Reducing annual maintenance by 50% and vegetation risk using LiveEO’s Treeline solution built on AWS with Hetch Hetchy

Hetch Hetchy delivers all of San Francisco’s water and produces power from four powerhouses, which it sells to entities like utility companies.

Hetch Hetchy’s lines run through high fire threat districts as well as beloved natural areas in Yosemite National Park and the Stanislaus National Forest. California regulations require vegetation to be at least 4 feet away from distribution lines and 10 feet away from 230,000-volt transmission lines in State Responsibility Areas to reduce wildfire risks. The company must balance preserving the natural environment with avoiding excess risk from vegetation overgrowth. 

Hetch Hetchy employs one forester to inspect all 45 miles of its distribution line and 324 miles of its transmission line once per year, referencing light detection and ranging (lidar) data for an overhead perspective and a handheld navigation device for wayfinding. The process was time consuming and required manual calculations using tools like range finders to determine whether vegetation needed to be removed or pruned. 

Hetch Hetchy wanted to increase efficiency after seeing a technology demonstration from satellite analytics company LiveEO. Built entirely on AWS, LiveEO’s Treeline solution uses artificial intelligence (AI) to analyze satellite data and visually indicate vegetation-related risks for companies in various industries, including utilities. Hetch Hetchy implemented Treeline, a software-as-a-service solution that is available from AWS Marketplace, where organizations can discover, deploy, and manage software that runs on AWS. It chose LiveEO because the solution was customizable and simple to use, with scalability, cost, and speed improvements compared with lidar.

With the scalability and comprehensiveness of AWS resources, Treeline can store the vast amount of satellite data required using Amazon Simple Storage Service (Amazon S3)—object storage built to retrieve any amount of data from anywhere—and support machine learning models and time-series data analysis. LiveEO uses Amazon Elastic Compute Cloud (Amazon EC2), secure and resizable compute capacity for virtually any workload, to process and analyze data. “Having all the data stored in the cloud and available to pull up on a phone, tablet, or computer is leaps and bounds ahead of the days where everything was on paper,” says Jim Payne, account manager at LiveEO. “It’s always with you using AWS.” 

Because lidar flights are subject to availability, weather delays, and visibility issues, Hetch Hetchy’s previous lidar flight took over 18 months to deliver insights. By the time the data was available, it was no longer actionable. Rather than relying on lidar flights, LiveEO can collect satellite images when cloud cover is low and deliver the data customized for Hetch Hetchy’s needs 8 weeks after imagery collection.

“Using Treeline, I can request the data in spring, get it in the summer, and hit the ground running before there’s a lot of change in the environment,” says Joy Mellera, right of way manager at Hetch Hetchy. 

Using predictive analytics, Treeline highlights vegetation that has the potential of growing into the regulated space around distribution and transmission lines in the next 1–2 years. Hetch Hetchy’s vegetation management team and tree crew can work more efficiently using the data, plan ahead to optimize resources, and remain in compliance. Treeline also highlights trees that could strike a line if they fell, and Hetch Hetchy’s forester assesses these trees every year, looking for risk factors like root uplifting or disease. “Treeline is much more visual, it’s quick, and it has a different risk color depending on the distance away from the conductor,” says Mellera. “It’s visually more appealing and informational than lidar points with classifications.”

In the 3 years since deploying Treeline, Hetch Hetchy reduced the number of annual outages due to vegetation by 50 percent. It also decreased the amount of annual maintenance for its tree crew by 50 percent because the company can stay on top of pruning and identify heavily vegetated areas to focus on to reduce future maintenance. Treeline saves the forester 260 hours of work every year, freeing up time for other tasks that reduce risks and increase reliability. He no longer needs to do manual calculations and can skip straight to tree health assessments. The AI analysis also reduces the margin for error from manual calculations. 

Prior to using Treeline, Hetch Hetchy’s crew would prune as they went, but there wasn’t a way to track decision-making or prove that the crew had inspected every distribution and transmission line. With Treeline, Hetch Hetchy can track progress over the years in an accessible way for stakeholders without forestry experience. “Our management team can use Treeline to understand how the vegetation management program is working and how it’s improving year to year without having to dive into spreadsheets and numbers,” says Mellera. “They can visually see that there’s less risk on the system and we’re doing what we should do.”

Hetch Hetchy uses Treeline’s data outside of vegetation management as well. If the company needs to shut off power in a remote section of its distribution line because of a wildfire, it can reference Treeline data to pinpoint where there is vegetation risk rather than conservatively shutting the power off further up the line. 

Hetch Hetchy also uses Treeline data to analyze areas where it would be advantageous to move lines underground. The company prefers to prune vegetation rather than remove trees whenever possible, but that means its crew prunes the same tree every few years. Using Treeline’s data, Hetch Hetchy can identify areas with high vegetation, strategically underground the lines, and reduce future maintenance.

Hetch Hetchy is considering analyzing satellite data twice per year and using it to address hazards on access roads to its facilities. It’s also exploring using AI to further mitigate risks by factoring in the impact of the tree species. For example, Hetch Hetchy is cataloging which trees are most hazardous, such as black oaks that experience sudden limb failure and eucalyptus trees that grow rapidly, so it can target maintenance effectively and anticipate risks. 

“The Treeline technology—built on AWS—gives us situational awareness,” says Mellera. “It helps us be less reactionary and more forward thinking.”