This subject has been a focus of research in Perth, Western Australia, since 1984-85. During the Perth Urban Water Balance Study (1983-87), while I was working at UWA's Centre for Water Research, it became apparent that there was no obvious way to incorporate the behaviour of shallow lakes and wetlands into regional scale groundwater flow models of the type being developed for the Swan Coastal Plain near Perth. There are hundreds of lakes and wetlands, and most are smaller than a typical cell or element in a numerical model, in those days finite elements 2 km2 in size.

Following preliminary modelling in vertical section by a final year engineering student at UWA (Oo Gin Hung in 1985), collaboration with Dr Malcolm Davidson at ANSTO led to solutions that define steady groundwater flow patterns near circular or elliptical lakes in plan, and semi-circular, semi-elliptical or infinitesimally thin lakes in vertical section. The solutions were obtained by a special kind of boundary integral method, where the solutions are expressed as a superposition of customised Green's functions that satisfy boundary conditions at all boundaries other than the boundary of the lake. The work was published by Townley and Davidson (1988).

Simon Nield started his MSc thesis at UWA in 1987. Whereas the work with Malcolm Davidson considered the effects of different aquifer fluxes upgradient and downgradient of a lake, Simon set out to include the effects of recharge, using my finite element package AQUIFEM-N in vertical section. Simon's MSc was completed in 1990. It led to identification of a number of different flow regimes that could be classified as "recharge" (where lake water recharges an aquifer over the whole lake bed), "discharge" (where groundwater discharges into a lake over the whole lake bed) or "flowthrough" (where groundwater discharges into a lake over part of the lake bed and lake water recharges the aquifer over the remainder). A brief paper by Townley et al. (1988) presented the first results.

I joined CSIRO in January 1988 and was asked to lead a small team, alongside Dr Jeffrey Turner, who was already leading a small team with expertise in field work, water chemistry and isotope analysis. Within a year, Jeff and I gained funding for a multi-year study that led to a 469-page report by Townley et al. (1993a). This report was scanned in June 2015, 22 years after completion of the Final Draft. The report is now available for download. The full report was summarised as one of a series of six reports on wetlands of the Swan Coastal Plain (Townley et al., 1993b). Townley et al. (1991) provided a brief introduction.

Dr Michael Trefry undertook most of the detailed analyses leading to extension of steady results to three dimensions. Although the numerical work was completed in about 1993, further analytical work was undertaken in 1996-97, leading to a paper by Trefry and Townley (1998). The 3D results were finally published by Townley and Trefry (2000). Supporting software for particle tracking and preparation for tessellation and rendering has been described by Trefry and Townley (1996).

Numerous field studies were undertaken to support our understanding of surface water - groundwater interaction. A paper by Turner and Townley (2006) was more than 15 years in the making. Most of the work was completed by 1997, so by publishing in 2006 we missed an opportunity to provide new results in a timely fashion. The paper combines modelling and field work carried out by a large team at CSIRO, over many years. More papers may follow, now that work completed in 1993 has been "rediscovered".

Throughout the 1990s, we worked with students and post-docs on projects aligned with our interests in surface water - groundwater interaction.

Andrew Brooker undertook an Honours thesis in Mathematics at UWA in 1990, developing a boundary integral element method (BIEM) for studying the flow of two fluids with different density through a causeway or levee. This work was motivated by studies of ponds used for solar salt production, in which different ponds contain fluids of different salinity, and there exists the possibility of bi-directional flow through a sufficiently permeable levee. Andrew joined us at CSIRO in 1990, and his work led to a paper by Brooker and Townley (1994), describing a flow process which has been observed through in the Great Salt Lake in Utah. Tony Barr also developed in-house water and solute balance software for ponds and lakes known as SALTPROD - but most modellers these days would use commercial software like GoldSim for this kind of modelling.

A study of Perry Lakes, directly across the road from CSIRO's laboratories in Floreat Park, led to another report by Townley et al. (1995); this work led to extensive field studies by John Rich as part of his PhD research through Murdoch University (John's PhD thesis was not completed until 2004). David Sim undertook an interesting study of Shenton Park Lake (now known as Lake Jualbup) for his 1995 Honours thesis in Natural Resource Management at UWA; this study demonstrated the annual cycle from a recharge regime during winter rains, to flowthrough regimes for most of the year, and possibly to a discharge regime when the lake nearly dries in late summer. Several other studies on tailings storage facilities, mine pit lakes and dredge ponds (in the mining of mineral sands) have contributed to our understanding, but these reports remain confidential to clients.

Townley's 1995 comment on a paper by John Wilson on induced infiltration from rivers into aquifers remains one of the few published "comments" in Water Resources Research (perhaps the only one) to which there was no published "reply". When asked by the editors to reply, John's response was that the comment was correct. The comment has never been cited. Further work by Tony Barr on this topic has yet to be published. It involves a detailed explanation of how FLOWTHRU can be used to explain flow patterns beneath rivers and streams when there are pumping wells nearby. A draft paper from 1995 has recently been found, with original figures, so the paper may soon be cleaned up and submitted.

In March 1994, I attended the Second International Conference on Ground Water Ecology in Atlanta. On the same trip, I gave a seminar at the US EPA's Robert S. Kerr Environmental Research Center in Ada OK, and taught a short course on surface water - groundwater interaction (using FLOWTHRU) at the USA EPA offices in Washington DC. By late 1996, shortly before the CSIRO Division of Water Resources merged with others to become CSIRO Land and Water, the Chief of the Division advised that there was "no future in surface water - groundwater interaction", and the team that had been working together from the late 1980s until the mid 1990s was disbanded. This led to my decision to leave CSIRO in 1997. Tony Smith joined CSIRO shortly afterwards, while still completing his PhD research at Murdoch University.

Tony Smith's PhD research in the late 1990s focused on computation and visualisation of dynamic groundwater flow patterns beneath shallow lakes and wetlands, in response to periodic forcing. By "periodic", we mean a combination of steady (or average) climatic conditions, and harmonic (or sinusoidal) fluctuations in climate with a particular amplitude and phase. In general, our focus has been on annual or seasonal fluctuations, though the same techniques apply to tidal or diurnal fluctuations. Dynamic animations of streaklines have been prepared by Tony and made available via a standard web browser. One paper has been finished on the influence of regional setting on steady lake-aquifer interaction (Smith and Townley, 2002). Another describes the methods by which periodic fluctations can be visualised (Smith et al.,2005).

Dr Bill Linderfelt came to CSIRO in Perth as a post-doc in 1995, joining a project that Jeff Turner and I had initiated studying interaction between shallow groundwater and saline surface water in the Swan Estuary near Perth. The Swan Estuary is fundamentally a fresh estuary, with river flows flushing the estuary during each winter (the wet season). In summer months, however, a seasonal seawater wedge propagates upgradient within the estuary. This means that the water column in parts of the river alternates seasonally between fresh and saline conditions. This has been the case for about ~100 years since a limestone bar at the mouth of the estuary was blasted, to allow larger ships to sail up the river from Fremantle to Perth. A number of conference papers were prepared and presented during the course of the research, and two excellent journal papers were later published in a special issue of Hydrological Processes. Smith and Turner (2001) describe density-dependent surface water - groundwater interaction and nutrient discharge in the Swan-Canning Estuary, adding density effects to concepts developed earlier by Nield et al (1994) and Townley and Trefry (2000), and for the first time showing how a seawater wedge develops asymmetrically in groundwater along a meandering river. Linderfelt and Turner (2001) describe the interaction between shallow groundwater, saline surface water and nutrient discharge.

Two other PhD theses were written by students at Murdoch University during this period of active research, and co-supervised at CSIRO. In 1997, Dr Jannette Nowell completed a PhD thesis entitled "Modelling biological responses to environmental variables in wetlands"; she used an early version of ExtendSim to simulate competition between typha and baumea (fringing vegetation) in response to fluctuating water levels. In 2000, Dr Alexander Alexandrov completed a PhD in mathematics, entitled "Mathematical modelling of groundwater and surface water interactions"; his focus was on fluctuating systems with sloping boundaries.