@CVP - High Density Transit Oriented Development Will Not Reduce Greenhouse Gases
Plan Bay Area’s High Density Multifamily Transit Oriented Development Won’t Reduce Greenhouse Gas Emissions
The following analysis of Marin's actual greenhouse gas emissions challenges was submitted to the Metropolitan Transportation Commission as a public comment on the Draft Bay Area Plan and its Draft Environmental Impact Report.
Since the publication of my book (The Best Laid Plans: Our Planning and Affordable Housing Challenges in Marin), I’ve been making presentations around the Bay Area about planning, affordable housing and the One Bay Area Plan. One of the most controversial parts of my presentation is when I state that single family homes produce a lower per capita MTCO2e (metric ton CO2 equivalent) per annum (GHG footprint) than high density, multifamily, transit oriented development (TOD).
The reaction is as if I’ve committed some unpardonable sin. But like many in the audience, The Bay Area Plan and the Draft Environmental Impact Report (DEIR) also make the unexamined assumption that high density TOD is categorically superior to single family housing development with regard to reducing per capita GHG emissions.
The assumption that there is a direct connection between high density TOD and GHG emissions reductions has been repeated as “Smart Growth gospel” for decades, and gone unchallenged in many “meta” studies on global climate change (studies compiling other study results). However, though it may be considered “heresy” by much of the environmental community to even suggest otherwise, a closer look at many original studies that support these assumptions, when compared with data from more recent evaluations, reveals that those studies were flawed and this assumption is simply not true.
There are studies and there are studies:
Most of us want to believe that scientific studies are “scientific.” However, like medical studies that one day “prove” something is good for us then prove that it’s bad for us the next day; science is unfortunately, by and large, the result of the goals of those funding the studies and the fundamental principal of “garbage in, garbage out.” And in fairness, as scientific knowledge has advanced, older studies have proven to be inadequate due to faulty logic and incomplete information.
In the 1970’s “sprawl” was an easy target of disdain for a new breed of young environmentalists who had grown up in suburbs, gone to good colleges and moved to cities where the available 24/7 access to activities better suited their lifestyle. In some ways the early environmental movement was a general attack on “white bread” suburbia and all its perceived false values and conspicuous consumption.
However, as much as urban centers are marvelously good economic environments and great social environments for certain demographic groups, urban development, as it exists today and as we still build it today, has yet to produce good environmental solutions. And when rated on a human health scale, urbanism scores very poorly in human health metrics, per capita, for disease and disorders of all kinds. GHG’s and air pollution in general are included in the possible reasons for that. With very few exceptions, we don’t find “disease clusters” in rural or suburban areas, as we do with urban environments, unless a specific toxic pollutant is present.
Even a cursory review of past studies reveals that very few appear to have actually gone back to original sources or brought a skeptical eye to the datasets they employed to justify their conclusions and projections. So let’s examine all the factors that should be included in a thorough analysis comparing GHG emissions.
There are six reasons why the conclusions of past studies, indicating that high density development produces lower GHG emissions on a per capita basis, are false. These are as follows:
The Definition of a “Unit” of Housing;
Common Areas and GHG Per Unit Calculations;
Urbanism’s “Heat Island” and “Cold Sink” Effects;
Urbanism’s GHG “Externalities;”
The Effects of Local GHG Sequestration.
GHGs produced by public transportation
Many of the studies have been developed to analyze and compare the GHG output of various housing densities and living configurations. Those undertaken in the 1970’s and early 1980’s, particularly, were overly simplistic and led to seemingly obvious but statistically incorrect conclusions. Because of this, the resultant “urban legend” about the beneficial relationship between GHG’s and urbanism was flawed.
As with all “science,” one has to ask who did the study, who paid for the study, and towards what end. During the early decades of the environmental movement there was great urgency to create the EPA, pass clean air and water legislation, endangered species laws, and address variety of other issues. Climate changing GHGs were not on the radar but the environmental report card of the nation was worse than it is today.
Many studies tried to show how bad things were in order to attract media and funding. They extrapolated trends that have not come true (mostly because of the legislation that was passed as a result). The six factors I’ve noted above are among the things that have taken decades to look at more carefully, and they have produced surprising results.
The Definition of a “Unit” of Housing: functional unit vs. living unit:
There are two definitions of a habitable unit used in different studies. A “functional” unit means a unit that can support an average family with all the amenities that are generally considered minimum standards for habitability. It does not factor in unit size, construction method, and so on. A “living unit” includes all the requirements of a functional unit but it is adjusted for square footage size (i.e. per person per square foot of living space) and sometimes for construction type.
However, many earlier studies through the 1990’s did not differentiate between these two definitions.
If one uses the functional unit definition to arrive at a per capita GHG calculation, it’s no surprise that high density units (which on average are smaller than single family homes) have lower energy usage and correspondingly lower GHG emissions per capita. However, as noted in Comparing High and Low Residential Density: Life-Cycle Analysis of Energy Use and Greenhouse Gas Emissions. J. Urban Plan. Dev., 132(1), 10–21. By Norman, J., MacLean, H., and Kennedy, C. (2006): “When the functional unit is changed to a per unit of living space basis the (beneficial) factor decreases to 1.0–1.5.” A factor of 1.0 indicates no advantage either way (and this is before any of the other considerations noted below).
In the Bay Area Plan, when trying to compare the GHG output of different Plan Alternatives that include both high density and low density single family, the use of the correct definition becomes relevant, and in the case of all of the suburban areas in the Bay Area (e.g. Marin County) it becomes extremely relevant.
Common Areas and GHG per Unit Calculations:
Up until recently, very few studies correctly factored in the “pro rata share” that each unit needs to include for common spaces in a multifamily, high density building. These would include the GHG burden to heat, light, cool and otherwise make habitable common spaces such as elevators, lobbies, community rooms, laundry areas, storage areas, swimming pools and recreational areas, hallways, and all other commonly shared areas.
Correctly factoring in typical high density common areas reduces the “advantages” that high density development has over detached single family development when calculating GHG emissions equivalents on a per capita basis. This would, of course, have differing impacts on the outcomes of the Bay Area Plan in different parts of the Bay Area: e.g. it would be very significant in calculating GHG emissions per capita in San Francisco, San Jose and Oakland, but less so in Marin, Sonoma and Napa.
Urbanism’s “Heat Island” and “Cold Sink” Effects:
Recent studies have begun to find that dense urban cores / high density developments that have so much concrete, steel, stone and other temperature variant materials, have a negative effect on energy consumption and GHG emissions. Heating and cooling effects, such as the “head island” effect (once an urban environment gets hot, it takes more and more MTCO2e to cool it down) and the “cold sink” effect (once an urban environment gets cold, it takes more and more MTCO2e to heat it up) must be considered for any analysis to be accurate (Note: According to the U.S. Energy Department, building operations are the biggest energy user, using 40 percent of the nation’s energy).
For example, according to a recent study done by the Lawrence Berkeley National Laboratory’s Heat Island Group, about these phenomena in the city of Los Angeles, they estimated that because of the heat island effect "the demand for electric power rises nearly 2% [more] for every degree Fahrenheit the daily maximum temperature rises." Note that the Plan Bay Area DEIR even acknowledges the effects of heat islands (page 25-21) but fails to apply its effects to its findings.
Correctly factoring in the heat island and cold sink effects would negatively alter the analysis of the per capita GHG emissions outcomes of TOD. In Marin, for example, where over 65 percent of the County is dedicated open space, there is a natural balance of development and natural topography that acts to eliminate the heat island and cold sink effects and offer a moderate climate throughout the year. This has beneficial effects on heating and cooling energy demands and reduces GHG emissions.
GHGs produced by public transportation:
What is equally missing from most studies of per capita GHG emissions and the GHG analysis in the Bay Area Plan DEIR is adding back the GHGs produced by public transportation, on a per capita basis, that replaces car and light truck use.If we build or add more public transportation to entice residents to reduce using their cars and light trucks, then that choice also has green house gas consequences.
Plan Bay Area and its DEIR completely ignore this in its analysis of impacts or per capita GHG assumptions.
It’s difficult to quantify these impacts or do the kind of detailed analysis I’ve done for other factors noted above. However, there is data we can look to in order to help us evaluate this. Richard Hall recently wrote a Patch blog analyzing this, Planning For Reality – Plan Bay Area Recipe For Transit Disaster. I won’t repeat everything he notes here, but generally, what he shows is that “studies” that are often referenced as proof of the positive effects of public transportation assume maximum ridership. Whereas, in reality, as evidenced in examples like Portland, Oregon, ridership of public transportation is extremely low and is generally unresponsive to investment that tries to “reengineer” how people live and work.
In fact, 40 percent of the cost of public transportation in the U.S. has to be taxpayer subsidized.
For example, Richard notes that “The city of Portland has conducted just the same kind of highly aggressive “compact infill development” policies as Plan Bay Area, combined with significant transit investment. The transit cost $3bn and the subsidies required to encourage building and habitation of this housing was another $2bn (guess who paid for that). The results were that in downtown Portland the share of weekday commuting on transit actually fell from 46% of trips to 36% during the past decade (according to annual surveys done by the city auditor).”
Thomas Rubin, the former Controller-Treasurer of the Southern California Rapid Transit District from 1989 until 1993, in his Critique of Public Transit Buses: A Green Choice Gets Greener, provides sound analysis to conclude that “on average in the U.S., moving a passenger one mile in an auto uses less energy, and produces less emissions, per passenger-mile (one person traveling one mile) than carrying that person one mile in an urban transit bus.”
An in-depth analysis comparing mass transit (buses and trains) to automobile energy utilization and emissions have concluded that, on average, auto use yields lower per capita, per mile GHG emissions than mass transit. The only exceptions to this might be in high density urban core cities with fully developed public transportation (e.g. New York City, Chicago, etc.). However, even then the “externalities” of this mass transit appears to never be factored into any studies that I’ve found.
Urbanism’s GHG “Externalities:
Proper analysis of GHG emissions externalities, or “exogenous” impacts and costs, has rarely been factored into any GHG calculations or studies, even those conducted by the EPA and CA EPA. The principle of external GHG impacts is simple. Everything that is required to service the habitability of development in any setting has external and largely unaccounted for “costs” that need to be factored into any per capita GHG emissions claims if we want to arrive at an apples to apples comparison.
Some of these factors would include the GHG loads required to provide fuel and energy, water, food, services such as garbage and sewage removal and treatment, and the unique demands of geographic location and micro-climates.
New York City recycles / repurposes less than 10 percent of its “trash.” Marin County recycles / repurposes almost 80 percent. Marin ships its remaining trash to local landfills, at a minimum distance. NYC’s trash travels thousands of miles, on average, to be dumped in landfills in the Western United States or sorted in the South before being shipped to landfills overseas, sometimes as far away as Asia. All of this has a GHG emissions cost that is not included in per capita energy consumption / GHG emissions metrics in studies or the DEIR.
This same principle applies to all the other categories. Power and water services to major metropolitan areas take significant energy to transport, and transmission loss boosting requirements for power and water evaporation both have measurable GHG emissions burdens that must be expressed in per capita metrics. Even food transportation has a quantifiable GHG cost that is significantly higher in urban environments than it is in places like Marin, where much of our food is locally grown.
In addition, a recent study, Greenhouse Gas Emissions Along the Urban-Rural Gradient, by Clinton J. Andrews, published in the Journal of Environmental Planning and Management, Vol. 51, Issue 6, 2008, notes that “Reflecting their central regional roles, municipalities… have higher per-capita emissions because they host both residential and commercial buildings. Buildings in urban areas typically contribute more emissions than personal transportation,” outweighing any other advantages that might exist.
A study conducted by the Australian Conservation Foundation, Housing Form in Australia and its Impacts on Greenhouse Gas Emissions (Oct. 2007), which did attempt to factor in all of the categories of variables (living unit definition, inclusion of common areas, the heat island and cold sink effects, the type and amount of driving and vehicle trips taken, GHGs from public transportation, and the GHG externalities), concluded that “reducing GHG emissions is not so simple as to be achieved through the urban consolidation agenda. Indeed, there is considerable evidence to the contrary.”
This study suggests that the Bay Area Plan’s transportation oriented development approach is flawed.
GHG per capita emission estimates from the recently published Australian Conservation Foundation Consumption Atlas. The data shows that “lower density areas, which rely more on automobiles, tend to produce less in GHG emissions than the high density, more public transport dependent areas that are favored by urban consolidation policies.” Their comparative findings about residential building types, resulting from this kind of comprehensive GHG per capita emissions analysis is even more eye-opening (chart below).
This research concludes that “low rise” high density development, the kind that is envisioned by the Bay Area Plan for Marin and many other parts of the Bay Area, produces 2.5 times the GHG emissions of single family home development and 3 times the GHG emissions of attached, single family townhouse development. High rise development produces 5 times the GHG emissions impacts of single family town homes. Even if these results were wrong by half (which they aren’t) they would still show a decided advantage to low density, suburban development.
If you consider these 5 factors, that are generally not accounted for, the “facts” and metrics that form the basis of the Plan Bay Area DEIR’s conclusions, that heavily favor high density TOD, are seriously flawed. Still, there is yet another aspect to calculating per capita GHG emissions correctly, to consider.
The Effects of Local GHG Sequestration:
In the environmental community there is a saying, “There is no such thing as ‘away.” This means that when we say we “threw it away” or it “went away,” we’re kidding ourselves. Everything is connected to everything else and has to go somewhere.
The final piece of a complete, per capita, data analysis that is required to accurately assess the true per capita GHG emissions impacts of various land use and planning scenarios, that was not even considered in the Australian Study, is the calculation of what portion of GHG’s produced are sequestered locally and what portion are unaccountably “exported” to neighboring counties or regions.
This is relevant inquiry because Plan Bay Area will influence land use patterns and increase density, impacting the local MTCO2e sequestration potential of Marin’s existing ecosystem. This analysis is also relevant to the entire premise of the Bay Area Plan. But there is no evidence that local MTCO2e sequestration was considered in the Plan’s DEIR when making claims about reducing GHG emissions from autos and light trucks.
To date, I have been unable to find a single study that combines the five other factors noted above with potential local sequestration MTCO2e variants that effect actual GHG impacts of various land use and development density scenarios. Yet, this data is vital to making sound planning and land use decisions and it weighs on the questionable efficacy of the Bay Area Plan.
Local Sequestration of Auto and Light Truck Emissions Compared in Urban and Suburban Locations (San Francisco and Marin County):
Automobile ownership in San Francisco County is presently 658 cars / light trucks per 1,000 people, or .66 per person. Auto ownership in Marin County is presently 756 cars / light trucks per 1,000, or .77 per person.
The population of San Francisco is 812,826 people. This equates to a total of 536,465 vehicles in San Francisco. The population of Marin County is 255, 031. This equates to a total of 196,734 vehicles in Marin County. These totals generally match DMV registration records.
According to the EPA, the average American car puts out 5.2 MTCO2 (metric tons of CO2) per year. Local auto sales figures would suggest that the Bay Area Region has a significantly lower average due to our early adoption of PZEV and ZEV vehicles (e.g. a Toyota Prius). However, for the sake of this analysis I will use the worse-case scenario national averages.
Using the EPA figure, this equates to:
San Francisco County produces 2,789,618 MTCO2 per year in GHG’s from auto and light truck usage,
Marin County produces 1,023,022 MTCO2 per year in GHG’s from auto and light truck usage.
According to the latest U.S. Census, San Francisco County, a dense urban development area, has a total of 329,700 occupied housing units of which 62,653 are single family detached homes and 267,047 are multifamily units (19 percent and 81 percent, respectively).
Marin County, a rural and suburban, low density development area, has a total of 100,650 housing units of which 63,656 are single family detached homes and 39,994 are multifamily units (63 percent and 37 percent, respectively).
On this per housing unit basis then, when comparing the GHG emissions from the use of autos and light trucks of San Francisco (high density urban development) and Marin County (low density rural and suburban development):
San Francisco produces an average of 8.46 MTCO2 per housing unit per year in auto GHG emissions;
Marin County produces an average of 10.16 MTCO2 per housing unit per year in auto GHG emissions.
Using this overly simplistic analysis based on only this one measure, one might conclude, as the Plan Bar Area DEIR apparently concludes, that dense urban development is superior to rural or suburban development with regard to auto and light truck emissions. However, that kind of analysis is inadequate to reach that conclusion.
Keep in mind that this part of the analysis is strictly breaking out auto and light truck GHG emissions when compared to housing unit counts and not factoring in all the other considerations presented above regarding the effects and impacts of unit sizes, definition of what a unit is, accounting for common areas in multifamily high density buildings, heat island and cold sink effects, public transportation GHGs, or GHG “externalities” that are exported to other regions.
However, continuing to use this simple measurement metric, we must now apply the impacts of local MTCO2 sequestration to properly compare the overall GHG impacts of urban environments to rural / suburban environments.
Local Sequestration Calculations:
San Francisco City/County covers 231.09 square miles or 147,898 acres of land. Of that approximately 10 percent is dedicated open space (mostly the land covered by Golden Gate Park, the Presidio and coastal areas and golf courses). The remainder is urban (90 percent).
Marin County covers 828 square miles or 529,920 acres of land. Of that approximately 65 percent is permanently dedicated open space and 15 percent is agricultural / recreational rural land. The remainder is approximately 5 percent fully developed land and 15 percent suburban.
The MTCO2 sequestration equivalencies for different types of land use are as follows (Sources: U.S. EPA Calculator, CA EPA, and CA Air Resources Board, which differ):
Forest and open vegetated land: more than 10 years old:
2.5 MTCO2 per year per acre.
Agricultural / Recreational grassland:
1.5 MTCO2 per acre.
Suburban land with a 40 percent lot coverage maximum:
1.0 MTCO2 per year per acre
Fully developed urban landscape: minimal vegetation
0.2 MTCO2 per year per acre
Comparing San Francisco County to Marin County:
90 percent urban developed land: 133,108 acres at 0.2 per acre equals sequestration of 26,622 MTCO2e per year.
10 percent forest and open vegetated land: 14,790 acres at 2.5 per acre equals sequestration of 36,975 MTCO2e per year.
TOTAL: San Francisco local sequestration equals 63,597MTCO2e per year.
65 percent forest / open land: 344,448 acres at 2.5 per acre equals sequestration of 861,120 MTCO2e per year.
15 percent is agricultural / recreational rural land: 79,488 acres at 1.5 per acre equals sequestration of 119,232 MTCO2e per year.
15 percent suburban land: 79,488 acres at 1.0 per acre equals sequestration of 79,488 MTCO2e per year.
5 percent urban developed land: 26,495 acres at 0.2 per acre equals sequestration equal 5,299 MTCO2e per year.
TOTAL: Marin local sequestration equals 1,065,139 MTCO2e per year.
Based on this analysis, Marin County, a rural / suburban development area, locally sequesters more than 100 percent of its locally generated auto and light truck MTCO2 emissions per year, whereas San Francisco only sequesters about 1.1 percent of its locally generated auto and light truck MTCO2 emissions per year.
This simple analysis resoundingly demonstrates that the entire premise of Plan Bay Area, the conclusions of the DEIR and the underlying premise of SB375 are completely false in asserting that high density, transit oriented development categorically results in an overall reduction in MTCO2e emissions for personal autos and light trucks.
Plan Bay Area’s premise only works if you ignore all the GHG’s and pollutants that are “exported” from urban regions to others. Whereas, this correct analytical method indicates that the denser a place becomes the worse the balance of GHG emissions and local sequestration gets. When you now factor in the other negatives of high density building types, noted above, the effects of increasing density is decidedly negative for overall GHG emissions per capita.
We need to be asking, what are the impacts on the efficacy of the Plan in achieving the goals of SB375 if the loss of land and the associated MTCO2e sequestration is accurately calculated?
The various facts presented in these analysis and the resultant conclusions provide evidence, without question, that when all factors are considered (the impacts of unit sizes, definition of what a unit is, accounting for common areas in multifamily high density buildings, heat island and cold sink effects, the GHGs produced by public transportation, unaccounted for GHG “externalities” exported to other regions, and local GHG sequestration) suburban, single family home development, as it is found in Marin, Sonoma, Napa and other parts of the Bay Area Region, is superior in reducing GHG emission on an overall basis and on a per capita basis than dense urban, TOD development found in San Francisco, Oakland and San Jose, or the developments contemplated and promoted by Plan Bay Area.
Neither the Plan nor the resultant DEIR acknowledge or in any way address or account for this data and findings presented here. What accurate and specific scientific evidence or data points then does the Bay Area Plan have to support its efficacy with regard to actually reducing auto and light truck driving mileage and GHG reduction on a per capita basis?
The building methods available to us today, even with token gestures like LEED certification, do not even begin to justify the claim that more high density, multifamily, transit oriented development is good for the environment. The truth is that development, TOD or otherwise, particularly in counties like Marin, Sonoma and Napa, only sets in motion an endless feedback loop the drives even more development to accommodate support services and our consumption driven economy, and ever more auto and light truck use and, more importantly, more shipping, trucking and other more impactful transportation demands as a result.
Examination of the Plan Bay Area Plan and its DEIR shows that this report fails to satisfy the requirements of SB375 and the technical requirements of the DEIR under CEQA because it fails to prove that any of the Alternatives will actually achieve the goal of reducing per capita or overall GHG emission from the use of autos and light trucks.
The DEIR analysis makes the common error of mistaking correlation with causation. It substitutes unscientific observations and unqualified statistics for proper scientific inquiry or demonstrable facts to arrive at what appear to be predetermined conclusions that are insupportable and inaccurate.
The DEIR attempts to persuade readers by inference and through anecdotal evidence rather than by doing the kind of specific and direct analysis as I’ve presented above. The DEIR offers a “take our word for it” approach but offers no detailed calculations or formulas, of any actual proof whatsoever to show the Plan’s efficacy in meeting the goals of SB375. Its statistical data relies on studies done by its partners (MTC, BAAMQ, etc.), whose objectivity and motivations must be questioned.
Based on the analysis presented here, the Bay Area Plan DEIR fails to fulfill the technical requirements under CEQA, and the Bay Area Plan and its Alternatives have failed to comply with the requirements and goals of AB32, SB375 and the SCS in reducing per capita or overall GHG emission. The analysis I’ve presented demonstrates that the Plan is actually more likely to increase per capita and overall GHGs than decrease per capita and overall GHGs.
For further discussion about the kinds of housing we really need in Marin, that will not result from the Plan Bay Area approach, please see The Bay Area Plan Fails to Solve Our Affordable Housing Needs In Marin County.
The Best Laid Plans:
Our Planning and Affordable Housing
Challenges in Marin
by Bob Silvestri