Sir - Dr Harbour's letter (August 31) raises serious issues of concern about deficiencies in the flood models used by the Environment Agency, and others, to assess flood risk and predict flooding. However the story of the impact of RWE npower's ash dump at "Lake H/I" at Radley on flooding does not end there, and has a disconcerting final twist. This "lake" may be missing from the models, but it was not overlooked when it came to assessing the flood risk associated with the proposed filling of Thrupp Lake.

"Lake H/I" is not a lake, but a massive solid pile of clay and ash sitting on the Thames floodplain just upstream of Abingdon. It was supposedly constructed, in 2002, according to Government planning policy guidelines (PPG25) which required that any development in the active floodplain should not impede flood water. This is the reason for the famous flood gates that are supposed to allow water to flow into the interior of the structure during floods.

As part of their planning application in 2006, to fill Thrupp Lake, npower commissioned Jacobs, its prime contractor, to carry out a flood risk assessment. This determined that the proposal did not increase flood risk, as Thrupp Lake and its companion lake, the Bullfield, are not part of the active floodplain, as defined by the model. Save Radley Lakes challenged this, pointing out several anomalies and inconsistencies in the modelling. In the end, the Environment Agency, who oversee the process, had to agree that the proposal did not constitute a flood risk, only because, in their opinion, the raised profile of "Lake H/I" both now, and post "restoration", would prevent flood water from the river from ever reaching the two lakes. If this is not an admission that H/I impedes flood water, I don't know what is.

Dr Basil Crowley, Chairman, Save Radley Lakes, Lower Radley