Modelling minewater flow and quality changes after coalfield closure

Abstract

The changes that have taken place in the British Coal Industry over the last five years have meant that in many coalfileds the last deep mines have closed. When a coalfield is abandoned and dewatering ceases, groundwater levels rebound, threatening surface waters with polluting discharges. However, the sparse data sets available limit modelling with existing techniques. A lumped parameter model GRAM (Groundater Rebound in Abandoned Mineworkings) has been developed. This model conceptualises a coalfield as a group of 'ponds'. Each pond is an area of the coalfield that has been extensively worked and can be considered as a single hydraulic unit. The ponds are connected by pipes which represent major inter-connecting roadways along which flow is assumed to be turbulent. Discharge to the surface is also represented using pipes. Flow through the pipes can be calculated using the Prandtl and Nikuradse of the Colebrook-White pipeflow equations. The storage coefficient can vary vertically to represent both worked Coal Measures and the intervening unworked strata. GRAM is able to predict the timing and volume of discahrges. An iron component gives an indication of the water quality evolution of the discahrges. Monte Carlo simulation allows the variables that have most error in their estimation to be represented by probability distributions. The Dysart-Leven Coalfield in eastern Fife, Scotland has not been mined since 1985. However, dewatering has continued to protect the workings in the Frances Colliery. In 1994 British Coal decided the Frances would never reopen, there is therfore no longer a need to continue dewatering. GRAM has been used to produce estimates of the quantity, timing and location of dischargges from the Dysart-Leven Coalfield should pumping cease. MODFLOW has also been applied to the coalfield with less success. Water quality modelling was also attempted using GRAM's iron component, however, conclusive results will not be obtained until the three variables over which ther is most uncertainty have been calibrated against existing discharges.