2020 ESA Annual Meeting (August 3 - 6)

PS 59 Abstract - Witness Tree Social Media Project: Lessons in science outreach from a novel climate ambassador

Shawna Greyeyes1, Clarisse Hart2, Kyle Wyche3, Andrew D. Richardson4 and Tim Rademacher4, (1)Coconino Community College, (2)Harvard Forest, Harvard University, Petersham, MA, (3)Biology, University of Pittsburgh, (4)Center for Ecosystem Science and Society; School of informatics, Computing and Cyber Systems, Northern Arizona University, AZ
Background/Question/Methods

The climate crisis is underway, but in the USA, the public remains largely unaware and inactive. Humans are imperfect climate sensors, and many tend to see climate change as either a partisan ideology or a geographically distant threat. To connect the public to the nature around them and how it is changing, we have pioneered a data-driven science communication technique called the Witness Tree Social Media project (@aWitnessTree).

In 2018, a team of post-docs, communication specialists, and undergraduates installed sensors on one 100-year-old red oak tree (Quercus rubra) at Harvard Forest in central Massachusetts. Open-source computer code analyzes the tree’s sensor data continuously, and specific conditions (a heatwave, heavy rainfall) trigger real-time posts to Twitter and Facebook following four themes: phenology, climate, community ecology, and tree physiology. The bot can contextualize current data with archival data, creating monthly/annual summaries and comparisons to 55+ year climate baselines.

Since the project launched in July 2019, it has attracted 8,700+ Twitter followers, many of whom engage directly with the bot, which can respond with selfies and answer a predetermined set of questions.

The project is both an outreach tool and a research environment. Initial design of the bot’s messages sought to test a range of science communication approaches, through posts with multiple, randomized, categorical treatments – presence/absence of multimedia, four content themes, and data-driven or conversational diction. Twitter analytics quantify resulting public engagement.

Results/Conclusions

We analyzed 65 Twitter messages posted by the bot between July 2019 and February 2020, focusing primarily on Twitter Engagement Rate (the number of times a user actively engaged with a post by clicking, sharing, or commenting, out of the total number of times the post passively appeared in news-feeds). Average Twitter engagement per post (2.8%) exceeded industry standards for non-profit social media accounts by an order of magnitude.

Multimedia in a post increased Twitter engagement (p = 1.61e-08), a result that aligns with prior social media research. Dominant content theme had no impact on overall Twitter engagement but altered the frequency of user replies (p = 1.57e-05), with climate themes eliciting more comments per post.

A 2020 grant will pilot deployment of 3 new Witness Trees at environmental education sites in greater Boston. We will assess public engagement online and explore qualitative impacts of the technology on climate change teaching in the K-12 classroom. The network will also enable data analysis on tree physiology along a land-use gradient.