about 53″ below the top of the dam and that our snow pack, for what its worth, has already melted, that it will take about 15″ of rain this spring to bring the lake up to its usual spring level. Clearly, that isn’t going to happen so we can look forward to extremely low water levels this spring and this coming summer. Another interesting thing we can glean from the chart is the rate at which the lake level falls when the dam gate is wide open. Now we know from past experience that when the lake is full or mostly full, with the gate fully open the lake level drops approximately 1 inch a day. However, as the lake level drops, not only does the pressure behind the dam (the “head”) fall but the friction of the water flowing down the channel increases. Both serve to slow the flow of water to the dam and the rate at which water flows through the gate. Eventually, without some runoff to replace the water being lost, the flow will finally stop. So the line on a graph showing the rate at which the lake level falls from the highest level (at or slightly above the top of the dam) to lowest level (wherever that is) would actually be a curve showing the level falling at 1″ per day at the top end to 0″ per day at the bottom end. But the real question everybody wants an answer to is, at what elevation does the lake level stop falling? How low can it go? This is the question people on Clary Lake have been asking for years! Since there’s almost always going to be some runoff to replace water flowing out of the lake, “as low as it gets” is going to vary from year to year depending on the weather. At a certain point the lake level will stop falling as the water flowing out of the lake equals the water flowing into the lake. What level this state of equilibrium corresponds to will depend on the level of the lake, the time of year, and the rate of runoff. It should be obvious that in a dry year it will go lower than in a wet year. This past fall we hit equilibrium at a level approximately 60″ below the top of the dam i.e., in mid-November, the lake level stopped dropping and stablized around 60″ below the top of the dam even as the dam remained open. In a drier year, it could and would go much lower, perhaps another 2′-4′ lower. I remember back in the summer of 1960 (I was 7 years old) the owner of the dam at the time- Chester Chase- decided to let the water out. It was a dry summer and I can distinctly remember the lake falling even lower than it did last fall. Some of you who have been around long enough will remember this as well. An analysis of the numbers shows that this past winter the lake level has fallen at an average constant rate of  a little over 1/2″ per day (0.56″) over an elevation range of over 20″ which suggests that the lake is no where near the lowest it can get, but that it is likely closer to hitting bottom than it is to hitting the top. Finally, for those people who think you can’t measure the water level of the lake in the winter because its frozen, relax. It may be counter-intuitive but the fact is that the lake being frozen does not measurably affect the water level. Water seeks its own level and the fact that the top 6″ or 12″ of it has frozen doesn’t matter. In other words, if the ice on the lake were to suddenly thaw overnight, the lake level would not rise. If you don’t believe me try this experiment: put 3-4 ice cubes in a tall glass then fill the glass with water, right up to the rim. Set it on the counter and wait for the ice to melt. Notice what happens to the water level in the glass. When we think of spring we think of ice melting and the lake level rising, but the two events are unconnected. The water level rise in the spring is the result of snow melt and and rain runoff, not the ice melting. The final thing I find interesting is how little precipitation we had this winter! In December 2011 precipitation was a whimpy 2.79″, January 2012 even less at 2.24″ but February 2012 was a paltry 0.90″. I’m not sure what normal is but I’m sure that ain’t it. I would be remiss in recounting my water level measurement adventures if I didn’t give honorable mention to David Hodsdon, long time Clary lake shore owner and a volunteer lake water quality monitor since 1975. David tirelessly collected water level measurements for many years, providing us with a valuable historical water level record for Clary Lake that doesn’t exist anywhere else. David didn’t have access to GPS surveying gear so his approach to the problem of accurately transferring an elevation a couple of miles from the dam to his property in Jefferson was simple and elegant: he assumed that when the dam was shut, on a calm day the water surface from the dam to the other end of the lake would define a level surface. So he waited for a calm day when the dam was closed and the water level was at or very near the top of the dam and then he went home and drove a stake into the shoreline by his house marking the height of the water at that time. He then used that stake as a bench mark, measuring the height of the lake surface above or below that stake. Simple. Elegant. As it turns out, according to my recent survey, the elevation reference point that he established was less than 4″ off, plenty close enough for his purposes and more than sufficiently accurate for the historical record. David’s commitment to the scientific process and his dedication to working to preserve the health and welfare of Clary Lake is one of the reasons why I established a benchmark in his yard; I wanted to give him an accurate elevation reference point to base his observations on should he decide to continue his water level measurements in the future. That’s about it. If you have any questions about all this, please feel free to contact me.