Effects of old landfills on groundwater quality. Phase 2,udinvestigation of the Thriplow landfill 1996–1997

摘要

Disused sand and gravel excavations overlying the major Chalk aquifer at Thriplow in Cambridgeshire haveudbeen filled with domestic waste in two phases. One area (Phase 1) was filled between 1957–77 with little compactionudof the refuse and was left uncapped, while Phase 2 was deposited between 1981–87 and capped withudclay. Aerial photography and surface resistivity surveys indicate that the site geometry is complex, with severaludphases of landfilling into excavations of differing depths. Drilling through the waste indicates that leachate productionudand waste stabilisation proceed at different rates in capped and uncapped landfills. Analysis of leachateudobtained by centrifugation or squeezing appears to give more insight into the pollution potential than do leachudtests with distilled water. The Biological Methane Potential (BMP) of the waste appears to be related to theudquantity of decomposable material but the chemical oxygen demand (COD) values are distorted by the presenceudof reduced metals. Too few boreholes have been drilled to define the leachate source in terms of its spatial distributionudand little is known of how its composition has changed with time. However, hydraulic conductivity measurementsudon the landfill caps suggest that it is sufficiently permeable for all rainfall to potentially infiltrate theudwaste.udBoreholes outside the landfill penetrate the Upper and Lower Chalk, and identify the Melbourn Rock and underlyingudPlenus Marls at the junction of the two formations about 20 m below ground level (bgl). Surface resistivityudsurveys using the BGS RESCAN system, confirm aerial photographs of the extent of the landfill and alsoudsuggest that leachate has migrated beyond the base of the landfill. Evidence of leachate migration in pre-existingudscreened boreholes completed above and below the Plenus Marls suggests that leachate is flowing above theudPlenus Marls. Hydraulic head measurements whilst drilling a borehole to the base of the lower Chalk approx. 70udm bgl revealed the potential for upward groundwater flow through the Plenus Marls. Thus, previously-drilledudboreholes penetrating the Plenus Marls are expected to recharge upwards into the shallow aquifer above theudPlenus Marls diluting any leachate in the upper aquifer and distorting the flow regime. Several of these boreholesudhave subsequently been modified to stem the flow across the Plenus Marls.udOne borehole down-gradient to the west of the site revealed a large thickness of drift composed of both sand andudclay rich material. This suggests the existence of a buried channel, the hydrogeological significance of which hasudyet to be assessed.udGroundwater chemistry appears to be influenced by three major factors. (a) the landfill leachate (b) the compositionudof shallow groundwater in the top 10 m of the Chalk, and (c) the composition of water from the LowerudChalk. Limited groundwater monitoring data appear to display a cyclic variation in chloride concentration. Theudorigin for this is not clear but it may correlate with cyclic variations in groundwater levels when the water tableudrises into the waste. Cyclic flushing of the landfill may release leachate into the aquifer giving rise to pulses ofudchloride. Alternatively changes in chloride may arise by the changing direction of groundwater flow which as yetudhas not been assessed.udA conceptual hydrogeological model in which flow is limited to above the Plenus Marls has been used touddevelop a more appropriate groundwater flow and solute transport model. However, the model lacks data onudaquifer properties, on contaminant inputs concentrations, fluxes and spatial variations, and there is a paucity ofudmonitoring data for calibration. Nonetheless preliminary transport modelling using an equivalent porous mediumudapproach shows that an effective porosity of about 5% best fits the regional data. Since this is much less than theudtotal porosity of about 40% for the Chalk, it would appear that only part of the Chalk is available for flow butudthat matrix diffusion could play an important role in leachate attenuation. Discrete fracture modelling using theudFRACTRAN code has allowed some scoping to be made of the hydraulic properties of the aquifer by comparisonudwith chloride hydrographs, but these again need to be better conditioned by in-situ measurement of fractureuddistributions and transmissivities.udA number of additional activities are required to improve the understanding of flow and contaminant transport atudthe site. These include better spatial definition of the waste distribution, improved data on the hydraulic propertiesudof the Chalk aquifer, and the use of automatic monitoring to record temporal changes in groundwater chemistryudand groundwater levels.