My projects to date have been research-focussed, requiring novel approaches to reach the desired project goal. Much of this effort is centered on my long-standing aim of developing appropriate methodology for spatial modelling from line transect data, an ambitious project on which I embarked for my PhD studies in 1995. The methods developed in my PhD were subsequently published in the Journal of Agricultural,  Biological and Environmental Statistics (Hedley and Buckland, 2004) ; this was awarded Best Paper in 2004-05. The ‘count method’ described in that paper is also now incorporated in the industry-standard software Distance 6 .

Despite these successes, I have never been totally satisfied with the methods, largely because of outstanding model selection issues, difficulties estimating variance, and smoother-related problems. To this end, I have been working on improvements to these methods with Mark Bravington from CSIRO, Hobart, using a particularly challenging dataset from over 20 years of International Whaling Commission surveys of minke whales in the Antarctic. (See Bravington and Hedley, 2009, for an up-to-date report on these methods and their application.)

We have gratefully enlisted the help of Simon Wood, author of the excellent R-package ‘mgcv’ for fitting generalized additive (mixed) models, and via a new ‘soap film smoother’ (Wood et al., 2008) , have made significant progress in tackling several of the undesirable behaviours of other smoothing functions, particularly at the edges of survey areas and/or in areas of intricate coastlines.