Lake Travis Water Level Forecast (8/1/2021)

 Lake Travis Water Level Forecast (as of 8/2/2021)

Author: David Ferreira (Resume)
Data Source:  Lower Colorado River Authority

This document is not intended for publication and is only a personal record of a forecast made on 8/2/2021, for tracking purposes.

This original forecast is provided free of charge to interested members of the general public as a courtesy. No warranty is made as to the accuracy.

Historical

Fig. 1:  Lake Travis water level decomposed: Actual (top), Trend-Cycle (2nd), Seasonal (3rd), and Error (bottom)

Forecast

Fig. 2: Lake Travis monthly mean water level 6-month forecast with confidence intervals

Fig. 3: Lake Travis monthly mean water level 1/2020 through 12/2021

Fig. 4: Lake Travis monthly mean historical with forecast

Forecast Data

	Mean	Lo.80	Hi.80	Lo.95	Hi.95	Actual	Δ	%Δ
Aug-2021	664.5	658.6	670.5	655.4	673.6	666.1	-1.6	-0.24%
Sep-2021	663.4	654.6	672.2	649.9	676.9	663.8	-0.4	-0.06%
Oct-2021	665.1	653.8	676.5	647.9	682.4	663.1	2.0	0.3%

Table 1: Lake Travis monthly mean forecast with results 

Frequently Asked Questions

Q1: Why do I do these Lake Travis forecasts?

A1: I generate these forecast on a quarterly or semi-annual basis for the following reasons:

• I have an ongoing recreational interest in the water level of Lake Travis.
  
  
• Public Service: My club (Austin Yacht Club) and all Lake Travis boaters have a keen interest in knowing future water levels.
  
  
• I derive entertainment and educational value from doing forecasts like this and seek other opportunities for similar challenges.
  
  
• I seek to re-channel my successful software engineering career toward a hybrid career that will merge my enjoyment of software engineering with my passion for econometric analysis and forecasting.

Q2: Doesn’t LCRA provide forecasts of Lake Travis water levels?

A2: Yes, they do provide 6-month “projections” each month. They also hold the keys to the dam. They have both control and fiduciary interest in the data being forecast. I would encourage readers to compare my forecasts to LCRA projections for accuracy. I would, however, caution the reader that LCRA projections are for water levels for specific dates (the 1^st of each month?). In contrast, my forecasts are for a monthly average water level for each month being forecast.

Q3: Can’t we just use Excel to forecast things like this?

A3: You can. And that may fit your particular need quite well. Another method would be to print out a graph and draw a line through both the same month in the first year of data and the last year of data and carry that line forward. These methods might both suite your needs (no sarcasm). I would encourage you to compare these forecasts against mine, those of LCRA, and observed results. I think you will find that not all methods are equally accurate over time.

Q4: So … you’ve studied Economics / Econometrics. How does that qualify you to forecast a hydrological phenomenon like the water level of a lake?

A4: The water level of Lake Travis is indeed a hydrological phenomenon. But it is also an economic phenomenon – it is impacted by statewide electrical demand, and the irrigation demands of farmers downstream. Forecasting the water level of Lake Travis is very much an econometric problem that involves some hydrology.

About the Author

I am currently enrolled in the MS Finance & Economics program at West Texas Univ. I am an experienced Software Development Eng., formerly of 3M, Dell, and Microsoft. I am also a former Peace Corps Volunteer and in a prior career, I was an Economist.