Determination of the Greenhouse Gas Emissions from Spontaneous Combustion

Spontaneous combustion is a potentially large source of greenhouse gas emissions from open cut coal mines. ACARP has previously supported a number of projects designed to develop a methodology to estimate the scale of these emissions, and while these have advanced the understanding of the problem, there have been practical problems which have limited their general applicability. In the current project, an approach was trialled where stationary monitors were placed at sites in the Hunter Valley and the Bowen Basin and inverse modelling applied to estimate area-wide fluxes of CO₂ from mines affected by spontaneous combustion. In the Hunter Valley, flux estimates were also made by traversing an instrumented vehicle through the CO₂ plumes from spontaneous combustion and the results were compared to those from the stationary monitors.

Inverse modelling was used to estimate area-wide CO₂ emissions fluxes from both sites. In the Hunter Valley, the technique provided an estimate for the region studied that ranged from about 1040 to 1600 kt CO₂ y^-1. This compared well with the estimates obtained from direct plume measurements using traverse-based techniques of around 820 to 1600 kt CO₂ y^-1. While there is a large spread in these estimates they are less than those found in Project C9062 which were based on infrared thermography and a relationship between the surface temperature and emissions from spontaneous combustion. This may have been a result of the large uncertainty in the ground temperature-emission relationship used for the thermal thermography calculations but also the infrared data were taken more than five years previously to the data used in the inverse model and the traverse based estimates.

The model utilised a detailed seeding map, which in the Hunter, was provided by the airborne thermal imagery obtained as part of Project C9062. No thermal images were available for the Bowen Basin mine and the inverse modelling was initialised with areas of known self heating provided by mine site staff. The model was able to identify those regions noted by staff as well as regions not originally identified, such as coal stock piles. Estimated emissions from the Bowen Basin mine ranged from 200 to 320 kt CO₂ y^-1, a range considerably smaller than found for the area within the Hunter Valley.

In their study of emissions from spontaneous combustion, Day estimated the emissions of greenhouse gases from six open cut coal mines in the Hunter Valley and in the Bowen Basin. Their estimates were based on data made available by the mines and also data that had been developed during previous ACARP projects (see Carras et al, 2000, 2002). In the Day study the estimates of greenhouse gas emissions from spontaneous combustion are significantly smaller than the estimates from the current study, which are approximately five times larger than for the largest individual mine studied by Day. However, in the current work the estimate of emissions for the Hunter Valley is for a larger region than the individual mines considered by Day. Nevertheless the results of this study suggest that for mines where spontaneous combustion is and has been an issue the emissions may be very significant.  

Experience gained in this study suggests a number of recommendations to enhance estimates. These include:

• There are advantages to using a detailed seeding value for initialising the inverse model. Thermal imagery was found to be a suitable methodology for the Hunter Valley where the data were available from a previous ACARP study.
• Where possible, direct plume measurements provide a reasonable method for estimating area wide fluxes. For an accurate estimate the method requires an understanding of the mixing processes at the point of measure. Numerical models such as TAPM can provide an estimate of the mixing processes while a more accurate approach would involve simultaneous measurement of vertical profiles at a fixed location using, for example, a tethered balloon.
• Simultaneous measurement of other species (eg CO, SO₂, NO_X ) may provide a mechanism, through ratios, to distinguish emissions from self-heating from other sources such as diesel machinery.
• Stationary monitors require installation for long periods of time to ensure sufficient data sets are collected for modelling purposes. Six months of data were sufficient in the Hunter Valley region where a monitor was placed in a location directly downwind of the source in the predominant valley flows. In the Bowen Basin mine site however, a longer sampling period would have been more appropriate due to smaller regions of self-heating and a predominance of very light winds associated with elevated concentrations which cannot be modelled.

All three methods used in this study showed significant overlap in the emissions estimated. However, all three methods require specialist knowledge and expertise for their implementation. Both the inverse method and the method based on plume traverses require specialist knowledge of air quality modelling, measurement of atmospheric gases and the interpretation of atmospheric field data. Such capability does not normally exist within coal mine staff. Use of these methods to estimate emissions from spontaneous would require the engagement of scientists and engineers with such specialist knowledge.

Finally, while the three methods used in this study have shown overlap there is still significant uncertainty associated with any single approach. Consequently the estimation of greenhouse gas emissions from spontaneous combustion still remains a difficult undertaking.