How much water is in the main Colorado River reservoirs?

Check out my California Reservoir Levels Dashboard
 

I based this graph off of my California Reservoir marimekko graph, because many folks were interested in seeing a similar figure for the Colorado river reservoirs.

This is a marimekko (or mekko) graph which may take some time to understand if you aren’t used to seeing them. Each “row” represents one reservoir, with bars showing how much of the reservoir is filled (blue) and unfilled (brown). The height of the “row” indicates how much water the reservoir could hold. Lake Mead is the reservoir with the largest capacity (at almost 29,000 kaf) and so it is the tallest row. The proportion of blue to brown will show how full it is. As with the California version of this graph, there are also lines that represent historical levels, including historical median level for the day of the year (in red) and the 1 year ago level, which is shown as a dark blue line. I also added the “Deadpool” level for the two largest reservoirs. This is the level at which water cannot flow past the dam and is stuck in the reservoir.

Lake Mead and Lake Powell are by far the largest of these reservoirs and also included are several smaller reservoirs (relative to these two) so the bars will be very thin to the point where they are barely a sliver or may not even show up.

Historical Data

Historical data comes from https://www.water-data.com/ and differs for each reservoir.

• Lake Mead – 1941 to 2015
• Lake Powell – 1964 to 2015
• Flaming Gorge – 1972 to 2015
• Lake Mohave – 1952 to 2015
• Lake Navajo – 1969 to 2015
• Blue Mesa – 1969 to 2015
• Lake Havasu – 1939 to 2015

The daily data for each reservoir was captured in this time period and median value for each day of the calendar year was calculated and this is shown as the red line on the graph.

Instructions:
If you are on a computer, you can hover your cursor over a reservoir and the dashboard at the top will provide information about that individual reservoir. If you are on a mobile device you can tap the reservoir to get that same info. It’s not possible to see or really interact with the tiniest slivers. The main goal of this visualization is to provide a quick overview of the status of the main reservoirs along the Colorado River (or that provide water to the Colorado).

Units are in kaf, thousands of acre feet. 1 kaf is the amount of water that would cover 1 acre in one thousand feet of water (or 1000 acres in water in 1 foot of water). It is also the amount of water in a cube that is 352 feet per side (about the length of a football field). Lake Mead is very large and could hold about 35 cubic kilometers of water at full (but not flood) capacity.

Data and Tools
The data on water storage comes from the US Bureau of Reclamation’s Lower Colorado River Water Operations website. Historical reservoir levels comes from the water-data.com website. Python is used to extract the data and wrangle the data in to a clean format, using the Pandas data analysis library. Visualization was done in javascript and specifically the D3.js visualization library.

How much water is stored in California’s snowpack and reservoirs?

California’s snow pack is essentially another “reservoir” that is able to store water in the Sierra Nevada mountains. Graphing these things together can give a better picture of the state of California’s water and drought.

The historical median (i.e. 50th percentile) for snow pack water content is stacked on top of the median for reservoirs storage (shown in two shades of blue).  The current water year reservoirs is shown in orange and the current year’s snow pack measurement is stacked on top in green.  What is interesting is that the typical peak snow pack (around April 1) holds almost as much water (about 2/3 as much) as the reservoirs typically do.  However, the reservoirs can store these volumes for much of the year while the snow pack is very seasonal and only does so for a short period of time.

Snowpack is measured at 125 different snow sensor sites throughout the Sierra Nevada mountains. The reservoir value is the total of 50 of the largest reservoirs in California.  In both cases, the median is derived from calculating the median values for each day of the year from historical data from these locations from 1970 to 2021.

I’ve been slowly building out the water tracking visualizations tools/dashboards on this site. And with the recent rains (January 2023), there has been quite a bit of interest in these visualizations.  One data visualization that I’ve wanted to create is to combine the precipitation and reservoir data into one overarching figure.

Creating the graph

I recently saw one such figure on Twitter by Mike Dettinger, a researcher who studies water/climate issues. The graph shows the current reservoir and snowpack water content overlaid on the historic levels.  It is a great graph that conveys quite a bit of info and I thought I would create an interactive version of these while utilizing the automated data processing that’d I’d already created to make my other graphs/dashboards.

Time for another reservoirs-plus-snowpack storage update….LOT of snow up there now and the big Sierra reservoirs (even Shasta!) are already benefitting. Still mostly below average but moving up. Snow stacking up in UCRB. @CW3E_Scripps @DroughtGov https://t.co/2eZgNArahy pic.twitter.com/YEH4IYKlnH

— Mike Dettinger (@mdettinger) January 15, 2023

The challenge was to convert inches of snow water equivalent into a total volume of water in the snowpack, which I asked Mike about. He pointed me to a paper by Margulis et al 2016 that estimates the total volume of water in the Sierra snowpack for 31 years.  Since I already had downloaded data on historical snow water equivalents for these same years, I could correlate the estimated peak snow water volume (in cubic km) to the inches of water at these 120 or so Sierra snow sensor sites.  I ran a linear regression on these 30 data points. This allowed me to estimate a scaling factor that converts the inches of water equivalent to a volume of liquid water (and convert to thousands of acre feet, which is the same unit as reservoirs are measured in).

This scaling factor allows me to then graph the snowpack water volume with the reservoir volumes.

See my snowpack visualization/post to see more about snow water equivalents.

My numbers may differ slightly from the numbers reported on the state’s website. The historical percentiles that I calculated are from 1970 until 2020 while I notice the state’s average is between 1990 and 2020.

You can hover (or click) on the graph to audit the data a little more clearly.

Sources and Tools

Data is downloaded from the California Data Exchange Center website of the California Department of Water Resources using a python script. The data is processed in javascript and visualized here using HTML, CSS and javascript and the open source Plotly javascript graphing library.

How does the current California snowpack compare with Historical Averages?

If you are looking at this it’s probably winter in California and hopefully the snowy in the mountains. In the winter, snow is one of the primary ways that water is stored in California and is on the same order of magnitude as the amount of water in reservoirs.

When I made this graph of California snowpack levels (Jan 2023) we’ve had quite a bit of rain and snow so far and so I wanted to visualize how this year compares with historical levels for this time of year. This graph will provide a constantly updated way to keep tabs on the water content in the Sierra snowpack.

Snow water content is just what it sounds like. It is an estimate of the water content of the snow. Since snow can have be relatively dry or moist, and can be fluffy or compacted, measuring snow depth is not as accurate as measuring the amount of water in the snow. There are multiple ways of measuring the water content of snow, including pads under the snow that measure the weight of the overlying snow, sensors that use sound waves and weighing snow cores.

I used data for California snow water content totals from the California Department of Water Resources. Other California water-related visualizations include reservoir levels in the state as well.

There are three sets of stations (and a state average) that are tracked in the data and these plots:

• Northern Sierra/Trinity – (32 snow sensors)
• Central Sierra – (57 snow sensors)
• Southern Sierra – (36 snow sensors)
• State-wide average – (125 snow sensors)

These stations are tracked because they provide important information about the state’s water supply (most of which originates from the Sierra Nevada Mountains). Winter and spring snowpack forms an important reservoir of water storage for the state as this melting snow will eventually flow into the state’s rivers and reservoirs to serve domestic and agricultural water needs.

The visualization consists of a graph that shows the range of historical values for snow water content as a function of the day of the year. This range is split into percentiles of snow, spreading out like a cone from the start of the water year (October 1) ramping up to the peak in April and then converging back to zero in summertime. You can see the current water year plotted on this in red to show how it compares to historical values.

My numbers may differ slightly from the numbers reported on the state’s website. The historical percentiles that I calculated are from 1970 until 2022 while I notice the state’s average is between 1990 and 2020.

You can hover (or click) on the graph to audit the data a little more clearly.

Sources and Tools

Data is downloaded from the California Data Exchange Center website of the California Department of Water Resources using a python script. The data is processed in javascript and visualized here using HTML, CSS and javascript and the open source Plotly javascript graphing library.

What is ScrabWordle? Wordle with a Scrabble twist

ScrabWordle is a game based on the hugely popular Wordle by Josh Wardle, where you try to solve a 5-letter puzzle. The Scrabble element of the game comes in where the point value of the target word is given based on the letter values in the game and is another helpful clue to help solve the puzzle.

I recently made WordGuessr and Tridle. So I’d already made the structure of the game and just needed to add the Scrabble elements. And it took me a little while to work out all the kinks and it also improved my CSS skills to get all the aspects of the look of the tiles correct.

How to Play ScrabWordle

If you know how to play Wordle, then the rules are the same:
1) the game reports the point value of the target word
2) just type in a guess and and press ‘Enter’, the score of your guess is updated in real time
3) the game will indicate how good your guess was. If any letter are in their correct position they will turn green, if they are correct but in the wrong place, they turn orange, and if they don’t show up in the word they turn black.
4) guess again, using the clues from the previous word
5) you can share your results with other folks or issue a challenge game
6) play all you want

The game is created using HTML, CSS and Javascript code to create interactivity and UI. Hope you enjoy it. I enjoy the extra element (word point values) to help me guess the word. Let me know in the comments how you like the new game play and if the difficulty levels are appropriate.

Over 5 Million games played! (December 2022).

What is Tridle? Triple Wordle

Tridle is a game based on the hugely popular Wordle by Josh Wardle, where you try to solve three puzzles (Triple Wordle) at the same time. It is also modeled after Dordle and Quordle, but a little different (3 instead of 2 or 4!)

Shortly after I started playing Wordle in early January, I wanted to play more games so I decided to make a Wordle game so I could play as often as I wanted. I made WordGuessr in a day or so. It ended up being fairly popular and lots of people still play.

Recently, I started playing Dordle (where you solve two puzzles simultaneously), which adds an extra layer of challenge onto the Wordle game. I thought that a triple Wordle game would be a fun game to code up and since I’d already made the structure of the single game, how hard could making 3 be. Well, it took me another day or so to work out all the kinks and it was also stretched my CSS skills to get all the look correct with all the various permutations of the game and generating three boards simultaneously.

If you know how to play Wordle, then the rules are the same:
1) just type in a guess and and press ‘Enter’.
2) the game will indicate how good your guess was for each puzzle separately. If any letter are in their correct position they will turn green, if they are correct but in the wrong place, they turn orange, and if they don’t show up in the word they turn black.
3) guess again, using the clues from the previous word
4) repeat until you get each of the words correct or fail to figure them all out after 8 tries.
5) you can share your results with other folks
6) play all you want

The game is created using HTML, CSS and Javascript code to create interactivity and UI.

This visualization lets you divide the US into 1,2,3,4,5,8 and 10 different segments with equal population and across different dimensions. The divisions are made using counties as the building blocks (of which there are 3143 in the US). There are numerous different ways to make the divisions. This lets you make the divisions by different types of geographic directions and divisions by population density.

Instructions

• Select a dimension on which to divide up the country – there are geographic dimensions, like north to south or east to west, or by population density
• For some geographic divisions (concentric rings or pie slices), you can choose the geographic center of the divisions
• You can also choose the number and color scheme of the divisions
• To show the divisions, either click the Animate Counties button or use the slider to add counties

If you can think of other interesting ways to divide up the US, please let me know and I can try to add them to this visualization.

Sources and Tools:
2018 county population data is from US Census Bureau. The map visualization is created using the Leaflet javascript mapping library and the data wrangling and user interface and interactivity are created using HTML, CSS and Javascript code.