Characterization of European cities’ climate shift – an exploratory study based on climate analogues

Guillaume Rohat (Institute for Environmental Sciences, University of Geneva, Geneva, Switzerland and Faculty of Geo-Information Science and Earth Observation, University of Twente, Enschede, The Netherlands)
Stéphane Goyette (Institute for Environmental Sciences, University of Geneva, Geneva, Switzerland)
Johannes Flacke (Faculty of Geo-Information Science and Earth Observation, University of Twente, Enschede, The Netherlands)

International Journal of Climate Change Strategies and Management

ISSN: 1756-8692

Publication date: 21 May 2018

Abstract

Purpose

Climate analogues have been extensively used in ecological studies to assess the shift of ecoregions due to climate change and the associated impacts on species survival and displacement, but they have hardly been applied to urban areas and their climate shift. This paper aims to use climate analogues to characterize the climate shift of cities and to explore its implications as well as potential applications of this approach.

Design/methodology/approach

The authors propose a methodology to match the current climate of cities with the future climate of other locations and to characterize cities’ climate shift velocity. Employing a sample of 90 European cities, the authors demonstrate the applicability of this method and characterize their climate shift from 1951 to 2100.

Findings

Results show that cities’ climate shift follows rather strictly north-to-south transects over the European continent and that the average southward velocity is expected to double throughout the twenty-first century. These rapid shifts will have direct implications for urban infrastructure, risk management and public health services.

Originality/value

These findings appear to be potentially useful for raising awareness of stakeholders and urban dwellers about the pace, magnitude and dynamics of climate change, supporting identification of the future climate impacts and vulnerabilities and implementation of readily available adaptation options, and strengthening cities’ cooperation within climate-related networks.

Keywords

Citation

Rohat, G., Goyette, S. and Flacke, J. (2018), "Characterization of European cities’ climate shift – an exploratory study based on climate analogues", International Journal of Climate Change Strategies and Management, Vol. 10 No. 3, pp. 428-452. https://doi.org/10.1108/IJCCSM-05-2017-0108

Download as .RIS

Publisher

:

Emerald Publishing Limited

Copyright © 2018, Guillaume Rohat, Stéphane Goyette and Johannes Flacke.

License

Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial & non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode


1. Introduction

It is by now widely acknowledged that climate change will pose significant threats to both urban systems and city dwellers (Bulkeley, 2013). Because urban areas hold more than half of the world’s population and most of people’s assets, it is of utmost importance to define adequate adaptation strategies (Lee and Lee, 2016). Their strict implementation at the urban level is supposed to significantly reduce the inhabitants’ vulnerability to climate change and ensure the quality of life for future generations. Nevertheless, despite an overwhelming scientific evidence of increasing climatic threats, urban adaptation strategies are more often absent than present, even in countries of the global North. Although a certain number of cities self-reported to be actively engaged in climate adaptation and mitigation at the local scale (Aylett, 2015), Reckien et al. (2013) found that 72 per cent of 200 European major cities have not yet implemented a climate adaptation plan. Such lack of political commitment is explained by numerous factors (Juhola, 2016), including insufficient funding, time-scale mismatches between political mandate and climate change (Bicknell et al., 2009; Hallegatte, 2009), underlying uncertainties of climate projections (Schneider, 2006) and misunderstanding of the forthcoming climate impacts (Van der Linden et al., 2014). Moreover, among the great number of factors identified as drivers of urban adaptation planning (Reckien et al., 2015), efficient and easy-to-understand scientific information and knowledge (Archie et al., 2014; Mycoo, 2015), involvement in climate-related cities’ networks and strong community engagement (Bulkeley et al., 2011) are identified to play an important role. Consequently, there is a growing need of new and innovative methods that: (i) raise urban residents and stakeholders’ awareness about the potential impacts of climate change; (ii) provide easily understandable scientific information about the future impacts and adequate adaptation options; (iii) foster cities’ collaboration within climate-related networks.

The climate analogues approach has the potential to address this need. This method – also known as the “climate twins approach” (Ungar et al., 2011) – is designed to match the future (or past) climate of a given location with the current climate of another location. This way, a pair of climate analogues is made of two different geographical locations sharing a significantly similar climate for a different time period. Such approach has been initially developed in the field of ecological studies, with the purpose of investigating climate change impacts on the shift of ecological communities and species habitat and the appearance of novel climate and ecoregions (Saxon et al., 2005; Peacock and Worner, 2006; Williams and Jackson, 2007, Veloz et al., 2012a, 2012b), as well as the implications of such shift for species’ survival and abundance (Anderson et al., 2013; Leibing et al., 2013). Climate analogues have also been used in agricultural studies to identify potential cultivars better suited to future climatic conditions (Webb et al., 2013) and to investigate adaptation solutions existing today, based on the assumption that the future of one farmer is similar to the present of another one, located in a different region (Ramirez-Villegas et al., 2011).

This approach has also shown a great potential for raising awareness about the magnitude and pace of climate change. For instance, Ungar et al. (2011), CSIRO-Bureau of Meteorology (2016) and Rohat et al. (2016) developed user-friendly climate analogue tools which provide an intuitive visualization of potential climate change impacts. In the same line, Kopf et al. (2008) and Climate Communication (2014) used climate analogues to communicate about the amplitude of climate change to a lay audience, whereas Beniston (2013) matched the past and current climates to provide easy-to-understand information about the celerity of climate change in the past decades.

However, the application of this approach in urban areas has largely been underused so far. The few climate analogues studies focusing on cities have shown that climate analogues can help assessing economic damages of climate change (Hallegatte et al., 2007) and identifying both adequate adaptation policies (Kellett et al., 2011) and best practices of climate adaptation (Rohat et al., 2016). Nevertheless, none of these studies used climate analogues to characterize the velocity of cities’ future climate shift – i.e. the speed and orientation of the geographical displacement over time – and to explore its potential implications on urban dwellers and on the design of adaptation strategies.

In this interdisciplinary effort, we propose a climate-matching method that reliably matches the current and future climates of any location worldwide, and we show how it can be used to assess the associated shift velocity. Employing a large sample – 90 different cities – we exemplify the applicability of this method and characterize the climate shift of European cities from 1951 to 2100. We then discuss the potential implications of such cities’ climate shift and provide insights into the possible use of the proposed approach, e.g. for raising awareness of both city dwellers and decision-makers about the pace, magnitude, and dynamics of climate change, for supporting the identification and implementation of adequate adaptation strategies, and for enhancing cities’ cooperation within transnational climate-related networks.

2. Methods and materials

2.1 Climate-matching approach

In the past few years, two main methods to match one climate with another have been described. One is based on the aggregation of different climate statistics within a similarity index – e.g. the CCAFS index (Climate Change, Agriculture and Food Security; Ramirez-Villegas et al., 2011; Leibing et al., 2013) or a simpler index using the standardized Euclidean distances (SEDs) (Williams and Jackson, 2007; Veloz et al., 2012a) – whereas the other is based on a comparison between a set of univariate climatic criteria and a set of arbitrarily established thresholds (Hallegatte et al., 2007; Ungar et al., 2011; Rohat et al., 2016). While the latter allows an easy control of the climate analogues’ quality – in terms of climatic proximity – the use of a similarity index allows ranking them and hence identifying the climatically closest one. Nevertheless, Grenier et al. (2013) showed that the uncertainty associated with the choice of climate models and scenarios is largely superior to the variation resulting from the use of different climate-matching approaches.

In this study, we applied a combination of the two foregoing methods to:

  1. match the climate of any location of interest (LOI) with other locations sharing similar climate – but at a different time period – which we named the LOI’s climate analogues; and

  2. determine the best climate analogue – i.e. the one sharing the most similar climate – for a given LOI and time period.

Although climate has been traditionally characterized by a specific combination of various variables (IPCC, 2001), matching one climate with another requires relaxing this definition. Because the climate-matching method developed in this study is used to investigate cities’ climate shift, we took into account climate variables that both represent the overall climate and have a major influence on the functioning of urban areas. Trade-offs have to be made between including the numerous climatic variables that determine a city’s climate and keeping low the number of climatic variables to identify a substantial number of climate analogues. This led us to select the five following quantities: monthly mean temperature and monthly mean precipitation, which are the two most essential climatic determinants (Holdridge, 1947); monthly minimal temperature for winter months (December, January and February) and monthly maximal temperature for summer months (June, July and August), which are, respectively, the indicators of cold and warm spells (Ungar et al., 2011); and annual total precipitation, which is an important climatic factor for water management in cities (Hallegatte et al., 2007). These variables were computed monthly (or annually in case of the annual total precipitation variable) and averaged over five 30-year periods, namely, P1 (1951-1980), P2 (1981-2010), P3 (2011-2040), P4 (2041-2070) and P5 (2071-2100).

To identify the climate analogues of a given LOI and time period, we first computed and averaged (as per grid points in the computational domain) the Euclidean distances between the LOI’s current climate (P1) and the future climate (P2, P3, P4 or P5) of all the grid points, for the five climatic variables (methodology available as Appendix). Second, we compared the averaged Euclidean distances (five per grid points) with specific thresholds. We arbitrarily fixed these thresholds at 1°C for the three temperature variables and 25 per cent of the LOI’s mean value (over the reference time period) for the two precipitation variables. If the averaged Euclidean distances for the five climate variables are under their respective thresholds, then the grid point’s future climate is considered as similar to the current LOI’s climate. Third, we applied an altitude filter to select the grid points that are located within a 200-meter range (above or below) of the LOI’s altitude. Although applying such altitude filter is uncommon in climate analogues studies (Hallegatte et al., 2007; Beniston, 2014), we argue here that it enables a more precise computation of the velocity of latitudinal climate shifts (Section 2.4). The remaining grid points – i.e. those which share significantly similar climate to the LOI and which have passed through the altitude filter – are considered as the LOI’s climate analogues (for a given future time period). Finally, we computed their similarity index based on an unweighted SED metric (Appendix) and ranked those to identify the best one, in terms of climatic proximity. Such workflow (Figure 1) is repeated for every LOI and for each of the four 30-year time periods (i.e. P2, P3, P4 and P5).

2.2 Climatic data

Data sets for the case study presented in this paper were extracted from the European project ENSEMBLES (2009), which provides daily values at a horizontal grid-spacing of 25 km, from 1951 to 2100, under the A1B scenario of Special Report on Emission Scenarios of the Intergovernmental Panel on Climate Change. To reduce the uncertainties associated with the use of a single regional climate model (RCM), we computed multimodel means of the five climatic variables used in the climate-matching method. Climatic projections originated from seven different RCMs, namely, CNRM-RM4.5 (CNRM, 2008), KNMI-RACMO2 (van Meijgaard et al., 2008), OURANOS-MRCC4.2.1 (Plummer et al., 2006), SMHI-RCA3 (Kjellström et al., 2005), DMI-HIRHAM5 (Christensen et al., 2007), GKSS-CCLM4.8 (Böhm et al., 2006) and METEO-HC-HadRMQ0 (Collins et al., 2006). The five climatic variables were computed monthly (and annually for the variable of annual total precipitation) for the five 30-year periods and for all the grid points (32,300 in total) of the 25-km grid-spacing computational domain.

2.3 Transects

According to studies applying the Köppen climate classification in Europe (de Castro et al., 2007; Gerstengarbe and Werner, 2008), the European historical climate is represented by a temperate climate in Western Europe, a continental climate in Eastern Europe and a subtropical climate in the Southern part. Jylhä et al. (2010) recently showed that European climates tend to move northeastwards. In this study, there is no attempt to assess the shift of European climatic zones, but rather the positional shift – mainly southwards – of European cities’ climate. Beniston (2013) showed that European isotherms have been moving northwards in the past decades, along several north-to-south transects. In the same line and following the existing studies assessing the European climate shift (Jylhä et al., 2010; Beniston, 2013, 2014, 2015), we developed three north-to-south transects, namely an Eastern Europe transect, a Continental transect, and a Maritime transect (Figure 2). These allow investigating the potential differences of climate shift over the European continent. Each of these is made of 30 different cities that have been chosen with regard to both their geographical location (proximity with a selected transect and distance with other cities) and regional importance (size of the population and administrative role). Overall, these 90 cities (Appendix Table AI) are located across 22 European countries, are distributed within several climatic zones, and host approximately 416 million inhabitants, i.e. more than half of the European population (Eurostat, 2012).

2.4 Southward velocity

To assess the southward climate shift velocity – i.e. the speed (in kilometres per year) of the expected southward positional change – of each city of the three transects, we first computed the latitudinal distance between the city and its best climate analogue (for each 30-year time period), using the Haversine formula (Sinnott, 1984). We then divided the latitudinal distance by the number of years between the reference period and the projected period, which can vary from 30 years up to 120 years. Applying this method, we estimated the southward velocity of every city’s climate shift for the seven following shifts:

  1. P1-P2: From 1951-1980 to 1981-2010 (30-year shift).

  2. P2-P3: From 1981-2010 to 2011-2040 (30-year shift).

  3. P3-P4: From 2011-2040 to 2041-2070 (30-year shift).

  4. P4-P5: From 2041-2070 to 2071-2100 (30-year shift).

  5. P1-P3: From 1951-1980 to 2011-2040 (60-year shift).

  6. P3-P5: From 2011-2040 to 2071-2100 (60-year shift).

  7. P1-P5: From 1951-1980 to 2071-2100 (120-year shift).

3. Results

3.1 Applicability of the method

Out of the 360 different attempts (90 cities and four future 30-year time periods) to identify a climate analogue, 304 were successful (success rate of 84 per cent), highlighting the applicability of this method over the European continent. Among the 90 investigated cities, 70 cities were found to have reliable climate analogues for each of the four 30-year future time periods. For the other 20 cities, no climate analogue was found for at least one future time period. Among them, two cities, namely Geneva (Switzerland) and Sofia (Bulgaria), did not have any climate analogues for the four future time periods. Most of these 20 cities are located at the edge of the European domain; hence their respective climate analogues are presumably located outside Europe. For instance, climate analogues of the cities located in the Iberian Peninsula (extreme south of the computational grid), e.g. Vigo (Spain), Faro (Portugal), and Porto (Portugal), are presumably located in North Africa. However, for other cities such as Geneva (Switzerland), no climate analogue was found simply because its future climate does not currently exist in Europe. This may be because of the appearance of novel climates in a changing climate context (Williams and Jackson, 2007).

3.2 Direction of shifts

Results showed that climate analogues are always located southwards of their respective city of reference. This rather expected result highlights the well-known equator-ward displacement of the climatic zones (i.e. from north to south in case of Europe). One of the added values of the method as applied here lies in the identification of transects of climate shift. To assess whether or not European cities’ climate shift follows the three predetermined transects, we computed the longitudinal distance between each climate analogue and their reference transect (Maritime, Continental, or Eastern Europe). Results showed that the longitudinal distance between climate analogues and their reference transect ranges from 0 to 218 km (68 km in average) for climate analogues of the Continental transect, from 0 to 715 km (88 km in average) for the ones of the Maritime transect, and from 0 to 428 km (average of 75 km) for the ones of the Eastern Europe transect. In addition to this great spatial proximity – in terms of longitudinal distance – between climate analogues and their reference transect, spatial analysis of the results showed that these three transects of climate shift very rarely overlap with each other (Figure 3). This emphasizes the future north-to-south transect-oriented shift of European cities’ climate.

3.3 Speed of southward velocity

Table I summarizes the main findings resulting from the southward velocity computation carried out for European cities’ climate shift, for the three north-to-south transects and for all the shift time periods (see Appendix Table AII for all detailed results). Overall, the southward velocity of European cities’ climate shift greatly differs depending on both their geographical location and the shift time period.

Among the four 30-year shifts, the slowest southward velocity was found for the cities of Jönköping (Sweden) and Cordoba (Spain), with a speed of 0.9 km year−1 for the P2-P3 shift. Zurich (Switzerland) and Cracow (Poland) exhibited the fastest southward velocity, 34.0 and 38.1 km year−1, respectively, for the P4-P5 shift.

Throughout the entire study period P1-P5 (i.e. from 1951-1980 to 2071-2100), Andorra-la-Vella (Andorra) and Berlin (Germany) showed the slowest (2.9 km year−1) and quickest (13.2 km year−1) climate shift, respectively. Such pace results in considerable displacement of climate in space over the European continent throughout the twenty-first century. As an example, Berlin’s climate in 2071-2100 (P5) will be located not less than 1,584 km southwards (South Spain) than its climate in 1951-1980 (P1).

When averaging the southward velocity of cities’ climate shift per transects, results showed that the cities of the Maritime transect tend to migrate southwards slower (7.3 km year−1 for the P1-P5 shift) than the cities of the two other transects (8.2 and 8.0 km year−1). Such conclusion would need to be strengthened by integrating more cities, although it corroborates findings from an earlier study based on rather similar transects (Beniston, 2013).

Results also indicated that the southward velocity of European cities’ climate shift is not constant from 1951 to 2100, and instead significantly accelerates throughout the twenty-first century (Table I and Figure 4). It starts from an average of 7.0 km year−1 for the P1-P2 shift and almost doubles to reach an average of 13.4 km year−1 for the P4-P5 shift. Computations of the averaged southward velocity for the two 60-year shifts also confirmed this finding. It starts from an average of 5.9 km year−1 for the P1-P3 shift and almost doubles to reach 11.3 km year−1 for the P3-P5 shift. Such doubling of speed is exemplified in Figure 5 for cities of the Continental transect.

A sensitivity analysis has also been performed – using the second best climate analogue of each LOIs rather than their best climate analogues – and showed similar results, in particular for the averaged result over transects and shift time periods (Appendix Table AIII).

4. Discussion

4.1 Implications

Up until now, most of the research on climate analogues and climate shift has aimed to assess the survival and abundance of species as well as the ecological changes of their habitat in response to the shift of climatic conditions. Recent findings (Ash et al., 2016) highlighted that species either shift their distribution to track climate change or adapt to the changes in their local environmental conditions. Similarly, climate shift in cities will threaten urban dwellers’ quality of life and alter cities’ functioning because of the new climate-related issues that will appear along the climate shift. Nevertheless, contrary to plants and other animal species, cities’ residents are unlikely to shift their distribution to track climate change, and hence will rather have to adapt to these changing conditions.

We have shown that European cities’ climate will shift southwards with an average speed of 7.9 km year−1 from 1951-1980 to 2071-2100 (P1-P5), under the A1B IPCC SRES scenario. This means that within one human generation (i.e. 25 years), European cities’ climate will shift 200 km southwards in average. Such rapid climate shift will undoubtedly have negative implications on the 416 million inhabitants of the 90 investigated cities, and potentially also on many more in similar cases. Moreover, cities’ residents will have to not only face the changing climatic conditions but also cope with the acceleration of the rate at which these changes occur, which is expected to double throughout the twenty-first century.

Such finding emphasizes on the strong dynamics of climate change, underlining that climatic conditions will change faster in the near future, without ever reaching an equilibrium state. Hence, an adaptation measure that is efficient at a certain period of time will not necessarily be efficient at another future time period. This brings attention to the fact that both dynamics and acceleration of the rate at which climatic conditions are changing must be taken into account when designing adaptation strategies in urban areas. Although we conducted the analysis over Europe only, findings are likely to be of similar magnitude in other continents.

4.2 Potential uses

In addition of being scientifically based, findings of climate analogues studies are thought to be easily understandable, hence can successfully raise awareness of a lay audience about climate change issues (Kopf et al., 2008; Jylhä et al., 2010; Beniston, 2014; Rohat et al., 2016). Despite being based on a rather complex method, our study is no exception. Its findings, straightforward and readily comprehensible, can potentially raise awareness of urban dwellers and decision-makers about both the magnitude and the pace of climate change, particularly when graphically displayed at city scale (Figure 6). Indeed, when cities’ residents and stakeholders visualize on a map that their city’s climate is shifting at several hundred kilometres southwards, they may immediately realize what climate change actually means in terms of the changing climatic conditions and what the magnitude of these changes is. As an example, Figure 6 shows that Berlin’s climate is shifting throughout Europe to reach North Spain by the end of the twenty-first century (2071-2100). Knowing that the climate in North Spain is much hotter and drier, with more frequent and intense heat waves, residents of Berlin could easily apprehend the magnitude of climate change and immediately envision the type of future climatic conditions they will have to cope with. Furthermore, displaying the different locations of Berlin’s future climate – at different future time periods – might also raise awareness about the pace and dynamics of climate change, emphasizing on the fact that the speed of change is greatly increasing throughout the twenty-first century.

In addition of being a potentially efficient tool to raise awareness and communicate about climate change to a lay audience, the approach described in this paper might also be of great use for decision-makers and urban practitioners in charge of designing and implementing adaptation strategies in urban areas. Indeed, by closely looking at the current climatic conditions of the cities located southwards – along a given transect – decision-makers can readily envision the future climate impacts and vulnerabilities that their respective city will face. In the same line, by looking at the adaptation options that are currently implemented in the cities located southwards, urban practitioners can immediately and easily identify the ones that would have to be implemented in their own city to be well-adapted to the future climatic conditions. Such use of the climate analogues approach as decision-support tool shows great potential (Rohat et al., 2016) but remains poorly explored. Hallegatte et al. (2007) showed that climate analogues could help assessing economic impacts of the future climate change and Kellett et al. (2011) demonstrated that such approach allows identifying adequate adaptation policies, although some limitations have been recently pointed out (Kellett et al., 2015). One of the added values of our approach lies in the fact that it allows identifying climate analogues for several future time periods. This means that urban practitioners and decision-makers can identify the future climate impacts and efficient adaptation options for different future time periods, by looking at their city’s climate analogue for these different periods, such as short-term (2011-2040), medium-term (2041-2070) and long-term (2071-2100) future. To exemplify this point, we computed the return periods of daily maximum temperatures for Aarhus, Berlin and Warsaw and their respective climate analogues for the four future 30-year time periods (Appendix Table AIV). Besides highlighting the great climate proximity between cities and their climate analogues, such results may be of valuable use for identifying future heat-related climate impacts and potential adaptation strategies – through knowledge sharing with climate analogues – for different time horizons of climatic challenges.

Finally, it is worth mentioning that using such approach as a decision-support tool for climate adaptation in urban areas would undeniably strengthen the collaboration among European cities, which is thought to be an important trigger for the implementation of efficient adaptation strategies in urban areas (Reckien et al., 2015). Particularly along a given transect, cities could share experience, adaptation options and best practices. Such transect-oriented network could be embedded within the existing climate-related cities’ networks, such as C40-Cities Climate Leadership Group or Covenant of Mayors.

While these potential uses appear promising and beneficial, limitations must be indicated at this point:

  • On the one hand, certain limitations are inherent to the climate analogues approach. By taking into account only a limited number of climate statistics (five in this study), such method simplifies the definition of what climate really is. For instance, these five climate parameters do not integrate wind speed and humidity (which both play a role in the perceived temperature) and do not account for the local parameters of urban areas – e.g. the urban heat island – which can largely influence the climate and its impacts. Moreover, climate models’ outputs that represent historical climate (1951-1980 and 1981-2010) are subject to spatially uneven biases that may lead to differences with weather stations data.

  • On the other hand, certain factors can hinder the exchange of adaptation strategies between cities sharing similar climate at different time periods. For example, two cities of a similar transect might be too different – in terms of size, population’s characteristics, functioning, shape, etc. – to efficiently share their adaptation options and best practices, which are often tightly linked to local characteristics. Moreover, the climate-related policies and infrastructures of a city might stem from political choices rather than from climatic conditions (Kellett et al., 2015), hence limiting the utility of sharing adaptation measures between cities. Finally, cities are not always well-adapted to their current climate, meaning that their adaptation strategies should not be taken as an example of good practices. In this case, knowledge sharing would only allow identifying future impacts and vulnerabilities.

5. Conclusion

This paper has introduced and described a climate-matching method that can potentially be applied worldwide. As long as reliable climatic projections are available, it allows identifying and ranking climate analogues of any cities. This method also enables assessing the velocity of their climate shift, for any future time periods.

We exemplified this approach for the European continent and applied it on 90 different European cities. We successfully determined (84 per cent of success rate) their climate analogues for four time periods, namely 1981-2010, 2011-2040, 2041-2070 and 2071-2100.

On the basis of the spatial analysis of these climate-matching results, we have shown that European cities’ climate will strictly shift southwards in the future. More specifically, this equator-ward climate shift appears to follow particular north-to-south transects, such as the three transects (Maritime, Continental and Eastern Europe) that we predetermined. Using these findings, we computed the southward velocity of this climate shift. Despite the heterogeneity of the results, the averages analysis has highlighted a significant rise (doubling) of the southward climate shift velocity throughout the twenty-first century. It reaches not less than 13.4 km year−1 in average for the shift from 2041-2070 to 2071-2100 (P4-P5). This finding is in line with other studies (Burrows et al., 2011; Chen et al., 2011; Diffenbaugh and Field, 2013). Such climate shift and increase of velocity have direct implications for urban inhabitants, who will have to adapt to this rapid climate shift and to the wide range of new climate-related issues that will occur in cities at a constantly growing pace.

One of the major added values of this study is that the findings are both scientifically sound and easily understandable by a lay audience. As a result, these can be used to raise awareness of both stakeholders and urban dwellers about the existence, the magnitude, the pace, and the dynamics of climate change. Knowing that their city’s future climate will be similar to the current climate of other cities located southwards and that such shift is expected to double throughout the twenty-first century, people may immediately envision what the future climate will look like in their city and at what pace these changes will occur. Furthermore, we have shown that such approach may also be used as a decision-support tool. It enables cities to learn from each other, in terms of future impacts and vulnerabilities as well as in terms of adaptation options, policies and best practices. Because of the southward climatic shift, knowledge transfer between European cities will be made from southern cities to northern cities, along the same transect. Such practical application of climate analogues might strengthen collaboration between cities and enhance their involvement in climate-related networks. Despite some limitations, mainly associated with the climate matching method and with the differences between cities’ characteristics, this exploratory study shows a great potential for future development, particularly regarding its applications as both a communication and decision-support tool in urban areas. On the basis of this exploratory study, further research could integrate different IPCC scenarios (such as the new set of representative concentration pathways) to assess the influence that different radiative forcing has on the southward climate velocity of cities. Further studies could also empirically test the efficiency of such approach as an awareness-raising tool. Finally, on the basis of specific case studies and with the help of cities’ stakeholders and policymakers, further research could also demonstrate the applicability of such approach as an efficient decision-support tool for designing and strengthening adaptation strategies in urban areas, at different time horizons.

Figures

Schema of the workflow that has been applied to identify the best – in terms of climatic proximity – climate analogue of each LOI, for each shift time period

Figure 1.

Schema of the workflow that has been applied to identify the best – in terms of climatic proximity – climate analogue of each LOI, for each shift time period

Map displaying the location of the 90 cities forming the three north-to-south transects, namely, Maritime transect (+), Continental transect (•) and Eastern Europe transect (▲)

Figure 2.

Map displaying the location of the 90 cities forming the three north-to-south transects, namely, Maritime transect (+), Continental transect (•) and Eastern Europe transect (▲)

Map displaying the investigated cities and their respective climate analogues for the four future 30-year time periods, according to the Maritime (+), Continental (•) and Eastern Europe (▲) transects

Figure 3.

Map displaying the investigated cities and their respective climate analogues for the four future 30-year time periods, according to the Maritime (+), Continental (•) and Eastern Europe (▲) transects

Southward velocity of the investigated cities for each 30-year shift time period

Figure 4.

Southward velocity of the investigated cities for each 30-year shift time period

Southward velocity of the Continental transect’s cities for the two 60-year shift time periods, namely, P1-P3 and P3-P5

Figure 5.

Southward velocity of the Continental transect’s cities for the two 60-year shift time periods, namely, P1-P3 and P3-P5

Climate shift over the European continent for the cities of Aarhus (Denmark), Berlin (Germany) and Warsaw (Poland), for the four 30-year shift time periods, namely, P1-P2 (from 1951-1980 to 1981-2010), P2-P3 (from 1981-2010 to 2011-2040), P3-P4 (from 2011-2040 to 2041-2070) and P4-P5 (from 2041-2070 to 2071-2100)

Figure 6.

Climate shift over the European continent for the cities of Aarhus (Denmark), Berlin (Germany) and Warsaw (Poland), for the four 30-year shift time periods, namely, P1-P2 (from 1951-1980 to 1981-2010), P2-P3 (from 1981-2010 to 2011-2040), P3-P4 (from 2011-2040 to 2041-2070) and P4-P5 (from 2041-2070 to 2071-2100)

Minimum, maximum and mean southward climate shift velocity for the three transects and the seven shift time periods

Transects
Shift time period Velocity (km year−1) Continental E. Europe Maritime Average (km year−1)
P1-P2 Min-max 1.9-15.0 2.0-20.0 0.8-22.3
Mean 7.3 7.8 5.9 7.0
P2-P3 Min-max 0.9-15.3 3.9-14.2 2.8-24.5
Mean 5.3 7.1 7.5 6.6
P3-P4 Min-max 2.0-29.6 5.0-30.1 2.8-34.3
Mean 11.3 11.3 10.4 11
P4-P5 Min-max 5.6-34.0 2.0-38.1 3.9-25.6
Mean 14.1 15.2 10.7 13.4
P1-P3 Min-max 0.9-11.4 2.4-11.3 2.0-12.1
Mean 5.5 6.3 5.9 5.9
P3-P5 Min-max 5.1-19.7 5.1-20.2 3.9-21.1
Mean 11.9 11.9 9.9 11.3
P1-P5 Min-max 3.0-13.2 5.4-11.3 4.0-10.9
Mean 8.2 8.0 7.3 7.9

List of the 90 investigated cities, with their coordinates, altitude and population

Transect City Latitude Longitude Altitude (m) Population (inhab.)
Continental Östersund 63.167 14.667 330 45,000
Falun 60.6 15.616 110 37,000
Örebro 59.26 15.22 40 100,000
Linköping 58.4 15.61 60 98,000
Jönköping 57.75 14.167 110 85,000
Växjö 58.86 14.8 170 56,000
Berlin 52.467 13.35 40 3,500,000
Leipzig 51.333 12.417 110 523,000
Dresden 51.05 13.73 120 524,000
Erfurt 50.97 11.03 190 206,000
Bayreuth 49.95 11.58 340 73,000
Frankfurt 50.1 8.683 100 681,000
Nuremberg 49.45 11.083 320 506,000
Stuttgart 48.776 9.17 250 607,000
Strasbourg 48.583 7.75 140 272,000
Basel 47.55 7.6 260 165,000
Zurich 47.378 8.54 410 379,000
Besancon 47.24 6.03 250 117,000
Geneva 46.217 6.15 370 189,000
Lyon 45.767 4.833 170 485,000
Clermont-Ferrand 45.78 3.08 410 140,000
Grenoble 45.187 5.72 220 156,000
Valence 44.93 4.89 130 62,000
Rodez 44.35 2.567 630 25,000
Toulouse 43.617 1.45 140 442,000
Andorra la Vella 42.51 1.523 990 23,000
Saragossa 41.65 −0.9 210 666,000
Valladolid 41.63 −4.72 700 307,000
Madrid 40.417 −3.717 650 3,165,000
Cordoba 37.883 −4.767 130 329,000
Eastern Europe St. Petersburg 59.917 30.417 20 5,222,000
Pskov 57.8 28.433 50 205,000
Velikiye Luki 56.33 30.533 100 99,000
Rēzekne 56.5 27.34 160 33,000
Daugavpils 55.875 26.53 110 90,000
Polotsk 55.483 28.8 130 83,000
Vitebsk 55.183 30.167 160 348,000
Kaunas 54.89 23.89 30 309,000
Vilnius 54.667 25.317 100 536,000
Minsk 53.92 27.49 210 1,894,000
Bialystok 53.34 23.166 130 295,000
Pinsk 52.12 26.1 140 135,000
Brest-Litovsk 52.08 23.7 170 323,700
Warsaw 52.217 21.005 110 1,715,000
Lodz 51.783 19.467 210 722,000
Lublin 51.248 22.57 190 348,000
Lutsk 50.75 25.335 190 212,000
Cracow 50.067 19.933 220 759,000
Lviv 49.85 24.02 290 724,000
Kosice 48.716 21.25 200 241,000
Budapest 47.498 19.04 110 1,732,000
Debrecen 47.529 21.639 120 205,000
Mukacheve 48.45 22.75 120 86,000
Cluj-Napoca 46.76 23.583 340 304,000
Timisoara 45.76 21.23 90 307,000
Belgrade 44.817 20.467 170 1,352,000
Prishtina 42.66 21.166 650 146,000
Sofia 42.7 23.33 550 1,212,000
Plovdiv 42.146 24.75 170 339,000
Skopje 41.997 21.433 250 537,000
Maritime Trondheim 63.6 10.383 10 179,000
Bergen 60.383 5.333 10 266,000
Haugesund 59.43 5.28 15 37,000
Stavanger 58.967 5.75 10 129,000
Kristiansand 58.15 8 10 88,000
Aalborg 57.05 9.919 10 204,000
Aarhus 56.15 10.22 20 320,000
Herning 56.762 8.317 50 87,000
Esbjerg 55.467 8.467 20 116,000
Bremen 53.083 8.8 10 547,000
Groningen 53.218 6.56 10 190,000
Amsterdam 52.38 4.9 10 780,000
Rotterdam 51.917 4.483 10 611,000
Calais 50.95 1.85 10 73,000
Lille 50.65 3.083 20 228,000
Rouen 49.433 1.083 20 111,000
Caen 49.18 −0.37 20 109,000
Rennes 48.1 −1.667 40 208,000
Brest 48.39 −4.48 50 140,000
Vannes 47.65 −2.76 30 54,000
Nantes 47.233 −1.538 10 293,000
Bordeaux 44.833 −0.567 10 244,000
San-Sebastián 43.32 −1.98 10 187,000
Bilbao 43.25 −2.933 20 347,000
Santander 43.46 −3.805 10 176,000
Gijón 43.53 −5.7 20 276,000
Vigo 42.23 −8.67 10 295,000
Porto 41.15 −8.617 50 231,000
Lisbon 38.733 −9.133 50 531,000
Faro 37.03 −7.91 20 65,000

Source: http://ec.europa.eu/eurostat/statistics-explained/index.php/Statistics_on_European_cities

Full results of the southward velocity computations for the 90 investigated cities and the seven shift time periods

Cities of each transect 1950-1980 to 1980-2010 1980-2010 to 2010-2040 2010-2040 to 2040-2070 2040-2070 to 2070-2100 1950-1980 to 2010-2040 2010-2040 to 2070-2100 1950-1980 to 2070-2100
Distance (km) Velocity (km/year) Distance Velocity Distance Velocity Distance Velocity Distance Velocity Distance Velocity Distance Velocity
Continental
Östersund 225.00 7.50 38.00 1.27 236.00 7.87 169.00 5.63 200.00 3.33 362.00 6.03 430.00 3.58
Falun 293.00 9.77 57.00 1.90 251.00 8.37 498.00 16.60 297.00 4.95 740.00 12.33 711.00 5.93
Örebro 89.00 2.97 235.00 7.83 125.00 4.17 257.00 8.57 294.00 4.90 308.00 5.13 600.00 5.00
Linköping 183.00 6.10 142.00 4.73 337.00 11.23 460.00 15.33 263.00 4.38 745.00 12.42 1007.00 8.39
Jönköping 88.00 2.93 28.00 0.93 729.00 24.30 225.00 7.50 104.00 1.73 930.00 15.50 1031.00 8.59
Växjö 191.00 6.37 127.00 4.23 167.00 5.57 195.00 6.50 242.00 4.03 352.00 5.87 594.00 4.95
Berlin 420.00 14.00 109.00 3.63 562.00 18.73 598.00 19.93 487.00 8.12 1102.00 18.37 1579.00 13.16
Leipzig 284.00 9.47 113.00 3.77 576.00 19.20 612.00 20.40 365.00 6.08 1181.00 19.68 1543.00 12.86
Dresden 256.00 8.53 137.00 4.57 389.00 12.97 531.00 17.70 393.00 6.55 893.00 14.88 1269.00 10.58
Erfurt 182.00 6.07 215.00 7.17 237.00 7.90 281.00 9.37 370.00 6.17 492.00 8.20 852.00 7.10
Bayreuth 115.00 3.83 183.00 6.10 441.00 14.70 203.00 6.77 296.00 4.93 496.00 8.27 779.00 6.49
Frankfurt 156.00 5.20 191.00 6.37 257.00 8.57 331.00 11.03 344.00 5.73 578.00 9.63 918.00 7.65
Nuremberg 112.00 3.73 223.00 7.43 293.00 9.77 408.00 13.60 326.00 5.43 699.00 11.65 1025.00 8.54
Stuttgart 199.00 6.63 239.00 7.97 460.00 15.33 708.00 23.60 438.00 7.30 1177.00 19.62 1536.00 12.80
Strasbourg 176.00 5.87 125.00 4.17 489.00 16.30 240.00 8.00 294.00 4.90 700.00 11.67 991.00 8.26
Basel 93.00 3.10 250.00 8.33 888.00 29.60 184.00 6.13 299.00 4.98 1071.00 17.85 1354.00 11.28
Zurich 343.00 11.43 86.00 2.87 59.00 1.97 1019.00 33.97 378.00 6.30 1060.00 17.67 1425.00 11.88
Besancon 177.00 5.90 263.00 8.77 161.00 5.37 679.00 22.63 440.00 7.33 794.00 13.23 1200.00 10.00
Geneva
Lyon 450.00 15.00 35.00 1.17 118.00 3.93 690.00 23.00 467.00 7.78 793.00 13.22 1063.00 8.86
Clermont-Ferrand 383.00 12.77 516.00 17.20 240.00 4.00 898.00 14.97 975.00 8.13
Grenoble 224.00 7.47 360.00 12.00 270.00 9.00 318.00 10.60 581.00 9.68 568.00 9.47 1141.00 9.51
Valence 635.00 21.17 368.00 12.27 187.00 3.12 891.00 14.85 946.00 7.88
Rodez 92.00 3.07 191.00 6.37 184.00 6.13 490.00 16.33 193.00 3.22 668.00 11.13 839.00 6.99
Toulouse 194.00 6.47 93.00 3.10 385.00 12.83 478.00 15.93 241.00 4.02 857.00 14.28 1018.00 8.48
Andorra la Vella 52.00 0.87 324.00 5.40 354.00 2.95
Saragossa 451.00 15.03 35.00 1.17 265.00 8.83 169.00 5.63 482.00 8.03 388.00 6.47 794.00 6.62
Valladolid 58.00 1.93 187.00 6.23 103.00 3.43 507.00 16.90 234.00 3.90 583.00 9.72 816.00 6.80
Madrid 229.00 7.63 458.00 15.27 161.00 5.37 267.00 8.90 685.00 11.42 407.00 6.78 864.00 7.20
Cordoba 390.00 13.00 28.00 0.93 418.00 6.97
Eastern Europe
St. Petersburg 206.00 6.87 126.00 4.20 383.00 12.77 115.00 3.83 239.00 3.98 475.00 7.92 708.00 5.90
Pskov 251.00 8.37 180.00 6.00 270.00 9.00 318.00 10.60 393.00 6.55 559.00 9.32 896.00 7.47
Velikiye Luki 246.00 8.20 185.00 6.17 246.00 8.20 349.00 11.63 363.00 6.05 577.00 9.62 859.00 7.16
Rēzekne 129.00 4.30 206.00 6.87 344.00 11.47 360.00 12.00 333.00 5.55 701.00 11.68 886.00 7.38
Daugavpils 204.00 6.80 129.00 4.30 257.00 8.57 353.00 11.77 319.00 5.32 609.00 10.15 859.00 7.16
Polotsk 179.00 5.97 193.00 6.43 295.00 9.83 370.00 12.33 349.00 5.82 661.00 11.02 922.00 7.68
Vitebsk 168.00 5.60 154.00 5.13 302.00 10.07 417.00 13.90 322.00 5.37 717.00 11.95 975.00 8.13
Kaunas 220.00 7.33 122.00 4.07 358.00 11.93 325.00 10.83 342.00 5.70 630.00 10.50 956.00 7.97
Vilnius 172.00 5.73 130.00 4.33 340.00 11.33 263.00 8.77 302.00 5.03 601.00 10.02 808.00 6.73
Minsk 61.00 2.03 116.00 3.87 313.00 10.43 342.00 11.40 177.00 2.95 644.00 10.73 754.00 6.28
Bialystok 106.00 3.53 175.00 5.83 473.00 15.77 557.00 18.57 275.00 4.58 802.00 13.37 1074.00 8.95
Pinsk 232.00 7.73 260.00 8.67 420.00 14.00 223.00 7.43 448.00 7.47 546.00 9.10 986.00 8.22
Brest-Litovsk 261.00 8.70 427.00 14.23 250.00 8.33 59.00 1.97 676.00 11.27 306.00 5.10 908.00 7.57
Warsaw 528.00 17.60 254.00 8.47 311.00 10.37 870.00 29.00 538.00 8.97 1076.00 17.93 1350.00 11.25
Lodz 438.00 14.60 198.00 6.60 300.00 10.00 931.00 31.03 573.00 9.55 1209.00 20.15 1318.00 10.98
Lublin 288.00 9.60 274.00 9.13 464.00 15.47 647.00 21.57 562.00 9.37 1086.00 18.10 1236.00 10.30
Lutsk 307.00 10.23 157.00 5.23 303.00 10.10 599.00 19.97 463.00 7.72 899.00 14.98 1002.00 8.35
Cracow 167.00 5.57 245.00 8.17 226.00 7.53 1143.00 38.10 412.00 6.87 1203.00 20.05 1087.00 9.06
Lviv 175.00 5.83 253.00 8.43 151.00 5.03 653.00 21.77 427.00 7.12 801.00 13.35 1226.00 10.22
Kosice 296.00 9.87 424.00 14.13 144.00 2.40 670.00 11.17 648.00 5.40
Budapest 474.00 15.80 197.00 6.57 265.00 4.42 654.00 10.90 872.00 7.27
Debrecen 82.00 2.73 139.00 4.63 277.00 9.23 346.00 11.53 215.00 3.58 619.00 10.32 821.00 6.84
Mukacheve 144.00 4.80 198.00 6.60 902.00 30.07 267.00 8.90 309.00 5.15 705.00 11.75 1265.00 10.54
Cluj-Napoca 281.00 9.37 316.00 10.53 369.00 12.30 487.00 8.12
Timisoara 77.00 2.57 357.00 11.90 304.00 10.13 407.00 13.57 423.00 7.05 560.00 9.33 952.00 7.93
Belgrade 599.00 19.97 342.00 11.40 153.00 5.10 577.00 19.23 488.00 8.13 579.00 9.65 1029.00 8.58
Prishtina 842.00 7.02
Sofia
Plovdiv 303.00 10.10
Skopje 706.00 23.53 802.00 6.68
Maritime
Trondheim 66.00 2.20 512.00 17.07 169.00 5.63 346.00 11.53 473.00 7.88 502.00 8.37 963.00 8.03
Bergen 112.00 3.73 119.00 3.97 749.00 24.97 517.00 17.23 228.00 3.80 1221.00 20.35 1311.00 10.93
Haugesund 133.00 4.43 734.00 24.47 115.00 3.83 317.00 10.57 728.00 12.13 390.00 6.50 1021.00 8.51
Stavanger 670.00 22.33 128.00 4.27 211.00 7.03 335.00 11.17 687.00 11.45 503.00 8.38 1114.00 9.28
Kristiansand 251.00 8.37 89.00 2.97 1029.00 34.30 481.00 16.03 311.00 5.18 1268.00 21.13 1186.00 9.88
Aalborg 38.00 1.27 365.00 12.17 548.00 18.27 320.00 10.67 390.00 6.50 866.00 14.43 1184.00 9.87
Aarhus 89.00 2.97 296.00 9.87 647.00 21.57 287.00 9.57 376.00 6.27 758.00 12.63 1125.00 9.38
Herning 281.00 9.37 192.00 6.40 232.00 7.73 117.00 3.90 461.00 7.68 281.00 4.68 738.00 6.15
Esbjerg 154.00 5.13 155.00 5.17 244.00 8.13 762.00 25.40 177.00 2.95 1005.00 16.75 1111.00 9.26
Bremen 206.00 6.87 302.00 10.07 281.00 9.37 207.00 6.90 493.00 8.22 480.00 8.00 972.00 8.10
Groningen 211.00 7.03 264.00 8.80 244.00 8.13 192.00 6.40 475.00 7.92 435.00 7.25 906.00 7.55
Amsterdam 263.00 8.77 395.00 13.17 140.00 4.67 203.00 6.77 596.00 9.93 336.00 5.60 851.00 7.09
Rotterdam 529.00 17.63 91.00 3.03 83.00 2.77 226.00 7.53 615.00 10.25 308.00 5.13 852.00 7.10
Calais 181.00 6.03 152.00 5.07 200.00 6.67 504.00 16.80 327.00 5.45 584.00 9.73 895.00 7.46
Lille 94.00 3.13 306.00 10.20 286.00 9.53 289.00 9.63 393.00 6.55 534.00 8.90 863.00 7.19
Rouen 174.00 5.80 85.00 2.83 139.00 4.63 269.00 8.97 259.00 4.32 401.00 6.68 659.00 5.49
Caen 96.00 3.20 154.00 5.13 192.00 6.40 306.00 10.20 250.00 4.17 471.00 7.85 593.00 4.94
Rennes 142.00 4.73 111.00 3.70 198.00 6.60 126.00 4.20 235.00 3.92 321.00 5.35 499.00 4.16
Brest 241.00 8.03 125.00 4.17 139.00 4.63 140.00 4.67 358.00 5.97 279.00 4.65 582.00 4.85
Vannes 91.00 3.03 338.00 11.27 180.00 6.00 393.00 6.55
Nantes 64.00 2.13 171.00 5.70 198.00 6.60 767.00 25.57 225.00 3.75 911.00 15.18 997.00 8.31
Bordeaux 75.00 2.50 115.00 3.83 817.00 27.23 199.00 6.63 136.00 2.27 919.00 15.32 1033.00 8.61
San-Sebastián 76.00 2.53 86.00 2.87 345.00 11.50 344.00 11.47 154.00 2.57 610.00 10.17 724.00 6.03
Bilbao 623.00 5.19
Santander 565.00 4.71
Gijón 23.00 0.77 101.00 3.37 117.00 1.95
Vigo 151.00 2.52
Porto 171.00 5.70
Lisbon 107.00 3.57 139.00 4.63 207.00 3.45 232.00 3.87 482.00 4.02
Faro 351.00 5.85
Average Continental 218.08 7.27 159.54 5.32 339.30 11.31 422.26 14.08 331.38 5.52 716.32 11.94 987.64 8.23
Average Eastern Europe 232.96 7.77 214.00 7.13 337.73 11.26 454.54 15.15 378.62 6.31 715.56 11.93 964.48 8.04
Average Maritime 177.24 5.91 224.42 7.48 312.21 10.41 321.43 10.71 354.30 5.90 591.96 9.87 873.96 7.28
Average all transects 209.54 6.98 198.24 6.61 330.32 11.01 402.79 13.43 353.90 5.90 678.43 11.31 944.30 7.87
Note:

Blank cells represent combinations of cities and shift time periods for which no climate analogue was found

Sensitivity analysis showing the absolute difference of minimum, maximum and mean southward climate shift velocity (for the three transects and the seven shift time periods) when computations are based on the second best climate analogues (for each LOIs) rather than on the best climate analogues

Velocity (km year−1) Transects
Shift time period Continental E. Europe Maritime Average (km year−1)
P1-P2 Min-max 1.5-0.5 1.2-1.4 0.4-0.7
Mean 0.2 0.3 0.1 0.2
P2-P3 Min-max 1.1-3.2 0.5-2.1 2.0-1.3
Mean 0.2 0.3 0.4 0.4
P3-P4 Min-max 0.9-0.8 1.9-3.1 0.8-2.4
Mean 0.2 0.4 0.2 0.3
P4-P5 Min-max 0.3-1.8 1.1-2.4 0.1-2.8
Mean 0.3 0.3 0.2 0.3
P1-P3 Min-max 0.7-2.5 0.5-3.1 1.5-1.7
Mean 0.2 0.1 0.1 0.1
P3-P5 Min-max 0.6-4.7 0.8-3.3 1.2-3.7
Mean 0.1 0.4 0.4 0.4
P1-P5 Min-max 1.1-4.2 0.8-3.8 1.4-2.5
Mean 0.2 0.3 0.3 0.3

Return periods in years of mean daily maximum temperatures for Aarhus (Denmark), Berlin (Germany) and Warsaw (Poland) and their respective climate analogues for the four future 30-year time periods

Cities Daily max temperature (°C) P1 city P2 city P2 climate analogue P3 city P3 climate analogue P4 city P4 climate analogue P5 city P5 climate analogue
Aarhus 29 23 11 11 9 9 5 3 2 2
31 78 33 30 26 22 13 8 4 4
33 261 98 85 79 60 34 25 9 9
35 884 290 239 240 161 86 74 21 23
Berlin 35 18 13 19 7 7 5 6 1 1.0
37 48 40 52 17 18 13 14 4 4
39 134 119 148 40 43 33 39 6 9
41 377 359 419 99 106 85 108 12 17
Warsaw 33 10 6 7 4 4 3 3 2 2
35 43 19 23 9 8 7 7 3 4
37 180 65 77 23 20 22 20 9 10
39 767 223 261 73 62 62 51 23 29

Appendix. Supplementary Material – Equations

Equation for computing the averaged Euclidean distances (EDavg) between the LOI’s future climate and the current climate of a given grid point, for the five climate variables:

EDavg=m=1nabs(A(m)f)B(m)c/n
where abs is the absolute value, A is the LOI’s value for the month m for the future time period f, and B is the given grid point’s value for the month m at the current time period c. For monthly mean precipitation and monthly mean temperature variables, number of months n = 12; for minimum winter temperature and maximum summer temperature variables, n = 3; for annual total precipitation variable, n = 1.

Equation for computing the similarity index of climate analogues, based on standardized Euclidean distances (SED):

SED=v=18[EDavg(v)XS/σS]
where EDavg(v) is the averaged Euclidean distance for the climate variable v, Xv and σv are, respectively, the mean and the standard deviation of the set of EDavg for all the grid points, for the climate variable v.

Table AI

Table AII

Table AIII

Table AIV

References

Anderson, A.S., Storlie, C.J., Shoo, L.P., Pearson, R.G. and Williams, S.E. (2013), “Current analogues of future climate indicate the likely response of a sensitive montane tropical avifauna to a warming world”, PLoS ONE, Vol. 8 No. 7, p. e69393, doi: 10.1371/journal.pone.0069393.

Archie, K.M., Dilling, L., Milford, J.B. and Pampel, F.C. (2014), “Unpacking the ‘information barrier’: comparing perspectives on information as a barrier to climate change adaptation in the interior Mountain”, Journal of Environmental Management, Vol. 133, pp. 397-410, doi: 10.1016/j.jenvman.2013.12.015.

Ash, J.D., Givnish, T.J. and Waller, D.M. (2016), “Tracking lags in historical plant species’ shift in relation to regional climate change”, Global Change Biology, Vol. 23 No. 3, pp. 1305-1315, doi: 10.1111/gcb.13429.

Aylett, A. (2015), “Institutionalizing the urban governance of climate change adaptation: results of an international survey”, Urban Climate, Vol. 14, pp. 4-16, doi: 10.1016/j.uclim.2015.06.005.

Beniston, M. (2013), “Exploring the behaviour of atmospheric temperatures under dry conditions in Europe: evolution since the mid-20th century and projections for the end of the 21st century”, International Journal of Climatology, Vol. 33 No. 2, pp. 457-462, doi: 10.1002/joc.3436.

Beniston, M. (2014), “European isotherms move northwards by up to 15 km year−1: using climate analogues for awareness-raising”, International Journal of Climatology, Vol. 34 No. 6, pp. 1838-1844, doi: 10.1002/joc.3804.

Beniston, M. (2015), “Ratios of record high to record low temperatures in Europe exhibit sharp increases since 2000 despite a slowdown in the rise of mean temperatures”, Climatic Change, Vol. 129 Nos 1/2, pp. 225-237, doi: 10.1007/s10584-015-1325-2.

Bicknell, J., Dodman, D. and Satterthwaite, D. (2009), Adapting Cities to Climate Change: understanding and Addressing the Development Challenges, Earthscan, London.

Böhm, U., Kücken, M., Ahrens, W., Block, A., Hauffe, D., Keuler, K., Rocker, B. and Will, A. (2006), “CLM – the climate version of LM: brief description and long-term applications”, COSMO Newsletter, Vol 6, pp. 225-235.

Bulkeley, H. (2013), Cities and Climate Change, Routledge, London.

Bulkeley, H., Schroeder, H., Janda, K., Zhao, J., Armstrong, A., Yi Chu, S. and Ghosh, S. (2011), “The role of institutions, governance, and urban planning for mitigation and adaptation”, in Hoornweg, D., Freire, M., Lee, M.J., Bhada-Tata, P. and Yuen, B. (Eds), Cities and Climate Change: Responding to an Urgent Agenda, World Bank, Washington, pp. 125-159.

Burrows, M.T., Schoeman, D.S., Buckley, L.B., Moore, P., Poloczanska, E.S., Brander, K.M., Brown, C., Bruno, J.F., Duarte, C.M., Halpern, B.S., Holding, J., Kappel, C.V., Kiessling, W., O’Connor, M.I., Pandolfi, J.M., Parmesan, C., Schwing, F.B., Sydeman, W.J. and Richardson, A.J. (2011), “The pace of shifting climate in marine and terrestrial ecosystems”, Science, Vol. 334 No. 6056, pp. 652-655, doi: 10.1126/science.1210288.

Chen, I.C., Hill, J.K., Ohlemüller, R., Roy, D.B. and Thomas, C.D. (2011), “Rapid range shifts of species associated with high levels of climate warming”, Science, Vol. 333 No. 6045, pp. 1021-1026, doi: 10.1126/science.1206432.

Christensen, O.B., Drews, M. and Christensen, J.H. (2007), “The HIRHAM regional climate model version 5”, Technical Report No. 06-17, Danish Meteorological Institute, Copenhagen.

Climate Communication (2014), “Climate communication: science and outreach”, available at: www.climatecommunication.org/ (accessed 15 March 2017).

CNRM (2008), “The scientific contents of ALADIN”, available at: www.umr-cnrm.fr/aladinold/scientific/scientif.html (accessed 15 March 2017).

Collins, M., Booth, B.B., Harris, G.R., Murphy, J.M., Sexton, D.M.H. and Webb, M.J. (2006), “Towards quantifying uncertainty in transient climate change”, Climate Dynamics, Vol. 27 No. 2, pp. 127-147, doi: 10.1007/s00382-006-0121-0.

CSIRO-Bureau of Meteorology (2016), “Climate analogues explorer tool”, available at: www.climatechangeinaustralia.gov.au/en/climate-projections/climate-analogues/about-analogues/ (accessed 15 March 2017).

de Castro, M., Gallardo, C., Jylha, K. and Tuomenvirta, H. (2007), “The use of a climate-type classification for assessing climate change effects in Europe from an ensemble of nine regional climate models”, Climatic Change, Vol. 81 No. S1, pp. 329-341, doi: 10.1007/s10584-006-9224-1.

Diffenbaugh, N.S. and Field, C.B. (2013), “Changes in ecologically critical terrestrial climate conditions”, Science, Vol. 341 No. 341, pp. 486-492, doi: 10.1126/science.1237123.

ENSEMBLES (2009), “Data distribution Portal”, available at: http://ensemblesrt3.dmi.dk/ (accessed 15 March 2017).

Eurostat (2012), “Eurostat regional yearbook 2012: focus on European cities”, available at: http://ec.europa.eu/eurostat/en/web/products-statistical-books/-/KS-HA-12-001-12 (accessed 15 March 2017).

Gerstengarbe, F.W. and Werner, P.C. (2008), “A short update on Koppen climate shifts in Europe between 1901 and 2003”, Climatic Change, Vol. 92 Nos 1/2, pp. 99-107, doi: 10.1007/s10584-008-9430-0.

Grenier, P., Parent, A.-C., Huard, D., Anctil, F. and Chaumont, D. (2013), “An assessment of six dissimilarity metrics for climate analogs”, Journal of Applied Meteorology and Climatology, Vol. 52 No. 4, pp. 733-752, doi: 10.1175/JAMC-D-12-0170.1.

Hallegatte, S. (2009), “Strategies to adapt to an uncertain climate change”, Global Environmental Change, Vol. 19 No. 2, pp. 240-247, doi: 10.1016/j.gloenvcha.2008.12.003.

Hallegatte, S., Hourcade, J.-C. and Ambrosi, P. (2007), “Using climate analogues for assessing climate change economic impacts in urban areas”, Climatic Change, Vol. 82 Nos 1/2, pp. 47-60, doi: 10.1007/s10584-006-9161z.

Holdridge, L.R. (1947), “Determination of world plant formations from simple climatic data”, Science, Vol. 108 No. 2727, pp. 367-368, doi: 10.1126/science.105.2727.367.

IPCC (2001), “Appendix I: glossary”, in Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K. and Johnson, C.A. (Eds), Climate Change 2001: The Physical Science Basis, Contribution of Working Group I to the 3rd Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, MA, pp. 787-798.

Juhola, S. (2016), “Barriers to the implementation of climate change adaptation in land use planning: a multi level governance problem?”, International Journal of Climate Change Strategies and Management, Vol. 8 No. 3, pp. 338-355, doi: 10.1108/IJCCSM-03-2014-0030.

Jylhä, K., Tuomenvirta, H., Ruosteenoja, K., Niemi-Hugaerts, H., Keisu, K. and Karhu, J.A. (2010), “Observed and projected future shifts of climatic zones in Europe and their use to visualize climate change information”, Weather, Climate, and Society, Vol. 2 No. 2, pp. 148-167, doi: 10.1175/2010WCAS1010.1.

Kellett, J., Hamilton, C., Ness, D. and Pullen, S. (2015), “Testing the limits of regional climate analogues studies: an Australian example”, Land Use Policy, Vol. 44, pp. 54-61, doi: 10.1016/j.landusepol.2014.11.022.

Kellett, J., Ness, D., Hamilton, C., Pullen, S. and Leditschke, A. (2011), “Learning from regional climate analogues”, NCCARF Publication 1/12, ISBN: 978-1-921609-41-1, National Climate Change Adaptation Research Facility, Gold Coast.

Kjellström, E., Bärring, L., Gollvik, S., Hansson, U., Jones, C., Samuelsson, P., Rummukainen, M., Ullerstig, A., Willén, U. and Wyser, K. (2005), “A 140-year simulation of European climate with the new version of the Rossby Centre regional atmospheric climate model (RCA3)”, SMHI Reports Meteorology and Climatology, No 108-SE-60176, Stockholm.

Kopf, S., Ha-Duong, M. and Hallegatte, S. (2008), “Using maps of city analogues to display and interpret climate change scenarios and their uncertainty”, Natural Hazards and Earth System Science Hazards, Vol. 8 No. 4, pp. 905-918, doi: 10.5194/nhess-8-905-2008.

Lee, T. and Lee, T. (2016), “Evolutionary urban climate resilience: assessment of Seoul’s policies”, International Journal of Climate Change Strategies and Management, Vol. 8 No. 5, pp. 597-612, doi: 10.1108/IJCCSM-06.2015-0066.

Leibing, C., Signer, J., van Zonneveld, M., Jarvis, A. and Dvorak, W. (2013), “Selection of provenances to adapt tropical pine forestry to climate change on the basis of climate analogs”, Forests, Vol. 4 No. 1, pp. 155-178, doi: 10.3390/f4010155.

Mycoo, M. (2015), “Communicating climate change in rural coastal communities”, International Journal of Climate Change Strategies and Management, Vol. 7 No. 1, pp. 58-75, doi: 10.1108/IJCCSM-04-2013-0042.

Peacock, L. and Worner, S. (2006), “Using analogous climates and global insect distribution data to identify potential sources of new invasive insect pests in New Zealand”, New Zealand Journal of Zoology, Vol. 33 No. 2, pp. 141-145, doi: 10.1080/03014223.2006.9518438.

Plummer, D., Caya, D., Côté, H., Frigon, A., Biner, S., Giguère, M., Paquin, D., Harvey, R. and de Elia, R. (2006), “Climate and climate change over North America as simulated by the Canadian Regional Climate Model”, Journal of Climate, Vol. 19 No. 13, pp. 3112-3132, doi: 10.1175/JCLI3769.1.

Ramirez-Villegas, J., Lau, C., Köhler, A.K., Signer, J., Jarvis, A., Arnell, N., Osborne, T., Hooker, J. (2011), “Climate analogues: finding tomorrow’s agriculture today”, CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS)”, Working Paper No. 12, Copenhagen.

Reckien, D., Flacke, J., Olazabal, M. and Heidrich, O. (2015), “The influence of drivers and barriers on urban adaptation and mitigation plans-an empirical analysis of European cities”, PLoS One, Vol. 10 No. 8, p. e0135597, doi: 10.1371/journal.pone.0135597.

Reckien, D., Flacke, J., Dawson, R.J., Heidrich, O., Olazabal, M., Foley, A., Hamann, J.J.-P., Orru, H., Salvia, M., De Gregorio Hurtado, S., Geneletti, D. and Pietrapertosa, F. (2013), “Climate change response in Europe: what’s the reality? Analysis of adaptation and mitigation plans from 200 urban areas in 11 countries”, Climatic Change, Vol. 122 Nos 1/2, pp. 331-340, doi: 10.1007/s10584-013-0989-8.

Rohat, G., Goyette, S. and Flacke, J. (2016), “Twin climate cities – an exploratory study of their potential use for awareness-raising and urban adaptation”, Mitigation and Adaptation Strategies for Global Change, doi: 10.1007/s11027-016-9708-x.

Saxon, E., Baker, B., Hargrove, W., Hoffman, F. and Zganjar, C. (2005), “Mapping environments at risk under different global climate change scenarios”, Ecology Letters, Vol. 8 No. 1, pp. 53-60, doi: 10.1111/j.1461-0248.2004.00694.x.

Schneider, S.H. (2006), “Climate change: do we know enough for policy action?”, Science and Engineering Ethics, Vol. 12 No. 4, pp. 607-636, doi: 10.1007/s11948-006-0061-4.

Sinnott, R.W. (1984), “Virtues of the Haversine”, Sky Telescope, Vol. 68 No. 2, pp. 158-159.

Ungar, J., Peters-Anders, J. and Loibl, W. (2011), “Climate twins – an attempt to quantify climatological similarities”, Environmental Software Systems, Frameworks of eEnvironment, 9th IFIP International Symposium, ISESS 2011, Brno, Czech Republic, Springer, Vienna, pp. 428-436.

van der Linden, S.L., Leiserowitz, A.A., Feinberg, G.D. and Maibach, E.W. (2014), “How to communicate the scientific consensus on climate change: plain facts, pie charts or metaphors?”, Climatic Change, Vol. 126 Nos 1/2, pp. 255-262, doi: 10.1007/s10584-014-1190-4.

van Meijgaard, E., van Ulft, L.H., van de Berg, W.J., Bosveld, F.C., van den Hurk, B.J.M., Lenderink, G. and Siebesma, A.P. (2008), “The KNMI regional atmospheric climate model RACMO version 2.1”, KNMI Technical Report TR-302, Copenhagen.

Veloz, S.D., Williams, J.W., Blois, J.L., He, F., Otto-Bliesner, B. and Liu, Z. (2012a), “No-analog climates and shifting realized niches during the late quaternary: implications for 21st-century predictions by species distribution models”, Global Change Biology, Vol. 18 No. 5, pp. 1698-1713, doi: 10.1111/j.1365-2486.2011.02635.x.

Veloz, S., Williams, J.W., Lorenz, D., Notaro, M., Vavrus, S. and Vimont, D.J. (2012b), “Identifying climatic analogs for Wisconsin under 21st-century climate-change scenarios”, Climatic Change, Vol. 112 Nos 3/4, pp. 1037-1058, doi: 10.1007/s10584-011-0261-z.

Webb, L.B., Watterson, I., Bhend, J., Whetton, P.H. and Barlow, E.W.R. (2013), “Global climate analogues for winegrowing regions in future periods: projections of temperature and precipitation”, Australian Journal of Grape and Wine Research, Vol. 19 No. 3, pp. 331-341, doi: 10.1111/ajgw.12045.

Williams, J.W. and Jackson, S.T. (2007), “Novel climates, no-analog communities, and ecological surprises”, Frontiers in Ecology and the Environment, Vol. 5 No. 9, pp. 475-482 doi: 10.1890/070037.

Acknowledgements

The authors wish to thank the anonymous reviewers for their comments that helped improving the clarity of this paper. The climate data used in this work was funded by the EU-FP6 Integrated Project ENSEMBLES.

Corresponding author

Guillaume Rohat can be contacted at: guillaume.rohat@unige.ch