Comparison of life-cycle GHG emissions per package delivered for drone and ground vehicle pathways under base case assumptions. The analysis focuses on the final delivery of the package, after the package is delivered to the regional warehouse. Emissions from battery and fuels production, as well as fuels combustion and electricity production required for transportation and warehousing, are included. The range of regional greenhouse gas (GHG) intensities of electricity in the US is represented by comparing results from low-carbon California to relatively high-carbon Missouri. Additional warehousing requirements for drone and van pathways are included. The results show that small quadcopter drones across all US regions have lower life-cycle GHG emissions than conventional delivery trucks powered by diesel and natural gas, electric vehicle (EV) trucks in most regions, and gasoline-powered vans. Large octocopter drones are shown to have lower GHG emissions than diesel and natural gas vehicles only when charged with low-carbon electricity. Both small drones and large drones are shown to have lower GHG emissions than use of a personal vehicle to pick-up a single package. Stolaroff et al. Click to enlarge.

The researchers set out to find if drone delivery is a promising idea for the environment or whether, like conventional overnight package delivery, it leads to much higher energy use and carbon emissions.

The researchers flew test campaigns with two commercial drones and developed an estimate of the energy needed to deliver a package in various scenarios. They also considered how battery technology and drone design will improve over time. Then, using life cycle assessment, they compared the drone scenarios with delivery by truck, van and passenger car. The technique accounts for upstream impacts, such as emissions from manufacturing the batteries or refining oil into diesel fuel.

The researchers found that the current practical range of multi-copter drones is about 4 kilometers, which means a new network of urban warehouses or waystations would be needed to support a drone delivery system. These warehouses, in turn, would take energy to run. Although drones consume less energy per package-mile than delivery trucks, the additional warehouse energy required and the longer distances traveled by drones per package greatly increase the life-cycle impacts.

Overall, the results are mixed, and the best choice depends on things such as the size of the drone, the weight of the package and the types of power plants on the regional electricity grid. Drones are favored in regions with relatively clean electricity, such as California.

A light package—say, a pair of sunglasses—flown by a small drone over a few miles, saves a lot of energy and greenhouse gas emissions compared to a delivery truck. But a larger package—say, a computer monitor—flown by a drone large enough to carry it, probably does worse than a delivery truck.

—LLNL scientist Joshua Stolaroff and lead author

To make larger drones more efficient than trucks and vans, charging could be done only with renewable and low-carbon electricity, said Stolaroff. Devising creative ways to deliver goods from existing retail stores rather than building additional warehouses could also help, he said.

The bottom line is to pay attention to life-cycle impacts when designing both the drone and logistics network.

—Joshua Stolaroff

The researchers recommend that regulators and companies looking to get an environmental benefit from drones should consider the systemwide impacts, and focus their efforts on small packages, with larger packages being left for trucks and vans.

The work was funded in part by the Climate and Energy Decision Making center through a cooperative agreement between the National Science Foundation and Carnegie Mellon University.

Resources

• Joshuah K. Stolaroff, Constantine Samaras, Emma R. O’Neill, Alia Lubers, Alexandra S. Mitchell & Daniel Ceperley (2018) “Energy use and life cycle greenhouse gas emissions of drones for commercial package delivery” Nature Communications Volume 9, Article number: 409 doi: 10.1038/s41467-017-02411-5