I was recently sent logger data from two observation wells monitored during a pumping test (Figure 1). Concern lay in the fact that the record from observation well OBH1 recorded highly variable conductivity readings up to 7500 uS/cm which were up to 10 times higher than those recorded in OBH2 (Figures 1). This could infer saline water being drawn toward the pumping well. If true this would cause serious problems in operating a large water supply and treatment system for the production well that had recently been upgraded at a cost of £1 million.

Logger Records from OBH1 and OBH2 (Figure 1)

Figure 1: Logging Record from Observation Wells OBH1 and OBH2

Measurements of water level, temperature and electrical conductivity were recorded every hour over a period of 35 days.  The two observation wells were located 750m south-west (OBH1) and 650 m south (OBH2) of the production well.

Could I help identify if the data were real or was there a problem with the instrumentation?

For clarity, these are not In-Situ loggers. In-Situ’s versionof these are the AquaTroll 100 and 200 which are high quality sensors with titanium housing.

There are a few simple observations that can be made from the logger records.

  • The changes in water levels are similar in both records and seem reasonable. (Note: the water level measurement displayed is “head of water” after barometric compensation as provided to me. This is the height of the water column above the submerged logger. Loggers were submerged to depths of approx 15m in OBH1 and 25m in OBH2 inferrring the water level to be at around 10.5 m below ground level in both wells).
  • The records for electrical conductivity are significantly different to each other. OBH2 shows little change, averaging around 800 μS/cm. That for OBH1 is very erratic recording values of up to 7,500 μS/cm.
  • When submerged temperature values are stable in both loggers and within 0.6°C of each other.


Comparing Loggers

Given that there was some anxiety from the client to understand the logger records quickly, and in the absence of any local information against which to interpret the data, I suggested that the loggers were strapped together and then lowered down and back up one of the observation wells.

The results of this exercise within well OBH1, which took less than 30 minutes to do, are shown in Figure 2. Logger L1 is the one previously used in OBH1 and L2 is the logger from OBH2. The loggers were programmed identically and set to record at 2 second intervals. Both should give identical records.

It is immediately clear in Figure 2 that logger L1 is producing erratic records and the data captured during the pumping test from this logger should not be used. There are however a few observations of detail worth making.

  • The head measurements were not barometically compensated. Consequently, in air, the total pressure recorded should be around 1 atmosphere or 1000 mbar or approx 10 m water equivalent – as is recorded in Logger L2. Logger L1 has a much more erratic barometric pressure record averaging less than 5 m (and for those who know these things, the offsets on both loggers were set to the same default value of zero). However, when submerged the magnitude of change in the water level recorded in L1 was identical to L2.
  • There is a loose, but poor correlation between temperature measurements when submerged. The temperature recorded in L1 is clearly in error, being 12 °C lower and recording negative values when submerged.
  • There is a simlarly vague correlation with EC when submerged, but EC values in L1 are 2 to 3 times greater than L2. When in air the remain positive and errratic in L1, (they should be zero).

Figure 2: Comparative logger results when profiling Well OBH1

Temperature Conductivity Profile of OBH1

Figure 3: Temperature and Conductivity Profile of Well OBH1

Having established that Logger L2 is working reliably, temperature and conductivity profiles can be produced for borehole OBH1 (Figure 3).

I leave the interpretation for the consultants concerned, but the key point from these is that is there is very little change in electrical conductivity within the borehole column and the water is not saline. At 500 μS/cm the water quality looks promising.

Concluding Comments

This case history is offered as an example of the importance of validating loggers against (a) common sense (b) other measurements such as manual depths to water level which would have been helpful in this case and (c) against similar instruments deployed together.

It is sometimes quicker and easier to do your own validation checks than to spend time and effort returning data and equipment to the manufacturer for their (often defensive) assessment of whether the equipment is at fault or not. In this case, the consultant and their client were able to move forward with their project very quickly by taking some simple steps which gave them the confidence to dismiss the possibility of saline intrusion backed up  with the necessary evidence to demonstrate this conclusion to others.



©Peter Dumble 2017
3 August 2017