A rail link to Dublin Airport should be just the first stage of a transport masterplan where all destinations in the city are within reach with justone change of train. And each line doesn't have to cost the billionsthat officialdom has projected, argues James Nix.

The Strategic Rail Review, due to be published shortly by the Department of Transport, is likely to impact on plans to extend Dublin's rail network. What is clear is that Dublin needs to work towards a masterplan where all destinations are reachable within just one change of train, as research has shown that each additional interchange imposed on customers carries with it a minimum 40 per cent reduction in demand.

Urban rail networks where passengers never need to change more than once operate in Kiev, Milan, Munich, Oslo, Prague, Rome, Sapporo, Shanghai and Washington, among other cities.

The accompanying diagram proposes this level of network integration for Dublin. Stage one comprises the airport link. There appears to be a growing consensus that the line should be routed along the Ballymun-Broadstone corridor serving DCU, the Botanic Gardens and Glasnevin.

Aided by advances in signalling technology, this alignment would allow the through-running of Cork-Dublin-Belfast trains in the future. It is also the most direct route between the city centre and the airport, offering passengers a journey time of just 12 minutes - even with stops en route.

The second project would be a rail interconnector between Heuston Station and Spencer Dock. It would run via MediaLab Europe in the Liberties, St Stephen's Green and Pearse Station (Westland Row), linking up with the existing DART line.

A later phase would route the Howth/Malahide DART line under both Connolly Station and Eden Quay, allow for intersection with the airport line at Dame Street, and continue the line to Tallaght.

Likewise, the southern leg of the coastal DART line could be put in a tunnel under Pearse Station, routed northwest via Eden Quay, the Ilac Centre, Glasnevin and on to Blanchardstown.

Dublin would then have a basic three-line network complemented by a distributor ring.

Turning to the Dublin Airport link, attention has shifted to securing the most cost-effective route and train specification and, in this context, it is worth looking at the Luas project for pointers. Luas consists of a total of 24 kilometres of track, all of it overground, at an anticipated cost of at least €800m, or €33m a kilometre.

Compared with light rail projects in Sheffield or Montpellier, the cost of putting trams back on Dublin's streets is high.

Michael Sheedy, the Luas project manager, candidly acknowledges that the diversion of electricity, phone, gas and sewer lines has taught rail-builders "an expensive lesson".

Put bluntly, taxpayers are footing a bill for work that has proved unnecessary in cities where a more organised approach to utility installation prevailed.

Against this backdrop, it must be asked whether tunnelling would be a cheaper option for future lines.

The most cost-efficient rail tunnelling projects of late are two extensions to Madrid's metro. The first project, comprising 56 kilometres built between 1995 and 1999, cost €28m a kilometre. The second scheme involves 55 kilometres of track and is on target to cost €46m a kilometre.

Despite their length, each scheme has been delivered within an extremely impressive time-frame of 40 months. Transport analysts such as Mike Smith, editor of online journal www.tunnelbuilder.com, attribute Madrid's success to a highly experienced project management team and acute cost control.

Dublin Airport to St Stephen's Green measures 11.5 kilometres via the route shown on the diagram. Approximately 65 per cent of the route is in tunnel.

Assuming that the construction conditions of Madrid can be replicated here - and there is no reason why not - a rail link to Dublin airport should cost €530 million in today's prices. (This compares with the €4.8 billion reported estimate arrived at by the Dublin Transportation Office and the Railway Procurement Agency for the airport line. However, the DTO and RPA have yet to detail the benchmark projects on which their estimates are based.)

Madrid's low-cost structure is attributable to a number of considerations which must be examined in an Irish context.

The first concerns train specification. In Dublin's case, this is essentially a choice between DART and Luas. Luas involves low-floor vehicles requiring more complex assembly and this pushes cost up. A DART carriage, on the other hand, is easier to build. The capital cost of a DART train carrying 1,500 passengers in 2000/2001 was €11.5m, compared with €12.8m to buy Luas rolling stock for the same number of passengers over the same period.

DART rolling stock is similar to that used in Madrid. Engineers in Madrid found that two parallel trains (both 2.8 metres wide) could be accommodated in a tunnel with an internal diameter of 8.43 metres. At 2.9 metres in width, DART vehicles will require a tunnel just marginally larger than that used in Madrid.

A further advantage of using DART is that Irish-based engineers and planners will acquire expertise for subsequent projects.

Using DART for the first rail tunnelling project means that identical boring machines, tunnel lining and trains can be employed for an underground link between Heuston Station and Spencer Dock. Adopting a standard tunnel profile and sticking to it accords very much with experience in Madrid.

Consideration must also be given to a railway's operating system. Automated trains are not used on the Madrid projects because the extensions feed into a highly developed city-centre network that is manually operated. Where new lines are constructed, automation is preferred.

Automated lines operate in Copenhagen, Lille, London Docklands, Lyon, Paris, Singapore, Toulouse, and Vancouver. The US cities of Chicago, Detroit and Jacksonville also use automation.

The driverless Meteor line in Paris (Line 14) has been such a success that automation of the entire Parisian metro is under consideration. Berlin is examining a similar strategy.

Automation does not mean unattended trains. Instead, "metro stewards" assume a wide-ranging role.

As well as monitoring train operation, they assist elderly customers, help people with directions, and see off potential anti-social behaviour. Indeed, research undertaken for the Copenhagen metro suggests that the very presence of stewards boosts patronage.

The use of automation on the Dublin Airport line would increase commissioning costs. However, not to adopt automation would be retrograde - especially as the airport is forecast to cater for 30 million passengers per annum by 2015.

Journey demand between the airport and city centre is set to be of a higher order than usership to and from Sandyford.

In time, however, Luas Line B - between St Stephen's Green and Sandyford - can be adapted and integrated into the airport line as population densities along the line increase in line with predictions.

This is precisely what happened with Line 10 in Madrid where a low capacity route has been reconfigured for wider trains. It is also what took place in 1840s' Ireland when the track of what is now the southern leg of the DART line was widened from 1.435 metres to 1.6 metres.

The capital cost of the network proposed here is no doubt significant, but project-by-project phasing would make it affordable.

Global experience shows that world-class transportation systems can be provided at reasonable cost.

By drawing on know-how from abroad, Dublin can have a world-class rail system in the years to come.

James Nix is pursuing an MPhil in the design of transport networks in the Dublin Institute of Technology and is a Barrister-at-Law degree at the King's Inns.