An eight-step simulation-based framework to help cities reach building-related emissions reduction goals

With buildings accounting for 40% of global carbon emissions, cities striving to meet sustainability targets aligned with the Paris Agreement must retrofit their existing building stock within 30 years. Previous studies have shown that urban building energy models (UBEMs) can help cities identify technology pathways—combinations of energy efficiency retrofits and renewable energy deployment strategies—to meet emissions reduction goals. UBEMs are currently limited by cost to only the largest cities but must be expanded to all cities if society is going to meet scientifically-identified emissions reduction goals. This manuscript presents an eight-step framework to scale technology pathways analyses using UBEMs to all communities in a repeatable, affordable manner. The roles and responsibilities of three key personas, the sustainability champion, GIS manager, and an energy modeler, for each step are identified. The eight-step process is tested with a case study of 13 100 buildings in Oshkosh, WI, USA. The case study identified a technically-feasible path to nearly net zero emissions for Oshkosh’s buildings. Constraints in the workforce, supply chain, and retrofit adoption to attain this goal were identified to inform policymakers. The case study suggests that the eight-step process is a blueprint for action in communities around the world.


Introduction
In the age of the global net zero transition to meet the intergovernmental panel on climate change (IPCC)'s 1.5 • C goal, cities need to eliminate the 40% or more of greenhouse gas emissions from their buildings [1,2].While many cities have set aggressive emissions reduction targets aligning with the IPCC's findings, few have developed concrete implementation plans [3].Most of today's buildings will still be in use in 2050, necessitating the retrofit of much of the existing building stock [4].While we know how to get individual buildings to net zero, tackling one building at a time will be too slow to achieve 2050 goals [5].Urban building energy models (UBEMs) are designed for exactly this purpose.
UBEMs were developed to simulate the energy use of buildings at the city scale.UBEMs can be used in multiple scenarios including identifying carbon reduction plans at the building-stock level [6].The carbon reduction plans usually involve several different technology pathways-combinations of building energy efficiency upgrades, end use electrification, and renewable energy deployment [7].Policymakers can then leverage this knowledge to craft implementation plans and policy interventions to support the necessary strategies to meet society's goals.
UBEMs have already been successfully demonstrated in different cities and climate zones around the world including Boston, London, and New York [8].Most previous UBEM studies have focused on the evolving methods used to model multiple buildings at a time.For example, Olkkonen et al developed optimization methods on top of the iterative 'what-if ' scenarios in their model of the entire Finnish building stock [9].Some tools are focused solely on calculating the current emissions from operational and embodied carbon in buildings [10].There are several excellent papers that provide a thorough review of the different UBEM tools [11], the strengths and weaknesses of different UBEM approaches [12], and the growth of research in this field [13].Many of these studies evaluate the effectiveness of differing technology pathways, identifying the most cost and carbon savings-effective approach [14,15].Due to the diverse nature of building stocks around the world, the technology pathways to achieve cities' emissions reduction goals varies widely.Thus UBEMs are needed for every community to chart a path to their emissions reduction goals [16].However, most existing software packages used to model UBEMs are highly technical and require one person to have expert-level knowledge of geographic information system (GIS), energy modeling, and even energy policy issues to implement and use [6].Furthermore, a large amount of data is needed to build an UBEM.For this reason, most case studies using UBEM today are limited to neighborhood-level studies such as in [17,18].The highly specialized knowledge and data needed to build an UBEM means that only the biggest cities that can afford expensive consultants or have a connection with academic researchers have developed city-scale UBEMs to date [19].
Recent advances in big data and computational tools such as UBEM.io are working to make city-scale UBEMs to carry out these technology pathway analyses accessible to all [20].UBEM.io is a web app that streamlines the integration of widely available shapefiles with a building template library to create an UBEM [20].It enables any municipality to quickly and cheaply assemble an uncalibrated UBEM to guide the development of emissions reduction strategies [20].Due to the law of large numbers, uncalibrated UBEMs have been shown to yield close results (usually less than 15% error in energy use intensity) compared to actual energy use at the building stock level [21].This level of accuracy is in line with American Society of Heating, Refrigeration, and Air Conditioning Engineers (ASHRAE) Standard 140 [22].These findings were recently confirmed by Bass et al who developed an uncalibrated UBEM with 51 000 buildings in Chattanooga, Tennessee and compared the accuracy of their model to measured 15 min electricity data [19].They found that at the archetype-level, results were consistent with the measured data, but at the individual building level results varied widely [19].For city-scale policy analyses, uncalibrated UBEMs are thus good enough to provide information to policymakers.
A key development in Ang et al was identifying the concept of three key personas critical to creating any UBEM: a GIS manager, a sustainability champion, and an energy modeler [20].This manuscript builds on the three key personas and experience gained developing UBEMs in eight cities around the world that illustrated the needs and challenges policymakers face when trying to achieve their stated emissions reduction goals [16].The authors leverage these experiences to prescribe an innovative framework for communities around the world to create and use uncalibrated UBEMs at the city-scale to develop retrofitting programs to meet their emissions reduction goals.The framework reduces the cost and complexity of a whole-city UBEM putting the models at the disposal of communities of all sizes, not just the largest cities.Through a case study of Oshkosh, Wisconsin, a small American city with 13 100 residential buildings and 66 000 residents, the authors detail how the eight-step process works and its scalability and accessibility to any community with the requisite data.In the U.S. alone, the approach used for Oshkosh opens the door to addressing emissions reductions in the 78 million housing units outside of major cities [23].
The manuscript is organized as follows: section 2 presents the framework for communities to use an UBEM for emissions reduction planning, section 3 discusses how this framework was applied to Oshkosh, and section 4 focuses on challenges with implementing the technology pathways and takeaways for future studies in other communities.

Methods
The emissions reduction goals, local climate, building construction, and requirements of communities around the world vary widely [16].Yet despite their differences, every community, from a small farming town to a big metropolis, can follow the eight-step framework outlined in figure 1 to create an UBEM that covers their entire building stock.The key innovation of the eight-step framework is defining the roles each of the three personas-GIS manager, sustainability champion, and energy modeler-plays in each step based on experience with case studies around the world.These steps are specifically chosen so that each persona works on tasks that are already under their purview and they are thus already familiar with.In this way, the framework streamlines the modeling process so communities save time, money, and resources.
The GIS manager is in charge of the community's GIS dataset, usually a shapefile, geojson, or a CityGML file [24].These positions exist because GIS data is used for everything from property taxes to life safety.GIS datasets contain accurate footprints of buildings so that the digital depiction of a property is accurate and thus property taxes are not over or under charged [25].
The sustainability champion can take many forms but they are also commonplace.In many mid-to-large cities, this is a salaried position whose job is to support the city and its residents in being more sustainable.In smaller communities this could be a volunteer, volunteer group, a function of the city council, or a secondary responsibility of a city staff member.In most jurisdictions, the sustainability champion is a generalist, dealing with everything from recycling and compost to buildings and transportation.The sustainability champions are a convener, they know the right people within the city government and within the broader community that need to be involved in any given project.They write grants to get funding for projects and are usually a respected voice in the broader community when it comes time to implementation.They are not usually, however, technical experts.Building emissions are just one part of their portfolio and they rely on others to inform them of best practices for building decarbonization.
The energy modeler is deeply familiar with individual building energy models.The field of building energy modeling has proliferated in the last decades with the growth of LEED and other similar building-rating systems.Modelers are familiar with building technologies and the construction practices of the local building stock.Either through their collaborations with local engineers or intrinsic knowledge from previous projects, they understand what can and cannot be physically implemented in their local communities.Through the eight-step process, they apply their individual building energy modeling skills to whole communities at one time.

Step 1: feasibility check
The first step for any community looking to meet their emissions reduction goals is to ensure they have the people and data that will be integral to their study.The sustainability champion initiates this step.Do they have baseline energy and emissions data for their building stock from a recent greenhouse gas (GHG) emissions inventory?If not, they will need to conduct a GHG study using city-wide energy use data (e.g.natural gas and electricity) and current emissions factors for these sources.These energy consumption data can usually be provided by the local utility in an aggregated form without causing data privacy issues.With a baseline established, does the community have specific emissions reduction targets defined?With this key background information in hand, they engage the GIS manager to ensure they have all the necessary GIS data.At this point, the community is well-situated to put out a bid for an energy modeler to join the team.

Step 2: define project scope
The scoping consultation is where the energy modeler, GIS manager, and sustainability champion formulate a series of questions they want to explore.Once the key building-retrofit related questions have been agreed upon, the sustainability champion can work with the energy modeler to transform these questions into strategies that can be studied with an UBEM.At this stage, the team is ready to define the scope of the study-e.g.what building types to include and which power grid decarbonization projections to use.This information should ideally come from a reputable source such as a local utility's goals or the National Renewable Energy Laboratory (NREL)'s Cambium dataset [26].

Step 3: GIS data preparation
In step 3, the GIS manager is tasked with preparing the city's GIS dataset to be used for the UBEM.While they have never built an UBEM before, all the requisite data, detailed in table 1, fall within their domain expertise.Leveraging the GIS manager's expertise in preparing the GIS file means that the energy modeler can focus solely on modeling building physics.This keeps costs on the project low since the GIS manager and sustainability champion are usually salaried positions in the city (or are volunteers).

Step 4: build the baseline UBEM
In table 1 the GIS manager already identified the common parameters used to segment building stocks [28].
The energy modeler's job is now focused on defining the simulation inputs (the equipment specification, construction properties, heating and cooling systems, and schedules that are part of any energy model) based on the previously defined segmentation.These data have typically been a chokepoint in creating UBEMs, oftentimes requiring hundreds of hours of pre-processing [29].The use of standardized archetypes describing representative building construction and use properties has helped streamline this process [30].
Creating archetypes previously required expert knowledge of the local building stock and a substantial amount of time, limiting the ability for small towns to afford UBEMs.Leveraging UBEM.io's library of building templates removes this barrier.Templates are currently available for anywhere in the U.S. based on U.S. Department of Energy prototypes and creating template libraries for other countries from national databases such as TABULA has been proven out in Buckley et al [31].Crucially, these templates only need to be created once for each region and then they can be accessed and used by all communities on UBEM.io or other repositories.With templates defining the building's simulation properties defined, selected, and assigned to the various geometries in the GIS file, the UBEM model can be created.The GIS footprints are extruded to the given height and assigned a template based on their segmentation characteristics.

Step 5: run the baseline UBEM
With these data in place, the model is run using urban building energy modeling software.For stock-level analysis, physics-based models such as EnergyPlus are most common and are more accurate than other model types such as regression-based statistical or reduced order models [24].UBEM.io is built to export to the urban modeling interface (UMI) although the eight-step framework outlined in this manuscript can be used with any other physics-based urban modeling software such as CityBES or TEASER [24].No matter the tool being used for analysis, when the model is run, it should be compared to measured baseline energy use data such as electricity and natural gas consumption to test for accuracy.To properly make this comparison, the model must be run with measured weather data for the same time period as the measured energy data (i.e. an annual meteorological year (AMY)) [32].
The resulting model is uncalibrated; the results will be based off the geometries and simulation parameters.As part of defining the templates, the energy modeler can use their tacit knowledge of the building stock to tweak the templates to better reflect the local context.While the parameters can be manually tweaked as the modeler sees fit, this is not meant to be a building-by-building level of effort.Using an uncalibrated model does not mean that the results will be inaccurate.As mentioned in section 1, uncalibrated UBEMs are usually within 15% of the measured data the archetype-level.These results are in part due to well-defined templates from national building stock surveys and in-part due to energy modelers' expertise and tacit knowledge of the local building stock.If the UBEM results are more than 15% off from the baseline inventory, then the emissions factor assumptions and templates need to be revisited.While previous studies have stressed the need to use calibrated models, the focus has usually been on providing detailed data at the archetype-level [33].An uncalibrated model provides sufficient information to answer policymaker questions at the stock-level [8].Calibration efforts and the data privacy issues that arise when using building-level measured energy data would unnecessarily increase the time, cost, and complexity of the UBEM study, limiting its scalability.

Step 6: create building upgrades
With a plausible baseline model established, the energy modeler turns their efforts to defining different building upgrade scenarios.These upgrades should represent feasible solutions that can be carried out in the majority of buildings in that category.Upgrades should be tailored to the building typology and region.
Once several different upgrade strategies are identified, they can be simulated and the process moves to step 7 to identify the best-performing options.Rapid iteration at this stage is critical.This is usually the most labor-intensive step for the energy modeler, yet most upgrades follow similar patterns for different regions and building types.UBEM.io has pre-defined common upgrade packages to help modelers speed up the process.Additionally, standardized upgrades could be developed at the state or national level, as has been done in the TABULA project [34].Once a few 'off the shelf ' scenarios have been tested, the energy modeler can identify further improvements that are specific to the community.Using the standard scenarios as a starting point greatly reduces study costs and opens UBEMs to every community.

Step 7: analyze scenarios
From the options studied in section 2.6, the energy modeler narrows down the recommended upgrades to a handful of strategies that enable the city to achieve its emissions reduction goals.Computational tools such as optimization can play a role in helping the team identify the most impactful and cost-effective options.This is a crucial step as too many options can overwhelm decisionmakers.Before the energy modeler presents these strategies to the sustainability champion and/or other local government representatives, they must be translated from technical terms into easy to understand concepts.Furthermore, the energy modeler should conduct some basic cost/benefit and payback analyses as part of this step to ensure that all recommendations presented are reasonable economically for all involved.

Step 8: present findings and develop implementation plans
With a simple message and a small number of potential retrofit options, the energy modeler presents their results to the sustainability champion.They must lay out the different technology pathways to achieve the city's emissions goals and the costs and savings associated with implementing these pathways.At this point, the energy modeler hands the project and data back to the sustainability champion for implementation.The sustainability champion can leverage the data from the modeler to draw conclusions about the size of the workforce needed to implement these technology pathways.The champion might involve the GIS manager for data visualization but in general they have now been given all the information they need to design programs and policies to implement the technology pathways.
The outputs of an UBEM are designed to communicate the need for collective building retrofits to meet a community's emissions reduction goal.Nolan et al showed that messages about neighbors' energy conservation behavior spurred people to conserve more energy [35].Jachimowicz et al showed that people will save energy if they think other people in their area care about saving energy [36].Furthermore, Alcott and Rogers showed that long-term reductions in energy consumption require repeated communication efforts in order to create lasting change [37].Thus to catalyze homeowner action, the sustainability champion will likely need to create a well-publicized demonstration project of the proposed retrofits while also streamlining the implementation for all residents.Furthermore, by educating the population that upgrading a home to be more energy efficient is both desirable and necessary to reach community goals, the uptake of the retrofit program should increase [38].
Finally, implementation needs to be accessible to all.This is where straightforward financing options through local banks or on-bill financing in collaboration with the local utility can make a big difference.If homeowners have easy access to capital and quick payback periods, they will be more inclined to carry out retrofits [39,40].Ultimately, successfully encouraging widespread program adoption will require a multi-pronged approach that makes building upgrades a simple and financially attractive process.

Results
The authors demonstrated the above described emissions reduction framework in the town of Oshkosh, Wisconsin, USA.Oshkosh is a midwestern city with 66 000 residents in ASHRAE climate zone 5A-cool humid [41].In many respects it is typical of any small American city, with an actively-engaged volunteer sustainability advisory board that sets goals and runs sustainability programs and a small paid planning department that manages the GIS data.For this analysis, the GIS manager was a town employee from the planning department and the sustainability champion was the town's volunteer sustainability committee.Like most communities of its size, Oshkosh does not have the resources to commission an energy modeler to build a traditional UBEM model from the ground up.Yet using the previously outlined eight-step framework, the authors, acting as the energy modelers, were able to work with city employees to gather the data and easily create and use an Oshkosh UBEM in a matter of tens of hours instead of hundreds of hours for a traditional UBEM [20].In conversations with practicing energy modelers who were introduced to the eight-step framework presented here, the cost estimate for such a study is approximately $15 000.While substantial for the operating budget of a small community, this cost is on-par with grants provided by non-profits and utility energy programs.

Step 1: Oshkosh feasibility check
The authors found that Oshkosh has access to all the necessary data to build an UBEM.Their planning department has a GIS shapefile with building footprints for the entire city and tax assessor data with building use type and age.Through ICLEI-Local Governments for Sustainability (a coalition of over 1700 city and state governments around the world), Oshkosh conducted a baseline GHG emissions inventory using measured gas and electricity consumption data for residential buildings from the local utility [42].Oshkosh and ICLEI used the emissions inventory to inform a series of emissions reduction targets: 25% by 2025, 40% by 2035, and 80% by 2050 [43].In terms of residential building emissions, this translates to city-wide residential building targets of 178 000 metric tons in 2025, 143 000 metric tons in 2035, and 47 500 metric tons in 2050.

Step 2: defining project scope
With the baseline data in place, the authors, acting as energy modelers, met with the Oshkosh sustainability champions.The team discussed the capabilities of UBEM tools and the questions that they wanted to explore.Ultimately, the questions focused on identifying cost-effective retrofits and renewable energy to meet their emissions reduction goal.These questions are aligned with that of most other sustainability-minded communities around the world.
The team agreed to use NREL's 2021 Cambium data for grid emissions to provide consistency between today's emissions and projected future emissions.Cambium contains state-by-state projections for the cost and emissions of electricity in the U.S. out to 2050.Cambium's 2022 (the first year in the database) long-run marginal emissions (0.59 kg CO 2 kWh −1 ) are roughly in line with the 0.54 kg CO 2 kWh −1 emissions factor provided by the local utility, Wisconsin Public Service [44].This is higher than the U.S. average of 0.40 kg CO 2 kWh −1 and the E.U. average of 0.23 kg CO 2 kWh −1 but lower than other areas of the world [45,46].The other main source of emissions in the study area is from natural gas furnaces for heating and hot water.The emissions factor for natural gas was assumed to be 0.18 kg CO 2 kWh −1 .
In consultation with the sustainability champion, the authors assumed that all new buildings built in Oshkosh will be efficient enough to not significantly impact emissions-i.e.all new buildings from 2022 onward were ignored.This assumption could be revisited at a later date as the progress of retrofits is reviewed, but it reflects the limited building stock growth rate of Oshkosh over the past ten years.The sustainability champion also narrowed the scope of the study down to only residential buildings.The reason for this decision was two-fold.First of all, residential buildings are the predominant building type in Oshkosh and their distributed ownership makes them a much more challenging sector to retrofit compared to the handful of owners of the commercial buildings in Oshkosh.Second, the narrowed scope provides a learning opportunity for the city that they can apply to other building types.

Step 3: preparing the GIS data
With the study scope defined the authors used UBEM.io to import all the GIS data provided by the GIS manager.Some key pre-processing was involved in this step, mainly assigning all sheds and similar auxiliary structures to the shading layer and determining building ownership.Owner-occupancy was identified by the sustainability champion as a key factor in whether a building will be eligible for retrofit given current incentive programs.Since the city's GIS file did not have a category denoting whether a structure is owner-occupied, the mailing address for the tax bill and the building's physical address were compared.If the two were the same, then the building was assumed to be owner-occupied.Across Oshkosh, 74% of residential buildings were calculated to be owner-occupied.Additional checks were carried out to ensure that building geometries did not overlap and there were unique ids for each geometry.

Step 4: build the baseline UBEM
The final UBEM included 13 100 residential buildings.Nearly all the residential buildings in Oshkosh are single family attached or detached homes, with a few multi-family buildings.Due to Oshkosh's low-density housing, the construction practices do not differ much between the single and multi-family low-rise residential buildings of the same vintage.Thus the residential archetypes are segmented only by age of construction.The three categories are: pre-1980 residential, post-1980 residential, and new residential (anything built after 2004).1980 in particular reflects the post-oil crisis implementation of building energy codes across the U.S. that led to standard energy-efficient construction practices.Each building in the city was assigned non-geometric properties along the divisions of these archetypes using the DOE's age-appropriate residential building data and some tacit knowledge of the building stock with salient characteristics documented in table 2 [47,48].In this study, due to the lack of additional data, all on-site combustion of fossil fuels is assumed to be natural gas.

Step 5: run the baseline UBEM
The authors used UMI to simulate the full model of Oshkosh in about five hours on a standard Windows desktop with 8 cores and 32 GB of RAM.The authors were able to obtain 2019 electricity and natural gas consumption data for residential buildings in Oshkosh from the local utility.Thus, the model was run using a 2019 AMY weather file created from measured data at the local airport using diyepw [49].
The model results were compared to the emissions from the carbon inventory and found to be within 9% of the measured data.The 9% error meets the ASHRAE 140 Standard and falls within the range of expected values (5%-15%) for an archetype-level study [8,19].While not absolutely conclusive, the alignment between measured and modeled data showed that using DOE templates to create archetypes predicts the energy use and emissions of Oshkosh's buildings well.

Step 6: create building upgrades
With a thus 'plausible' baseline model, the authors focused on different technology pathways that could be combined to meet Oshkosh's emissions reduction goals.It is key to note in figure 3 that natural gas is the predominant source of residential emissions.Consequently, even if the electric grid decarbonizes significantly as the Cambium dataset suggests, Oshkosh will not be able to meet its emissions goals without transitioning away from fossil fuels for heating, hot water, and other end uses.The baseline model also showed that pre-1980 and post-1980 residences account for nearly all of Oshkosh's emissions.Consequently, in consultation with the sustainability champion, post-2004 residences were not considered for retrofits.Three retrofit strategies were defined.The first is focused on energy efficiency, the second takes the efficiency retrofit and electrifies equipment and heating, and the third includes all of the energy efficiency and electrification retrofits and adds photovoltaics.
The energy efficiency upgrade strategies investigated are based on the DOE's ENERGY STAR Certified Home program [50].This nationwide program defines prescriptive insulation and airtightness goals by climate zone for new construction that greatly decrease a building's energy consumption but do not rise to passive house standards.This means homes could be further upgraded if desired, but the goal is to use strategies that are low-cost and scalable.While meant for new construction, the prescriptive requirements work well in retrofits as well.The insulation upgrades (listed in table 3) generally require a layer of continuous external insulation that would coincide with siding replacement and adding insulation in wall cavities if they are un-insulated (e.g. in pre-1980 residences).The standards further specify air sealing all cracks, adding further insulation in the roof, and installing some underfloor insulation between the basement and first floor.The program also requires all lighting to be LEDs and all appliances to be ENERGY STAR certified.Finally, the furnace needs to be upgraded to an ENERGY STAR certified 95%+ efficient unit.
The second stage retrofit consists of electrifying the heating system using a cold climate heat pump for heating/cooling and a heat pump water heater for hot water in addition to the energy efficiency upgrade package.This retrofit enables Oshkosh to eliminate the buildings' natural gas consumption.
While the electrification upgrades provide the potential that a decarbonized electricity grid will help Oshkosh meet it is goals, the city and its residents can be actively involved in this work by deploying distributed energy resources such as rooftop PV.This has the further added benefit of being a good financial choice for most buildings.To this end, the final upgrade analyzed differing amounts of rooftop PV.Using the  EnergyPlus PV module with a conservative 15% efficient PV panel and a 90% efficient inverter, the team simulated the electricity production potential of all the rooftop area in Oshkosh.The validated EnergyPlus PV module uses a full solar radiation analysis that accounts for shading, reflections, and temperature-dependence [51].Based on the simulated average yearly electricity production potential from this PV across all of Oshkosh, the required cumulative PV array size was scaled to the remaining emissions reduction needs after the electrification retrofit.

Step 7: analyze scenarios
The three technology pathways developed in step 6 require increasing amounts of effort but also lead to decreasing energy use intensity.As seen in figure 2, the efficiency upgrade leads to a 61% decrease in energy use, mostly from heating, and the efficiency and electrification upgrade leads to an 84% energy use reduction.
The PV retrofit includes approximately 78 MW of installed PV.This requires approximately 30% of the rooftop area in Oshkosh, although it could be installed on a mixture of rooftops and the ground or procured through a power purchase agreement.An array this size produces 30 kWh m −2 of electricity across Oshkosh annually, resulting in a nearly net zero EUI and Oshkosh achieving its 2050 target.

Step 8: present findings and implementation
While the energy use reductions presented in step 7 are laudable, the city is ultimately concerned with the emissions reduction potential of the technology pathways that are feasible and economical for residents.Persuading homeowners to upgrade their homes will require a combination of economic and social capital.On the economic side, the authors tabulated the costs of each upgrade based on RSMeans and NREL Electrification Futures Study data using the approximately 220 m 2 DOE prototype single family home for area-dependent costs [48,52,53].Using the energy savings based on simulations and utility costs, the authors calculated the savings and payback periods for the different upgrade packages that each homeowner can choose from.This analysis is required to ensure that technically feasible pathways are economically feasible for homeowners.
In carrying out a payback period analysis, the energy modelers (the authors) found that although heat pumps lead to substantial emissions reductions, the efficiency and electrification upgrade has a payback period of 43 years for pre-1980 residences.This is longer than the lifetimes of heat pumps so the electrification upgrade does not make economical sense in Oshkosh at the moment.This currently occurs  because the cost of natural gas is so low compared to the cost of electricity.These economics are community-specific and may change as the cost of electricity and natural gas shift in the coming years.
Consequently, the best course of action for homeowners in Oshkosh is either the efficiency or efficiency + electrification + solar retrofit (where the low-cost solar electricity negates the economic issues with electrification).The final combinations of the strategies presented to Oshkosh are shown in figure 3.These strategies empower Oshkosh to take charge of achieving its emissions reduction goals through local retrofits but also place them in the context of emissions reduction efforts on the power grid.Given 2050 projected grid emissions, the third retrofit actually enables Oshkosh to meet nearly net-zero emissions goals.
A clear takeaway from these results is that Oshkosh cannot meet its emissions reduction goals through efficiency and grid decarbonization alone.They must electrify their end uses in order to achieve their 2050 goals.Additionally, while this goal is technically feasible, current retrofitting rates hover around 1% a year, making achieving them by 2050 unlikely [4].The key for Oshkosh will thus be motivating homeowners to partake in these upgrades.An example 'back of the envelope' calculation to present the technology pathways to owners of pre-1980 residences is shown in table 4.This information could be turned into graphics and provided as part of the implementation process.It is meant to show homeowners that their individual contributions matter and make economic sense.They also urge action and would be distributed via local channels such as town meetings and the planning office.
The additional savings for post-1980 residences versus the pre-1980 residences detailed in table 4 come from the 10% increase in the average floor area of these newer homes across Oshkosh.This trend toward bigger houses is common across the country, and while it raises retrofit costs a little bit, the resulting savings are more pronounced.The costs factor in federal tax credits available in 2022 and assume aggregation of PV installations to attain a commercial-scale installation price of $1720/kW [54].

Discussion
The eight-step framework outlined in this paper has been successfully applied in Oshkosh, WI, at cost and effort levels that are scalable across the U.S. wherever GIS data sets are available.
The Oshkosh case study highlights a few key takeaways for applying the eight-step process in other municipalities.First, the goals need to be clearly set to guide all decision making.Second, the boundary conditions must be agreed upon at the outset of the study.This includes what types of buildings to include, what emissions are counted, what baseline data set is used, and how the power grid emissions are expected to evolve.The grid emissions, in particular, have an outsized impact on the results.Another key takeaway from the Oshkosh case study is that although end use electrification is not always economical today, its emissions reduction potential as the grid decarbonizes will be critical to meeting emissions reduction goals.
These facts are not unique to Oshkosh and will be a challenge for communities around the world that will need to be accounted for in any emissions reduction planning.
There is an opportunity to use this framework to engage whole regions at a time.For example, the suburbs of Boston are all similar in composition and one energy modeler could build an UBEM for the entire area and engage with communities' respective sustainability champions to tweak results presentations as needed.This efficiency in modeling is enabled by breaking down the modeling process into the discrete tasks listed in this manuscript.To further demonstrate the validity of this approach, models were developed in collaboration with six communities in North America and four in Europe: Petaluma, CA, USA; Bristol, VT, USA; Sandy Springs, GA, USA; Codman Square, Boston, MA, USA; San Pedro Garza Garcia, Mexico; Calgary, Canada; Zagreb, Croatia; Porto, Portugal; Lisbon, Portugal; and Rotterdam, The Netherlands.The UBEMs successfully modeled the communities and the respective case studies can be found at www.ubem.io/case-studies.

Putting the retrofits in perspective
Reinhart et al calculated the job creation potential of similar efficiency and electrification retrofit packages in Oshkosh and four other cities around the U.S. [55].The energy efficiency package was estimated to require 91 hours per home and the efficiency, electrification, and PV package 157 hours per home based on RSMeans data [52,55].Using these numbers, an assumption of 25 years of retrofits and 1560 on-the-job hours per year per full-time equivalent worker, at least 31 and 53 workers respectively will be needed in Oshkosh alone each year to implement these packages.According to utility rebate data from 2015 through May 2021, less than 1500 heat pumps were installed in Wisconsin in total, with only 136 contractors across the whole state performing these installations [56].Even then, nearly 25% of these systems were installed by the same five contractors [56].Oshkosh accounts for just half a percent of the state's housing stock, so even if equally distributed there is likely only one contractor in the city that can perform the requisite retrofits [57].On the other hand, Vermont, with 1/8th of the population of Wisconsin, installed over 10 300 heat pumps in 2020 [56,58].Based on a twenty-five year retrofitting program, 525 retrofits would be needed each year, which is nearly equal to the number of heat pumps installed across the whole state in 2020 [56].Clearly, the workforce required to implement retrofits in Wisconsin will need to grow exponentially in the coming years.Yet current efforts across the state are lagging behind-a recent press release lauded the funding of 42 trainees for clean energy and water efforts in 2023, not even enough to meet the needs of Oshkosh, let alone the rest of the state [59].This issue goes beyond Wisconsin, with the U.S. and Europe already constrained on heat pump installations by the lack of skilled craftspeople to install them [60].
Furthermore, with so few heat pumps currently being installed the supply chains to support this scale of retrofitting will need to be ramped up.One specific challenge is the raw materials required.Heat pumps require supply-constrained components such as Copper, Nickel, Aluminum, steel, and several microprocessors for the control panels, pumps, and fans [60,61].The semiconductor challenge is more systemic, with heat pumps often being de-prioritized when supply is short [60].Even with the raw material, the International Energy Agency has found that current manufacturing capacity for heat pumps is 60% below what it needs to be to achieve 2050 goals [60].While some of the systemic supply chain issues might affect long-term retrofitting, the U.S. installed 4.3 million heat pumps in 2022.This pace, which, if kept up, puts retrofitting at least 75% of the U.S. building stock with a heat pump in 25 years in the realm of possibilities [62].While these numbers are specifically focused on heat pumps, similar issues in energy efficiency measure installation (e.g.insulation) will also need to be addressed.
Finally, there is the issue of adoption (or lack thereof) of building retrofits in rental units.Due to the Landlord-Tenant problem, renters are unlikely to invest in any efficiency upgrades and the landlords have little incentive to invest either [63].Consequently, rental units are unlikely to be retrofitted and significant emissions reduction potential is missed.As shown in figure 4, this leads to high emissions in some neighborhoods in 2050.This leads to a significant decrease in emissions reduction from the full technical potential.Whether by regulation or through creative on-bill financing tied to the address not the renter, there are pathways to overcome this challenge but they have yet to be addressed in Oshkosh.[64].Area-normalized emissions from residential buildings at the census block level in Oshkosh.Red areas will need to be focus areas for new programs and policies that engage rental units in retrofits.

Conclusion
The authors present an eight-step UBEM-based framework for communities to identify technology pathways for their building stock to reach previously defined carbon reduction targets.The eight steps were developed and refined through previous case studies at the neighborhood level in cities around the world.Its application at the city-scale is illustrated in this manuscript through a case study of Oshkosh, WI, USA.By clearly defining the role for the GIS manager, sustainability champion, and energy modeler in each step, these powerful models can be built at the city-scale for approximately $15 000, less than a tenth the cost of bespoke UBEMs built by consultants for large cities.This cost level makes them accessible to diverse communities, no matter the size.The Oshkosh UBEM was used to identify economically-feasible retrofit strategies that can be rolled out across the building stock to achieve their emissions reduction goals.UBEMs can also quantify the lost emissions reductions potential across a city when systemic barriers such as the lack of retrofits in renter-occupied units are accounted for.They can also identify the most cost-effective and coordinated way to approach retrofits in a specific city.Finally, as shown in this manuscript, UBEMs can also be used to quantify additional impediments to retrofit adoption such as a lack of a trained workforce or adequate supply chains.The ultimate goal of these quantification efforts is to motivate policymakers to find creative solutions to longstanding issues to drive building retrofit adoption to the levels required to reach communities' 2050 goals.

Figure 1 .
Figure 1.Eight steps to meeting a community's emissions reduction goals.The key personas for each step are defined.A sustainability champion (in yellow), a GIS manager (in blue), and an energy modeler (in green).

Figure 2 .
Figure 2. The energy use intensities by end use of the baseline model for the two technology pathways in Oshkosh.The PV pathway is excluded as the consumption EUI does not change from the electrification upgrade.

Figure 3 .
Figure3.The strategies for Oshkosh to meet its emissions reduction goals.It is only through a combination of all three strategies (energy efficiency, electrification, and photovoltaics) that Oshkosh can meet its 2050 goal.The addition of grid decarbonization lets Oshkosh achieve nearly net zero by 2050.

Figure 4 .
Figure 4. Area-normalized emissions from residential buildings at the census block level in Oshkosh.Red areas will need to be focus areas for new programs and policies that engage rental units in retrofits.

Table 1 .
Required data for building simulation.

Table 2 .
Oshkosh baseline building archetype characteristics.These values are based on the U.S. DOE residential prototype buildings and ResStock data.

Table 3 .
Oshkosh upgrade requirements by archetype.Note the windows were not upgraded because their payback period is over 20 years.

Table 4 .
Back of envelope calculations for pre-and post 1980 homes to undergo retrofits.