“The Effect of Building Retrofit Measures on CO2 Emissions Reduction – A Case Study with U.S. Medium Office Buildings.” The effect of building retrofit measures on CO 2 emissions reduction – A case study 5 with U.S. medium office buildings

13 Building retrofits have great potential to reduce CO 2 emissions since buildings are responsible for 36% 14 of emissions in the United States. Several existing studies have examined the effect of building retrofit 15 measures on CO 2 emission reduction. However, these studies oversimplified emission factors of electricity 16 by adopting constant annual emission factors. This study uses hourly emission factors of electricity to 17 analyze the effect of building retrofit measures on emission reduction using U.S. medium office buildings 18 as an example. We analyzed the CO 2 emission reduction effects of eight building retrofit measures that 19 related to envelope and mechanical systems in five locations: Tampa, San Diego, Denver, Great Falls, and 20 International Falls. The main findings are: (1) estimating CO 2 emission reduction with constant emission 21 factors overestimates the emission reduction for most measures in San Diego, while it underestimates the 22 emission reduction for most measures in Denver and International Falls; (2) The same retrofit measure may 23 have different effects on CO 2 emission reduction depending on the climate. For instance, improving lighting 24 efficiency and improving equipment efficiency have less impact in emission reduction in cold climates than 25 hot climates; and (3) The most energy efficient measure may not be the most efficient emission measure. For example, in Great Falls, the most energy efficient measure is improving equipment efficiency , but the 27 most efficient emission measure is improving heating efficiency .


30
The United States (U.S.) is the second-largest contributor to CO 2 emissions [1] and reducing emissions 31 in the U.S. is necessary to mitigate the risk of catastrophic climate change. Intergovernmental Panel on 32 Climate Change (IPCC) declared that the CO 2 emissions humans spew into the atmosphere leads to climate 33 change. By the end of the 21 st century, the current CO 2 emissions will cause global warming to around 1.5-34 2 °C if we do not drastically limit CO 2 emissions by mid-century and beyond [2]. Global warming is 35 associated with many physical and biological damages, such as receding glaciers, bleached corals, 36 acidifying oceans, killer heat waves, and hurricanes [3] factors of electricity as an uncertainty variable and investigated the optimal set of building measures to 49 minimize emissions for the Swiss building stock. An average CO 2 emission factor of electricity in Spain 50 was adopted by Garriga et al. [15] to study the optimal carbon-neutral retrofit of residential communities 51 in Barcelona Kong was adopted in this research. An average emission factor of electricity in the last five years in Finland 54 was used by Niemelä et al. [17] to determine the cost-optimal renovation from the CO 2 emission reduction 55 potential perspectives. Life-cycle CO 2 emission reduction of retrofit measures in new commercial buildings 56 was studied by Kneifel and a state-level annual emission factor of electricity was adopted in this study [18]. 57 However, the CO 2 emission factor of electricity is oversimplified in existing studies and a constant 58 factor throughout the whole year is adopted. if solar power generation is prevalent in one area, CO 2 emission factors of electricity will be low during the 61 daytime and high at nighttime. If a region has extensive hydropower generation, emission factors of 62 electricity will be lower during the rainy season than the dry season. As a result, using a constant average 63 emission factor may underestimate or overestimate the emission reduction of some building retrofit 64 measures. 65 The above literature review shows that there is a lack of study on the emission reduction of building 66 retrofit measures with dynamically changing electricity emission factors. Existing research adopted a 67 constant emission factor, while electricity emission factors are dynamically changing. The impact of 68 electricity emission factors on building emissions is significant since electricity is the major energy source 69 of buildings. Therefore, it is crucial to investigate the emission reduction difference between using 70 dynamically changing emission factors and a constant factor. 71 In this study, hourly CO 2 emission factors of electricity are adopted to analyze the effect of building 72 retrofit measures on emission reduction. U.S. medium office buildings are used as an example in this study. 73 This paper is organized as follows: Section 2 introduces the design of the case study including location  74  selection, building retrofit measures selection, and the method to estimate the emission reduction effect of  75  individual measures. Section 3 presents the hourly CO 2 emission reduction by applying individual measures  76 using one location as an example. And the annual CO 2 emission reduction effect of individual measures in 77 all locations is analyzed in Section 3. Section 4 discusses the impact of climates on emission reduction 78 effect, the difference between energy efficient measures and emission efficient measures, and the difference 79 between using the hourly CO 2 emission factors of electricity and the annual factor. Finally, interesting 80 findings are concluded in Section 5. 81

CO2 emission reduction
117 The CO 2 emission reduction effect of the individual measure ( ) can be obtained using the following 118 formula: 119 where, is CO 2 emissions of baseline building model; and is CO 2 emissions of retrofit building 120 model by applying the retrofit measure . The 0 and can be obtained using the following formula, 121 which is also illustrated in Fig. 3. 122 where, , is CO 2 emissions at time for the building with retrofit measure . For the baseline building, 123 = 0. The is the total number of hours in a year, which is 8784 in this study. The , is CO 2 emissions 124 from electricity at time for the building with retrofit measure . The , is CO 2 emissions from natural 125 gas at time for the building with retrofit measure . The , is electricity consumption at time for the 126 building with retrofit measure . The is electricity CO 2 emission factor at time . , is natural gas 127 consumption at time for the building with retrofit measure . is natural gas emission factor, which is a 128 constant value. 129  Table 3 shows the model input values of baseline models and retrofit models, which result in 45 models (5  134 locations × (1 baseline model + 8 retrofit models)). The objective of this study is to investigate the emission 135 reduction effect due to building retrofit measures on different locations. Therefore, the embodied emissions 136 of building retrofit measures are not involved in this study. 137

CO 2 emission estimation 149
Using the electricity and gas consumption data obtained in the subsection 2.3.1, this subsection 150 introduces the method to estimate CO 2 emissions of baseline models and retrofit models. As shown in Fig.  151 3, CO 2 emissions from electricity are calculated by multiplying hourly electricity consumption with hourly 152 emission factors of electricity, and CO 2 emissions from natural gas are calculated by multiplying hourly 153 natural gas consumption with one natural gas emission factor. Hourly for the baseline models and retrofit models. We use the baseline model in Great Falls as an example to 173 illustrate the hourly electricity and natural gas consumption, as shown in Fig. 6 and Fig. 7. To make the two 174 types of energy consumption comparable, the unit of natural gas consumption is converted from MJ to kWh. 175 Fig. 6 (a) and Fig. 7 (a) shows that the electricity consumption is much higher than the natural gas 176 consumption in Great Falls. Electricity consumption is relatively even throughout the year, while natural 177 gas consumption primarily concentrates in winter. Fig. 6 (a) and Fig. 7 (a) also shows that there is a periodic 178 change in the electricity and natural gas consumption: electricity and natural gas consumption is intensive 179 during the workday, while they are almost zero over the weekend. Fig. 6 (b) and Fig. 7 (b) shows that 180 electricity consumption is concentrated from 7:00 to 22:00 in winter and 8:00 to 16:00 in summer; natural 181 gas consumption is concentrated from 8:00 to 22:00 in winter and almost no consumption in summer. 182 The emissions in Great Falls mainly occur on some days during winter while almost always zero during 199 summer. On the contrary, Fig. 6 (a) shows that electricity consumption is intensive during the whole year 200 in Great Falls. This inconsistency is due to time-variant emission factors: hourly CO 2 emission factors of 201 electricity in Great Falls are almost always zero during summer and high in winter, as shown in Fig. 4. As 202 a result, the emissions from electricity consumption in summer are almost always zero despite the amount 203 of electricity consumption. Emissions from natural gas are also almost always zero during summer due to 204 low natural gas consumption as shown in Fig. 7. Therefore, total CO2 emissions in Great Falls during 205 summer are almost always zero. 206 For a period of one whole day in winter, the variation of CO 2 emissions (Fig. 8) is consistent with 207 energy consumption ( Fig. 6 and Fig. 7): emissions from the building mainly happen during the daytime, as 208 shown in Fig. 8, and energy consumption from the building also mainly happens during the daytime, as 209 shown in Fig. 6 and Fig. 7. This is because hourly emission factors of electricity in Great Falls on one whole 210 day are relative constant (Fig. 4) and the natural gas emission factor is a constant value. It is worth noting 211 this phenomenon may not occur for other locations, such as San Diego, where electricity is largely provided 212 by solar. 213 Fig. 9 shows the annual CO2 emissions of baseline building models and retrofit building models in five 214 studied locations. "MEASURE_e" represents emissions from electricity and "MEASURE_g" represents 215 emissions from natural gas. There are some interesting findings among different locations. 216 This is because San Diego and Great Falls have high renewable energy penetration, which is 46% and 97% 220 respectively. 221 Moreover, International Falls has the largest CO 2 emissions from natural gas, followed by Great Falls, 222 Denver, San Diego, and Tampa. CO 2 emissions from natural gas increase as the climate gets colder since 223 natural gas is used for heating. When the climate gets colder, heating loads increase accordingly [40][41]. 224 So, natural gas consumption for heating increases when the climate gets colder, which leads to the increase 225 of CO 2 emissions. 226 The CO 2 emissions from natural gas only account for a small part of total emissions in Tampa, San  227 Diego, Denver, and International Falls, but they account for more than 30% of total emissions in Great 228 Falls, as shown in Fig. 9. One of the reasons is that natural gas consumption in Great Falls is large due to 229 the cold climate feature mentioned above. Another reason is that hourly emission factors of electricity in 230 Great Falls are very low due to the high penetration of hydropower and wind power. in Great Falls reduce CO 2 emissions in winter due to the high emission factors of electricity; (2) HEATING 237 reduces CO 2 emissions more significantly than the other seven measures since natural gas is used for heating; 238 (3) COOLING hardly reduces CO 2 emissions since emission factors of electricity in summer are almost 239 zero when cooling is needed; (4) SWH also has little impact on CO 2 emissions because only a little amount 240 of energy is used for service water heating; (5) by improving the efficiency, LIGHT and EQUIP reduce 241 Note: Renewable energy (RE) penetration is obtained from [31]. electricity consumption and related internal heat gain. This can reduce the cooling load in the cooling season 242 but increase the heating load in the heating season. As a result, they reduce CO 2 emissions in the spring and 243 fall when cooling is still needed and electricity comes from fossil fuel, and they increase CO 2 emissions 244 when natural gas is used for heating; and (6) by reducing the solar heat gain and increasing insulation, 245 WINDOW reduces the cooling load but increases the heating load. Therefore, it reduces CO 2 emissions in 246 the spring and fall, and increases CO 2 emissions when heating is needed. in emission reduction than hot climates. Fig. 11 shows that the CO 2 emission reduction effects of LIGHT 262 and EQUIP in International Falls (cold climate) are 4.4% and 5.1% respectively, while they are 6.6% and 263 8.8% respectively in Tampa (hot climate). 264 Using EQUIP as an example, Fig. 12. shows the hourly CO 2 emission factors of electricity, the 265 reduction of electricity consumption, the reduction of natural gas consumption, and the reduction of CO 2 266 emissions in Tampa and International Falls. Both locations have similar emission factors in electricity 267 generation (Fig. 12 a). However, the reduction of electricity consumption by applying EQUIP is more 268 effective in hot climates, such as Tampa (Fig. 12 b), since it also reduces the cooling load due to the reduced 269 internal heat gain from the equipment. For cold climates, like International Falls, additional heating will be 270 needed when internal heat gain resulted from equipment is reduced. This also leads to an increase of gas 271 consumption in the cold climate location, as shown in Fig. 12 (c). As a combined effect, Fig. 12

276
Due to the variability of CO 2 emission factors, the most energy efficient measure is not necessarily the 277 most efficient emission measure. For instance, the most energy efficient measure in Great Falls is EQUIP 278 (Fig. 13) while the most efficient emission measure is HEATING (Fig. 11). Improving equipment efficiency 279 reduces electricity consumption and related internal heat gain. This can reduce cooling loads but increase 280

Tampa (hot humid) International Falls (very cold)
(a) Hourly CO 2 emission factors of electricity (b) Electricity reduction (c) Natural gas reduction (d) CO 2 emission reduction heating loads. Therefore, improving equipment efficiency in Great Falls mainly reduces electricity 281 consumption in summer. However, this large energy reduction does not lead to corresponding emission 282 reduction because electricity in Great Falls in summer mainly comes from hydropower with zero emissions. 283 On the contrary, natural gas is used for heating in Great Falls, improving heating efficiency can directly 284 reduce emissions so that it becomes the most efficient emission measure. 285 A different example is San Diego, whose most efficient emission measure is the same as the most 286 energy efficient measure: EQUIP, as shown in Fig. 11 and Fig. 13 reduction. Fig. 14 shows the estimation bias on emission reductions using the constant emission factor by 305 comparing with the one using hourly factors. 306 Fig. 14. Estimation bias on CO 2 emission reduction using the annual emission factor 308 To quantitatively compare the difference of emission reduction by using hourly emission factors and 309 constant emission factor, Table 4 shows the CO 2 emission reduction by using these two methods and their 310 difference. Fig. 14 and Table 4 shows that using the constant emission factor tends to overestimate the 311 emission reduction in San Diego (up to 1550 kg), underestimate in Denver (up to 692 kg) and International 312 Falls (up to 1165 kg), both over-or underestimating in Tampa and Great Falls. The largest difference occurs 313 in San Diego and the smallest difference in Tampa. 314 As shown in Fig. 2, San Diego has plenty of solar power during the daytime, thus, hourly CO 2 emission 317 factors during daytime are lower than both the hourly emission factors during nighttime and the annual 318 factor (Fig. 5). This will lead to an overestimated emission for energy used in the daytime if the annual 319 factor is adopted. As a result, it will also overestimate the emission reduction for the proposed energy 320 efficiency measures since they mainly reduce energy consumption in the daytime. 321 On the contrary, hourly emission factors in Denver and International Falls during daytime are higher 322 than both the hourly emission factors at nighttime and the annual factors (Fig. 5). Since electricity 323 consumption mainly occurs during the day, applying annual emission factors to the reduced electricity 324 consumption will underestimate the CO2 emission reduction. 325 As shown in Fig. 2, Tampa's electricity source is dominated by natural gas (78%) and nuclear (12%), 326 which leads to relative constant hourly emission factors (Fig. 5). Thus, using hourly or annual emission 327 factors only results in a relatively small difference in the predicted emission reduction. 328 Although estimating CO 2 emission reduction with the constant annual emission factor can produce 329 biases, it takes less time for data collection and processing. The existing method (adopting annual factor) 330 is still applicable for locations where fossil fuel is dominated because using constant annual emission factor 331 in these locations only produce minor biases. However, our proposed method (adopting hourly factors) is 332 suggested for locations where renewable energy is dominated because using constant annual emission factor 333 in these locations leads to large biases. 334

335
This study analyzed the CO 2 emission reduction of building retrofit measures that related to envelope 336 and mechanical systems in five locations: Tampa, San Diego, Denver, Great Falls, and International Falls. 337 Instead of using the constant annual CO 2 emission factor of electricity, this study adopted hourly emission 338 factors. We found that using the constant emission factor cause estimation bias: it overestimates the 339 emission reduction for most measures in San Diego, while it underestimates the reduction for most 340 measures in Denver and International Falls. Another finding is that the same retrofit measure may have 341 different CO2 emission reduction depending on the climates: improving lighting and equipment efficiency 342 has less impact on CO 2 emission reduction in cold climates than hot climates. Furthermore, the most energy 343 efficient measure is not necessarily the most efficient emission measure: in Great Falls, the most energy 344 efficient measure is improving equipment efficiency, but the most efficient emission measure is improving 345 heating efficiency. Those finding are applicable only for medium office that natural gas is used for heating 346 and electricity is used for cooling. 347 The innovation and contribution of this study mainly lie in the following two aspects. Firstly, it reveals 348 that hourly emission factors should be adopted in CO2 emission reduction analysis for locations where 349 renewable energy is dominated. Secondly, the method of estimating CO 2 emission reduction of building 350 retrofit measures proposed in Section 2.3 can be applied to other building retrofit cases. Using this workflow, 351 future studies can estimate their CO 2 emission reductions by providing electricity emission factors together 352 with their estimated building energy consumptions and retrofit measures. 353 This study analyzes the CO 2 emission reduction effect of building retrofit measures based on one-year 354 simulation data. However, the composition of electricity generation may change over time, and CO 2 355 emission factors will change accordingly. Thus, if a building retrofit measure reduces electricity 356 consumption, emission reduction resulting from it may change over time. With the increased penetration 357 of renewable energy in electricity generation, the annual reduction of emissions due to the building retrofits 358 will likely decrease. Since the effects of building retrofit measures will last for a few decades, it would be 359 interesting to study the CO2 emission reduction effect of building retrofit measures over a longer time frame. 360