1. Introduction

Article 2 of the Paris Agreement (^{United Nations, 2015}) suggests “holding the increase in the global average temperature (∆T) to 2 ºC and pursuing efforts to limit the temperature increase to 1.5 ºC above pre-industrial levels.” Hence, this paper analyses likely human bioclimatic conditions in Mexico for climate change scenarios with time horizons in which global average temperature (∆T) reaches 1.5 ºC and 2 ºC. The 1 ºC time horizon has been added as an extra tool for decision makers to consider when discussing regulations on residential electricity consumption for air conditioning on the 30 metropolitan areas that had more than half a million inhabitants in 2010 (Table I).

Since the goal of this paper is to estimate the human bioclimate conditions and their effects on residential electricity consumption in Mexico for the rest of the century, warming due to urbanization was added to the global climate change scenarios.

Increase of heat in cities because of urbanization, i.e., urban heat island (UHI), adds up to thermal gain due to climate change. The air in the urban canopy of UHIs is usually warmer than that of the surrounding countryside (^{Oke, 1987}). The combination of these effects will lead to an increase in demand of electricity for air conditioning (AC) to cool indoor areas. Cities that so far can manage without AC will have to add them gradually. The results of ^{Wan et al. (2011)} for different climates, especially for Australia (^{Wang et al., 2010}), California (^{Xu et al., 2012}), and Canada (^{Berardi and Jafarpur, 2020}) show that heating will decrease in cities with cold or temperate climates during the summer and, most likely, in all cities during winter.

Previous research in Mexico has estimated changes in human bioclimatic conditions under climate change scenarios. ^{Jáuregui and Tejeda-Martínez (2001)} estimated human comfort conditions in Mexico City for a possible doubling of atmospheric CO₂ but the UHI effect was not considered. ^{Tejeda-Martínez and Rivas (2003)} analysed similarly south-eastern Mexico while ^{Rodríguez et al. (2004)} and ^{García and Tejeda-Martínez (2008)} analysed the state of Veracruz. ^{García et al. (2010)} studied three cities in north-western Mexico. ^{Luyando and Tejeda-Martínez (2010)} dealt with the expected bioclimate halfway through the 21^st century in five urban regions in Central Mexico, located 2200m above sea level. ^{Luyando (2016)} assessed bioclimatic adjustments in Mexico City for the second half of the 21^st century as a result of global climate change and urban heating.

Figure IV.2.2 (p. 167) of the 4 ^th National Mexican Memorandum about Climate Change (^{SEMARNAT-INE, 2010}) is a summary of ^{Tejeda-Martínez et al. (2011)}, which measured the average thermal stress conditions in ten urban areas with more than one million inhabitants in the last decade of the 20^th century. It shows estimates of electricity consumption by user in the first decade of the 21^st century; projections for 27 cities are presented for time horizons in 2020, 2050, and 2080.

This paper uses the notion of thermal comfort as described by the American Society of Heating, Refrigerating and Air-Conditioning Engineers, which defines it as “[joint] condition[s] of mind which expresses satisfaction with the thermal environment” (^{Berglund et al., 1997}). A common, indirect way to assess thermal sensation is through bioclimatic comfort indices.

^{Epstein and Moran (2006)} and ^{Blazejczyk et al. (2012)} concur in classifying the indices into three groups: 1) rational indices, based on the general equation of thermal balance; 2) empirical indices, based on objective and subjective strain; and 3) direct indices, based on direct measurements of environmental variables. The first two groups depend on many variables, some of which require uncommon specific measurements such as radiant temperature of the environment, globe temperature, inner walls’ temperatures, and occupants’ skin temperatures. The third group is easier to use and apply because it depends on environmental variables. Although these direct indices are relatively old, they are still in use in human bioclimate studies, and perform satisfactorily when compared to sophisticated indices.

Many human bioclimate studies coincide that room temperature, atmospheric humidity, solar radiation, radiation emitted by the surroundings, and wind are the main environmental elements in human thermal sensation (^{Auliciems and De Dear, 1998}). Only indices that depend solely on the temperature and relative humidity are useful enough, since wind and solar radiation can be controlled indoors.

More than 160 indices have been used to assess human bioclimate in the last 100 years (^{De Freitas and Grigorieva, 2015}). The most sophisticated ones simulate the physical and physiological conditions present in the feeling of human comfort. In contrast, the simplest ones use only room temperature and atmospheric humidity as input data. Apparent Temperature, Effective Temperature, and the Humidex are among those that depend on these last two variables (^{Anderson et al., 2013}; ^{Chirico and Magnavita, 2019}; ^{Binarti et al., 2020}).

^{Blazejczyk et al. (2012)} found a high coefficient of determination (97%) between Effective Temperature - a direct index- and the Universal Thermal Comfort Index (UTCI) - a sophisticated index that focuses on physiological adaptation -. According to ^{Auliciems and De Dear (1998)}, the PMV index and Effective Temperature are the best options to quantify the physiological human response to the thermal environment.

^{De Freitas and Grigorieva (2017)} recently classified and evaluated 165 bioclimatic indices using six evaluation criteria: validity, usability, transparency, sophistication, completeness, and scope under different environmental conditions, assigning up to a maximum score of 30. Effective Temperature stands out with a score of 23, followed by Humidex with 15. Although the Humidex has an average rating, it is found in a number of recent publications on climate change (^{Oleson et al., 2015}; ^{Orosa et al., 2014a}; ^{Orosa et al., 2014b}; ^{Kum and Celik, 2014}; ^{Giannakopoulos et al., 2011}), heat waves (^{Stewart et al., 2017}; ^{Lee et al., 2016}; ^{Hamdi et al., 2016}; ^{Kravchenko et al., 2013}; ^{Kershaw and Millward, 2012}; ^{Mastrangelo et al., 2007}), urban environments (^{Roshan et al., 2017}; ^{Charalampopoulos et al., 2013}), and the energy sector (^{Yildiz et al., 2017}; ^{Miller et al., 2017}; ^{Marvuglia and Messineo, 2012}; ^{Beccalli et al., 2008}).

^{Jáuregui et al. (1997)} and Moran and ^{Epstein (2006)} showed that if some variables in a complex index are parameterized, its sensitivity and variability are reduced to those of a simple index. Therefore, bioclimatic evaluations based on temperature and humidity variables are still valid.

Hence, this research will use the Effective Temperature index and the Humidex. Preliminary results to those shown in this article appeared in ^{INECC-PNUD (2017a)} and ^{SEMARNAT-INECC (2018)}.

2. Data and Methods

3. Results

Estimated AC needs were calculated from the Humidex, with results from the GFDL-CM3 and HADGEM2-ES models and two radiative forcings (4.5 and 8.5 Wm^-2). The results shown here correspond to the combination of the Humidex with HADGEM2-ES and a forcing of 8.5 Wm^-2. This combination was chosen because HADGEM2-ES showed greater sensitivity and the forcing of 8.5 Wm^-2 for future cooling needs has precautionary value.

The thermal increases by UHI (from Equation 9) range from 1.1 ºC, for the metropolitan area of the Valley of Mexico, to 0.6 ºC for Poza Rica. Thus, they are lower than the climate change horizons (1.0 ºC, 1.5 ºC and 2.0 ºC).

Figure 1 shows the annual cooling needs in CDH for all 30 urban centers for the base period, while Figure 2 shows the annual increases in consumption for cooling per inhabitant at horizons ΔT = 1.0 ºC, 1.5 ºC and 2.0 ºC, with and without the UHI. The largest hygrothermal rise is expected in coastal metropolitan areas or near the eastern and southern coasts (Fig. 1). Increases in consumption for cooling are expected to be higher in the metropolitan areas of Mexicali, Reynosa-Río Bravo, Cancún, Villahermosa, Veracruz, Tampico, Monterrey, Mérida, and Ciudad Juárez when converting to energy units per inhabitant (Fig. 2). Note that Ciudad Juárez is located at 1120m above sea level, while all others are below 1000m above sea level. These are followed by Tuxtla Gutiérrez, Acapulco, and Poza Rica, also located below 1000m above sea level.

Fig. 1 Annual cooling needs (cold degree hours, CDH) for base period (climatology 1979-2012 and population 2010). 

Fig. 2 Annual increases in consumption for cooling (kWh/inhabitant) for time horizons that will reach increases in the average global temperature (∆T) of 1 ºC, 1.5 ºC and 2 ºC. a) without UHI, b) with UHI. 

Increases will grow in time, from one ΔT horizon to another, and it is expected to be even larger in the presence of UHIs. These increases will be higher from March to August (summer) and will decrease slightly from September to February (winter).

Figure 3, similar to Figure 2 but applied to the entire population in each metropolitan area, shows increases in cooling needs. Monterrey has a larger expansion because of its population. While usage in metropolitan areas in central Mexico will increase minimally, none of them display a reduction in consumption for cooling.

Fig. 3 Annual increases in consumption for cooling (GWh/area) for time horizons that will reach increases in the average global temperature (∆T) of 1 ºC, 1.5 ºC and 2 ºC. a) without UHI, b) with UHI. 

Table III presents the percentage increase in consumption needs for cooling compared to the base period (averages from 1979 to 2012, population from 2010) in warm or semi-warm metropolitan areas for the climate change scenario ∆T = 2.0 ºC. It is clear that population growth contributes the most to increases in energy consumption.

Figure 4 shows the base period of annual heating needs in Heat Degree Hours (HDH). These values were used to calculate the percentage reductions in Figure 5. As noted above, it was not possible to find energy equivalences because of a lack of conversion from HDH to energy units in the literature.

Fig. 4 Annual needs of heating (hot degree hours, HDH) for base period (climatology 1979-2012 and population 2010). 

Fig. 5 Decrease in heating needs (%) for time horizons that will reach increases in the average global temperature (∆T) of 1 ºC, 1.5 ºC and 2 ºC. a) without UHI, b) with UHI. 

Acapulco, Villahermosa, and Cancún did not use heating during the base period and, thus, show no reduction in future heating use; similarly for Veracruz during the summer months (March-August). The decline in heating needs increases as time horizons progress (ΔT = 1.0 ºC, 1.5 ºC and 2.0 ºC).

It is predicted that the largest decrease in heating needs will occur during summer when considering UHI in all metropolitan areas considered. The metropolitan areas of Tampico and Poza Rica show a 100% decrease in heating needs from time horizon ΔT = 1.0 ºC for summers with UHI; Tuxtla Gutiérrez will reach the same decline from horizon ΔT = 1.5 ºC. This means that the small heating demands during the early morning hours -usually solved by the building’s thermal inertia or clothing- will disappear completely in summer. Similar results are found for the metropolitan areas of Valley of Mexico, Guadalajara, Monterrey, León, Mérida, Cuernavaca, Saltillo, Reynosa, and Oaxaca. These areas are expected to have a significant gradual decrease close to or over 60% in heating use for time horizon ΔT = 2.0 ºC when the UHI is considered. This reduction means less need for shelter and heating use in cool mornings.

All analysed metropolitan areas show a decrease of at least 15% for heating needs for winter (results not shown) without considering the UHI; decreases will be less when considering the UHI.

The metropolitan areas that combine both an increase in cooling needs and a decrease in heating needs are Reynosa, Monterrey, Tampico, Mérida, Veracruz, Poza Rica, and Tuxtla Gutiérrez, mainly during summer and considering the UHI. All the analysed metropolitan areas will experience increases in both minimum and maximum temperatures, leading to higher cooling needs and a decrease in heating use.

Table IV shows the decrease in heating needs for ΔT = 2.0 ºC compared to the base period. Metropolitan areas that in the base period did not have heating needs, show a 0% decrease. Those with a decrease of 100% or close to 100% can gradually abstain from heating.

4. Discussion

The methods and results of this study -decreasing heating needs and increasing cooling needs due to climate change- are consistent with those published by ^{Petri and Caldeira (2015)} for 25 cities in the United States, ^{Mourshed (2011)} for Dhaka and Bangladesh, ^{Christenson et al. (2006)} for four cities in Switzerland, and ^{Schoetter et al. (2020)} for Toulouse and Dijon in France.

Table V shows the increase intervals in electricity consumption for cooling per inhabitant. This is expected to exceed the 100 kWh per inhabitant in metropolitan areas, and ranges from a minimum increase (ΔT = 1 ºC without UHI) to a maximum one (ΔT = 2 ºC with UHI). The largest increase will occur in northern Mexico and in urban coastal areas of the Gulf of Mexico and the Caribbean. As previously mentioned, the results of the works cited in the introduction and those of this work correspond in order of magnitude to the percentage of consumption per inhabitant.

Time trends in decadal consumption for four representative areas are shown in Figure 6. As an example, Puebla-Tlaxcala displays a similar behaviour to other temperate zones, such as Toluca, León, Morelia, San Luis Potosí, Aguascalientes, Querétaro, Tijuana, and even the Valley of Mexico. The main difference is the population size in each of them. The trend seen for Monterrey is representative of the pattern seen in other northern warm areas during summer such as Mexicali, Ciudad Juárez, La Laguna, Saltillo, and Reynosa. Note that cooling needs in these northern areas are almost independent of the UHI. This is slightly different in metropolitan areas of the humid tropics like Cancún, Veracruz, Tampico, Acapulco, Tuxtla Gutiérrez, Mérida, and Villahermosa. Three areas are an exception: Celaya, Poza Rica, and Pachuca. CONAPO estimates that population in those areas will decline before mid-century.

Fig. 6 Four examples of decadal consumption trends: a) Puebla-Tlaxcala, b) Monterrey, c) Cancún, d) Celaya. 

Heating needs are certainly expected to decrease under global and local warming scenarios due to urbanization. Results of this study show that the percentage reduction (over 50%) will be greater in warm places like Mérida, Tampico, Veracruz, Tuxtla Gutiérrez, and Poza Rica. Acapulco and Cancún already used no heating in the base period. Cold areas like Toluca, Pachuca, and Tijuana will decrease use by 15 to 30%, considering the UHI. The decrease in the remaining areas will be between 25 and 60%. However, it is important to note that temperate metropolitan areas (Aguascalientes, Celaya, Cuernavaca, Guadalajara, León, Morelia, Oaxaca, Pachuca, Puebla-Tlaxcala, Querétaro, Saltillo, San Luis Potosí, Tijuana, Toluca, Xalapa, and Valley of Mexico) experience extreme minimum temperatures below 5 ºC in the base period, and it is very likely that cold waves will continue to affect them in the future.

5. Recommendations

Some regions of the country have already experienced a shortage of electricity due to excessive demand derived from AC use (^{Loret de Mola, 2017}; ^{Arias, 2019}; ^{ADN, 2019}). This study shows that the situation will worsen in the future. Moreover, cold urban areas, like Toluca, Pachuca, and Tijuana, have also experienced heat waves increasing AC demand. Recommendations resulting from this analysis point towards a need to reduce AC energy consumption through measures inside buildings and in urban environments, as well as to develop awareness campaigns.

There is a sustainable building framework in Mexico for residential areas (^{CONAVI, 2010}) but it is not mandatory. There are also two official Mexican standards issued with the purpose of reducing energy consumption from AC in buildings: NOM-018-ENER-2011 (Thermal Insulation for Buildings. Features and Test Methods), and NOM-020-ENER- 2011 (Energy Efficiency in Construction. Building Enclosures for Residential Use). Although both are mandatory, their compliance is not truly monitored. There are also some voluntary Mexican technical standards: NMX-AA-164-SCF1-2013 (Sustainable Building: Minimum Environmental Criteria and Requirements), NMX-AA-171-SCFI-2014 (Requirements and Specifications of Environmental Performance in Lodging Facilities), and NMX-AA-SCFI-157-2012 (Sustainability Requirements and Specifications for Site Selection, Design, Construction, Operation, and Abandonment of Tourist Property Development Sites in Coastal Areas of the Yucatán Peninsula).

Local government needs to embrace these regulations and, together with state and federal governments, promote, monitor, and implement those regulations. It is also recommended that energy providers (particularly, the Federal Electricity Commission, CFE) get involved in monitoring compliancy with NOM-018 and NOM-020 official standards, ensuring that all new and renovated properties comply with these regulations.

Important measures have been implemented in the last decade. For example, the National Workers Housing Fund (INFONAVIT) started the Green Mortgage Program, which grants an additional credit for eco-technologies that save water, electricity, and gas. These technologies include solar heaters, thermal insulation, energy saving lighting, etc. The National Housing Council launched “This is Your Home”, a program that links subsidies to a series of criteria that guarantee property’s sustainability.

The government of Mexico City (formerly the Federal District) issued a Sustainable Building Certification Program in 2008. It established a standard to reward residential and commercial buildings with tax or financial incentives. However, most of these buildings have been built by large commercial developers and have been certified by international standards, namely LEED (Leadership in Energy and Environmental Design) rather than by individuals. LEED and similar labels, do not correspond with Mexico’s socioeconomic, cultural, and technological aspects and cannot be applied nationwide.

It is recommended that each metropolitan area incorporates new regulatory codes to current urban development regulations. Construction licenses should include compliance with energy efficiency and sustainable building measures. For example, a building’s thermal load or an energy simulation could be listed as prerequisite to obtain licenses for building construction or remodelling. This is similar to the Building Energy Certificate required for developers in the European Union. It is, therefore, paramount that municipal authorities, research centers, professional associations, and industrial chambers of construction and housing work together towards achieving the same goal.

UHI intensity tend to increase morbidity and mortality in children and the elderly. In Mexico City, for example, children between the ages of 0 and 14 suffer acute diarrhoea as a consequence of the UHI and climate change (^{Vargas and Magaña, 2020}). In Barcelona, Spain, from 1992 to 2015, elderly men and women were the most vulnerable groups (^{Ignole, et al, 2020}).

Mitigation of UHIs requires a higher density of green areas. The aforementioned standard NMX-AA-164-SCF1-2013 (for Sustainable Building) establishes that the percentage of free areas must be 10% larger than the minimum value established by local regulation, without parking areas. Moreover, these common areas for users and visitors should allow water infiltration.

Ventilation should also be promoted. Rows of tall buildings on the seashore often disturb the local sea breeze regimes in coastal cities. New real estate developments should therefore consider fresh wind corridors to enhance ventilation of residential areas. However, in humid climates, adding more humidity to the environment can increase the intensity or duration of hygrothermal discomfort, which could lead to increased installation of AC systems. Particularly in urban environments where promoting ventilation may be difficult, mitigating the UHI’s intensity would not only lead to energy savings, but would also ease the strength of heat waves and promote better quality of life.