In the ongoing debate on Auckland's transport, a rail link to Auckland Airport sits between two other major rail project proposals - the $2 billion central city tunnel loop and a North Shore link, which could cost up to $4 billion.

Auckland has long entertained the suggestion of a rail link between downtown and Auckland Airport. But it becomes clear there is a far more affordable and possibly efficient option we should be debating before saying yes to rail.

A rail link from the central business district using an inner-city tunnel from Britomart and the Onehunga line (which would need to be double-tracked) and crossing the Manukau Harbour on a route adjacent to State Highway 20 is estimated to cost at least $2 billion, assuming a viable route can be protected which doesn't add significantly further to the cost.

A link that loops to the airport from the main rail line at Manukau is alone estimated at upward of $500 million.

Besides the huge initial cost, the far bigger unresolved issue with a dedicated rail link to the airport is whether it would be used sufficiently - even if doubling as a freight rail route - to justify the likely high operating costs and subsidy such a link will require.

On some estimates a subsidy of $15 per passenger per trip would be required, which would be paid for by taxpayers.

Unchallenged evidence tabled at the Auckland Regional Council some years ago projected the passenger mode share for a CBD-Airport rail service would be, at best, about 6 per cent. This is similar to airports in Australia and the United States and makes no difference to the numbers of cars going to the airport.

The same reports suggested a much more cost effective and efficient option would be an enhanced CBD-Airport bus service but remains untested.

The report, by Australian consultants Parsons Brinkerhoff, given to the ARC in 2006 showed airports with a rail connection like Sydney and Brisbane, and many in the United States, still have upwards of 70 per cent mode share based on cars.

Even cities like Portland in the United States and Melbourne which give high priority to public transport have a car mode share of around 85 per cent, compared to Auckland's 82 per cent with the 18 per cent balance by bus.

The bus mode share at all airports in the survey was significantly higher than that for rail.

Worldwide there are two key problems with CBD rail services to airports.

The first is on services where air passengers share with commuters and find them difficult to board and unreceptive to luggage - this is a major constraint against Sydney's airport rail service.

The second issue relates to rail services where passengers have to transfer to another mode to complete their journey. In contrast, in Brisbane the airport rail service, while of marginal profitability, appears to be working well on lines where transfers are not required, including long-distance trips for tourists.

For many years Melbourne, like Auckland, debated putting a rail service to the airport. However in 2005, they undertook a study comparing the viability of an enhanced bus service with rail.

This study projected the rail mode share would be between 3 per cent to 6 per cent and be similar to an enhanced bus alternative patronage put forward at the same time - taking the 6 per cent 7 per cent at the time of the study to about 10 per cent.

The Victoria State government concluded such a small diversion to public transport was insufficient to justify the high cost of a new rail link.

Instead, Melbourne has opted for an enhanced bus service to the airport.

This includes a new fleet with specially designed low-floor buses that facilitate passenger access with luggage, and priority access to the freeway for quick access to inner Melbourne. Melbourne's airport patronage has nearly doubled since these innovations were introduced. Given other priorities for improving Auckland's transport system, a serious look at Melbourne's example with an enhanced bus service to the airport, including roading, operational and marketing improvements deserves serious attention.