Supervising Scientist Division

Supervising Scientist Annual Report 2003 - 2004: Jabiluka

Supervising Scientist, Darwin, 2004
ISBN 0 642 24391 3
ISSN 0 158-4030

2 - Environmental assessments of uranium mines (continued)

2.3 Jabiluka

2.3.1 Developments

The highlight of 2003-04 was the completion of works associated with Long Term Care and Maintenance (LTCM) at Jabiluka involving:

These works followed MTC agreement (on 30 June 2003) to continue with the irrigation programme. Approval for the civil works was given by the Northern Territory Minister for Business, Industry and Resource Development on 28 July 2003 and works commenced immediately.

Figure 2.15: View of the Jabiluka Interim Water Management Pond after LTC&M site works

Figure 2.15: View of the Jabiluka Interim Water Management Pond after LTC&M site works

Some milestones on the project included:

Site demobilisation and handback to ERA occurred on 5 December 2003.

All infrastructure (such as stockpiles, demountables, gate house, water tanks, turkey nest pond, workshop and the fuel farm) has been removed from the site. Core samples from the site have been moved to Djarr Djarr camp and are currently stored in shipping containers.

No other developments occurred at the Jabiluka Site during the reporting period. The site as a whole held up well throughout the reporting period with only minor works to sediment traps and the access road required post wet season.

2.3.2 On-site environmental management

Water management

The site is now maintained as a passive discharge site. The Interim Water Management Pond (IWMP) has only its own catchment and therefore is a net loser of water. ERA may supplement the pond water storage with potable water from the potable bore to protect and increase the life of the liner in the future, but this was not undertaken during the reporting period. Water level during the wet season reached a maximum depth of approximately 1 m in the pond.

An overflow point has been constructed in the existing concrete choke structure in the unlikely event that the pond fills and passive discharge is required.

Radiological exposure to employees

The new Jabiluka Authorisation, which took effect in July 2003, removed the statutory requirement of quarterly reporting of radiological monitoring data for Jabiluka. Therefore, the last quarterly report submitted under the old Authorisation was the March 2003 report. This report shows that, as in previous quarters, no employee going underground at Jabiluka has received a radiation dose above the dose limits for members of the public.

The original Jabiluka Radiation Management Plan had been drafted for the development of the decline and does not account for the change of operations on the Jabiluka minesite to a care and maintenance state, which took place in 1999. The new Authorisation now only requires reporting of radiation monitoring data if any ground disturbing activities involving radioactive mineralisation occur on site.

Between 18 August and 27 September 2003 all the mineralised material that was stockpiled on-site, including stockpile cover and lining material, pond water that was not spray-irrigated and silt from the pond were returned to the underground workings. Results of the radiological monitoring of employees and monitoring of environmental areas during this operation were reported in the ERA Radiation Monitoring Program Report (RSRP-0311-03).

The maximum total effective dose to a designated employee working as a bogger driver was 0.37 mSv during the operation. The underground boggers are used to push up and compact the material that was transported and dumped underground. Some 60% of the total effective dose was due to exposure to direct γ-radiation, 30% due to the inhalation of radon decay products and the remaining 10% due to the inhalation of radioactivity trapped in or on dust. This is below the limit of 5 mSv for non-designated employees and much smaller than the occupational radiation dose limit of 20 mSv recommended by the International Commission on Radiological Protection.

Audit outcomes

The mid term review, which is a follow up to the 2003 Annual Audit of the Jabiluka operation (reported in the 2002-2003 Annual Report of the Supervising Scientist) was undertaken on 17-18 November 2003. The review team consisted of representatives of oss, NLC and DBIRD.

The procedure was the same as in previous years with the protocol being developed by the audit team to conform with the appropriate ISO standards and then submitted to ERA two weeks prior to the interview and site inspection.

The Annual Audit occurred between 24-28 May 2004.

The audit team found there were no Category 1 or Category 2 non-conformances and 9 conditional issues relating to commitments yet to be undertaken or completed. The team made one observation (Table 2.10).

Table 2.10: Observation from the 2004 Jabiluka environmental audit
1. It was noted that there was a lack of signage and lack of locks on the core containers at Djarr Djarr camp.
Minesite Technical Committee

The Jabiluka Minesite Technical Committee did not meet during 2003-04.

Authorisations and Approvals

Changes to, and approvals under, the Authorisation during 2003-04 are listed in Table 2.11.

Table 2.11: Jabiluka authorisation changes/approvals
Date Issue
31 July 2003 Variation 0140-01 to consolidate previous Authorisations, modify theenvironmental monitoring programme and approve placement of the project in long-term care and maintenance mode. This includes backfill of the decline.Cosmetic changes are also included to reflect the change of relevant NT legislation from Uranium Mining (Environmental Control) Act to the Mining Management Act.
11 May 2004 Approval of Jabiluka Rehabilitation Plan #7

There were no reportable incidents at Jabiluka during the year.

2.3.3 Off-site environmental protection

Surface water quality

In accordance with the Jabiluka Authorisation, ERA is required to monitor a range of surface and groundwaters on the lease and to demonstrate that as a result of activities within the project area the environment remains protected. Specific water quality objectives (criteria thresholds are described in Section 2.2.3), based on the approach of the ANZECC and ARMCANZ Water Quality Guidelines, must be achieved. In addition to the ERA programme, the Supervising Scientist conducts a routine environmental monitoring programme including chemical, physical and biological monitoring in Ngarradj (Swift Creek).

Chemical and physical monitoring of Ngarradj (Swift Creek)

The first water chemistry sample for the Supervising Scientist's 2003-04 wet season surface water monitoring programme was collected from the Ngarradj downstream statutory compliance point on 23 December 2003. Water samples were collected weekly for the first seven weeks of the wet season. After considering the reduced risk to the environment posed by the site (which is now in a care and maintenance mode) and the high water quality measured in the early part of the season, the sampling of water for chemical analyses was reduced to fortnightly from 4 February 2004. The last samples for the season were collected on 25 April 2004.

Key water quality data from the SSD routine monitoring of Ngarradj are reported on the Internet at The highlights of the monitoring season are summarised below.

The on-site activities carried out over the 2003 dry season (returning the stockpiles underground and irrigation of ponded water) do not appear to have affected metal and ion concentrations in Ngarradj during the 2003-04 wet season (Table 2.12), with all measured parameters showing (i) similar trends at both upstream and downstream sites, (ii) similar values to those seen in previous years, and (iii) compliance with the guidelines/limits set for environmental protection (apart from the first flush period discussed below).

Concentrations of uranium at the downstream site, less than 0.4% of the limit, were similar to the concentrations measured concurrently at the upstream site, and at both sites in previous years (Figure 2.16).

While no change is obvious in the turbidity measurements recorded through the routine monitoring programme (Table 2.12) an event based programme of monitoring (a method under development to investigate long-term transport of sediments) detected elevated turbidity levels at the downstream site following storm events. The increase, while detectable, was within the range of results recorded for Ngarradj previously and the results of biological monitoring show that the increase did not impact on the downstream ecosystem.

The only time any parameters measured were above the guideline/limit was at the start of the wet season when EC and sulfate were high at both the upstream and downstream sites. As the downstream measurements were almost identical to those upstream, it was concluded that the limit exceedance was not related to mining. The water quality limits and guidelines for Ngarradj were met through the remainder of the season providing reassurance that the aquatic environment did not suffer adverse effects from activities at the Jabiluka site. The results of the biological monitoring programme confirm this.

Table 2.12: Summary of SSD Ngarradj (Swift Creek) water quality measurements upstream and downstream of Jabiluka Since the 1998-99 wet season
  Median Range
Parameter Limit Year Upstream Downstream Upstream Downstream
pH 3.92-6.00* 2003-2004
21 2003-2004
- 2003-2004
NO3 (as NO3) (mg/L) 1.26 2003-2004
Sulphate‡ (mg/L) 1.5 2003-2004
0.76* 2003-2004
5.8 2003-2004

* guideline only, ‡ dissolved (<0.45 µm);
Nitrate was only measured in the early part of the 2003-04 season. As nitrate values are generally higher in the early stage of the season the median for the 2003-04 season is biased toward a higher value. The maximum values for the 2003-04 season are actually lower than for previous seasons.

Figure 2.16: Uranium concentrations in Ngarradj since the 1998–99 wet season

Figure 2.16: Uranium concentrations in Ngarradj since the 1998-99 wet season

Source: SSD data

Biological monitoring in Ngarradj (Swift Creek)

Benthic macroinvertebrates and fishes have been selected as key 'indicator' groups to monitor and assess potential impacts upon aquatic ecosystems arising from the Jabiluka mine. Strategic baseline sampling of both indicator groups has been conducted in the past six wet seasons (1998-99 to 2003-04 wet seasons) with results of the macroinvertebrate studies reported in all Supervising Scientist's Annual Reports published since 1999-2000.

Monitoring using macroinvertebrate community structure

Benthic macroinvertebrate results are being used to assess whether or not mine contaminants reaching Ngarradj (Swift Creek) downstream of the Jabiluka minesite during the wet season are adversely affecting the aquatic biota. Such contaminants may include suspended solids, or residual contaminants such as nutrients (from explosives) or uranium from waste water spray-irrigated on land during the dry season.

The design and approach of the macroinvertebrate study for Jabiluka are similar to those described above for the Ranger macroinvertebrate study. Macroinvertebrate samples were gathered from sites in Ngarradj Creek (draining the minesite) upstream and downstream of Jabiluka, and also from paired upstream and downstream sites in three adjacent streams currently unaffected by any mining activity at Jabiluka (control streams). For the three wet seasons, 1998-99, 1999-2000 and 2000-01, samples were collected from each site at three to four weekly intervals for the period of creek flow. From the 2001-02 wet season onward, this sampling effort was reduced so that only Ngarradj and one of the control creeks were sampled monthly during the wet season, while all four streams were sampled on the final sampling occasion in March. For each sampling occasion and for each pair of sites for a particular stream, a dissimilarity index was calculated - see explanation of this index in Section 2.2.3 of this Annual Report, 'Biological monitoring in Magela Creek'. Sample processing and data analysis have now been completed for wet seasons up to and including 2002-03 with results plotted in Figure 2.17.

As may be seen from Figure 2.17, dissimilarity values for Ngarradj over the five wet seasons are reasonably constant and similar to values reported in adjacent control streams over the same time period. While single-factor ANOVA shows significant differences in the mean dissimilarities amongst the four streams over the five-year period of study (P=0.008), this result is due to the relatively low dissimilarity values observed in Catfish Creek compared with values for the other three creeks (see Figure 2.17). For any paired grouping of sites in a stream, macroinvertebrate communities tend to be more similar the closer the sites are together and the greater the degree of water permanence (eriss unpublished data). Of the four streams sampled in the Jabiluka study, Catfish Creek sites are closest together while this is the only creek that either flows, or holds water, throughout the year. These factors may explain the generally lower dissimilarity values found in Catfish Creek. Relatively high dissimilarity values observed in Ngarradj Creek on occasions over the five-year period (indicating differences in types of macroinvertebrate taxa and/or their abundances between the two sites) were discussed in the 2002-03 Supervising Scientist Annual Report.

Figure 2.17: Paired upstreamdownstream dissimilarity values (using the Bray-Curtis measure) calculated over four wet seasons for community structure of macroinvertebrate families in several streams near the Jabiluka minesite.

Figure 2.17: Paired upstreamdownstream dissimilarity values (using the Bray-Curtis measure) calculated over four wet seasons for community structure of macroinvertebrate families in several streams near the Jabiluka minesite.

Note: Missing data indicate absence of flow (7J Creek, 3rd wet season) or no sampling (7J and North Magela creeks, 4th and 5th wet seasons).

Of note, dissimilarity values observed in Ngarradj in 2002-03 are relatively constant and near average. These results support conclusions drawn in previous Annual Reports that up to and including the 2002-03 wet season, there were no adverse effects of runoff from the Jabiluka minesite upon macroinvertebrate communities downstream in Ngarradj.

Monitoring using fish community structure

Baseline sampling as well as monitoring of the Jabiluka minesite have included measures of fish community structure in small streams using the same design and sampling locations as the Jabiluka macroinvertebrate study. Thus, counts of the number of individuals of different fish species are made by bankside visual observations of fish at sites in Ngarradj (draining the minesite) upstream and downstream of Jabiluka, and also from paired upstream and downstream sites in three adjacent streams currently unaffected by any mining activity at Jabiluka (control streams). Sampling has been conducted at each site in March and April of each of the past five wet seasons, 1998-1999 to 2003-04, though data collection has been prevented on some occasions in the past due to unsuitable field conditions (see Supervising Scientist Annual Report, 2002-03). Sampling of Ngarradj in March 2004, for example, was not possible due to extended rain, overcast conditions and greatly reduced water visibility.

The addition of a February sampling for 2003 in Ngarradj and Catfish Creeks was to examine seasonal variation in fish assemblages and evaluate the possibility of including an extra monthly sample in the routine monitoring programme. This sampling event was almost entirely dominated by juvenile Black-banded rainbowfish and Spangled grunter, indicating sampling in February was too early to enable time for migration of fish from downstream recruitment areas. Because of these reduced species counts, community study and analysis are limited (see Supervising Scientist Annual Report 2002-03 for further discussion).

A dissimilarity index was calculated for each pair of sites within a stream for each sampling occasion. The paired-site dissimilarity values for Ngarradj and three control streams are compared for the six-year period 1999 to 2004 in Figure 2.18. In the exposed stream, Ngarradj, and the three control streams, Wirnmuyurr, North Magela and Catfish Creeks, the paired-site dissimilarities for 2004 were within the range of values observed for previous years and were very similar to the average dissimilarity calculated for those years (Figure 2.18).

Figure 2.18: Figure 2.18 Paired upstream–downstream dissimilarity values calculated for fish community structure in four streams near the Jabiluka minesite over the past several wet seasons.

Figure 2.18: Figure 2.18 Paired upstream-downstream dissimilarity values calculated for fish community structure in four streams near the Jabiluka minesite over the past several wet seasons.

Note: Calculations made using Bray-Curtis measure with log (x+1) transformed abundance data. Not all sites were sampled on each sampling occasion.

Extremes in dissimilarity values are more often associated with February and March samplings. Given the difficulty experienced in recent years of successfully obtaining visual data during the wettest months, February to March, future monitoring should focus on samples in April when water levels are declining and visibility is rarely a problem.

However, even when only April data are considered, unusual within-stream dissimilarities can occur from factors unrelated to mining activity. The largest variation occurred in North Magela Creek in 2003. The marked decline in dissimilarity recorded then was due to a natural occurrence of a late upstream fish migration (see the Supervising Scientist's Annual Report for 2002-2003). Late migrations were not encountered in 2004 and the dissimilarity of upstream and downstream sites returned to values near the average of the baseline data.

The constancy of paired-site dissimilarity values for April samples in Ngarradj and two of the control streams over the past several years supports the conclusion that there were no adverse effects of runoff from the Jabiluka minesite on fish communities in Ngarradj. Further, the 2004 values indicate that earthworks associated with the decommissioning of the site during the 2003 Dry season did not alter fish communities in Ngarradj. Unusual changes in paired-site dissimilarities from Ngarradj and North Magela Creek appear to have been chance results relating to coincidence of sampling with fish recruitment behaviour.

Implicit in this design is the comparison of fish community structure between the paired sites in Ngarradj with similar paired-site data in control streams. However, this approach is confounded by possible movement of fish between the two sites in each stream. Greater reliance will need to be placed, if data accrue in future, on the comparisons between sites in Ngarradj and those in independent control streams.

Fish studies associated with Jabiluka are conducted in billabongs as well as streams. Two of the billabongs studied for monitoring of Ranger mine are also strategically located for detection of effects of any future mining at Jabiluka. These are Wirnmuyurr (potentially exposed) and Cathedral (control) billabongs. As reported in Section 2.2.3, the structure of the fish communities in these billabongs in 2004 was within their natural range of variation indicating no mining-related impacts (Figure 2.13).

Radiological exposure of the public

The critical group that may receive a radiation dose due to activities at Jabiluka are the 60 or so inhabitants of Mudginberri, a small community approximately 10 km south of Jabiluka. At Four Gates Road radon station, a few kilometres west of Mudginberri, the Supervising Scientist has a permanent atmospheric research and monitoring station. Radon decay products (progeny) and long lived alpha activity (LLAA) concentrations are measured on a monthly basis. In addition, radon gas is continuously measured at the station with radon data being recorded every 30 minutes.

Figure 2.19 shows the radon decay products (progeny) and long lived alpha activity (LLAA) concentrations measured at Four Gates Road radon station. In addition, the monthly average radon decay product concentration as calculated from the average radon concentration, in Becquerel per m³, using an equilibrium factor ƒ of 0.5 and a dose conversion factor of 0.0056 mJ/Bq, is shown. Radon progeny and long lived alpha activity concentrations are very small and comparable with natural background levels.

Figure 2.19: Radon decay products (RdP) and long lived alpha activity (LLAA) concentrations measured at Mudginberri Four Gates Rd radon station (SSD) and on top of Jabiluka Hill (ERA).

Figure 2.19: Radon decay products (RdP) and long lived alpha activity (LLAA) concentrations measured at Mudginberri Four Gates Rd radon station (SSD) and on top of Jabiluka Hill (ERA).

Note: Radon decay product concentrations, as calculated from radon concentration data measured continuously at Four Gates Rd, using an equilibrium factor ƒ = 0.5 and a dose conversion factor of 0.0056 µJ/Bq are shown as well.