Interesting Standard Deviation Query

#1
Hi,

My boss just asked me if we could use standard deviation for the following:

A water treatment plant is recieving flow (F), which is continously increasing with time. We have one data point for each day of the year (from 1995 to 2006). Using this data we need to come up with flow projection numbers for 2025.

For average flow condition we just extrapolated the linear trendline. However, for our purpose we need to come up with high and low numbers along with the average for the 2025 flow.

We need to figure out the most logical and valid way to figure out the range for the anticipated flow in 2025. Can someone please try to guide on this?

Thanks!
 

JohnM

TS Contributor
#2
If you do a linear regression, it will give you confidence bands and/or prediction limits around the extrapolated line.
 
#3
Hello
I think you have to deseasonalize your data before fitting the trend line.After predicting the values for 2025 reseasonalized it. You may also include a confidence bands for the forecast.