North Wales is an area of glaciated mountains with deep valleys descending westwards to the Irish Sea coast. The region experiences high annual rainfall, principally from mid latitude cyclones, and as a consequence supports areas of temperate rain forest and upland peat blanket bog of international ecological importance. A number of small towns lie within the region and are frequently at risk from flooding through saturation excess overland flow into rivers.

A study of rainfall patterns across one of the major catchments of North Wales was carried out between 2002 and 2006 as part of a larger PhD hydrology study. An array of twenty-two data logging rain gauges was installed at locations from valley floors to mountain summits, augmented by three data logging weather stations at valley locations to provide humidity, wind speed and wind direction data.

Observed rainfall patterns differed with the direction of approach of frontal systems, with two patterns dominating: 1) Rainfall maxima along mountain summit ridges when mid-altitude air flows were sub-parallel to mountain orientation, 2) Rainfall maxima inland of mountains, on lower ground at the heads of valleys, when mid altitude air flows were cross-cutting to mountain ridges. The patterns of rainfall recorded in the field were closely replicated by computer models using MM5 and WRF software with a 1km grid spacing.

It is concluded that rainfall type 1 along mountain ridges is generated by a simple orographic uplift process, whilst rainfall type 2 results from seeder-feeder enhancement. Type 2 rainfall is dependant on high velocity moist airflows within the deep valleys which are forced to rise at steep glaciated valley heads. The extents of rainfall enhancement at particular locations during storm events have been estimated.

Ongoing work is using geometrical filtering within a digital terrain model of the wider region of North Wales, to identify areas of likely seeder-feeder rainfall enhancement for frontal systems with different approach vectors. Data is being compared initially to WRF models of storm events, and will subsequently be supported by field rainfall recording on a high resolution grid.