Surface Water Chemistry Monitoring Program

The Supervising Scientist Division (SSD) undertook a substantial review of its surface water chemistry monitoring program over the 2010 dry season, resulting in some significant changes in the monitoring program from the 2010-11 wet season onwards. This change followed five years of comparative evaluation of continuous monitoring and discrete grab sampling methods.

The collection of water samples for chemical analysis changed from weekly grab sampling to event-based sample collection using automatic samplers. These samplers are triggered when pre-programmed thresholds are detected by the continuous monitoring of EC (a measure of concentrations of dissolved salts) and turbidity (a measure of suspended sediment). This change involved significant infrastructure upgrades in the field and in the office (detailed below), together with updated operating procedures and manuals. Each monitoring site has a dedicated manager who has overall responsibility for all aspects of site maintenance and data validation.

Monitoring infrastructure

Improvements to monitoring infrastructure completed since the 2010 dry season include:

  • Installation of duplicate sets of continuous monitoring equipment at the upstream and downstream Gulungul Creek sites;
  • Construction of a new custom built monitoring pontoon for the Magela Creek downstream site (Figure 1);
  • Installation of two Gamet auto-samplers on the Magela Creek downstream pontoon;
  • Various improvements to the Gulungul Creek monitoring sites to facilitate calibration and replacement of sondes under high-flow conditions;
  • A substantially improved data download and telemetry system;
  • New anchor points were installed at Magela Creek downstream and Gulungul Creek upstream monitoring sites during 2012 dry season; and
  • Sondes have been fitted with new turbidity sensors.
MCUGT (Magela Creek upstream of the Ranger mine) monitoring station.

Photo: Magela Creek downstream monitoring pontoon

Continuous data

Continuous electrical conductivity (EC), turbidity and stream height data is automatically downloaded daily, or more frequently on a manual basis if required, from the monitoring stations and imported into SSD's water quality database. Any events exceeding pre-set EC or turbidity levels will result in immediate notification (via mobile phone) to key staff members, facilitating a rapid response.

Continuous data for all sites will be posted to the SSD website weekly in arrears, once the data has been checked and validated. This frequency will decrease to monthly during recessional flow conditions after the 1st of May.

From 2012-13 onwards the continuous EC data will be assessed and interpreted against the Ecotoxicology Electrical Conductivity-Magnesium Pulse Framework.

Automated sample collection

The upstream and downstream monitoring sites in both Magela and Gulungul Creeks are equipped with auto-samplers. These samplers are programmed to collect a 1 litre water sample in response to the occurrence of pre-specified EC or turbidity conditions. The responsible staff members are notified of sample collection via an SMS message from the mobile phone network.

The downstream Magela Creek site is equipped with two samplers. One sampler is used for collection of higher EC samples and the other is for collection of higher turbidity samples. The use of two samplers at this site ensures sufficient capacity for collection of both EC and turbidity events, and backup redundancy in the event of malfunction of one of the samplers.

Wherever possible the samples of collected water will be retrieved from the automatic samplers within 24 hours of collection. However this might not always be possible owing to dangerous weather or flow conditions, and unforseen limitations on staff availability.

Analysis of Water Samples

From the 2010-11 wet season onwards, all water samples are analysed for total metal concentrations (dissolved metals plus those weakly bound to suspended particulate matter). This contrasts with the previous weekly grab sample program where all samples were filtered in the field immediately after collection, and analysed for filterable (dissolved) metals only.

igure 1

Figure 1. Proportions of the total analyte concentration associated with the dissolved (filterable through a 0.45 µm filter) or particulate fractions. Error bars represent the standard deviation.

The event-based samples are subject to a variable period of standing (mostly less than 24 hours) before they are able to be retrieved from the auto-sampler and processed in the laboratory. By analysing the total metal concentration (as distinct from dissolved metal concentration), the proportion of dissolved metals that would typically become "lost" due to adsorption to the surface of particulate matter during the standing period, will be accounted for.

Over two wet seasons (2008-09 and 2009-10), the SSD analysed event-based water samples collected over a range of EC and turbidity values for both the total (dissolved metals as well as those weakly bound to suspended particulate matter) and filterable (dissolved metals only) metal concentrations. The data are compared in Figure 1.

The results in Figure 1 (also summarised in Table 1 below) show that concentrations of magnesium (Mg) and sulfate (SO4) are predominantly associated with the dissolved fraction. This is consistent with their chemically non-reactive nature. For uranium, approximately 30% was associated with the particulate fraction and approximately 70% with the dissolved fraction. The behaviour of manganese (Mn) was highly variable.

Table 1.  Percentage of the total metal concentration in a sample that is expected to be associated with either the dissolved or particulate fractions.
  Mg Mn SO4 U
Dissolved (%) 95 (± 11) 42 (± 30) 95 (± 11) 67 (± 14)
Particulate (%) 5 (± 11) 58 (± 30) 5 (± 11) 33 (± 14)

Data from 94 water samples.

It is clear that the majority of Mg, SO4 and U in samples collected from the Magela Creek downstream site are present in the dissolved fraction. The variation (one standard deviation) in the proportion of dissolved and particulate concentrations is shown in brackets in Table 1. Using the information in Table 1, the relative proportions (dissolved or particulate) of the total concentration likely to be present in samples measured during the wet season can be estimated. This will enable a comparison to be made with the dissolved data measured prior to 2010-11.

The use of totals for monitoring of water quality provides a more conservative (i.e. more protective) assessment of the concentrations of metals present, noting that the guideline values used for compliance assessment in Magela and Ngarradj Creeks (Tables 2 and 3 below) are based on the dissolved (more bioavailable) concentrations present.

Table 2.  The Water Quality Objectives for Magela Creek 2013
Parameter Trigger Values
Focus Action Guideline Limit
pH 5.9 – 6.5 5.6 – 6.7 5.0 – 6.9 N/A
Turbidity (NTU) 5 10 26 N/A
Electrical Conductivity (µS/cm) 18 30 N/A 42
(≥72 hour limit)
Magnesium (mg/L) 1.0 2.0 N/A 3.0
(≥72 hour limit)
Sulfate Use EC
as proxy
Use EC
as proxy
Use EC
as proxy
Manganese (µg/L)
Applicable when flow > 5 cumecs.
35 45 N/A 75
Uranium (µg/L) 0.3 0.9 N/A 6
Total Ammonia Nitrogen (mg/L) 0.1 0.3 0.7 N/A
Radium-226 N/A N/A N/A 10 mBq/L
wet season
median difference
Table 3.  The Water Quality Objectives for Ngarradj (Swift Creek) as determined by the Jabiluka Minesite Technical Committee on 21 September 2001
Parameter Trigger Values
Focus Action Guideline Limit
pH 4.61 – 5.31 4.27 – 5.65 3.92 – 6.0 N/A
Electrical Conductivity (µS/cm) 15 18 21 N/A
Magnesium (mg/L) 0.37 0.5 0.76 N/A
Nitrate (mg/L) 0.3 0.63 1.26 N/A
Sulfate (mg/L) 0.6 0.9 1.5 N/A
Uranium (µg/L) 0.02 0.03 N/A 5.8

Reports and documents which determined the above water quality guidelines:

Quality Assurance/Quality Control and Radium

Each monitoring site manager will visit their assigned sites fortnightly to conduct a quality assurance/quality control (QA/QC) check. This will involve taking an in-situ measurement of EC and turbidity using a calibrated hand held instrument and performing instrument calibrations and other maintenance as required. A range of QA/QC samples will be collected monthly from each site, including an auto-sampler blank, an auto-sampler duplicate and an in-situ filtered grab sample (using the old grab sampling method). These data will enable rigorous assessment of the quality of the auto-sampler samples as well as a comparison between the old grab sampling and the new event-based sampling methods.

During these regular QA/QC site visits a sample for routine radium analysis will also be collected from the Magela Creek stations. Two fortnightly radium samples will be combined to make a four week composite sample from each site. This monthly routine monitoring program is complemented by radium analysis on composite event-based samples from the autosampler, which will provide a more comprehensive assessment of the variation in radium concentrations over the wet season.

Water Quality Monitoring during Recessional Flow

After 1st May and if recessional flows have been confirmed in Magela Creek from the hydrograph (sustained flow at < 5 cumecs) and water chemistry traits (slowly rising and converging EC at upstream and downstream monitoring sites) then the four-weekly QA/QC sample collections will cease. Fortnightly continuous monitoring site/mini-sonde checks and radium sample collection will continue. Website updates will change to monthly unless there are notable events/changes in water quality, which require earlier reporting to stakeholders.

If recessional flows have not occurred on 1st May then routine continuous monitoring will continue until recessional flows have become established.

Magela Creek

Gulungul Creek

Ngarradj (Swift Creek)