RESEARCH PATHWAY: personal reflections on a career in research

Geographer and climatologist Gerald Mills (University College Dublin) reflects on a long research career investigating urban climate. He considers how the field has evolved from measuring and modelling to understanding the influence of the local context (local buildings and urban context) and its impact on indoor temperatures in adjoining buildings and outdoor conditions. A key aspect is linking urban climate knowledge to building design and urban planning.

As a physical geographer and climatologist with a focus on cities and their climate effects, at all scales, I see myself as part of a geographical tradition that is concerned with humans and their environment, which for most of humanity is the city. Urban climatology is (and has been) an obvious area of research for a physical geographer. Moreover, in contrast to other areas of climate, the urban climate effect is attributable to human decisions on the urban form and functions of cities. Urban form describes the physical make-up of the city, the extent of the paved landscape, the materials used to construct roads and buildings and the three-dimensional layout. Urban functions describe the flux of resources that are needed to sustain the population and their activities. While form alters the natural exchanges of energy, momentum and mass, function generates wastes (heat, materials, etc.) that are mixed into the overlying air. The net result is that the urban atmosphere is profoundly modified at a hierarchy of scales and cities contribute significantly to regional and global climate changes. My work contributes to the creation of an Urban Climate Science (UCS) that integrates across scales (building to planet) and across disciplines.

My research on urban climate spans 40 years (since the mid 1980’s) and has evolved from an initial focus on modelling and measuring the energy exchanges in a city street (an urban canyon). My work now involves describing urban landscapes for modelling, linking urban climate knowledge to design and planning, and developing global urban databases to support climate research. This span has coincided with the transformation of urban climatology from a mostly descriptive to analytic science, the evolution of numerical models and of observation systems suited to urban studies, and the integration of new technologies (e.g. satellites) and data sources (e.g. low-cost sensor networks). This period has also seen the expansion of the number and diversity of disciplines studying urban climates to include branches of urban planning/design, biometeorology, engineering, etc.

In the mid-1980’s, urban climatology was a relatively small field and was dominated by few in geography and meteorology. The dominant topic of study was the urban heat island (UHI) and the common methodology was measurement and comparison of air temperature within and outside the city. Much of this work could be described as descriptive, case-study based and focussed on hypothesis generation. I entered this field as a novice, but it was clear (from reading scientific articles) that the future of the field had to shift toward an understanding of processes (energy, mass, and momentum exchanges) rather than responses (temperature, wind, etc.) and that this would require skills in computer programming and modelling. I had completed an undergraduate and research master’s in Geography at University College Dublin (UCD), which, at the time, was a small department that supported PhD by substantial thesis with no coursework. I was sufficiently aware to know that I could not engage in modern climate research in this context and this led to my decision to study at a US institution (Ohio State University), which had a structured programme that included the opportunity to take diverse courses across disciplinary boundaries, which was not possible in my home university.

Critically, this coursework gave me the scientific foundation time to formulate a PhD research proposal on understanding the energy exchanges within a street (an urban canyon) and between the street and the overlying air. This research adopted the approach of parsing the complex urban landscape into simplified repeatable forms, such as urban canyons. The doctoral work created a numerical model and evaluated it with observations gathered at a selected street site. The thinking was that by understanding the processes in these simple environments, we could extrapolate the results to the entire landscape. Unlike modern PhDs that are often funded by established research projects, I had the freedom to create my own project. This meant that I had to critically read the literature and identify the ‘gaps’ that my work could then address.¹ For many researchers this gap has already been identified and their role is to clarify the literature and then implement the project according to predetermined deadlines, often linked to specific papers that will be published. Although I see the benefits of this approach, I am grateful for the freedom I had to develop my own research area.

The type of doctoral training that I received has informed a great deal of my subsequent work. For a start, it meant that I was knowledgeable about how numerical models work and the types of decisions/compromises that the modeller must make. Moreover, having direct experience of making observations and data processing made me aware of issues concerning the exposure of instruments and calibration. I would encourage a starting researcher to acquire contemporary skills that are similar, such as Python programming or Machine Learning. As we leave an era of data paucity for one that is data rich, the challenge will be to extract useful information.

A common refrain among urban climatologists has been how little the knowledge acquired has been used by urban planners/designers and architects. In the years following my PhD, I did a great amount of reading outside Geography (and climatology). At that stage in my academic career, I was employed on a part-time basis to teach at an American university (UCLA) and, as I was not part of the tenure system, I was able to pursue my urban interests with little interference. UCLA had a strong architecture and urban design programme that had produced seminal work in the 1970’s on passive urban design and housed a superb library of materials. My reading in this field showed that climatologists and designers focussed on different urban climates. While the former concentrated on solar energy and daytime heating, much of the latter focussed on the near-surface air temperature effect (the UHI) and night-time cooling. Building layout played a role in each by regulating solar access (energy gain) and sky view (energy loss). The work of Knowles (1981) was insightful as it used the concept of a solar envelope to generate urban layout that ensured access to the sun. It seemed to me that urban climate and passive architecture looked at the same object (the city) from two vantage points, one associated with daytime solar gain and another associated with night-time cooling. This thinking directed me toward a different type of numerical model using cube-shapes to represent the urban landscape. This model allowed me to change urban layout by modifying the street width, thereby changing solar access and sky access. To evaluate this model, I used techniques that were developed in architecture, specifically a scaled-physical model to examine the role of neighbouring buildings on the indoor temperatures of these simple cube-shaped buildings. I think that this work arose from reading broadly and not confining myself to the climate literature. I always encourage students to read outside the specific area of interest as the gaps are often between, rather within disciplines. I include these ‘classic’ architecture readings on my physical geography classes as I think that they are still relevant to modern concerns for sustainable and low-carbon cities.

Given the development of the field before 2000, I was fortunate that the literature available to me was relatively small, and it was possible to harvest ideas from many disciplines. (Many of these papers are difficult to find and, even in university libraries, are not available for free.) Some of these papers are exceptional in terms of their articulation of ideas. Moreover, as many were based on experimental work that was difficult to do, they have not really been repeated. It is a great pity that much of the earlier literature doesn’t get cited, as it is still valid. One of the advantages of the modern citation systems is that it allows searches for ‘forward’ citations. These citation systems are extraordinary at harvesting articles on any given topic and have contributed to a new genre of review, the meta-analysis of a body of literature on a topic. However, they contribute to a form of over-referencing in articles that do not really help the reader to identify the core ideas upon which the work is built.

Sometimes, more research does not shed more insight! As an example, consider the enormous amount of published material on the UHI, which is now commonly used as a metric for urban policies that seek to reduce the urban effect. This literature can be divided into two phases. The first phase was based on fixed and mobile weather stations to record the air temperature variation across the urbanised landscape, using a base station located in a rural area as a benchmark to evaluate the magnitude of the canopy-level UHI. This type of work dominated the field from 1950 to 2000 and we have a very good understanding of the causes of the effect. What gaps remained were largely to do spatial coverage as much of the work had been completed for mid-latitude, high-income cities. In other words, apart from educational exercises, there is usually little to be learnt from a new canopy-level UHI study. The second phase, which started in the 1990s consists mostly of satellite-derived surface UHI work. Satellites can provide a synoptic view of the UHI, but it can only provide it at specific times (overpass) when the surface is visible. Moreover, it only ‘sees’ some of the urban surface, that which is dominated by rooftops and the ground. While the canopy-level UHI is a night-time phenomenon, the surface UHI is both a day and night phenomenon, although the locations of maximum surface and canopy UHIs do not always correspond. These two types of UHI are not well integrated in the literature and, generally, too little attention is paid to the limitations of each. In fact, sometimes they are confused leading to incoherent policy recommendations. Few papers refer to Lowry (1977), who provided a methodological framework for isolating urban climate effects. Moreover, not enough consideration is given to the impact of the UHI, which is not the same for all cities. I am reminded of the closing paragraph of Chandler’s (1965) book on the Climate of London:

Londoners live in a profoundly man-modified climate. A few of the changes wrought by the widespread substitution of houses and factories for fields and woods, and surfaced roads for cart-track, might be considered favourable. Such are the higher autumn, winter and spring night-time temperatures which reduce heating costs and lengthen the frost-free period, but these advantages must surely be outweighed by increased pollution and decreased sunshine.

It is advantageous to reflect on the indoor/outdoor divide in climate studies. One of the most rewarding research topics for me is how the indoor climates of buildings affect (and are affected by) neighbouring buildings in a dense urban context. As a climatologist, living in a moderately cool climate with a dominant heating demand, I was surprised to learn of building cooling demands in cities such as London. I became aware of this only in the last decade when I was introduced urban building energy models that showed how high internal energy demands in well insulated buildings can create daytime cooling demands, even in winter. This led me to think a great deal more about the indoor-outdoor interactions and to think about how building information can be incorporated more explicitly into urban climate research. This will require the relevant specialist disciplines to develop a common language to permit effective communication and data sharing. For example, the current definition of the urban canopy layer (UCL) used by climatologists is defined as the outdoor space at and below the mean roof level; this excludes part of the building envelope (the roof) and the indoor air space. A consequence of this is that the information on building envelopes (e.g. the material layers and ventilation factors), and functions (e.g. heating and cooling systems and occupation patterns) are not accounted for in most urban climate modelling. Similarly, most building energy models do not account for the urban effect on the typical weather files in assessing the heating/cooling needs. There is now the potential for multi-scale modelling to support ‘hyper-local’ simulations. At that point it will be possible to connect building/street scale models to larger scale weather/climate models. This work must be supported by an integrated research framework with shared concepts, language and data infrastructure to make urban climate science a reality.

I would encourage all early career researchers to join a learned society and establish their personal networks. My ‘awakening’ to many of these research opportunities is really the outcome of conversations with those outside of a narrowly conceived urban climatology. I have been very fortunate to be a member of the International Association for Urban Climate (IAUC) since its foundation in 2000, which was established to provide a forum for multi-disciplinary study on cities. This has provided me with much of the intellectual stimulus for the work that I do. From my experience, sustaining a research profile in an environment where you are the only one focussed on a topic is challenging as there is a temptation to shift your attention to research where you are more likely to be rewarded. One of the worst pieces of advice I was given after my PhD was to re-focus my work on global climate change. Fortunately, I was mature enough to recognise that I could contribute little new to that field of research at that time, whereas I felt I could make significant contributions to urban climatology.

My final advice to a starting researcher today is to develop a core area of research that you think provides a fertile space for intellectual development and don’t be dissuaded from pursuing it.

Note

1. For many researchers this gap has already been identified and their role is to clarify the literature and then implement the project according to predetermined deadlines, often linked to specific papers that will be published.

References

Chandler, T.J. (1965). The Climate of London. Hutchinson.

Knowles, R. L. (1981). Sun, Rhythm, Form. MIT Press

Lowry, W.P. (1977). Empirical estimation of urban effects on climate: a problem analysis. Journal of Applied Meteorology and Climatology, 16(2), pp.129-135.

Oke, T.R., Mills, G., Christen, A. and Voogt, J.A. (2017). Urban Climates. Cambridge University Press.

Stewart, I.D. and Mills, G. (2021). The Urban Heat Island. Elsevier.

Share: