Evolution and paradigm shift in forest health research: A review on global trends and knowledge gaps

1 1 Evolution and paradigm shift in forest health research: A review on 2 global trends and knowledge gaps 3 4 Cristina Acosta-Muñoz 1,2 *, Rafael M. Navarro-Cerrillo 2, Francisco J. Bonet-García 1, Francisco 5 J. Ruiz-Gómez 2, Pablo González-Moreno 2. 6 7 1 Department of Botany, Ecology and Plant Physiology. Ecology Area. University of Cordoba, 8 Campus de Rabanales, Crta. IV, km. 396, E-14071 Córdoba. Spain. 9 2 Department of Forestry Engineering, Research Group Evaluation and Restoration of 10 Agroforest Systems - ERSAF. University of Cordoba, Campus de Rabanales, Crta. IV, km. 396, E- 11 14071 Córdoba. Spain. 12 13 * Corresponding author: Cristina Acosta-Muñoz, [email protected] 14 15 ORCID: 16 Cristina Acosta-Muñoz 0000-0002-9796-6367 17 Rafael M. Navarro-Cerrillo 0000-0003-3470-8640 18 Francisco J. Bonet-García 0000-0002-4627-1442 19 Francisco J. Ruiz-Gómez 0000-0002-1999-3415 20 Pablo González-Moreno 0000-0001-9764-8927 21 22 23 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 2 24 Abstract 25 Forests provide key ecosystem services to human society, and the ability to provide these 26 services depends on their overall health. Forest health is an attractive and interesting concept 27 in forestry research, which environmental, social and political interests have shaped. Assessing 28 forest health is crucial, but finding a single definition of the concept is complex. It is determined 29 by the aim of the forest study, different areas of knowledge, scales of work, technology, 30 methodologies, historical moment or source of funding, among others. With almost a century 31 of scientific evidence, the aim is to identify and contextualise temporal changes in the relevance 32 of this key concept. Trends are analysed through the construction of three main descriptors 33 (state variables, drivers and methods) and the main conceptual subdomains (themes). This 34 review reveals the significant geographical bias in the research, which the Global North 35 predominantly conducts. We observe the evolution of forest health research driven by diverse 36 needs and interests, ranging from air pollution to the multifaceted impacts of climate change. 37 Methodologies applied in this field have also evolved from traditional crown condition 38 inventories to the use of advanced tools such as remote sensing or ecophysiology, improving 39 the characterisation of forest health patterns at both global and individual scales. Forest health 40 research has evolved towards more holistic and multidisciplinary approaches, reflected in the 41 broadening and integration of methodologies and technologies, influenced by historical context, 42 which influence what is being researched today and future scenarios. We identified key 43 knowledge gaps in the scientific literature, in particular the concepts of ecosystem services, 44 Essential Biodiversity Variables (EBVs) and the concept of ‘One Health’. These findings highlight 45 the need for future research to incorporate these critical but often overlooked areas, potentially 46 reshaping future directions and scenarios for forest health research. 47 Keywords: Research trends; multidisciplinary approaches; global change drivers; global 48 environmental challenges; technological advances. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 3 49 1. Introduction 50 The transformation of forests by human activity underscores the imperative need to focus on 51 their preservation and health (1). Forests are essential for sustaining fundamental ecosystem 52 services for biogeochemical cycles and humanity (2). This link between forest health and the 53 capacity of forests to provide such services highlights the importance of understanding and 54 maintaining the ecological integrity of these ecosystems. Recognising and responding on forest 55 health is crucial to ensure their continued contribution to environmental and human well-being. 56 The concept of forest health is an umbrella concept encompassing a wide range of conceptual 57 subdomains (3,4), adopted by practitioners to understand health status (5). This reflects the 58 complexity inherent in the investigation of forest ecosystems, their interaction with human 59 activities and environmental changes. As a result, researchers have used different study 60 perspectives, definitions and research terms over time depending on the focus, scale of work 61 and other aspects considered such as priorities [6], [7]. 62 Researchers have adopted various terms related to forest health such as forest dieback, forest 63 decline or forest decay, associating these processes with the presence of diseases or pests, 64 observing symptoms at tree level (8). At larger scales, forest managers and researchers have 65 traditionally focused on characterizing the potential causes and spatio-temporal patterns (9). 66 Thus, the terms reflect not only tree mortality, but also a general loss of vigour and yield that is 67 spread over relatively large areas and is often related to high environmental stress (10). 68 Recently, the term forest health has evolved including also structural and functional aspects 69 (11). For instance, some authors define a healthy forest as one that includes a mosaic of 70 successional patches representing all development stages (12). At the same time, other more 71 holistic terms such as forest condition, forest state and forest integrity have been proposed. 72 Among them, forest integrity has been one of the latest suggestions, defining the overall 73 capacity of a forest system to sustain composition, structure and function within the historical 74 range of variation (13,14). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 4 75 Beyond the different conceptual subdomains we have listed above (used to define or assess 76 forest properties), we argue that the science and research has changed over time, with different 77 themes and associated terminology. We determine significant changes in recent decades based 78 on the following three main descriptors: a) attributes to measure forest condition; b) drivers 79 impacting on forest health condition (i.e., biotic or abiotic); and c) technologies and 80 methodologies associated with measurement and analysis of the two previous aspects. 81 With the expansion of science, the ever-deepening knowledge and the rapid pace of publication 82 seen in almost all scientific disciplines (15), this review emphasises the critical importance of 83 understanding what has brought us to the present. Understanding the historical context of the 84 discipline lends depth to current perceptions of forest health and is crucial to addressing the 85 challenges ahead. We not only value the foundations of our knowledge, but also recognise the 86 importance of following a deliberate and informed path for the future and innovation in forest 87 ecosystem research. 88 The inherent dynamic of constantly evolving research approaches is also related to changes in 89 the methodologies and technologies. The most frequent measurements have been crown 90 condition and tree damage (e.g. defoliation and discoloration), or growth in terms of biomass 91 and diameter increments (16,17). More holistic approaches go beyond the tree level to 92 characterize population, community, and ecosystem properties such as biodiversity and 93 regeneration dynamics. Regarding the drivers, forest health is affected by several disturbance 94 agents of different origin, and which can impact forest systems in a complex and interactive way 95 (11,18). In addition, context-dependency of the relevance of different abiotic and biotic agents 96 affects the overall research outputs, with bias towards scientist’ geographical regions and 97 specific taxa (19). 98 Forest researchers have put enormous effort into forest observation and monitoring to 99 understand forest health in relation to forest condition and related drivers (20). These 100 encourage the development of a wide range of methodologies aiming to characterize forest .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 5 101 ecosystem trends to inform policy and management decisions. These methodologies include 102 direct measures of vegetation such as physiological (e.g. photosynthesis, pigments, water 103 transport, respiration), structural measures derived from traditional forest inventories (e.g. 104 growth, dendrochronology), measures related to external agents, but with implications on 105 vegetation (e.g. drought, changes in land cover and land use), or measurements related to the 106 role and functioning of forest as an ecosystem (e.g. nutrient cycling and productivity) (9,21). 107 Lately, an increasing relevance of measurements derived from remote sensors deployed at 108 satellite and unmanned aerial vehicles has been observed for the detection of non-visible 109 phenomena in the forest (22–24). The convergence of these advanced methodologies together 110 with technological innovation allows for an ever deeper understanding of forests and their 111 dynamics. 112 Searches in the main scientific information databases (e.g. Web of Science or Scopus) show that 113 current knowledge on forest health is fragmented across several research disciplines (forestry, 114 environmental sciences, ecology, entomology, plant sciences, remote sensing, biodiversity 115 conservation, geosciences, agricultural and biological sciences, earth and planetary sciences, 116 social sciences, computer sciences, biochemistry, genetics and molecular biology, and others). 117 Several attempts have been made from different disciplines to review and describe these 118 changes in forest research (25) and to synthesize the conceptual frameworks around forest 119 health (11) without having a complete picture of the temporal dynamics of the concept. 120 Systematic and bibliometric reviews of scientific literature is key to synthesize a research field 121 and to understanding the conceptual trend. We used this approach to understand the temporal 122 and regional trends, research conceptual subdomains and methods used on forest health 123 assessment and monitoring at global scale. 124 Exploring the evolution of approaches to scientific research in forest health, to better 125 understand trends, developments and challenges, and the implications this has for the way 126 science is conducted globally, is of relevance [8]. In our study, we aimed to i) contextualise the .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 6 127 status of and approaches used in forest health research (i.e. scientific output, main contributors, 128 issues and keywords), ii) assess the temporal evolution of recurrent terms in forest health 129 research (i.e. a complete temporal map of relevant keywords), and iii) understand temporal 130 trends in the main conceptual subdomains encompassing forest health (i.e. topics) and the three 131 descriptors of forest health introduced in this study: condition (variables used to measure the 132 state or condition of forests), drivers (abiotic or biotic agents causing changes in forest 133 condition) and methodologies (techniques used to assess forest health). 134 We conducted a scientific literature search in academic databases covering a wide spectrum of 135 forest health terminology (26). Data mining was applied to extract information and patterns 136 from large bibliographic datasets that qualitatively, quantitatively and graphically allow a deeper 137 understanding of scientific production (27–29). Using a systematic review, we contributed to 138 temporally characterise the forest health concept to provide a holistic definition. Finally, we 139 discuss the gaps and future potential conceptual subdomains and descriptors that seem to arise 140 in the research field. 141 2. Material and methods 142 The workflow carried out for the analysis included (Fig 1): data collection, scientometric and 143 bibliometric analysis and visualization (all of these are detailed below). Specifically, to generate 144 the results of the first objective we performed a descriptive analysis of scientific production, a 145 bibliometric analysis of maps in terms of co-occurrence of keywords, and an analysis of 146 publications and contributions. To achieve the second objective, we provided a time trend 147 analysis of recurrent terms. And for the third objective we performed a temporal trend analysis 148 on a semantic clustering of the keywords (obtained from the previous objectives) in the three 149 forest health domains established in this study: forest condition, drivers and methods. 150 Figure 1. Graphical summary of the bibliographic and bibliometric review workflow for the study of forest 151 health concepts evolution. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 7 152 2.1. Scientific Data Base Search 153 The Web of Science (WoS) and Scopus database was chosen as high-impact search engines for 154 formal scientific publications, excluding non-conventional literature (grey literature). In January 155 2024, we carried out a preliminary search with a single keyword linear strategy using “forest 156 health” in Title, Abstract and Keywords fields, for all recording times in the database. The 157 temporality of the search was not limited, as the aim was to know completely all the existing 158 records in the databases from their origin to the present. The search was refined to include only 159 disciplines related to biological, environmental, forestry, earth sciences or methodological 160 sciences, excluding humanities or medical sciences. 161 Titles and abstracts of the 100 most relevant articles from each year were read and reviewed 162 for the state of the art of forest health research and scientific production. Potential articles that 163 could contain definitions or different terminology of forest health were read in detail (i.e. 164 reviews and highly cited papers). From this preliminary review, a list of conceptual subdomains 165 for the umbrella term forest health was compiled and used in the final database query. 166 The following search strategy was used to obtain the corpus: Title / Abstract / Keyword = ["forest 167 health" OR "forest mortality" OR "tree mortality" OR "forest integrity" OR "forest state" OR 168 "forest decline" OR "forest decay" OR "forest dieback"] 169 2.2. Descriptive analysis of the status of forest health research 170 The records obtained from the final query were analysed using the quantitative bibliometric 171 analysis algorithms of the R package Bibliometrix and the associated application Biblioshiny (26). 172 First, we conduct a descriptive analysis that characterises the scientific production over time, 173 main contributing authors, the co-authorship network, country publication impact, 174 geographically contextualising the main journals and funding agencies. Secondly, Sankey 175 diagrams were used to focus the analysis on keywords to identify patterns and trends in the 176 main terms used to describe forest health research in the above context (30). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 8 177 2.3. Temporal evolution in forest health issues 178 An analysis of relevance and development in research topics based on the initial systematic 179 review separated the selected keywords into four main groups: a) “motor themes”, that is, 180 themes well developed and important for the structure of the research field, b) “emerging or 181 declining topics” when they are both weakly developed and marginal, c) “basic and transversal 182 topics” which are important for a research field but are not developed and d) “niche with a 183 specialized character”, which are peripheral and specific topics for the research field. 184 Temporal changes in the concept and application of forest health were assessed by an automatic 185 keyword network analysis using WOSviewer (29), based on the previous review and debugging 186 of retrieved words. We used co-occurrences of keywords that appeared together in the title, 187 abstract or keyword list, and that were mentioned at least 10 times, giving a total of 1,731 188 keywords. We then plotted the top 1,000 keywords in a network and recurrence map. Finally, 189 this network was overlaid with the year of publication to identify temporal trends in keyword 190 association. 191 2.4. Clustering of conceptual sub-domains and temporal trends 192 We also implemented an approach considering a manual semantic keyword clustering to 193 understand temporal trends across the three domains of forest health considered in this study 194 (e.g., condition, drivers and methods) and main topics or definitions of forest health. From the 195 bibliographic search and subsequent download of references, for each year, the 50 most 196 relevant author and recommended keywords were extracted (keyword PLUS - index of terms 197 automatically generated from the titles of cited articles). This set of keywords went through a 198 process of cleaning up duplicates, normalising or removing special characters, reviewing 199 compound words and reducing some words to their basic roots. From this list of keywords for 200 all years, duplicated words were removed and grouped semantically. These terms were 201 classified into descriptors of topic (theme or discipline in forest health concept), condition (i.e., .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 9 202 variables used to measure forest state or condition), drivers (abiotic or biotic agents causing 203 changes in the forest condition) and methodologies (techniques used to assess forest health). 204 Within each descriptor type, we grouped the terms according to similar semantic meaning 205 (Table S1 Supplementary Material). In addition, when reviewing the list, other words were 206 proposed for being particularly relevant for the analysis based on the preliminary review 207 indicated in section 2.1. For each term, we compiled their occurrence in the abstracts of each of 208 the scientific records retrieved and calculated their frequency per year. All analysis were carried 209 out with R version 4.3.2. (31). 210 3. Results 211 3.1. Descriptive analysis of scientific production in forest health research 212 3.1.1 General findings 213 Scientific evidence in forest health from 1934 to 12/2023 shows an exponential publication 214 trend (as in most scientific disciplines), with an annual growth of 7.8%, although there was a 215 decline during the COVID-19 pandemic (Fig S1; Table S2 Supplementary Material). We analysed 216 10,338 papers from 1,511 sources, with 26,025 authors, highlighting that the 20 most prolific 217 authors account for 9.37% of the publications, indicating that forest health research is highly 218 diversified in terms of researchers involved in this topic. The top 4 authors stand out, with more 219 than 60 publications each (Fig S2 Supplementary Material). The leader in publications, Camarero 220 J.J., uses growth ring analysis and remote sensing to study the interaction between forests and 221 their environment, underlining the importance of longitudinal studies for conservation policies. 222 The first recorded publication was by Veblen T.T. in 1983, focusing on forest instability and tree 223 mortality using dendrochronology. The most cited authors are Allen D.C. and Breshears D.D. in 224 2010, for investigating global forest decline and drought. 225 3.1.2. Publication impact countries .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 10 226 Considering the 10 countries with the highest publication record on forest health, USA marks a 227 significant difference with the rest of the world throughout the whole period studied (Fig S3 228 Supplementary Materials). Next, Germany was the country with the longest track record in 229 related research. Canada increased its relevance in last decades, overtaking Germany in 2007, 230 and reaching the second position. Since 2010, China obtained an exponential increase in the 231 number of publications, reaching currently the third top position. 232 In terms of cross-country collaborations, 5 clusters were observed. A first cluster (Fig S5 233 Supplementary Materials) related by spatial continuity and ecological similarities, including 234 countries in North America (USA, Canada, or even Mexico); they in turn also related to other 235 countries by their latitude (Russia), or by the problems raised and methodological challenges to 236 cover large countries (such as China or Australia). A clear cluster was observed where South 237 American countries seem to be very closely aligned and related to the United Kingdom, the 238 Netherlands and Japan. The last distinct cluster related Eastern and Northern European 239 countries to New Zealand. 240 Despite the bibliometric reflection, the literature review shows that the geographic origin of 241 affiliation of the main authors of the publications does not determine the area of study in the 242 research. Collaborative research and international co-authorship favours diversification of the 243 study regions addressed beyond their own borders. Origin of funding agencies was largely 244 consistent with the most productive countries (Fig S6 Supplementary Materials). 245 3.1.3. Research approaches and issues 246 The search returned 13,677 keywords proposed by authors and 10,399 KeyWords Plus. The 247 Sankey diagram (Fig 2) shows boxes of different sizes and colour intensities allowing to identify 248 the areas of greatest activity and connection, among the 10 most common keywords (themes), 249 the 10 countries with the highest scientific output and the top 10 thematic journals (Fig S7 250 Supplementary Materials). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 11 251 The results highlight that the top countries publish most frequently in journals related to forest 252 management and also ecology. There are variations in keyword priority between countries, with 253 a general focus on climate change, although Spain and Australia also highlight drought. 254 Fig 2. Sankey diagram showing the relationships between frequent journals (left) from the top 10 255 publishing countries (middle) and the top 10 most mentioned keywords (right) in forest health-related 256 research. 257 Co-occurrence analysis showed the 50 most frequently mentioned keywords (Fig S9, supplementary 258 material), revealing "mortality" as the most common term, linked to climate change and drought, and 259 associated with forest vulnerability, climate responses, forest health and growth. Clusters were identified 260 focusing on forest decline due to stress, nitrogen deposition and soil problems, specifically related to pine 261 and spruce. Another cluster emphasises forest management dynamics, impacts and biodiversity. A final 262 cluster addresses the effects of fire and pests on specific species and sites. 263 3.2. Temporal keyword analysis 264 We grouped the keywords on the x-axis showing the relevance of the topics and on the y-axis the degree 265 of research development (Fig 3). We found that the core group of most relevant topics contains research 266 related to ‘climate change’, ‘tree mortality’, ‘drought’ and ‘fire’. Another group of core themes, although 267 less detailed (compared to the previous one), are ‘forest health’, ‘remote sensing’, ‘dendrochronology’, 268 ‘bark beetle’ and ‘biodiversity’. Of the relevant core themes with the highest level of development, the 269 keyword ‘disturbance’ is the most developed, followed by ‘forest management’, ‘wildfire’, ‘prescribed 270 fire’ and ‘Pinus ponderosa’. The results show that in general, the most developed topics with a high level 271 of specialisation are those related to the physiological processes of the forest, containing keywords such 272 as: ‘water stress’, ‘hydraulic failure’, ‘photosynthesis’, ‘carbon starvation’. Finally, in the group of 273 underdeveloped or unused keywords was ‘forest decline’ when talking about more concrete process 274 words, as well as ‘atmospheric pollution’ and ‘ozone’, issues that were relevant but already little studied. 275 Fig 3. Word grouping map by themes of relevance and development in research on forest health: a) 276 “motor themes”, well developed and important for the structure of the research field, b) “emerging or .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 12 277 declining topics” weakly developed and marginal, c) “basic and transversal topics” important but not 278 developed and d) “niche with a specialized character”, peripheral and specific topics for the research field. 279 The temporal network analysis of the top 1000 keywords showed a total of four clusters over 280 the study period (1934-12/2023) (Fig 4). Although the entire period analysed was included in 281 the graphical representation, due to the low number of publications recorded before the mid- 282 1980s in the databases consulted, the first cluster (purple colour) appears from this initial stage 283 of scientific production. It was closely linked to the concept of “forest decline”, especially in 284 topics related to “air pollution” (“nitrogen”, “ozone” or “acidification”). Around late 1990s and 285 early 2000s, the diagram showed a second broad cluster (in blue-green) highly associated with 286 the term "growth" and "forest health". These terms were mostly related to monitoring and 287 development-related terms (e.g. “stands”, “competence”, “deforestation”, or issues related to 288 “biomass” and “carbon sequestration”). Approximately in 2010, a new cluster (in green) appears 289 with quite wide range of terms of similar importance but related to ecosystem processes and 290 characteristics: (e.g. "biodiversity", "dynamics", "disturbance", "management, "conservation", 291 “restoration” and topics associated with “fire ecology”). This cluster seem to converge into the 292 concept of "tree mortality", peaking around 2015. Finally, in the most recent period (in yellow), 293 the research activity focused on “climate-change”, showing a great interest in the variables 294 measured and the tools related to “ecophysiology” and “remote sensing”. 295 Fig 4. Temporally normalised co-occurrence of most frequent keywords related to forest health research 296 (VOSviewer graph). 297 3.3. Clustering and trends in forest health descriptors 298 Manual classification of keywords in forest health reveals four sets of issues clustered into the 299 four established categories: topic, condition, drivers and methodologies (Fig 5), with fluctuations 300 in the 1970s to mid-1980s due to low production and thematic dispersion (removed from the 301 graphical representation), stabilising since the 1990s. Interest in "forest decline" has declined, .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 13 302 being replaced by "tree mortality", and "forest health" had a peak around 2005, followed by a 303 recent decline in the last decade. The attributes measured to characterise forest condition 304 reflect evolving environmental concerns, from "pests and diseases" and "air pollution" to a 305 growing interest in "climate" and "fire", which have become a constant concern. 306 Methodologically, understanding the functionality of organisms based on "ecophysiology" has 307 declined in relevance, while tools such as "inventories" remain constant. Since the 1990s, the 308 use of "Geographic Information Systems" and "remote sensing" has grown significantly, just as 309 modelling-based methodologies have increased in importance in recent years, although to a 310 lesser extent than the former. 311 Fig 5: Proportion of occurrence of keywords per year considering four different aspects of forest health: 312 5a) topic (theme or discipline in forest health concept), 5b) condition (variables used to measure forest 313 state or condition), 5c) drivers (abiotic or biotic agents causing changes in the forest condition) and 5d) 314 methodologies (techniques used to assess forest health). 315 4. Discussion 316 Over the last 90 years, we have seen a remarkable increase in the publication of research on 317 forest health, which underlines the need for an assessment of its evolution. The study 318 synthetically contextualised the accumulated body of knowledge, identifying substantial 319 changes in the ways in which forest research is approached, studied, measured and the 320 technologies associated with it. The results obtained showed aspects related to how science has 321 been done so far, detecting new emerging concepts or trends. Particularly noticeable are the 322 changes in scientific production, together with variations in the concepts and methods 323 employed by the forest health research community, reflecting how these aspects have been 324 linked to the main environmental concerns of each period, to funding or to the geographical 325 area of influence of the researchers. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 14 326 Following this bibliographic and bibliometric review, we aim to further explore the rationale for 327 temporal trends, highlighting their importance and implications at each point in time. The study 328 allows us to identify both significant advances and gaps in knowledge, contributing to the 329 configuration of new lines of research that respond to emerging challenges in forest health. 330 4.1 Scientific production on forest health 331 As with many other concepts in various disciplines (32,33), research on "forest health" is 332 experiencing exponential growth in terms of the number of articles published. Leaving aside the 333 general trend in academic production, the exponential growth observed denotes that the 334 concept under study remains dynamic and continues to be of interest to researchers. This 335 vigorous increase in the production of scientific literature reflects not only the growing global 336 concern for the state of our forests but also the recognition of the complexity and 337 multidimensionality that characterises them (34). As such, there remains a need for a collective 338 effort to understand and mitigate the impacts of threats such as climate change, tree diseases 339 and deforestation (35,36). 340 The affiliation origin of the researchers revealed important information on how the knowledge 341 is created regarding the topics about forest health. Our findings show that USA, Canada, and 342 Germany are the countries of affiliation origin for most of the researchers working in the target 343 topic. These results unveil two important biases: 344 First, there is a mismatch between the most publishing countries and those which harbour a 345 higher percentage of forests worldwide. According to the latest Food and Agriculture 346 Organisation (FAO) report on global forest resources (37), more than half of the world's forests 347 are concentrated in 5 countries: Russia (815M ha - 20%), Brazil (497M ha - 12%), Canada (347M 348 ha - 9%), USA (310M ha - 8%) and China (220M ha - 5%). Only Canada and the USA are among 349 the countries with large areas of forest with substantially more forest health articles compared 350 to the others. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 15 351 Second, there are not countries located in the tropics in the list of the higher publishing records. 352 This situation is more remarkable if we consider that the largest percentage of forest (45%) is in 353 tropical areas. This uneven distribution of publications between the Global North and the Global 354 South has been described in other disciplines such as ecology (38): most of the research is done 355 in the Global North although both the biodiversity and the forests are mainly in the Global South. 356 This geopolitical situation impacts very deeply in the completeness of the “forest health” 357 concept since it does not consider the views of researchers from the countries with more forest 358 cover. 359 4.2 Evolution of the most relevant keywords on forest health research 360 4.2.1. Integrating Keywords: Uncovering patterns 361 The total number of keywords found in “Climate change”, “tree mortality” and “drought” were 362 the topmost common keywords mentioned in forest health related research (Fig 3 and S9 363 Supplementary Materials). In fact, a large set of forest health studies have built on the delicate 364 situation of forest ecosystems worldwide with large-scale mortality processes driven by climate 365 drivers (9,39). Interestingly, the relevance and development analysis considered these terms as 366 "Basic Theme" showing a high relevance and a medium degree of development, which indicates 367 their current popularity but also further room for development compared to themes such as 368 disturbance or wildfire. 369 In this "Basic Theme" group, the analysis also highlighted terms such as "remote sensing", 370 "defoliation", "dendrochronology" and "biodiversity", revealing a multidisciplinary and multi- 371 scale approach to capture the complexity and dynamism of forest ecosystems. This approach 372 demonstrates a broad perspective on integrated, multi-scale forms of forest measurement: such 373 as defoliation as a measure of forest response at the leaf level, dendrochronology as a measure 374 of growth rate at the tree level, the use of remote sensing allowing extensive monitoring of .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 16 375 forests at the landscape level, or biodiversity as a manifestation of forest structure and 376 functioning at the community level (40–42). 377 Among the “Niche Themes” (high density and low relevance), we identified three main groups 378 that seem rather peripherical or with regional interest to the research field. From the initial 379 reading and bibliographic review, we found that the research field of invasive species and beetle 380 outbreaks mostly concentrated in North America on conifer forests (43–45) and on the other 381 side pure ecophysiological studies (46). The former group reinforces the idea of the bias toward 382 the Global North: P. ponderosa is a heavily timbered species typically found in temperate areas 383 of North America (47). Besides, concerns about bark beetles and prescribed fire are a 384 management activity also frequently used in temperate areas of the Northern Hemisphere (48). 385 Regarding the “Emerging or Declining Themes” with low development and relevance, it is 386 remarkable how the clustering process identifies forest decline, pollution and ozone as themes 387 that are no longer mainstream regarding forest health. These topics refer mainly to the events 388 of acid rain that were relatively common in Europe and North America during the second half of 389 20th Century and even nowadays in China (49). 390 4.2.2. Origins and context of paradigm shifts 391 The temporal change in the proportion of keywords tells a history very useful to understand the 392 research topic of “forest health”. The analysis of the temporal evolution of keyword clusters 393 reveals two main patterns (Fig. 4): a) there is a consistent trend towards a higher level of 394 knowledge integration across the time series and b) there is a clear link between the evolution 395 of global research and environmental challenges at each point in time and the changes in forest 396 health research. Based on this analysis, we identify four different temporal clusters that have 397 occurred sequentially (Fig.6): 398 Fig 6: Historical and thematic analysis of nearly a century of advances in forest health research, illustrating 399 key international environmental and socio-political milestones (to the right of the figure) that align with .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 17 400 shifts in prevailing themes and scientific terminology (evidenced by the word clouds to the left of the 401 figure). 402 I) The arise of global environmental problems linked to atmospheric pollution. At the 403 beginning of the time series, we found monocausal approaches to forest health 404 disturbances, where the most important drivers were “pests and diseases” as well 405 as “air pollution”. This earliest cluster contains concepts, which are attributable to 406 well defined scientific disciplines: “nutrients”, “forest soils”, “fertilization”, “ozone”, 407 “seedlings”, “calcium”, etc. This pattern indicates the low level of discipline 408 integration that the target concept experienced prior to 1990. Furthermore, it 409 shows clear links with the first modern environmental movements worldwide took 410 place between the 1960s and 1970s, focusing on nature conservation and 411 environmental protection. Predecessor events are the book "Silent Spring" by 412 Rachel Carson (1962), which denounced the harmful effects on the environment of 413 the massive use of chemicals such as pesticides. The first "Earth Day" (1970), the 414 United Nations Conference on the Human Environment in Stockholm (1972), and 415 the "Energy Crisis of 1973" awakened awareness of the dependence on oil and the 416 search for alternative sources. A central forest health topic in this cluster is “forest 417 decline” with strong links to acid deposition, air pollutants, and ozone. In fact, 418 “forest decline” was a terminology commonly used to depict the research concern 419 about forest deterioration due to air pollution mostly in Northern Europe and North 420 America (51). This forest problem gained international relevance in the 1980s with 421 “The Geneva Convention on Long-Range Transboundary Air Pollution” (1979), “The 422 Vienna Convention for the Protection of the Ozone Layer” (1985) or the signing of 423 “The Montreal Protocol on Substances that Deplete the Ozone Layer” (1987). This 424 environmental problem kept research strongly active until the end of the 20th 425 century. It is from the 1990s onwards that the evidence on the effects of pollution .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 18 426 began to be related to human health and ecosystems, and although the global 427 burden of pollutants has been increasing in the first two decades of the 21st century, 428 efforts are being made to continue reducing them (49). 429 II) Global environmental conservation. The second cluster is dominated by the decline 430 and physiology of the forest, appearing in late 1990s and early 2000s. Thus, the 431 methodologies mainly used are related to ecophysiology and forest inventories. This 432 period shows concepts with a higher level of integration among disciplines and 433 knowledge bodies: “growth”, “competition”, “forest health”, “ecosystems”, 434 “carbon sequestration”, etc. It also reflects the arise of current environmental 435 problems such as carbon emissions and deforestation. This may be mainly due to 436 the events that took place during the 1990s, where concerns with a more holistic 437 and multidisciplinary view of nature conservation and the environment began to 438 broaden. At this time, among others, the most famous world summits took place: 439 “The United Nations Conference on Environment and Development” or better 440 known as “The Rio de Janeiro Summit" (1992), laid the first foundations for the 441 signing of the United Nations Framework Convention on Climate Change and the 442 signing of the treaty “The Convention on Biological Diversity”, being the first global 443 agreement to promote aspects of international cooperation in the conservation and 444 sustainable use of biodiversity. The famous “Kyoto Protocol” (1997) on the 445 reduction of greenhouse gases that cause climate change is also approved in this 446 period. Almost simultaneously, the FAO publishes a report that highlights the 447 deforestation of large tracts of tropical forests in Latin America, Africa and Asia (52). 448 At the same time, the achievements of the international policy on air pollutants 449 reduced to some extent the pressure of air pollution on forest ecosystems (53) and 450 consequently in the forest health research field. In summary, this temporal cluster 451 represents the initial foundations for a more holistic and larger-scale view of the .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 19 452 planet’s global problems, evidenced among the key words in scientific publications 453 of the time, leading to the use of more multidisciplinary, integrative and 454 comprehensive concepts. 455 III) Multi-causality and tree mortality. The research that begins with the 21 st century 456 shows a multi-causal thinking in the problems that occur in the deterioration of 457 forests and the environment. This group shows a wide range of concepts, where the 458 words that stand out the most are “patterns”, “dynamics” and “disturbances”. Now 459 the forest problems are based on multi-causality, a more complex vision that can be 460 studied not only at the tree level but at different scales and in a multidisciplinary 461 way. The characterization of ecosystem dynamics is based on classification, offering 462 scales, intensities or patterns that measure diversity, fragmentation, deforestation, 463 succession, competition, susceptibility, regeneration, among other processes (54– 464 56). Furthermore, other concepts such as “management” and “restoration” also 465 emerge as a key concept suggesting a more applied vision in the forest health 466 research agenda (57). 467 At the end of this period, the research agenda converge on the topic of “tree 468 mortality” with numerous links to a wide range of concepts. Other highly integrative 469 concepts also appear (e.g. restoration, resilience) reinforcing the paradigm of multi- 470 causality in forest health research (58,59). “Wildfires”, its consequences, and some 471 methods used to monitor them, are present here to explicit the environmental 472 issues addressed in that time. It is no longer enough to quantify the causes of 473 disturbances in the system, but rather the effects of the disturbances themselves, 474 in search of solutions and to assess both the damage and the improvement in the 475 global balance. 476 IV) Climate change driven-research. The most recent cluster contains mainly concepts 477 related to “climate change” (e.g. “vulnerability”, “adaptation”, “change impacts”, .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 20 478 etc.). Interestingly, this cluster seems to reduce the degree of knowledge integration 479 as scientists are focusing mostly on understanding the consequences of climate 480 change on forests; although this challenge is much more complex than those 481 described previously. This is evidence of the greater environmental awareness, both 482 social and political, in the mitigation of climate change. One example is the approval 483 of the "The Paris Agreement" (2016), which establishes a global framework on 484 climate change focused on concrete aspects such as curbing global warming and 485 achieving carbon neutrality before the end of the century, where the use of the best 486 available science and technology are directly included to improve the conditions of 487 the planet. In fact, “climate change” can be considered a so called wicked problem 488 (60): multifaceted problems with fuzzy definition, elusive and complex solutions. 489 This explains why the current distribution of words within the drivers becomes more 490 equative. It is also the boom in technological development that derives part of these 491 efforts in generating instruments, methods and measurement and evaluation 492 techniques that are increasingly more accurate, reliable, and accessible. The 493 emergence of portable electronic equipment or geospatial technologies like remote 494 sensing, were a breakthrough to obtain continuously and efficiently data across 495 different spatio-temporal scales. This idea is supported by the presence within this 496 cluster of methods of assessment (e.g. carbon-isotope, dendroecology) and a great 497 amount of forest condition variables (e.g. evapotranspiration, water-use efficiency, 498 stomatal conductivity, hydraulic failure, etc.) currently measured with sophisticated 499 ecophysiological sensors (e.g. “gas-exchange” related to Eddy covariance towers or 500 photosynthesis sensors). 501 4.3 Trends topics, concepts and methodologies on forest health research .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 21 502 We also found similar temporal patterns from the temporal analysis of the four different 503 descriptors (topic, condition, drivers, and methodologies; Fig 5). First, we have identified a 504 temporal trend towards higher complexity. In the case of drivers, there is a clear trend from 505 monocausal to multicausal drivers of interest. From the predominance of “air pollution” and 506 “pest and diseases” during the first decades, to the emergence of other concepts: 507 “competition”, “land use”, “management”, and “fire”, which ultimately end in steep increase of 508 climate related drivers related to the “climate change” paradigm (Fig 5C). This could mean that 509 the drivers of change in the “forest health” research domain are more complex now than they 510 used to be decades ago. A similar pattern can be found in the group of “topics” words (Fig 5A). 511 The predominance of forest decline leads to a richer scenario where tree mortality, forest health 512 and still forest decline have certain importance. Regarding the condition words, the last decades 513 show also the rise of terms that imply a richer and more integrative approach: community, 514 mortality, growth, etc (Fig 5B). They coexist in a scenario more equitable than the existing at the 515 beginning of the time series. 516 On a second note, methods have followed the divergence and increased in complexity of the 517 other forest health aspects (Fig 5D). Finding appropriate methods to measure forest condition 518 has been always a major challenge in forest health research. Different types of methodologies 519 and techniques to assess forest status have been continuously evolving. Historically and up to 520 the present day, classical inventories have been an objective measure of forest species 521 composition, quantity and distribution of trees, as well as tree quality based on simple structural 522 measures (61,62). These inventories have become more complex as measurement tools have 523 evolved, although in general they are techniques that require little instrumentation, they are 524 limited in the amount of land that a team of people can cover. In this sense, important forest 525 monitoring programmes emerged in the 1980s. 526 Some of these programmes are the International Co-operative Programme on Assessment and 527 Monitoring of Air Pollution Effects on Forests (ICP Forests) of the United Nations Economic .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 22 528 Commission for Europe (UNECE), since 1985 (63); or the Forest Health Monitoring (FHM) of the 529 United States, which began in 1990 (64). Many of these inventories were accompanied by 530 physiological measurements of the plants as reliable methods of direct observation to monitor 531 the "vital signs" of the plants (65,66). Figure 4 (D panel) shows this first stage with the 532 predominance of ecophysiology methods, in which definitions of physiological factors and 533 vegetation damage, pollution prevalence, pests and nutrients appear as drivers of forestry 534 research in the literature (67,68). 535 On the other hand, the spatio-temporal perspective of forest health is currently under 536 development, constantly incorporating new methodologies mainly focused on the “massive 537 data approaches” at spatial level. These massive data monitoring tools does not only refer to 538 the use of remote sensors for landscape scale assessment, but also other methodologies 539 developed in the last decades in field of ecophysiology and the molecular biochemistry, such as 540 the assessment of gas fluxes at ecosystem level (Eddy co-variance towers), the use of high 541 throughput molecular techniques for microbial communities evaluation or genomic approaches 542 at individual and community levels (e.g., soil proteome, biogechemical cycling, etc…) (69,70). 543 Forest modelling, GIS and remote sensing are needed to manage efficiently and in a sustainable 544 way forest resources (71). These methods allow us to explore the spatial dimension of forest 545 health. In turn, forest modelling allows us to explore the temporal dimension of forest health 546 via long-term and short-term forecasting processes. All these modeling methods have 547 particularly increased in the last two decades becoming especially useful to improve forest 548 management in areas with scarce economic resources. 549 4.4. The way forward: future vision for the forest health concept 550 In this section we envision how forest health research might evolve in the coming years based 551 on similar disciplines and the gaps found. First, we did not find any article combining the idea of 552 forest health with the concept of “essential biodiversity variable” (EBV) (72). This is one of the .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 23 553 most prolific frameworks in the last decade for ecosystem monitoring, but it has not been found 554 among the relevant keywords of our analysis. We believe that the research field of forest health 555 would be very benefited from embracing the EBV framework, especially when considering the 556 description of forest condition. Using EBVs to describe forest health can be useful to increase 557 the comparability of studies carried out in different places. 558 Similarly, the term “ecosystem services” is also missing in the forest health literature. This 559 concept was introduced in the scientific literature several decades ago (73,74), but it seems to 560 have gone unnoticed in the “forest health” research field (75). We believe that the link between 561 “healthy forests” and their capability to provide ecosystem services might emerge as a new and 562 interesting field of research. While EBVs can help to homogenize the ways of assessing forest 563 health, ecosystem services can contribute to standardizing how we quantify the outcomes 564 provided by forests. 565 Finally, to integrate the current meaning of the target concept and to transcend it using the 566 above-mentioned proposals, we envision a conceptual and operational alignment between the 567 concepts of “forest health” and “one health” (76). The concept of one health has reached a very 568 holistic meaning in the present time. It used to be focused on single aspects of the health: pain, 569 infection, symptoms, etc. The current meaning put the focus on the concept of health far beyond 570 the absence of illness. One health aims to put together human, animal, and environmental 571 health. This holistic view is slowly moving from the academia into practice (77). This process 572 requires to increase our efforts in transdisciplinary collaboration (78). 573 Despite the large number of scientific articles related to forest health, the initial literature review 574 found that although they use the term, few authors dare to give a clear and comprehensive 575 approach in their manuscripts. After what we have learned, we recognise that this is a qualitative 576 concept that encompasses the overall state of a complex system studied from various 577 disciplines. According to this holistic approach, we might agree in defining forest health as the .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 24 578 capacity of a forest to sustainably provide a wide range of ecosystem services while maintaining 579 biodiversity, natural rhythms and resilience to disturbances inherent in forest dynamics (79–82). 580 Therefore, whether a forest is healthy or not will depend on its natural functioning, buffering 581 capacity and resilience, for which integrated monitoring and management with a vision of 582 conservation of the vital constants and functions of the system is essential. As forest ecosystems 583 are complex systems, assessing and understanding the totality of their functioning is a constant 584 challenge. This means that research on forest health continues to change according to the 585 knowledge needs and concerns observed in forests by scientists and experts, according to 586 available techniques and technologies, policies and social concerns, and the availability of 587 resources, mainly. 588 5. Conclusions 589 Forest health research has experienced exponential growth in the number of authors and 590 publications, reflecting its relevance and dynamism within the scientific community. We have 591 found a geographical bias in knowledge creation and research focus, as it does not align with 592 the countries that host the largest percentage of the world's forests (such as tropical countries). 593 This disparity between the Global North and the Global South raises concerns about the integrity 594 and inclusiveness of this field of study. 595 Concepts and research in forest health demonstrate an evolution towards integration of 596 knowledge over time, and global environmental challenges. Keyword analysis revealed a 597 thematic paradigm shift from the effect of air pollution to current interests in climate change 598 impacts, tree mortality and drought, increasingly integrated with remote sensing technologies 599 and specialised topics such as invasive species and ecophysiology. 600 The temporal analysis of clustering by descriptors reveals a transition towards complexity and 601 multidisciplinary approaches, showing an evolution from mono- to multi-causal factors. This 602 reflects an effort to understand interactions in complex systems. The integration of advanced .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 25 603 methodologies, including remote sensing, forest modelling and big data analysis, has 604 significantly improved the capacity for forest monitoring and management. 605 Finally, we envisage a conceptual and operational alignment between the concepts of "forest 606 health" and "one health". The holistic “one health” perspective, which integrates human, 607 animal, and environmental health, can provide a comprehensive approach to forest health. 608 Transdisciplinary efforts and collaboration are needed to bridge the gap between academia and 609 practical implementation. 610 Funding: 611 This research was funded by: 612 - Project DesFutur funded by Fundación Biodiversidad del Ministerio para la Transición Ecológica 613 y el Reto Demográfico (MITECO) and European Union (“NextGenerationEU”/PRTR). 614 - Grant RYC2021-033138-I, funded by MCIN/AEI/10.13039/501100011033 and European Union 615 (“NextGenerationEU”/PRTR). 616 - Project Evidence (ref 2822/2021) funded by Red de Parques Nacionales (OAPN y MITECO). 617 Contributions: 618 CAM: Conceptualization, Methodology, Investigation, Formal analysis, Data Curation, Writing - 619 Original Draft, Writing - Review & Editing, Figures 1,2,3,4,6 and Supplementary Material, 620 Supervision; RMNC: Conceptualization, Writing - Review & Editing; FJBG: Conceptualization, 621 Resources, Writing - Original Draft, Writing - Review & Editing; FJRG: Writing & Review; PGM: 622 Conceptualization, Methodology, Investigation, Formal analysis, Writing - Original Draft, Writing 623 - Review & Editing, Figure 5, Supervision. 624 Acknowledgments: 625 Thanking Pablo Salazar-Zarzosa for his detailed reading and constructive comments for 626 improvement. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 26 627 Conflict of Interest: 628 The authors declare that they have no conflict of interest. 629 Declaration of competing interest: 630 The authors declare that they have no known competing financial interests or personal 631 relationships that could have appeared to influence the work reported in this paper. 632 Supplementary Materials: 633 Table S1: Clustering of the main keywords of the publications search results for forest health, 634 Fig S1: Evolution of the annual scientific production related to Forest Health, Table S2: Main 635 information of the results of the bibliographic search in Web of Science related to Forest Health 636 (Bibliometrix R information), Fig S2: The most relevant 20 authors with the highest number of 637 publications and their evolution of the scientific production evolution, Fig S3: Spatial distribution 638 of the number of publications in different countries on forest health 1934-12/2023 (Graph R 639 from Bibliometrix), Table S3: Scientific production related to forest health 1934-12/2023 by 640 continent, Fig S4: Spatial distribution of the number of publications in different countries on 641 forest health 1934-12/2023 (Graph R from Bibliometrix), Fig S5: Evolution of scientific 642 production related to forest health in the 5 main countries (Graph R from Bibliometrix), Fig S6: 643 Thematic categories of the journals to forest health 1934-12/2023 (WoS graph), Fig S7: Most 644 relevant funding agencies in forest health studies by number of works supported (WoS), Fig S8: 645 Co-occurrence network to 50 most mentioned keywords in the scientific literature (1934- 646 12/2023) (WoSviewer). 647 Data availability: 648 The data used to generate the analysis in this article are fully reproducible. They were extracted 649 from the Web of Science database through the query and time period indicated in the 650 methodology. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 27 651 6. References 652 1. Camarero JJ. The drought‒dieback‒death conundrum in trees and forests. Plant Ecol 653 Divers. 2021;14(1–2):1–12. doi: 10.1080/17550874.2021.1961172. 654 2. Deuffic P, Garms M, He J, Brahic E, Yang H, Mayer M. Forest Dieback, a Tangible Proof of 655 Climate Change? A Cross-Comparison of Forest Stakeholders’ Perceptions and Strategies 656 in the Mountain Forests of Europe and China. Environ Manage. 2020;66(5):858–72. doi: 657 10.1007/s00267-020-01363-9. 658 3. Bailin A. Worlds in worlds: Assigning inferences to subdomains. J Lit Semant. 659 2004;33(2):93–109. doi: 10.1515/jlse.2004.33.2.93. 660 4. Cobo MJ, López-Herrera AG, Herrera-Viedma E, Herrera F. An approach for detecting, 661 quantifying, and visualizing the evolution of a research field: A practical application to 662 the Fuzzy Sets Theory field. J Informetr. 2011;5(1):146–66. doi: 663 10.1016/j.joi.2010.10.002. 664 5. Teale SA, Castello JD. The past as key to the future: a new perspective on forest health. 665 For Heal An Integr Perspect. 2011;3–16. 666 6. Urquhart J, Marzano M, Potter C. The Human Dimensions of Forest & Tree Health-Global 667 Perspectives. 2018. 668 7. Mack P, Kremer J, Kleinschmit D. Forest dieback reframed and revisited? Forests 669 (re)negotiated in the German media between forestry and nature conservation. For 670 Policy Econ. 2023;147. doi: 10.1016/j.forpol.2022.102883. 671 8. Ostry ME, Venette RC, Juzwik J. Decline as a disease category: is it helpful? 672 Phytopathology. 2011;101(4):404–9. doi: 10.1094/PHYTO-06-10-0153. 673 9. Hartmann H, Bastos A, Das AJ, Esquivel-Muelbert A, Hammond WM, Martínez-Vilalta J, 674 et al. Climate Change Risks to Global Forest Health: Emergence of Unexpected Events of 675 Elevated Tree Mortality Worldwide. Annu Rev Plant Biol. 2022;73:673–702. doi: 676 10.1146/annurev-arplant-102820-012804. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 28 677 10. Menezes-Silva PE, Loram-Lourenço L, Alves RDFB, Sousa LF, Almeida SE da S, Farnese FS. 678 Different ways to die in a changing world: Consequences of climate change for tree 679 species performance and survival through an ecophysiological perspective. Ecol Evol. 680 2019;9(20):11979–99. doi: 10.1002/ece3.5663. 681 11. Trumbore S, Brando P, Hartmann H. Forest health and global change. Science. 2015; 349, 682 814-18. doi: 10.1126/science.aac6759 683 12. Edmonds RL, Agee JK, Gara RI. Forest health and protection. 2nd Edition. Waveland Press; 684 2011. 685 13. Tierney GL, Faber-Langendoen D, Mitchell BR, Shriver WG, Gibbs JP. Monitoring and 686 evaluating the ecological integrity of forest ecosystems. Front Ecol Environ. 2009;7:308– 687 16. doi: 10.1890/070176. 688 14. Rempel RS, Naylor BJ, Elkie PC, Baker J, Churcher J, Gluck MJ. An indicator system to 689 assess ecological integrity of managed forests. Ecol Indic. 2016;60:860–69. doi: 690 10.1016/j.ecolind.2015.08.033. 691 15. Weingart P. The moment of truth for science. The consequences of the “knowledge 692 society” for society and science. EMBO Rep. 2002 Aug;3(8): 703-6. doi: 10.1093/embo- 693 reports/kvf165. 694 16. Dobbertin M. Tree growth as indicator of tree vitality and of tree reaction to 695 environmental stress: A review. Eur J For Res. 2005;124:319–33. doi: 10.1007/s10342- 696 005-0085-3. 697 17. Randolph KDC, Dooley K, Shaw JD, Morin RS, Asaro C, Palmer MM. Past and present 698 individual-tree damage assessments of the US national forest inventory. Environ Monit 699 Assess. 2021;193:116. doi: 10.1007/s10661-020-08796-z. 700 18. Doblas-Miranda E, Martínez-Vilalta J, Lloret F, Álvarez A, Ávila A, Bonet FJ, et al. 701 Reassessing global change research priorities in mediterranean terrestrial ecosystems: 702 How far have we come and where do we go from here? Glob Ecol Biogeogr. 2015;24:25– .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 29 703 43. doi: 10.1111/geb.12224. 704 19. Pyšek P, Richardson DM, Pergl J, Jarošík V, Sixtová Z, Weber E. Geographical and 705 taxonomic biases in invasion ecology. Trends Ecol Evol. 2008;23(5):237–44. doi: 706 10.1016/j.tree.2008.02.002. 707 20. Abas A. A systematic literature review on the forest health biomonitoring technique: A 708 decade of practice, progress, and challenge. Front Environ Sci. 2023;11:910730. doi: 709 10.3389/fenvs.2023.970730 710 21. Shukla P, Skeg J, Buendia E, Masson-Delmotte V. Climate Change and Land: an IPCC 711 special report on climate change, desertification, land degradation, sustainable land 712 management, food security, and greenhouse. Cambridge University Press; 2019. 713 22. Bayr C, Gallaun H, Kleb U, Kornberger B, Steinegger M, Winter M. Satellite-based forest 714 monitoring: spatial and temporal forecast of growing index and short-wave infrared 715 band. Geospat Health. 2016;11(1):310. doi: 10.4081/gh.2016.310. 716 23. Lausch A, Erasmi S, King DJ, Magdon P, Heurich M. Understanding Forest Health with 717 Remote Sensing -Part I-A Review of Spectral Traits, Processes and Remote-Sensing 718 Characteristics. Remote Sens. 2016;8:1029 8, 1029. doi: 10.3390/RS8121029 719 24. Ecke S, Dempewolf J, Frey J, Schwaller A, Endres E, Klemmt H-J, et al. UAV-Based Forest 720 Health Monitoring: A Systematic Review. Remote Sens. 2022;14(13). doi: 721 10.3390/rs14133205. 722 25. Pautasso M, Stenlid J, Oliva J, Menkis A. Scientometrics of Forest Health and Tree 723 Diseases: An Overview. Forest 2016;7(1):17. doi: 10.3390/f7010017. 724 26. Dervis H. Bibliometric analysis using bibliometrix an R package. J Scientometr Res. 725 2019;8(3):156–60. doi: 10.5530/jscires.8.3.32. 726 27. van Eck NJ, Waltman L. Visualizing Bibliometric Networks. Measuring Scholarly Impact. 727 2014. 285–320 p. doi: 10.1007/978-3-319-10377-8_13. 728 28. Aria M, Cuccurullo C. bibliometrix: An R-tool for comprehensive science mapping .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 30 729 analysis. J Informetr. 2017;11(4):959–75. doi: 10.1016/j.joi.2017.08.007. 730 29. Iliescu AN. Conceptual Atlas of the Knowmad Literature: Visual Mapping with VOSviewer. 731 Manag Dyn Knowl Econ. 2021;9(3):379–92. doi: 10.2478/mdke-2021-0025. 732 30. Riehmann P, Hanfler M, Froehlich B. Interactive sankey diagrams. Proc - IEEE Symp Inf 733 Vis INFO VIS. 2005;233–40. doi: 10.1109/INFVIS.2005.1532152. 734 31. R Core Team. R: A language and environment for statistical computing. R Foundation for 735 Statistical Computing. Vienna, Austria; 2021. Available from: https://www.r-project.org/ 736 32. Pautasso M. Publication Growth in Biological Sub-Fields: Patterns, Predictability and 737 Sustainability. Sustain 2012;4(12):3234–47. doi: 10.3390/su4123234. 738 33. Chavalarias D. What’s wrong with Science?: Modeling the collective discovery processes 739 with the Nobel game. Scientometrics. 2017;110(1):481–503. doi: 10.1007/s11192-016- 740 2109-9. 741 34. Bréda N, Peiffer M. Vulnerability to forest decline in a context of climate changes: new 742 prospects about an old question in forest ecology. Ann For Sci. 2014;71(6):627–31. doi: 743 10.1007/s13595-014-0411-3. 744 35. Pautasso M, Schlegel M, Holdenrieder O. Forest Health in a Changing World. Microb Ecol. 745 2015;69(4):826–42. doi: 10.1007/s00248-014-0545-8 746 36. Wigand ME, Timmermann C, Scherp A, Becker T, Steger F. Climate Change, Pollution, 747 Deforestation, and Mental Health: Research Trends, Gaps, and Ethical Considerations. 748 GeoHealth. 2022;6(11). doi: 10.1029/2022GH000632. 749 37. FAO. Global Forest Resources Assessment 2020 - Key findings. Rome. 2020. doi: 750 10.4060/ca8753en. 751 38. Nuñez MA, Barlow J, Cadotte M, Lucas K, Newton E, Pettorelli N, et al. Assessing the 752 uneven global distribution of readership, submissions and publications in applied 753 ecology: Obvious problems without obvious solutions. J Appl Ecol. 2019;56(1):4–9. doi: 754 10.1111/1365-2664.13319. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 31 755 39. Allen CD, Macalady AK, Chenchouni H, Bachelet D, McDowell N, Vennetier M, et al. A 756 global overview of drought and heat-induced tree mortality reveals emerging climate 757 change risks for forests. For Ecol Manage. 2010;259(4):660–84. doi: 758 10.1016/j.foreco.2009.09.001. 759 40. Stone C, Turner R, Verbesselt J. Integrating plantation health surveillance and wood 760 resource inventory systems using remote sensing. 2008;71(3):245–53. doi: 761 10.1080/00049158.2008.10675043. 762 41. Pause M, Schweitzer C, Rosenthal M, Keuck V, Bumberger J, Dietrich P, et al. In 763 Situ/Remote Sensing Integration to Assess Forest HealthA Review. Remote Sens. 764 2016;8(6):471. doi: 10.3390/rs8060471. 765 42. Ruiz-Benito P, Vacchiano G, Lines ER, Reyer CPO, Ratcliffe S, Morin X, et al. Available and 766 missing data to model impact of climate change on European forests. Ecol Modell. 767 2020;416:108870. doi: 10.1016/j.ecolmodel.2019.108870. 768 43. Abella SR, Hausman CE, Jaeger JF, Menard KS, Schetter TA, Rocha OJ. Fourteen years of 769 swamp forest change from the onset, during, and after invasion of emerald ash borer. 770 Biol Invasions. 2019;21(12):3685–96. doi: 10.1007/s10530-019-02080-z. 771 44. Dudney JC, Nesmith JCB, Cahill MC, Cribbs JE, Duriscoe DM, Das AJ, et al. Compounding 772 effects of white pine blister rust, mountain pine beetle, and fire threaten four white pine 773 species. Ecosphere. 2020;11(10). doi: 10.1002/ecs2.3263. 774 45. Baer KC, Gray AN. Biotic predictors improve species distribution models for invasive 775 plants in Western U.S. Forests at high but not low spatial resolutions. For Ecol Manage. 776 2022;518:120249. doi: 10.1016/j.foreco.2022.120249. 777 46. Petit-Cailleux C, Davi H, Lefèvre F, Verkerk PJ, Fady B, Lindner M, et al. Tree Mortality 778 Risks Under Climate Change in Europe: Assessment of Silviculture Practices and Genetic 779 Conservation Networks. Front Ecol Evol. 2021;9:582. doi: 10.3389/fevo.2021.706414 780 47. Lowery DP. Ponderosa Pine (Pinus ponderosa Dougl. ex Laws.). FS-254. 1984. Available .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 32 781 from: https://www.fs.usda.gov/research/treesearch/32194. 782 48. Glitzenstein JS, Brewer JS, Masters RE, Varner JM, Hiers JK. Fire Ecology and Fire 783 Management of Southeastern Coastal Plain Pine Ecosystems. In: Greenberg CH, Collins 784 B, editors. Fire Ecology and Management: Past, Present, and Future of US Forested 785 Ecosystems. Managing Forest Ecosystems; vol. 39; Springer; 2021. p. 63–104. doi: 786 10.1007/978-3-030-73267-7_3. 787 49. Fowler D, Brimblecombe P, Burrows J, Heal MR, Grennfelt P, Stevenson DS, et al. A 788 chronology of global air quality. Philos Trans R Soc A Math. Phys. Eng. Sci. 789 2020;378(2183). doi: 10.1098/rsta.2019.0314. 790 50. Carson R. Silent Spring. New York: Fawcett Crest; 1962. 791 51. Monserud RA. Simulation of Forest Tree Mortality. For Sci. 1976;22(4):438–44. doi: 792 10.1093/forestscience/22.4.438. 793 52. FAO. Forest resources assessment 1990: tropical countries. Rome; 1993. 794 53. Grennfelt P, Engleryd A, Forsius M, Hov Ø, Rodhe H, Cowling E. Acid rain and air pollution: 795 50 years of progress in environmental science and policy. Ambio. 2020;49(4):849–64. doi: 796 10.1007/s13280-019-01244-4. 797 54. Stephenson NL, Van Mantgem PJ, Bunn AG, Bruner H, Harmon ME, O’Connell KB, et al. 798 Causes and implications of the correlation between forest productivity and tree mortality 799 rates. Ecol Monogr. 2011;81(4):527–55. doi: 10.1890/10-1077.1. 800 55. Zhang X, Cao Q V., Duan A, Zhang J. Modeling tree mortality in relation to climate, initial 801 planting density, and competition in Chinese fir plantations using a Bayesian logistic 802 multilevel method. Can J For Res. 2017;47(9):1278–85. doi: 10.1139/cjfr-2017-0215. 803 56. McDowell NG. Deriving pattern from complexity in the processes underlying tropical 804 forest drought impacts. New Phytol. 2018;219(3):841–4. doi: 10.1111/nph.15341. 805 57. Diggins C, Fulé PZ, Kaye JP, Covington WW, Diggins C, Fulé PZ, et al. Future climate affects 806 management strategies for maintaining forest restoration treatments. Int J Wildl Fire. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 33 807 2010;19(7):903–13. doi: 10.1071/WF09109. 808 58. Fan C, Tan L, Zhang P, Liang J, Zhang C, Wang J, et al. Determinants of mortality in a mixed 809 broad-leaved Korean pine forest in northeastern China. Eur J For Res. 2017;136(3):457– 810 69. doi: 10.1007/s10342-017-1045-4. 811 59. Muzika RM. Opportunities for silviculture in management and restoration of forests 812 affected by invasive species. Biol Invasions. 2017;19(11):3419–35. doi: 10.1007/s10530- 813 017-1549-3. 814 60. Crowley K, Head BW. The enduring challenge of ‘wicked problems’: revisiting Rittel and 815 Webber. Policy Sci. 2017;50(4):539–47. doi: 10.1007/s11077-017-9302-4. 816 61. Fitts LA, Domke GM, Russell MB. Comparing methods that quantify forest disturbances 817 in the United States’ national forest inventory. Environ Monit Assess. 2022;194(4):1–17. 818 doi: 10.1007/s10661-022-09948-z. 819 62. Klockow PA, Edgar CB, Windmuller-Campione MA, Baker FA. Stand Inventories as an Early 820 Detection System for Forest Health Threats. For Sci. 2023;69(1):1–9. doi: 821 10.1093/forsci/fxac046. 822 63. Lorenz M, Fischer R. Chapter 2 - Pan-European Forest Monitoring: An Overview. Dev 823 Environ Sci. 2013;12:19–32. doi: 10.1016/B978-0-08-098222-9.00002-9. 824 64. Tkacz B, Riitters K, Percy KE. Chapter 4 - Forest Monitoring Methods in the United States 825 and Canada: An Overview. Dev Environ Sci. 2013;12:49–73. doi: 10.1016/B978-0-08- 826 098222-9.00004-2. 827 65. Benner P, Sabel P, Wild A. Photosynthesis and transpiration of healthy and diseased 828 spruce trees in the course of three vegetation periods. Trees. 1988;2(4):223–32. doi: 829 0.1007/BF00202377. 830 66. Ibanez I, Katz DSW, Peltier D, Wolf SM, Barrie BTC. Assessing the integrated effects of 831 landscape fragmentation on plants and plant communities: the challenge of 832 multiprocess–multiresponse dynamics. J Ecol. 2014;102(4):882–95. doi: 10.1111/1365- .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 34 833 2745.12223. 834 67. Hyink DM, Zedaker SM. Stand dynamics and the evaluation of forest decline. Tree Physiol. 835 1987;3(1):17–26. doi: 10.1093/treephys/3.1.17. 836 68. Curtis PS, Wang XZ. A meta-analysis of elevated CO2 effects on woody plant mass, form, 837 and physiology. Oecologia. 1998;113(3):299–313. doi: 10.1007/s004420050381. 838 69. Millar DJ, Ewers BE, Mackay DS, Peckham S, Reed DE, Sekoni A. Improving ecosystem- 839 scale modeling of evapotranspiration using ecological mechanisms that account for 840 compensatory responses following disturbance. Water Resour Res. 2017;53(9):7853–68. 841 doi: 10.1002/2017WR020823. 842 70. Fasanella M, Suarez ML, Hasbun R, Premoli AC. Individual-based dendrogenomic analysis 843 of forest dieback driven by extreme droughts. Can J For Res. 2021;51(3):420–32. doi: 844 10.1139/cjfr-2020-0221. 845 71. Hirons S, Matilda Collins C, Singh M. Assessing variation in the effectiveness of IUCN 846 protected area categorisation. What remotely sensed forest integrity and human 847 modification reveals across the major tropical forest biomes. Ecol Indic. 848 2022;143:109337. doi: 10.1016/j.ecolind.2022.109337. 849 72. Lausch A, Borg E, Bumberger J, Dietrich P, Heurich M, Huth A, et al. Understanding Forest 850 Health with Remote Sensing, Part III: Requirements for a Scalable Multi-Source Forest 851 Health Monitoring Network Based on Data Science Approaches. Remote Sens. 852 2018;10(7). doi: 10.3390/rs10071120. 853 73. Ehrlich, P.R. and Ehrlich AH. Extinction: The Causes and Consequences of the 854 Disappearance of Species. Random House. New York; 1981. 855 74. Ehrlich PR, Mooney HA. Extinction, Substitution, and Ecosystem Services. Bioscience. 856 1983;33(4):248–54. doi: 10.2307/1309037. 857 75. Patel A, Rapport DJ, Vanderlinden L, Eyles J. Forests and societal values: Comparing 858 scientific and public perception of forest health. The Environmentalist. 1999;19(3):239– .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint 35 859 49. doi: 10.1023/A:1026402812084. 860 76. Atlas RM. One Health: Its Origins and Future. Curr Top Microbiol Immunol. 2013;(365):1– 861 13. doi: 10.1007/82_2012_223. 862 77. Couto R de M, Brandespim DF. A review of the one health concept and its application as 863 a tool for policy-makers. Int J One Heal. 2020;6(1):83–9. doi: 10.14202/IJOH.2020.83-89. 864 78. Schneider MC, Munoz-Zanzi C, Min K, Aldighieri S. “One Health” From Concept to 865 Application in the Global World. Oxford Res Encycl Glob Public Heal. 2019. doi: 866 10.1093/acrefore/9780190632366.013.29. 867 79. Slovik S. Early needle senescence and thinning of the crown structure of Picea abies as 868 induced by chronic SO2 pollution. I. Model deduction and analysis. Glob Chang Biol. 869 1996;2(5):443–58. doi: 10.1111/j.1365-2486.1996.tb00094.x. 870 80. Burger JA. SOILS BIOLOGY AND TREE GROWTH | Soil and its Relationship to Forest 871 Productivity and Health. Encycl For Sci. 2004;1189–95. 872 81. Gauthier S, Bernier P, Kuuluvainen T, Shvidenko AZ, Schepaschenko DG. Boreal forest 873 health and global change. Science. 2015;349(6250):819–22. doi: 874 10.1126/science.aaa9092. 875 82. Millar CI, Stephenson NL. Temperate forest health in an era of emerging 876 megadisturbance. Science. 2015;349(6250):823–6. doi: 10.1126/science.aaa9933. 877 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 5, 2024. ; https://doi.org/10.1101/2024.06.03.597256doi: bioRxiv preprint