October 2018

PLUMECAST – a predictive modelling system

  • By Fardausur Rahaman, Pricipal – Air Quality and Graeme Starke, Technical Discipline Manager – Air Quality, SLR Consulting

A case study exploring the economic and environmental benefits of adopting fine-scale meterological and dispersion forecasting systems in open cut mining operations


The voice of the community regarding environmental impacts is becoming increasingly strong in Australia and mining operations are facing economic and environmental challenges to satisfy regulatory bodies and local communities that they are operating within compliance limits and are effectively managing potential public health risks. Air quality and noise are the two most prominent issues that have potential to upset the community and can lead to complaints and scrutiny from environmental regulators. Costs associated with the incorporation of mitigation measures in the mine plan (eg bunds, haul road design, buffer zones) and the implementation of operational mitigation measures (eg watering, chemical dust suppressants) also have potential to significantly impact on a mine’s financial position.

Trigger action response plans (TARPs) are becoming more widely used to assist in managing onsite operations and optimising mitigation practices to reduce the risk of non-compliances and complaints. However, these systems are still reactive to the real-time data being recorded by the monitoring network and the measurements are limited by the number and location of monitoring sites. As a result, many mines may still struggle to ensure compliance at surrounding residences due to limitations in the information available on current and future changes in local weather and dispersion conditions. This is especially true for mining operations in relatively complex terrain, where spatial variations in meteorology can be significant over the mine site and its surrounding areas. In addition, there is potential for significant financial, operational and safety implications associated with unscheduled changes to onsite operations that may be required at short notice if a trigger level is exceeded.

Considering the above, this article explores potential benefits of using predictive meteorological and dispersion modelling systems to assist sites in making informed decisions regarding the level and duration of mitigation required at a particular location/source to avoid significant offsite impacts or non-compliances.

Predictive modelling systems

Meteorological and plume dispersion prediction relies upon computer simulations of atmospheric physics and fluid motion, using inputs from global radiosondes and weather satellite data. These simulations establish three-dimensional grids representing the globe and incorporate the effects of synoptic wind patterns, heat transfer, solar radiation, relative humidity and hydrology to provide predictions of local-scale future meteorology that can then be applied to predicting future air quality impacts.

Several attempts have been made in the past by both academic and non-academic organisations to develop a computational tool with the ability to predict changes in local micrometeorology and dispersion conditions for near future. This would assist mine operators in meeting the challenges of controlling air emissions from their site based on upcoming meteorological and dispersion conditions while minimising impacts on the production schedule; however, these modelling systems generally failed to meet the expectations of mine operators. This was mainly due to the use of relatively simple meteorological and dispersion models, which were selected as computational power at the time was inadequate for the adoption of more robust models. However, with the increased availability of more powerful computers and faster download speeds, several forecasting modelling systems have been developed in recent times by different organisations, which use advanced meteorological (WRF/CALMET) and dispersion (CALPUFF) models.

Figure 1. Meteorological data validation.

The major influence on the accuracy of the air quality impact prediction is the accuracy of the meteorological model (Baklanov, 2002). A review of publicly-available literature relating to WRF model settings for the Australian context has been performed. Based on this review, the model was tested with different combinations to identify the optimum combination of WRF parameters for forecasting meteorology in Australia, New Zealand and Papua New Guinea. Based on the findings of this analysis, the following key principles have been identified to ensure the accuracy of the forecast:

  • care should be taken in selecting the options for each individual parameter (eg turbulence, radiation etc) as use of inappropriate options may result in poor prediction of local meteorology; therefore, the model outputs should be validated against concurrent meteorological measurements at locations near the mining activity to ensure that the appropriate options have been selected
  • a three-dimensional dispersion model such as CALPUFF should be utilised to predict the impacts of topography, land use and pollutant travel time between source and receptor
  • emission estimates for onsite operational activities must be based on best available information to maximise the accuracy of the downwind concentration predictions
  • where possible, the downwind concentration predictions should be validated against monitoring data
  • forecasts should be updated at six-hourly intervals.

If the above principles are followed, a combination of the WRF and CALMET models can provide reliable meteorological forecasts in the Australian context. Based on an analysis of model predictions given by WRF/CALMET with available observational data, it has been established that the model predictions validate well with the observational data, providing an appropriate combination of the different WRF options in conjunction with adequately fine-scale topographical and land use data is used.

Case study

A number of case studies at different mine sites operating in Queensland have been conducted by SLR using this modelling system (hereafter referred to as PLUMECAST) to assess the quality of the forecast as well as to assess whether the model outputs provide adequate information to mine operators for identifying and mitigating offsite impacts associated with different dust/gaseous emissions. The outputs of one trial are presented below.

The WRF model was configured with four nested domains with 27 km, 9 km, 3 km and 1 km grid resolutions. The WRF output of the innermost domain (1 km resolution) was used as input to CALMET to forecast the local meteorology at three mine sites in Queensland, including one mine site where SLR is providing hourly forecasts of blasting plumes (particulates and deposition) using PLUMECAST. One of these mine sites is located in complex terrain with locally observed meteorological data available. This site was therefore selected for meteorological forecast validation, given that if the modelled forecast validates well in complex terrain, it can be expected to provide a reliable forecast for sites located in simpler terrain.

A comparison of the forecasted and actual wind data for three consecutive days is presented in Figure 1. This comparison shows that the forecasted wind parameters align well with the observed meteorological data for the day of analysis.

The dispersion model of the PLUMECAST system was used to investigate whether the output of the model can provide useful information to the mine operator in making informed decisions regarding the next day’s work schedule. To make an informed decision, the mine typically requires the forecasted cumulative (including background) impact from the mine to assess the potential for any exceedances at surrounding sensitive areas. If the model predicts exceedances, the operator needs to know the contribution from each significant dust-generating activity to the predicted exceedance to optimise mitigation resources.

Figure 2. Typical output of PLUMECAST – hourly average PM10 concentrations at a particular hour. Note: The base map presented in this figure is not representative of the mine site modelled. A separate base map was sourced from another area and has been used for demonstration purposes.
Figure 2. Typical output of PLUMECAST – hourly average PM10 concentrations at a particular hour. Note: The base map presented in this figure is not representative of the mine site modelled. A separate base map was sourced from another area and has been used for demonstration purposes.

A sample output from PLUMECAST is provided in Figure 2. The output shows the cumulative impacts as shaded contours with the increment from each significant mining activity as different colour contour lines. In this example, the plot shows the wind pattern across the modelling domain (as arrows), as well as the predicted cumulative and incremental PM10 concentrations.

From the information presented in Figure 2, it can be seen that:

  • the mining operation at this hour is likely to contribute significantly to dust levels at the receptors to the north, mainly as a result of emissions from hauling, with minimal contribution from the coal handling preparation plant.
  • if blasting was performed during this hour, there is potential for significant impacts at the nearest receptors.

If the forecast for subsequent hours on this day predicted a wind shift to northerly winds, a more appropriate time could be scheduled for the blast. In addition, while more intensive watering on the northern haul road may be advisable for this time of the day, later, when the wind shifted to the north, the water trucks may be relocated to the southern haul roads.


Investigations by SLR have demonstrated that PLUMECAST, or similar WRF/CALMET forecasting tools with appropriate design, can be used to provide reliable hourly meteorological and dispersion forecasts that can provide valuable information to mine operators to assist them in making informed decisions regarding mitigation measures and daily scheduling. This may lead to reduced operational costs and enhanced productivity, while minimising the potential for complaints and scrutiny from the surrounding community and regulatory bodies. 

A brief summary of the potential advantages of adopting such forecasting techniques in mine operations is presented in Table 1.

Table 1. Click for larger image.


Baklanov, A, Rasmussen, A, Fay, B, et al, 2002. Potential and shortcomings of numerical weather prediction models in providing meteorological data for urban air pollution forecasting, Water, Air, & Soil Pollution: Focus, 2:43, https://doi.org/10.1023/A:1021394126149.

Share This Article