You don't see many Montana plates in New Jersey, let alone one from Petroleum County.1 Of course it was a pickup - a F-150 that had clearly seen serious duty. I waited about ten minutes for the driver and his wife to come out. We chatted for awhile - long enough for my accent to return - and I left wondering what they thought of the change in population density as they crossed the country.
Winnett is the county seat of Petroleum County. With less than two hundred people it is hardly a metropolis. The entire county has fewer than 475 people. With an area of 4,336 km2 the county's population density is approximately nothing and hugely different from their next stop - Brooklyn.
It isn't a hard rule of thumb, but mass transit begins to become practical when the population density exceeds 5,000 to 10,000 people per square kilometer. Brooklyn blows through this at about 91,000 and Manhattan is weighs in at nearly 180,000. These dense regions can't deal with many single occupant cars - there isn't enough space on the roads or parking lots.
Autonomous vehicles are widely heralded as the agent that will change everything - car ownership, insurance, safety, pollution, on so on. While they do represent major change, most of the analyses in the popular and tech press ignore infrastructure, geography, sociology, population density and synchronization issues. Some of these may change easily, others will be rigid. Regionality in terms of vehicle type and use seems likely.
Dense urban areas are better served by public and active (eg. human muscle) transit. Autonomous vehicles may replace taxis, but they're unlikely to go much beyond that in very dense regions as the streets aren't big enough, and there won't be enough vehicles given parking constraints, to handle rush hours. Bus rapid transit (BRT) systems that eliminate many of the issues with bus routes have emerged and are offering subway-like performance at under a tenth a subway's price tag. China has been plagued with enormous city planning failures rooted in Soviet and American examples, but that is changing. Emerging cities are now being designed like Guangzhou's Liuyun Xiaoqu neighborhood - a very livable, diverse and self-contained city core. European cities have been taking different paths that may lead to similar results. Many of these point towards a Jeff Speck walkable city model. He has proposed simple changes for existing cities, but new cities offer a fresh chance.
Ideally, at least for transportation, we would live in urban areas with dense mass transit and active transit infrastructures. For good or bad much of a cities' infrastructure and layout changes slowly - very very slowly. Beginning in the early 1920s the private automobile brought that coincided with a large population increase and migration. The inertia of the infrastructure and its geography has made city and suburban planning relatively stagnant for about fifty years.
At some point population density falls to a level where mass and active transit aren't practical. These regions are very common in the US - from old first ring monoculture suburbs near major cities to sprawling suburbia to rural America. The types of autonomous cars and ownership will be partly dependent on population density. There is the logistics problem of making sure a car will show up in some brief amount of time. As population density drops further a car may be a few miles away when you request it or it may have other users in the queue. What if you have a family emergency? While cases like these may be rare, people tend to think in terms of use cases that might happen - witness the fact that a pickup is the most popular personal vehicle in the US. Modeling this is difficult as so many variables including social attitudes come into play, but I suspect the population density sweet spot for autonomous vehicles may be in the 1,000 to 5,000 person per square mile range. Owner driven pickups are a safe bet in rural Montana for some time.
To get insight into vehicle type and use models it is necessary to look at other factors in addition to population density. A major area is energy type and pollution. A few variables to consider:
° Carbon intensity Well-to-wheel carbon emissions per unit of energy (grams carbon dioxide per kilowatt-hour or grams carbon dioxide per gallon of gasoline). Tar sands oil is particularly bad. Burning coal to make electricity for an electric car can be worse than burning gasoline in the car's engine, producing hydrogen from hydrocarbons for a fuel cell electric vehicle is far from clean... Getting these numbers right can be tricky - corn ethanol being an excellent example. There will be an accelerating trend of vehicle electrification with enormous infrastructure implications as the shift takes place. It is likely the change won't be smooth everywhere.
° Energy intensity The energy required to move a vehicle a distance - kilowatt-hours per mile, liters of gas per 100 kilometers, miles per gallon (actually its inverse). We've seen improvement in vehicle powertrains since the 80s. Much of this was erased by increasing vehicle weight from safety requirements and expectations along with a perceived need for larger vehicles. Advances are still coming, but the easy steps have been taken. One interesting possibility is the advent of the one or two seat urban/suburban vehicle for speed limits under 35 km/hr (about 22 mph). At such low speeds crashes are more survivable and vehicles can become vastly lighter and, as a result, more efficient. Some planners are talking about even slower speeds for autonomous vehicles. Novel autonomous traffic control at slow and accelerations speeds in dense regions promise gains in efficiency with little or no increase in transit time.
° Trips and distance Simple - sort of - how many trips and how long? This varies widely causing a variety of expectations and vehicle compromises.
° Occupancy It makes sense to carpool with autonomous vehicles during rush hours. Van-like structures may be important. One workshop suggested 12 to 16 passenger vans for London with transfers for longer commutes. Integration with mass transit in major urban areas is essential.
° Embedded energy How much and what type of energy was used to make the car? Very roughly a gasoline fueled sedan requires about the same amount of energy to make as it uses in the first year or two of operation. The energy used to build it is often very polluting and carbon intensive.
All of these need to be considered together. One issue with current ride sharing and most practical shared autonomous vehicle schemes is an increase in vehicle miles traveled even though the number of vehicles drops. One study looked at a factor of ten drop in vehicles and found total miles travel increased by over ten percent. Even now some areas this has been shifting people from mass transit to Uber/Lyft/... with passenger occupancy rates not much above 0.5 passengers and an increase in pollution.
This is a rich and deep area and I haven't touched on many other issues. Working through them allows you to flesh out scenarios and begin to understand where they may or may not apply. The only solid predications I can make is if you are currently driving your own car, it is likely you will be doing the same ten years from now. The complexity of the problem and fiefdoms of cities and countries insures detailed predictions of future that suggest single models are likely wrong. I can imagine at least three very different classes of vehicle and none resembles the current family car. China is certainly ground zero from about 2025 to 2040, but I should think India and several African nations - most notably Nigeria - may be equally important in the long run.
Feel free to get in touch if you're interested in greater detail..
___________
1 Montana has 56 counties and tag numbers are assigned by a county prefix based on the population rank of the county in about 1930. Many natives carry county number ID cards and a fair number, I was once one, know all 56 by heart. Jeri will notice this is not the current design - it is from a plate generator page. The number 42 is obvious.
__________
Recipe Corner
Late Spring has brought Summer heat. That means ice cream and its relatives. I had an avocado sorbet in NYC .. that got me experimenting and this came out even better.
Avocado Sorbet
Ingredients
° about 1-1/2 ripe avocados (one of mine was damaged) peeled and pitted
° 1 can (13.5 oz) Thai full-hat coconut milk
° 1/2 tsp salt (a starting point)
° 1/4 cup cocoa powder ... this means the real thing
° 1/2 cup maple syrup. I used grade A, grade B should be just as good and perhaps more flavorful
Technique
° blend all but the maple syrup in a food processor and then add the maple syrup until it is mixed in
° pour into your ice cream maker and follow the machine's instructions
° this doesn't freeze well, so eat it immediately. That shouldn't be too much of a problem