Author: nick

This is part three of a three-part series exploring the environmental lifecycle analysis completed by Divergent. It will get into some technical detail, but is intended for anyone who has an interest in total system environmental and health damages for manufacturing.

Before reading this section, please read our lifecycle analysis overview, Part 1, and Part 2.

Modeling philosophy: change as few variables as possible

There are obviously many ways to create and modify models. Thankfully, by using GREET and AP2, we were able to leverage decades of existing work. These peer-reviewed, published, and transparent models have inputs, calculations, and outputs that have been highly vetted.

When modifying an existing model, it behooves the modeler to make as few changes as possible. That is the philosophy we used at Divergent when making modifications to include our vehicles in the existing models. By reducing changes, you’re reducing chances for error and reducing the number of assertions you need to make. It’s also easier to delineate the full set of changes you made so that others can review and understand your work.

Similarities and differences between our analysis and NAS analysis

We based our modeling on the method used in the 2009 report by the National Academy of Science (NAS), Hidden Costs of Energy: Unpriced Consequences of Energy Production and Use. They used GREET for lifecycle emissions calculations and APEEP (the predecessor to AP2) for health and environmental system damage cost calculations.

We used updated versions of those same two models, GREET and AP2. The models have not fundamentally changed in structure since 2009, though some of the content has been updated with more recent calculations (e.g., the environmental cost of CO2 emissions).

We updated GREET to include modern electric vehicles with 85 kWh batteries and to also include the Divergent vehicles, both compressed natural gas (CNG) and gasoline.

cents per vmt graph-1

Delineation of variable changes

In keeping with the modeling philosophy above, we limited the set of changes in the GREET model to only those changes necessary to include modern EVs and our Divergent vehicles. We used the existing lightweight car vehicle type in GREET and modified it to suit the Divergent Blade. No changes were made to AP2, as it takes inputs from GREET. Below is a table delineating the exact changes we made.

That’s all. By limiting the number of changes, we have a strong and defensible modeling technique. Divergent vehicles drastically reduce the environmental and health damages compared to existing, traditional vehicle manufacturing. Keep in mind that we modeled the Divergent Blade, a 0-60 in 2.5 seconds 700 horsepower supercar. We can expect that future Divergent vehicles will have even lower total system damage numbers.


Thank you for joining us in our mission to reduce the environmental impact of automobile manufacturing. We hope this blog series has been informative, and welcome your thoughts and comments in the space below.

This is part two of a three-part series exploring the environmental lifecycle analysis completed by Divergent. It will get into some technical detail, but is intended for anyone who has an interest in total system environmental and health damages for manufacturing.

Before reading this section, please read our lifecycle analysis overview and Part 1 of this series.

Introduction to AP2

In Part 1, we introduced the Argonne National Lab’s GREET model. Developed over the last 20 years, it is a complicated and deep model that outputs greenhouse gas and particulate emissions.

AP2, formerly called APEEP, is a model developed by Nicholas Muller, Associate Professor of Economics at Middlebury College and Visiting Associate Professor at Carnegie Mellon University. AP2 takes the outputs from GREET and translates them into a dollar figure – the environmental and human health cost of those emissions.

From the AP2 website: “The Air Pollution Emission Experiments and Policy analysis (APEEP) model is an integrated assessment model that links emissions of air pollution to exposures, physical effects, and monetary damages in the contiguous United States. The model has been used in many peer-reviewed publications,” including a paper in Science in August 2014.

For example, in our recent analysis of various vehicle types, AP2 received a total CO2 count from GREET for the manufacture and operation of a vehicle. AP2 translated this CO2 count to an environmental and health system damage cost for the lifetime of the vehicle. To standardize this number, we then divided the cost across the total miles that car will be driven in its lifetime (estimated to be 160,000 miles per vehicle). The output from AP2 is then a cents (US dollars) per vehicle mile traveled (VMT).

History of GREET + AP2

The 2009 National Academy of Sciences report “Hidden Costs of Energy: Unpriced Consequences of Energy Production and Use” used this same dual-model approach, connecting GREET and APEEP to show the combined damages. This report was one of the first to show the full lifecycle emissions of various types of vehicles and their environmental and health system cost. It revealed costs that were previously hidden, bringing into question some of the policies and trends in vehicle manufacturing.

This report was one of the founding inspirations for Divergent. Seeing this report, we believed we could and should do better: build a more environmentally sustainable car.

Connections between the GREET and AP2

GREET and AP2 have a fairly simple point of connection. We analyzed six vehicle types:

  1. Gasoline (25 mpg)
  2. Hybrid gasoline (40 mpg)
  3. Plug-in electric car (85 kWh)
  4. Plug-in electric SUV (85 kWh)
  5. Divergent compressed natural gas
  6. Divergent gasoline

For each of these vehicles, we used GREET to calculate greenhouse gas and particulate emissions over the lifetime of the vehicle, both for manufacture and operation. The chart below shows the conceptual connections between the models.

GREET and AP2 connections

We then used the outputs from GREET as inputs to AP2. The table below shows the most important outputs from GREET.

table of outputs from GREET

Using AP2, costs for each of the various pollutants were calculated and then tabulated. We also include the cost of the greenhouse gases directly, estimated by the US government at $39 per ton of CO2 equivalent.

table of outputs from AP2

The Vehicle production, Fuel production, and Operation lines under the Health Damages (per VMT) are the numbers that are used in the graph below.

cents per vmt graph

What’s next in this series?

In the next (and last) posting in this series, we’ll discuss the input variables to GREET, as well as our philosophy of changing as few of those variables as possible.

This is part one of a three-part series exploring the environmental lifecycle analysis completed by Divergent. It will get into some technical detail, but is intended for anyone who has an interest in total system environmental and health damages for manufacturing.

Before reading this section, please read our lifecycle analysis overview, which sets the context for the discussion here.


My name is Nick Hofmeister. I have a computer science degree from Northwestern and an MBA from MIT. I’ve worked in fairly diverse industries, starting at Microsoft, followed by Bain & Co., an algae biofuels company, and three biotech/software startups I’ve founded. Over the years, my teams and I have raised about $260 million for early stage companies. I started making environmental analyses about 7 years ago to demonstrate the lifecycle value of a fuel made directly from a photosynthetic plant.

Why do we care about lifecycle analyses?

Around 2 billion cars have been built over the last 115 years; twice that number will be built over the next 35-40 years (Emmott, Stephen, Ten Billion, p. 95, New York, New York: Random House, Second Edition, 2014). The environmental and health impacts will be enormous. The automotive industry has $9 trillion a year in revenue and employs 11 million people.  And that doesn’t include oil & gas, a $4 trillion industry employing another 1.3 million people. These giants greatly affect the world’s material flows and all the consequences thereof. Unfortunately, very little has changed in the manufacture and use of vehicles in the last 100 years – our cars are getting bigger, heavier, and often less efficient at using energy. At Divergent, we intend to change this trajectory.


There are many types of environmental analyses, and they vary in both scope and kind. For this discussion, we are focused on environmental lifecycle analyses. Lifecycle analysis follows the path of an object through a system for its full lifetime, looking at its environmental impact at each stage of life. Lifecycle analyses can be performed on cars, toasters, a strawberry, or even a human being.

The exact scope of a lifecycle analysis is one of the trickiest questions: which parts of the life cycle do you include, and which not? For the strawberry, do you include the cost of making its fertilizer, the cost of its transportation to the grocery store, the cost of its disposal (e.g., the green tops, the rotten berries), and the cost of its packaging? The total cost of environmental and health system damages varies widely depending on the scope.

The gold standard in vehicles – GREET

In vehicles, thankfully, the Argonne National Lab has set the gold standard through a model called GREET (Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model). The first use of GREET was in 1995, and it has been peer- and industry-reviewed very thoroughly in the intervening 20 years. The scope of this model is quite wide – it looks at both vehicle operation and manufacture (vehicle cycle) and fuel manufacture and use (fuel cycle). It is also quite deep, capturing emissions down to the extraction and manufacture of materials (e.g., aluminum, lithium), as described in the image below:


The Argonne GREET model was also used for the 2009 National Academy of Sciences report “Hidden Costs of Energy: Unpriced Consequences of Energy Production and Use”, mentioned in the introductory article of this series).

What is the scope of GREET?

This model is very detailed and robust. It contains more than 100 fuel production pathways, such as corn ethanol, gasoline from algae, and cellulosic ethanol from switchgrass. It contains more than 70 different vehicle types, from a standard gasoline car, to an electric car, to a lightweight compressed natural gas vehicle.

GREET also looks backward into the manufacture of a vehicle quite deeply. As an example, consider the chassis of a vehicle. In GREET, even the transmission system (i.e., gearbox) is broken down into 8 major material types, each of which is accounted for separately in the model:  steel, copper, cast iron, magnesium, wrought aluminum, cast aluminum, average plastic, and rubber.

This level of detail, both in breadth and in depth, is necessary to create an accurate picture of the emissions of the vehicle across both its manufacture and its operation.

Inputs and outputs

GREET is free and publicly available for download. It is also, essentially, open source: if you download the Excel version, you can change any variable in the system that you wish.


However, to reduce the complexity in the model and variations in the answers, there are a limited number of inputs that are commonly changed. These include:

  1. Vehicle type
  2. Fuel type
  3. Total weight
  4. Balance of materials

Vehicle type has options such as standard internal combustion gasoline vehicle, electric vehicle, and lightweight compressed natural gas vehicle. Fuel type is tied with vehicle type, but there are many variations for each vehicle type. For example, an electric vehicle can use only electricity, but that could be electricity from the U.S. National Grid or from a coal-heavy region like the Midwest. A gasoline vehicle can use fuel with 15% corn ethanol or 5% cellulosic ethanol.

Total weight is a high-level, summarized factor that greatly affects the outcome of the GREET model. A Tesla Model S weighs approximately 4,700 pounds: every foot traveled by that vehicle is a result of moving 4,700 pounds of material. And at least 4,700 pounds of material needed to be manufactured in order to make the car (in reality, due to wasted materials, it is much more than that). In contrast, the Divergent Blade is 1,400 pounds, which makes it easier to move and easier to manufacture from an environmental standpoint.


There is also a common set of outputs created by GREET, showing the emissions of a vehicle across both its fuel cycle and its vehicle cycle. These are calculated on an energy/mile or grams/mile basis, depending on the output. Here are the most common outputs:

  1. Total energy
  2. Fossil fuels
  3. Coal
  4. Natural gas
  5. Petroleum
  6. Water consumption
  7. CO2 (VOC, CO, CO2)
  8. CH4
  9. N2O
  10. Total greenhouse gases (GHG)
  11. Volatile organic compounds (VOC)
  12. CO
  13. NOx
  14. Particulate matter large (PM10)
  15. Particulate matter small (PM2.5)
  16. SOx
  17. Black carbon (BC)
  18. Organic carbon (OC)

If you look at the examples included in GREET, it’s worth noting that a gasoline car uses a fair amount of natural gas. Why? Natural gas is used to make electricity, and electricity is used to make the materials that comprise a gasoline car and is used for its assembly. An electric vehicle typically uses a fair amount of coal, since coal is a major factor in the U.S. electricity grid.

What’s next in this series?

In the next posting in this series, we’ll discuss AP2 (formerly known as APEEP), which translates the output of GREET into environmental and health system damages.