The species awareness index as a conservation culturomics metric for public biodiversity awareness

Although threats to global biodiversity are well known, slowing current rates of loss remains a challenge. The Aichi targets set out 20 goals on which the international community should act alleviate decline, 1 (Target 1) aims raise public awareness importance biodiversity. conventional indicators for Target low spatial and temporal coverage, conservation culturomics metrics show how can be quantified at scale. Following methods used Living Planet Index, we devised species index (SAI) measure change in based Wikipedia views. We calculated this page level 41,197 listed by International Union Conservation Nature (IUCN) across 10 languages >2 billion views from July 2015 30 March 2020. Bootstrapped indices page-level SAI showed that overall increased marginally over time, although there were differences among taxonomic classes languages. Among classes, fastest reptiles slowest amphibians. languages, Japanese Chinese German users. as whole was significantly higher traded species, January 2016 through 2020, appeared not strongly related whether is or pollinator. As data source awareness, could integrated into Biodiversity Engagement Indicator. El Índice de Sensibilización Especie como Medida Culturomia la Conservación para Pública por Biodiversidad Aunque las amenazas biodiversidad mundial son bien conocidas, reducir tasas actuales pérdida todavía es un desafío. Los objetivos establecieron metas cuales debe actuar comunidad internacional aliviar declinación biodiversidad. Una estas (Objetivo busca sensibilizar al público sobre importancia los indicadores convencionales del Objetivo tienen una baja cobertura espacial y temporal, medidas culturomia conservación muestran cómo sensibilización puede cuantificarse escala global. Seguimos métodos utilizados el Planeta Viviente diseñar índice especie (ISE) medir cambio en con base vistas Wikipedia. Calculamos este nivel página especies incluidas listas Unión Internacional Naturaleza (UICN) diez diferentes idiomas más 2 mil millones entre julio marzo índices arranque ISE mostraron que general incrementó ligeramente tiempo, aunque hubo diferencia clasificaciones taxonómicas idiomas. Entre taxonómicas, rápido lento anfibios. idiomas, usuarios japonés chino alemán. su totalidad fue significativamente alta comercializadas, enero 2020 pareció no estar relacionado fuertemente si polinizador o comercializada. Como fuente información pública biodiversidad, podría ser integrado dentro Indicador Participación Biodiversidad. 尽管全球生物多样性面临的威胁已众所周知, 但要减缓目前生物多样性丧失的速度仍然是一项挑战。 “爱知目标” 提出了 个目标来要求国际社会采取行动以缓解生物多样性下降, 其中一个目标 (目标 旨在提高公众对生物多样性重要性的认识。虽然目标 的传统指标的时空覆盖度较低, 但保护文化组学指标展示了如何在全球范围内量化生物多样性意识。我们效仿地球生命力指数 (Living Index) 的方法设计了一个物种认知指数, 基于维基百科的浏览量来衡量人们对物种认识的变化。针对国际自然保护联盟 列出的 个物种, 我们在 种语言的维基百科上基于 年 7 月 日到 3 日间超过 亿次页面浏览量计算了这个指数。基于网页的物种认知的引导指数显示, 尽管在类群和语言之间存在差异, 但人们对生物多样性的总体意识随时间推移都稍有提高。在不同类群中, 人们对爬行动物的整体认识增长最快, 而两栖动物的增长最慢。在不同语言中, 日语使用者的总体物种认知增长最快, 而中文和德语使用者的则增长最慢。尽管人们对物种的认识整体上有所提高, 且对贸易物种的认识明显更多, 但从 月到 月物种认知的变化似乎与该物种是否受贸易或是否为传粉者无关。物种认知指数作为公众生物多样性意识的数据来源, 可以纳入国际生物多样性保护参与指标之中。【翻译 :胡怡思 ;审校 :聂永刚】 challenge (Mace et al. 2018; IPBES 2019). Two problems requirement transformational behavioral economic (IPBES 2019) difficulty leveraging Strategic Plan 2011−2020, underpinned Targets, represents an effort guide these changes (UNEP CBD 2010). Specifically, Targets decline Three have sufficient suitable (6, 9, 11), 4 intermediately (4, 7, 12, 14), insufficient (1, 5, 8, 10, 13, 16, 17–20), none (Mcowen 2016). Concerned with biodiversity, states aware value Conventional (i.e., barometer [UEBT 2019]) coverage (Leadley 2013; Mcowen 2016) do incorporate itself species). Without robust capturing evidence toward achieving target 1, determining has been met will hard. emerged field concerned digitized human–nature interactions (Ladle 2016; Sherren 2017). Quantifying area active interest. Using sources, such Twitter, Facebook, Flickr, Wikipedia, Google Trends, number researchers shown online improve understanding perceives environmentalism (e.g., Mccallum & Bury Roberge 2014; Papworth 2015). More recently, explored sources combined build single indicator awareness. For example, Cooper (2019) examined frequencies keywords social media, newspapers, internet searches, reasoning relative reflect issues. A significant step forward applying culturomic approaches development indicators, provided framework future research. Given their focus issues, potential improvement changing itself. represent powerful quantifying Page quantify interest (Roll phenology (Mittermeier In context valuable pages linked explicitly scales. Pages exist taxa multiple levels, red-list statuses, ecological systems unambiguous link between taxon identity reveal response natural history documentaries, demonstrating informative long-term (Fernández-Bellon Kane 2020). Moreover, because characteristics provide mechanistic ecosystem services, particular proxy its contribution. increasing contribute pollination trade indicate greater importance. specifically, particularly important, given reported declines animal pollinators Hallmann 2017; Powney using without limitations caveats (see Discussion), it provides basis useful new indicator. An metric thought analogous Index (LPI). LPI aggregation vertebrate population trends (Loh 2005; Collen 2009; McRae 2017), showing average rate populations. Treating size, method similarly applied derive Multiple search infer specific Fink 2020; Fukano Lenda 2020), but far know, one aggregated evaluated approach frequency Red List (hereafter IUCN species) call (SAI). then variation metric, aiming assess changed. pooled 6 distinct classes—including core pollinating groups (insects, birds, mammals) most heavily vertebrates (reptiles, mammals, amphibians, ray-finned fishes)—and each class top (by user) (Arabic, Chinese, English, French, German, Italian, Japanese, Portuguese, Russian, Spanish). modeled function class, language, contribution, contribution pollinator derived academic literature named-entity recognition. considered might other more holistic view API (application programing interface), software written Python, download daily user period 2015–31 (downloaded 16–21 April downloaded all group reptiles, fishes (Actinopterygii), insects, amphibians language projects retrieved our list OneZoom (Rosindell Wong Rosindell map species’ scientific name, identifier (ID), Wikidata Q identifier, main name language. Downloading only excludes redirect thus controls caused URL reach page. Each ID unique database (IUCN whereas appears. refer individual distinguish use page, visitor recorded human, excluding automated bots). Mittermeier (2019), able retrieve before date archived page-view API. returned, month, kept those series represented months Appendix S6 complete series). rather than total does always return days month. To account Wikipedia's popularity use, also random 11,000 Random API, requests pages. same manner again time series. From views, removed any appearing initially sampled maximize remaining after removing incomplete explore varied built combining text analysis manual inspection Millard [2020] S1 detailed methods). released Scheffers Food Agriculture Organization (FAO) fisheries statistics (FAO 2020) compile squamate fish. merged onto make either Schefers FAO Before calculating SAI, briefly absolute defined 31 generalized linear mixed-effects models: model log10 article (yes no), interaction trade, effect pollination, Rather attempting find parsimonious model, determined full predicted values, AIC values candidate null models described Appendices S28 S29. measurement per Because measures within weighted equally irrespective popularity, meaning highly viewed dominate SAI. Hereafter, level, SAIs bootstrapped (Fig. 1). monthly period. All above tortuosity S12), smoothed loess regression (span = 0.3) transforming back change. After smoothing above, (across languages) averaging P. tigris Q19939, (irrespective language) identified Q19939 tigris. SAIs. Bootstrap confidence intervals taking 2.5th 97.5th percentiles 1000 step. check extent influenced jackknifed marked trend (Appendix S11). averaged above. influence class-level trends, S13). calculate combination. 2016–January This yearly periods control seasonality. robustly differed various groups, constructed fitted interaction. One interaction, status (Appendices S30, S31, S32). With exception made analyses processing carried R 4.0.3 (R Core Team series, initial included approximately 2.23 These 1735 Views 24.92 million Arabic 1.08 English S4). highest 13,571 lowest insects 2743 S4; breakdown S6). subset proportion specifically (approximately 35%) 38%). Most had least 80% grouping present set, consisted 2.82 113,622 S8), days. 629.85 87.94 S8). 3486 9174 S9). consistently S16). contrast, lower bottom 8 Some uniquely high classes. Across (total views) (F 15206.44, p < 0.001) S5 S24), different 0.3869, 0.5339) S23). markedly affected inclusion French S11), so excluded level. exclusion declined June 2017 May 2018 2b). increase largely variable baselines S22), negative mid-2017 mid-2019. At jackknifing affecting fishes, birds increase, decrease 2a). Birds experienced peak early 2b), driven 3). Awareness mammals steadily increased, both dropped considerably start mid-2016 increasing, cause unclear. seasonal peaking July–August year; notable Average S27). fastest, slowly (with Wikipedia). Portuguese Wikipedias Wikipedias. S26 S19) S25 S19). It enables investigation variety variables, taxonomy, geographic distribution, service provision. found marginally; increases overall, some insects) decreased marginally. Furthermore, results suggest likely Wikipedia-derived additional source, explicit scales family, class). culture perceived widely recognized (Daniel 2012; Roll 2019; Ladle 2019), drivers complex. Overall capture many making difficult isolate causes decrease. shows consistent 5 hypothesized fish may consumption; seafood demand China recent years However, brief analysis, difference nontraded S18). indicates consumption alone. note, concordant surveys Japan largest percentage point familiarity term (UEBT unclear what driving Counterfactual scenario modeling relationships, demonstrated studies (Acerbi Fernández-Bellon Veríssimo complex, conspicuous did differ nonpollinating animals. relate value. But relationship appears weak nonexistent. reasons, case. First, impact dependent; associated larger Second, unlike direct food), indirect people, benefit less intuitive. Third, nature often deliberately sought who central role work required understand provision makes publicized (Smith Saunders 2016), would expect if relatively service-providing species. Indicator (BEI) independent metrics, emphasize platforms inferring Particularly itself, changing. Combining BEI presents challenges. units measure. scaled 0–100 scale Trends. benchmark inspired LPI. Namely, country cuts countries scale, simple. solution rescale disaggregate users 5). weighting language-level format amenable web scraping (Wikimedia Traffic Analysis Report, 2018). Such solve differing scales, national newspaper, Trends scores Despite providing novel approach, online-derived subject S3). Primarily, proxies grouping, exposure species-related information inferred. cannot tell why visited retained. On-site Analytics suite intimate reasons visit (Soriano-Redondo site publicly available. Text mining help type exposed similarity reference indication salience Similar climate invasive species; threat-related terms threat (Jarić penetration continues digital become important. marginally, inconsistent believe realm. thank Y. J. mapping T. Newbold, R. Gibb, CBER Journal Club Data Club, D. Redding discussions comments earlier drafts manuscript. Correia anonymous reviewers during review. Please note: publisher responsible content functionality supporting supplied authors. Any queries (other missing content) directed corresponding author article.