Letter to the Editor

Fourth water issue

Tuesday, September 22, 2009

Dear Editor,

Problem No. 4 with the Model -- Predicted vs Actual

The software much of the world uses to model aquifers was written by Michael McDonald. Nebraska hired McDonald/Morrissey Associates as Nebraska's experts in the dispute with Kansas. Nebraska has hired many experts and attorneys. Almost all of them are under non-disclosure agreements and will not tell anyone what they did or how they did it. However, in April of 2006, McDonald/Morrissey Associates released a review of the accuracy of the Kansas/Nebraska Model they helped develop.

The full report can be read on the WaterClaim site, where there are interactive maps that let you see the actual vs. the predicted. The report reviews 73 locations and compares the Model calculated results with the observed results. The report does not quantify the percentage of locations that are different; however, summing the results shows that 56 percent of the well locations have trends that are different than what the Model predicted. That is a 56 percent error rate.

According to the report summary, the Model is only wrong 15 percent of the time. The summary of the report disagrees with the details in the rest of the report. Since most people only read the summary, one could easily get the wrong impression.

The report concludes that the individual well records are not important. It says what is important is the accuracy of the observed vs. the calculated base flow numbers at the 24 locations that the Model uses to measure compliance. Unfortunately, there are no actual gauges at most of the 24 points used to measure compliance.

A review of the report, made by the creators of the Model software, shows that the Model has failed to accurately predict aquifer levels in much of the Basin and that most of the areas of error are in the eastern portion of the Nebraska part of the Republican River Basin. In many cases, the actual aquifer level, goes up, yet the Model says it went down. The simulation, however, is the official measurement, and the actual level is ignored.

When the Model says the aquifer drops 2 feet, when in reality it stayed level, then the Model will say that there is less water flowing from the ground into the stream than what actually happens. This causes Nebraska to be reported as having consumed more water than it actually has.

We have placed all of the data online where you can see the information for yourself. Everything is documented so that anyone wanting to verify the information can do so.

To correct this problem would require a recalibration of the Model until its predictions match reality. This would require the Nebraska DNR representative to the technical committee to raise the subject and cause it to be corrected. The DNR answers to the Governor and follows his instructions. If the Governor finds the information we have presented to be correct, it is his responsibility to cause the problem to be fixed.

He is the only elected official who can start the process to correct the problem. The NRDs can encourage a correction, but not cause it to be made.

Steve Smith,


Respond to this story

Posting a comment requires free registration: