Why Does Blue Mesa Reservoir Fall Faster as It Gets Lower?

I have a tendency to notice something, wonder why it is happening, and then spend entirely too much time figuring out whether my assumption is actually correct.

Recently, that something has been the water level at Blue Mesa Reservoir.

Blue Mesa is Colorado's largest reservoir and an important part of the Gunnison and larger Colorado River system. During the summer, I started watching its daily elevation, inflow and outflow numbers. At first, I had a simple question:

How quickly is Blue Mesa dropping?

That soon became a more interesting one:

Does the same amount of net water leaving the reservoir cause a larger drop in elevation as the reservoir gets lower?

My assumption was yes.

A Reservoir Isn't a Swimming Pool

A rectangular swimming pool contains roughly the same volume of water in every vertical foot. Remove the same amount of water and the level falls by roughly the same amount whether the pool is full or half empty.

Blue Mesa is a flooded river valley.

As its elevation falls, the reservoir generally becomes narrower. A vertical foot near the top should therefore contain more water than a vertical foot farther down.

If that's true, the same volume of water leaving Blue Mesa should cause progressively more elevation loss as the reservoir declines.

That was my hypothesis.

Looking at Net Outflow

Simply comparing Blue Mesa's outflow with its elevation change doesn't quite work because water is simultaneously flowing into the reservoir.

For this exercise, what matters is:

Net Outflow = Outflow − Inflow

I started comparing that number with the daily change in reservoir elevation.

From roughly July 11 through July 24, Blue Mesa's actual daily decline slowed from about 0.44 feet per day to 0.27 feet per day.

That initially looks contrary to my hypothesis, until you account for the fact that net outflow was also declining substantially.

Once normalized for net outflow, the pattern became much clearer.

Early in the period, Blue Mesa was losing approximately:

0.400 feet of elevation per day for every 1,000 cfs of net outflow.

By the end of the period, it was closer to:

0.423 feet per day for every 1,000 cfs.

That's roughly a 5–6% increase in less than two weeks.

The reservoir was dropping more slowly because less water was leaving it, but each unit of net water loss was producing a slightly larger change in elevation.

That's exactly what I expected to see.

Then I Kept Watching

One short dataset isn't terribly convincing, so I kept collecting the numbers.

By August 5–11, Blue Mesa had fallen from approximately 7,433.84 feet to 7,431.39 feet, with net outflows generally between 777 and 985 cfs.

Running the same calculation produced approximately 0.438 to 0.447 feet of daily elevation decline per 1,000 cfs of net outflow.

So our progression looked something like this:

Mid-July: ~0.400 ft/day per 1,000 cfs

Late July: ~0.423 ft/day per 1,000 cfs

August: ~0.44–0.45 ft/day per 1,000 cfs

As Blue Mesa's elevation declined, the amount of elevation lost for a comparable net outflow continued to increase.

What Are We Actually Seeing?

Reservoir operators already have detailed elevation-storage-capacity curves for Blue Mesa. I'm certainly not suggesting we've discovered something new about reservoir geometry.

What interested me was whether that geometry could be observed using nothing more than publicly available daily elevation, inflow and outflow numbers.

Apparently, it can.

There are limitations. Evaporation, precipitation, groundwater interactions, measurement timing and the difference between instantaneous and averaged flow data can all affect the calculation. I wouldn't treat any single day's ratio as an exact measurement.

The trend is what interests me.

We started around 0.400 feet of decline per 1,000 cfs of net outflow. Later we saw 0.423. By August, we were seeing roughly 0.44–0.45.

The reservoir has continued downward, and each comparable unit of net water loss appears to be producing more elevation loss.

The Ruler Is Linear. The Reservoir Isn't.

People naturally look at reservoir elevation in feet. If a reservoir drops ten feet, another ten-foot decline feels like it should represent roughly the same amount of water.

But Blue Mesa isn't shaped that way.

As the reservoir retreats into the narrower portions of the old Gunnison River valley, each vertical foot represents less water than the foot above it.

This whole exercise started because I looked at the numbers and thought:

"Wait a minute. A reservoir isn't a cylinder."

So ChatGPT and I started crunching the daily data.

The question now is how quickly this relationship changes as Blue Mesa continues to fall. If the trend holds, the same 1,000 cfs of net outflow that caused roughly 0.40 feet of daily decline in mid-July should produce progressively greater elevation loss at lower reservoir elevations.

I'll keep watching.

Sometimes the most interesting data analysis starts with nothing more complicated than noticing that something doesn't behave quite the way our intuition says it should.