Base Flows

Base flow is the portion of the stream that comes from the aquifer seepage.  That is the portion of the stream flow that is not caused by precipitation events.  How this number is derived has huge implications for Nebraska.  The entire Model is based on this concept. 

This is a very detailed and complicated subject.  Those that take the time to understand it will understand one of the fundamental components of the Model and how it works.  Once you understand the assumptions that go into the computer simulation, then you can see what is needed to fix the problem.

There is a report by McDonald Morrissey Associates that talks about base flows.  McDonald Morrissey and Associates was the firm that Nebraska hired as its experts to operate the computer simulation used by the three States in the Republican River Compact settlement.  The report is very limited in what it says about base flows.  The report can be read here.

http://www.waterclaim.org/presentations/problems/mcdonaldmodelreview_060411.pdf

The report by McDonald does not cite the source for the observed base flow data that they use nor do they report how they calculate it.

No two hydrologists will necessarily calculate the same base flow for a particular area.  So in 1995 Tony Wahl released a program to help make base flow determinations a little more objective.  It is not known if McDonald uses this method or not.  The Wahl method is referenced on the US Bureau or Reclamation web site and can be found via http://www.usbr.gov/pmts/hydraulics_lab/twahl/index.cfm

The image here is taken from his program.

How is base flow calculated?

In English, that means hydrologists look at total stream flows for a number of days (usually between 3 or 5).  The program identifies the “turning points” where there is a change in the volume attributed to base flow.  There is no actual measurement of the water in the stream.  Instead a mathematical smoothing method is used.  Illustration 1 shows the daily stream flow and base flow for 2004 for gage 06823000.  The base flow was calculated using the Wahl program.  The data came from the USGS at http://tinyurl.com/2nrvoy. 

From North Fork Base Flow 06823000.xls

Illustration 1

The blue area is what hydrologists say comes from the aquifer.  The red area would come from run off. 

As can been seen, base flow for the Republican River fluctuates by a great deal.  I assume McDonald uses a Wahl type method as it is the standard for the field. 

Illustration 2 shows a stream flow detail with a single precipitation event.   

Illustration 2

Illustration 2 clearly shows that base flow, as calculated by the standard method, responds within hours or days to a precipitation event. 

The standard method of calculating base flows works fine for determining water flows from the soil directly associated with the stream.  It cannot measure contributions from the aquifer that are further away from the stream.  That is what the MODFLOW program attempts to do that was designed by McDonald. 

The industry does not appear to differentiate between base flows contributed by the soil very close to the stream and the aquifer that is further away.  Because the industry uses the same terms for two different things it gets confusing.

Over the last 60 years, stream flows have dropped by large amounts across the Republican River Basin.  The politicians say this drop is caused by pumping far from the stream.  They say base flows are down dramatically and that groundwater pumping is the cause of the problem.  However, as this analysis shows, most of the base flows are very closely tied to precipitation events and stream flow.  If stream flow is reduced for any reason such as conservation practices that delay or stop water from getting to the stream then base flows are also reduced because base flows are simply a function of stream flow. 

The computer simulation for the Republican River Basin shows the changes in the system every second starting in 1918.  At any one point in time the operator can look at the system and see what is happening.  To make life easier, the administrators of the Republican River Basin simulation decided to watch 25 points out of the 30,224 that exist.  On an annual basis they would determine how much water from aquifer seepage there was at those 25 points. They would also determine how much water should be there if no irrigation wells were operated.  The difference between the two numbers are the depletions caused by groundwater irrigation.  Note, that there is no physical measurement of water at these points.  It is just simulated water in a computer program with real world input placed into the simulation at a variety of other points.

Reports out of the simulation show that the vast majority of the depletions at those 25 points are caused by pumping from prior years.  Illustration 3 shows that even if all pumping was turned off, that the effects from prior pumping would continue to have a negative impact on the stream for decades due to the lag effect.

Illustration 3

The trick is to reconcile the information we have from the industry standard method of computing base flow, which shows that the vast majority of the base flow is directly related to precipitation events close to the stream, and the computer simulation used by the RRCA which shows that almost all (94%) of the depletions to the stream are caused by pumping that happened far from the stream and that do not affect the stream until decades after the pumping occurs.

The USGS has created what is called a Base Flow Index.  It has computed the average base flow for every stream gage in the United States.  It boils the daily data from many years down into one number for each gage.  According to the USGS, the BFI for the gage you see in Illustration 1 above is 80%.  That means that on average 80% of the water that flows in the stream comes from the aquifer.  The word average is very important.  For the year 2006, the actual base flow percent was 90%.  Small fluctuations in the base flow have major implications for Nebraska's ability to stay in compliance with the agreement it made with Kansas.

See the full image at http://www.waterclaim.org/charts/surfaceflow/baseflowrrb.jpg.   This is a large file so it will take time to load.  The image was compiled by WaterClaim.  Click on the image to enlarge it and scroll right to see all of it.

I have also mapped all of the USGS BFI numbers on Google Earth, which is even easier to read.  You can find that at http://tinyurl.com/36hlaj.  

Stop and think about how this system takes into consideration the various effects on the stream. If conservation intercepts water and holds it away from the stream does it increase or decrease the stream flow?  Does it increase or decrease base flow.  Can you see the effects of conservation in this methodology?  Most people that believe there is too much irrigation say that conservation simply changes when water gets to the stream but does not change the total getting there.  Is that true?  Or does conservation make more water available to the plants on the land where the conservation is at and hence increase evapotransporation, meaning less water in the stream now or in the future.

How is pumping far from the stream reflected in this methodology?  As you can see, the assumption used with this method is that a good portion of the water in the stream comes from the aquifer, while only the spikes in flow come from run off events.  Yet, the amount of water that enters the river from the aquifer fluctuates by a very large amount on a daily basis.  The only way that is possible is if the aquifer that is in under consideration is the alluvial aquifer.  The alluvial aquifer is the portion of the aquifer that is connected to the stream and that can have an effect on the flow of the stream within a few days or weeks.  The alluvial aquifer is usually located within several hundred feet of the stream.  Geology determines what is alluvial and what is not.  In the Republican River Basin there is a shale ridge that separates much of the Ogallala formation from the River.

Ann Bleed, director of the Nebraska Department of Natural Resources, says that all aquifer water is connected to the stream.   That might be true if one looks at the system over several thousand years.  However, for the measurable future it is not true. 

The BFI varies from gage to gage.  It is not a constant number as you might have heard.  According to the USGS and according to the data used by the Model, the Republican River and its tributaries are not primarily run off streams as has been stated by several influential water officials.

Here are a few of BFI numbers.  The non-weighted average for the entire basin is 53%.  You will notice that the base flow percentage is much higher in the west than it is in the east.

The base flow number is the most critical number in the entire computer simulation.   It determines whether Nebraska is in or out of compliance. 

A close look at the base flow numbers show that the entire computer simulation is based on a mathematical smoothing formula for the alluvial aquifer.

The Model has a very high error rate (56%) in predicting what will happen with the aquifer.  The errors are especially prevalent in the eastern end of the Basin.  See Inaccurate Aquifer Level Predictions

The Model makers claim a high degree of accuracy in matching the Model to the observed base flows at the accounting points.  However, it appears that the Modelers use the same data and methods to compute both sides of the equation. 

The Model needs to be adjusted to cause it to accurately model aquifer levels at each of the measurement points. 

A method of measuring the effects of conservation, vegetation, and non-alluvial wells effect on base flow needs to be developed.  Nebraska and Kansas are working on that now but are several years away from reporting their findings.

The contribution to the stream by distant or deep wells should not fluctuate on a daily basis. 

There needs to be a separation of base flow information by type of aquifer.  (alluvium vs non-alluvium)

Reports by water officials at both the NRD and DNR level need to accurately reflect the complexity of the stream/aquifer interaction.  The simplifications used in statements about the Republican River Basin are misleading because they often omit the time scale or the large standard deviation in the averages that are cited.

In summary, the fundamental foundation that the computer simulation is based on is the estimation of base flows.  Because "actual" base flows are determined by a smoothing formula of actual stream flows, there is no consideration by the system of what might cause reduced stream flows other than groundwater pumping.  The assignment of reduced stream flows to groundwater pumping ignores other major causes of depletions to the stream and hence inaccurately assigns a major portion of the blame to the wrong cause.   That means when we craft solutions based on a system that ignores the primary causes of the problem, that the solution will not give us the results we hope for.  The key point is that the method used to calculate base flows might be an accurate reflection of alluvium flows but it is not and cannot be used as an accurate method to calculate base flows by non-alluvial base flows.