Karst Hydrogeology Application in Kentucky:

The karst aquifer discharging water to Royal Spring in Georgetown Kentucky has its headwaters approximately 24 km southeast of Royal Spring in urban Lexington, Kentucky. The principal conduit draining the groundwater basin has no known, humanly accessible, entrance. Furthermore, the potentiometric surface is several meters of water above the conduit. The Kentucky Geological Survey has been asked to locate the cave/conduit at three sites. The farthest down gradient site is at the confluence (divergence?) of geographic and hydrogeologic, and hydrologic conditions. The site is on public property where the conduit continues north and the surface overflow channel (now rarely carries water) diverges to the west, and at the county line between Fayette County and Scott County. The purpose of this effort is to monitor water quality as it leaves the Lexington/Fayette Merged Government County. 

KGS has drilled into the cave.  The inside diameter of the cased hole is about 13 cm, and the depth from ground level to the cave ceiling is 18.3 m. The height of the cave passage is about 80 cm. There are three holes into the cave, # 20, #23, and #25. Submersible lights are visible with a down hole TV camera from 20 to 25, but not from 23. The drilling for hole #20 went into the floor of the cave about 10 more meters and 2 meters for hole #23. The deeper drilling created a sloping, conical depression in the mud on the floor of the cave. The mud is about 80 cm thick and if absent the total height of the cave would be at least 1.5 m. The width of the cave is approximately 4 meters.

A crucial component of the groundwater monitoring method is to be able to calculate the TMDL (Total Maximum Daily Load) or flux of various constituents.  To do this the discharge of water past the monitoring station must be measured. The Norteck Aquadopp Current Profiler is the crucial instrument for accomplishing this task. We have two objectives 1) determine the distance and azimuth to the highest amplitude reflection and assume that reflection is the wall to calculate the cave cross sectional area, and 2) measure the average flow velocity across the cross-section to calculate the discharge continuously. Therefore we are using both amplitude and velocity data from the instrument.

This long, elaborate discussion of project background leads to my question. I am having trouble positioning the Doppler sonar in the borehole so that the sensors are not too shallow or too deep, and are measuring the water flow. I have caretully measured the length of the cable and marked it at critical intervals, but stil seem to get data that are from inside the borehole, or too deep, and inside the crater in the floow. How much will the communications cable stretch if it is bearing the full weight of the instrument plus the accruing weight of 18 meters (60 ft.) of cable? The weight of the lower 20 feet of cable is offset by the bouancy of the cable and unit in water. What would be the best indicator from the data or combination of readings that the instrument was positioned with the sensors centered vertically in the flow, out side of the borehole, but near the ceiling?

Jim Currens

Hi Jim,

I unfortunately don't have any information on how much the cable stretches. I'll check on this for you.

In the meantime, I would recommend using a secondary line to actually deploy and retrieve the instrument from the borehole. You could use a low stretch synthetic rope for this and allow the rope to hold the weight of the instrument and communication cable. This would probably a slightly better option than using the communication cable itself as I don't believe the standard cables are designed for use as a deployment line.

As for data indicators, a few ideas come to mind. Do you have a good estimate of what the water table elevation is? This should let you know at what height in the borehole the instrument will first see water. There should be a very dramatic change in the data quality at this point, and the amplitude data should look much better. I suspect the borehole is too narrow for any meaningful measurements, but simply being immersed in water will make things look much better. Hopefully this matches up with the expected water elevation in the borehole.

Second, assuming the elevation matches the expected water table elevation, you can hopefully use the pressure sensor to estimate the position below the water table, letting you know when the sensor head is within the cave.

As far as data within the cave, I would play a little with the measurement cell size and look at the amplitude data. Set this to a fairly small value (say 5 cm or so) so you can see changes with a little more resolution. This will result in very noisy, probably useless, velocity data but should let you at least determine if you are within the cave as a secondary measure. Since you are cabled you can then switch to a setup for velocity measurement and get the required data for discharge calculations.

I would also examine the pitch, roll and heading information. If the flow is not too strong, the instrument should hang vertically in the water and hopefully not rotate much. This will let you know quite a bit about mapping the cave. Realize the compass update rate can be different than the amplitude and velocity data rate and can be set in the software.

Since the cave height is about the length of the instrument, I suspect you'll be able to tell when it touches bottom because there will be much less tension on the retrieval line.

This sounds like an interesting but definitely possible measurement. Best of luck to you.

P.J.

Dear Mr. Rusello,\

Thanks for the quick reply.  I am already doing most of the things you recommend above. I have sleeved the sonde with 8 cm inside diameter PVC pipe, a little longer than the sonde but exposing the sensors at the “lower” end. The sleeve is held in place with set screws that are aligned with the upper edge of the groves machined into the sonde body.  We originally had braided synthetic fiber rope attached to the sleeve, but found it too stretchy.  I replaced it with some vinyl clad wire rope. Although the wire rope is a very small diameter, it doesn’t stretch as much as the rope or the communications cable. I am aware of the vulnerability of the communication cable.  A mechanically strong communications coupling, or a built in stress relief point on the sonde body would be a nice addition to the sonde.

I have discussed positioning the sonde based on water depth with other KGS staff but we think it impractical because the water level will change significantly with increased precipitation and we are not crystal clear on the length of bore hole from the water table to the cave ceiling Although a simple algorithm (arithmetic) would solve for meters of water expected over the sonde.  I might give it a try next set of observations. One idea I had was to build some kind of switch that is closed by the proximity to the bore hole, but breaks the connection as the sensors emerges from the bottom few meters of uncased hole. A 12 volt battery and a light bulb at ground level would complete the system. This would be simple to do and would give accurate and precise positioning of the sonde.

Based on the estimated height of 0.8 meters by 4 meters would a cell size of 5 cm g(iving a maximum range of 1 meter, correct me if I misunderstand the range versus the number of cells) be large enough? I will try a cell size of 20 cm next time I deploy the unit and we will look at those data. For velocity I will reset to 0.5 meters unless the data at 0.2 meters look useful.

I have two additional questions. On our unit the sensors output file indicates that the“Distance Measurements” DISABLED.  Is there a direct output of distance to the echo object available? If yes, how do I activate it?  Second, and this sounds incredibly ignorant, how do I determine the average velocity of my profiler data? Simple arithmetic mean of the individual cell velocities, or some process more sophisticated?

 Jim C.

I like the idea of a switch letting you know when you'v exited the sleeve of the borehole. Seems like it would be fairly simple to build. Based on this though, what is the borehole lined with? If it's a magnetic material (such as steel) please be aware the Aquadopp's compass will be affected. There are a few threads on the forums regarding calibrating the compass for interference, but I am not experienced in this matter unfortunately.

Are you using a High Resolution or standard Aquadopp? This will influence the cell size you can and should use for your measurements. I don't use the standard Aquadopp very often, and it seems the smallest cell size is 0.1 m.

I think the distance measurement is used by an Aquadopp equipped with our wave measurement firmware. I don't think the standard Aquadopp or the HR variant will determine a distance based on the echo intensity.

For averaging, this will again depend on your instrument type (standard or HR). I'm typically interested in the mean in time, so I'll treat each cell as a time series and calculate my desired statistic at each cell to generate a profile. For you application, it sounds like you're more interested in a mean flow, so averaging in time is appropriate and you might find working with earth coordinates (east, north and up) is the simplest method. If you use earth coordinates (ENU), you can do some averaging on the instrument but please see my warning regarding compass interference , as this will affect this calculation.

P.J.

Additional information for Karst application, URGENT.

Dear Mr Rusello,

We have solved most of the problems disccused in the previous postings, including the vertical positioning of the Aquadoppler in the cave.  Maximum flow velocities are about 1 m/sec. Because we have other locations where we also want to use the Aquadoppler, we would like to install a Marsh-McBirney 201-D magnectic induction flow meter in the cave. Both the Aquadoppler and 201-D have long cables. The 201-D instrument  is also less of a finacial risk in case of loss due to any number of potential catastrophies.

The first step towared deploying the 201-D is to simultaneously deploy the two instruments for developing a rating curve. We have checked for magnetic interference from thje 201-D with the compass and found it minimal. It would be most convenient if we could log the 201-D analog signal with the Aquadoppler.The 201-D output ranges from 0.1 to 1 volt. (0.1v/ft.sec).  So far we have only been able to record the first signal from any channel when the Analog input is active.  I have the analog channel 1 set on profile.  We are not logging ANY of the data generated by either insturment and the sensor readouts, etc do not update. The sampling interval has been set on 10 min, 1 min, and 0.1 min.  Logging resumes when I disconnect the 201-D and restart the Aquadoppler. 

Please advise.

Cordially,

Jim Currens 

Hi Jim,

I don't have too much experience with the analog inputs, but I think the problem may be your sample setup. Please search the forums for "analog input" and you'll find quite a few posts related to this type of situation. Relevant to your setup is this thread:

http://www.nortek-as.com/en/knowledge-center/forum/system-integration-and-telemetry/148657565#424086869

I would also double check the wiring just to make sure everything is working correctly. You can also hook up a small DC power source (a couple of volts) and see if this is logged correctly.

P.J.

Hello P.J.,

I'm having more problems with my application of the Aquadoppler to ground water flow in a karst conduit.   Please see the attached EXCEL spread sheet.  We did have success in getting the logger to record the analog signal from the Marsh-McBirney 201-D electromagnetic flow velocity meter. The sonde for the 201-D was positioned in the same hole, 10 to 12 cm below the Aquadoppler sensors.

The deployments have been characterized by two phenomena.  The first is probably my fault but I can't see the error.  We know from previous work that the bearing of the conduit is approximately 175 deg up flow.  The bearings on the velocity seem to be correct, yet the posting of the location of the maximum velocity for the beam plots 180 degrees from where it should be.  The second problem, however, I have no clue as to the cause.  For the last 3 deployments we've been getting wildly varying data from Sensor #1. The velocity is cycling between a maximum of about 1 m/sec and -0.2 m/sec, plus the plots of the maximum velocity a widely scattered, although the unit was deployed in a fixed position. A deployment with the sensors facing down flow resulted in similary variability in sensor 1.  There may be a hydrologic explaination of the cycling but I would think changes in flow would affect both sensors and certainly not the same single sensor. The instrument bearings, roll and pitch confirm that the unit is mostly stationery.  I can send additional data sets if you want them.

Thanks,

Jim Currens

The attachement was too large. Please send an email address and I will forward the spread sheet first thing tomorrow morning.

Hi Jim,

As said by e-mail I have received a .prf from you. This one seems to be corrupted in some way. Either had the file transfer corrupted it, or the way of retrieving the data is formating it in some way. To be more specific it seems like all 00 are converted to 20. I am pretty sure other errors are present, but they are not that easy to spot. Is this retrieved by a data logger? Could this be converting something? Please try to upload the file to ftp://ftp.nortek.no/Incoming/ 

I think this just might explain the cycling in velocities.

For the 180 degrees switch; which coordinate system are you using? If you are measuring in XYZ the sign of the y-axis flips relative to the site since this coordinate system relates to the instrument body.

Best regards

Jonas Røstad

Hi Jim,

The data came finally through uncorrupted. I have been looking at W2522501.prf

The cycling in velocities is because two of the beams are obstructed. In order to have good velocities the beams must have clear sight. 

The attached picture is showing the three amplitude profiles at the same 10 times. As you can see beam 1 and 3 has got bumps that are reflections from hitting something. Beams 2 has no bumps. Further down the time series of the beam velocities are shown. Beam 1 and 3 has the described cyclic pattern while beam 2 has got more stable readings.

Since the ENU velocities is a product of all the beams together with tilt and compass the north velocities are corrupted as a consequence of beam 1 and 3. If I am right this is a head with two horizontal beams and one vertical. Of this reason the main contribution to ENU coordinates comes from beam 1 and 2. In this dataset I would say the first 4-5 is to be trusted.

Best regards

Jonas Røstad

Jonas,

I've read your email and the above post and have some further questions.

First, you are correct in that sensor 3  is facing vertically downward and 1 and 2 are horizontal.  Second, these data are from a cave. From down hole cameral inspections the dimensions are roughly a meter high by 6 to 8 meters wide.  Obstructions are to be expected and their location is important. We have, however, observations made before the data sets I sent that do not have the cycling. All data sets are at the same general site and most are in well 25. All of the data I sent are from 25, I believe, and the set you have looked at are from 25.  All of the wells are less than 5 meters apart.

Other relevant facts are as follows.

1.     The cyclic problem began relatively recently and does not affect older data sets.

2.     The cycling occurs in the same sensor regardless of the bearing of the instrument.

3.     The cycling never occurs in sensor 2.

How do these facts fit in with the explanation of an obstruction?

I do appreciate the effort you have put into this.

Jim Currens

Hi Jim,

Your confirmation about the head is helpful. If you look at the velocities in the previously posted .jpg you'll see that the North velocity signature is almost identical, but opposite to Beam 1 velocity.

The other datasets with other bearings are also hitting things. In KYH2501 Beam 2 is the one that are most affected.

Regarding the other questions of why other data sets are OK - it is hard to say. If you provide me with one good data set I will hopefully be able to tell you the difference.

To sum up. An acoustic doppler instrument needs un-blocked blocked beams to give trustable data.

Best regards

Jonas