PaperlayerView exact revision on arXiv

Source-linked AI summary

Geospatial Dynamics of Online Attention to Nuclear Power in Japan Before and After Fukushima

Yifan Sun, Ken Kurosaki, Benjamin C. McLellan

arXiv:2610.29924v1cs.SI

TL;DR

Surveys rarely capture what people want to know in the moment, and Japan lacked systematic evidence on nuclear-related search demand over time. This study analyzes prefecture-level Google Trends searches from 2004–2024, finding routine host-locality attention and a post-Fukushima distance pattern, with search framing varying by proximity.

  • Problem

    Japan lacked systematic, data-driven evidence on what people sought about nuclear power and how that demand changed over time, limiting evidence for matching public information to demand.

    Source: Introduction · p. 2

  • Method

    The authors analyze 2004–2024 prefecture-level Google Trends searches, linking them to reactor status to map information demand across space and time.

    Source: Introduction · p. 2

  • Results

    Routine attention was higher in reactor-host prefectures with little neighboring spillover; after Fukushima it declined with distance, and query types varied by proximity.

    Source: Discussion · p. 8

  • Takeaways & Limitations

    Search patterns can inform public engagement and crisis communication, while accessible foundational content during routine periods may support education about concepts relevant in emergencies.

    Source: Introduction · p. 2, Discussion · pp. 8–9

Abstract

from arXiv · show

Nuclear power is increasingly seen as necessary for sustainable development, but this role depends on informed public engagement and support. Building this support requires understanding what citizens want to know, as reflected in their spontaneous online searches. To map this demand, we present the first long-term (2004-2024), prefecture-level analysis of Google Trends data on nuclear power in Japan. The results indicate that in routine periods, search interest followed a host-locality pattern, being significantly higher in prefectures hosting nuclear plants, with little spillover to their neighbors. Immediately after the Fukushima Daiichi accident in March 2011, this pattern changed, and interest instead declined with distance from Fukushima, before the host-locality pattern returned from 2012. During March 2011, searches leaned heavily toward general and live information rather than guidance on what to do. Their framing also varied with proximity, as those near Fukushima prioritized live updates and formal countermeasures, while others leaned toward background information and personal-scale questions. These patterns and the shift between them provide a behavioral evidence base for designing communication strategies in Japan and countries weighing the use of nuclear power. More broadly, the finding that searches vary with proximity can inform public communication in other geographically anchored crises.

Introduction

Survey evidence describes attitudes toward nuclear power, but does not capture what people seek during unfolding events; Japan also lacked a systematic analysis of nuclear-related searches over time. This study uses two decades of prefecture-level Google Trends data to map online information demand across space and time.

Introduction

  • Before Fukushima, nuclear power supplied close to 30% of Japan’s electricity; the post-accident shutdown increased reliance on imported fossil fuels and raised fuel costs.The shutdown posed challenges for Japan’s energy security, economy, and environment.Source: Introduction · p. 1
  • Surveys document attitudes and framing effects, but can be affected by reporting bias and are poorly suited to capturing information demand during unfolding crises.Voluntary searches offer a complementary signal made at the moment of need and continuously recorded.Source: Introduction · p. 2
  • Google Trends has been used in health, polling, and economic research, but nuclear-related search studies were limited to international Fukushima reactions and qualitative searches for “radiation.”The cited examples include influenza-like-illness estimates one to two weeks ahead of official reports.Source: Introduction · p. 2
  • The study asks how routine prefectural attention is distributed, how Fukushima reshaped it, and what people searched for afterward across places.These are the paper’s three research questions.Source: Introduction · p. 2
  • The analysis links prefectural searches to reactor status and uses consistent, continuous Google Trends data from 2004–2024 to trace demand across space and time.The authors present this as the first long-term, prefecture-level analysis of nuclear-power searches in Japan.Source: Introduction · p. 2
  • The findings are intended to help governments, host communities, and operators match information to public searches, and may inform engagement in other geographically anchored crises.The paper also frames the analysis as relevant to countries weighing nuclear power use and expansion.Source: Introduction · p. 2

Results

Search interest shifted from a host-locality pattern before Fukushima to a distance-from-Fukushima pattern in 2011, then returned to host-locality from 2012. March 2011 searches also differed by place and by whether people sought explanations or guidance.

Routine period (2004–2010): Localized attention

  • Before the accident, host prefectures had significantly higher search interest than non-hosts, and elevated interest generally did not spill over into neighboring prefectures.Border status alone had no significant effect; the authors describe this as a host-locality pattern.Source: Routine period (2004–2010): Localized… · pp. 3–4, Figure 1 · p. 3

Fukushima 1F accident period (2011): Proximity-driven attention

  • In 2011, search interest declined significantly with distance from Fukushima (Spearman’s ρ = −0.54, p < 0.001), while host and non-host prefectures no longer differed significantly.Interest fell to roughly half its post-accident peak within about a week and around 20% within five weeks, similarly across distance groups, and returned close to pre-accident levels within a year.Source: Fukushima 1F accident period (2011):… · pp. 4–5, Figure 2 · p. 4

Recovery and restart period (2012–2024): Reversion to localized attention

  • From 2012, the host-locality pattern re-emerged outside Fukushima and strengthened after reactor restarts began in 2015.Host prefectures had significantly higher interest in all three periods; the probability of higher interest rose from 0.857 to 0.925 to 0.956 across the shutdown, initial restart, and recent periods.Source: Figure 3a · p. 5, Figure 3 · p. 5

Content Analysis of Search Queries

Explanatory queries

  • Update queries dominated nationwide during the first three days, then contracted toward Tohoku and Kanto and later Miyagi and Fukushima, while definition queries led farther away.Definition queries focused on nuclear energy and radiation; update queries covered topics including transportation, electricity, and gasoline. Because the earthquake and nuclear accident occurred together, the demand for updates cannot be attributed to either event alone.Source: Explanatory queries · pp. 6–7, Figure 4 · p. 6

Action-oriented queries

  • During March 11–17, Fukushima was the only prefecture where countermeasure queries had the larger share; how-to queries were more prominent elsewhere.The spatial pattern was less distinct than for explanatory queries, and the two action-oriented query types covered many of the same themes.Source: Action-oriented queries · p. 7, Figure 5 · p. 7, Table 2 · p. 7

Discussion

Online attention shifts from a routine host-locality pattern to Fukushima-distance effects after the accident, then returns to host-locality; search content and specificity also vary during the crisis. These findings inform geographically targeted communication and baseline education.

Discussion

  • Routine attention is substantially higher in prefectures hosting operational or planned reactors, with generally small spillover; during and after Fukushima, the spatial pattern shifts with distance and later returns to host-locality.After 2011, Fukushima retained the highest interest, while the proximity-driven pattern did not last; host-locality re-emerged and strengthened as restart decisions returned nuclear policy to local agendas.Source: Discussion · p. 8
  • Routine public engagement should define its audience by who is close enough to be affected, including across prefectural borders, and host-prefecture channels matter again as reactors restart.Tottori’s high interest among prefectures without reactors is attributed to part of the prefecture lying within 30 km of the Shimane plant.Source: Discussion · p. 8
  • During March 2011, searches near Fukushima favored live updates and formal countermeasures, while farther prefectures leaned toward background information and personal-scale how-to questions.The post-accident surge faded at a similar rate everywhere, so proximity shaped search intensity rather than how long the surge lasted.Source: Discussion · p. 8
  • March 2011 searches were often foundational rather than technical, and queries grew more specific as the month progressed.The authors suggest emergency communication should not assume high technical literacy and should keep pace with deepening information demand.Source: Discussion · p. 8
  • Search behavior during emergencies can inform educational efforts that build accessible foundational knowledge during routine periods and prioritize the few concepts that matter most.The authors suggest this approach may also apply to other fields with complex concepts that are seldom needed but can become critical during emergencies.Source: Discussion · pp. 8–9

Methods

The analysis uses Google Trends search measures across prefectures and time windows, estimates distance to Fukushima, and applies nonparametric group comparisons and rank correlations.

Google Trends data

  • Search Interest for the topic “原子力発電所” measures its relative frequency among searches, normalized to 100 within each time window and geography; Google Trends topics group related queries and cross-language variants.The topic was selected after comparison with alternatives such as “nuclear energy” and “nuclear power.”Source: Google Trends data · p. 9
  • The study collected prefecture-level Search Interest for five nonoverlapping periods and weekly 2011 values, plus March 2011 comparison-mode data for explanatory and action-oriented term pairs and related queries.Period values were normalized to the highest-interest prefecture in each window; weekly 2011 values were normalized to each prefecture’s own 2011 peak.Source: Google Trends data · p. 9

Geo-location and nuclear status

  • Distance from Fukushima was measured as GRS80 geodesic distance between prefectural government offices, and historical reactor status came from Japan Atomic Industrial Forum data.Source: Geo-location and nuclear status · p. 9

Statistical analysis

  • Mann–Whitney U tests compared Search Interest distributions between prefecture groups, while Spearman rank correlations assessed the relationship between Search Interest and distance from Fukushima.Both tests were two-sided with α = 0.05; the group comparisons also report common-language effect size, termed probability of higher interest.Source: Statistical analysis · p. 9
Loading 2610.29924v1…