Treefall risk assessment in an urban green area for a hypothetical 30-year return period storm using damage data from Typhoons Faxai and Hagibis in 2019

Treefall risk assessment in an urban green area for a hypothetical 30-year return period storm using damage data from Typhoons Faxai and Hagibis in 2019 | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Treefall risk assessment in an urban green area for a hypothetical 30-year return period storm using damage data from Typhoons Faxai and Hagibis in 2019 Kohei Katayama, Yukira Mochida, Fumito Koike This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9147742/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Urban trees support human well-being, yet their benefits must be balanced against the risk of tree failure. Here, we propose a simple quantitative approach that practitioners can use to assess and manage urban treefall risk. We developed a simple mechanical model based on readily measured trunk diameter and height to calculate a hypothetical maximum compressive stress in the trunk for a given wind speed. This stress-based index was treated as a latent variable linking the mechanical and stochastic components of our framework. We then fitted a statistical model that predicts treefall probability from the latent stress index using empirically observed treefall records from Typhoons Faxai and Hagibis (2019). Unmodeled sources of variability and other unaccounted-for factors were implicitly incorporated through statistical calibration. The fitted model showed a statistically supported association between the latent stress index and treefall probability (broadleaved tree AUC = 0.94, conifer AUC = 0.77). Using extreme value analysis, we estimated the 30-year return level of maximum wind speed at the study site and assessed treefall risk for individual trees under this scenario. The model indicated higher risk for tall non-native trees (relative to local native species) and for conifers. Although available treefall data remain limited, to our knowledge, this study is among the first to predict future treefall probability using empirically observed treefall records and to provide a quantitative framework for practitioners to manage treefall risk. Biological sciences/Ecology Earth and environmental sciences/Ecology Earth and environmental sciences/Natural hazards latent variable maximum compressive stress mechanical model return periods statistical model Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction Urban trees play an essential role in environments by supporting human well-being through mental health benefits (Kurose and Koike 2020; Bratman et al., 2015), mitigating high temperatures (Bowler et al., 2010), and reducing flood risk through enhanced rainfall interception and infiltration (Xiao & McPherson, 2002; Bartens et al, 2008; Öztürk et al. 2025). Despite these benefits, fallen trees during storms can cause injuries or fatalities, obstruct traffic, and damage electrical cables and other urban infrastructure (Jim & Liu, 1997; van Haaften et al., 2016). Reliable and feasible assessment of treefall risk for practitioners is essential for balancing the benefits and risks associated with urban trees (Zur et al. 2025). One typical risk assessment is to survey many aspects such as trunk decay, species, and tree architecture, and estimate the risk by expert judgement (Kane & Ryan, 2004; Smiley et al., 2017; ANSI 2023). Such qualitative risk assessment by specialists worked significantly; however, reliability is limited (Koeser et al., 2020; Koeser et al., 2023). Among many causes of tree failures (van Haaften et al. 2021), storm wind is the most important cause in subtropical and warm-temperate regions with Typhoon and Hurricane damage (Koeser et al., 2020; Salisbury et al., 2023). Material and structural mechanics have been applied to assess the likelihood of tree failure by wind (James et al., 2014; Luo and Ai 2022). However, wood strength differs significantly within a tree (Grazide et al. 2015; Luo and Ai 2022), and in urban landscapes, each tree has a unique, complicated form of trunk, branch, and crown (Koike 1985). It is unrealistic to measure the detailed structure of every tree in a city (Jackson et al. 2019) and assess treefall risk using purely mechanistic methods. Because practitioners can readily obtain tree height and diameter, we developed a simple mechanical model to estimate a hypothetical maximum compressive stress for each individual tree as a mechanically derived index of failure susceptibility. Peak wind speed in urban areas varies with local factors such as buildings and other artificial structures (Huang et al. 2024). In most cities, however, the only readily available wind data come from the nearest meteorological station, and site-specific wind conditions at individual trees are difficult to obtain. Given these uncertainties, we adopted a hybrid framework (Luo and Ai 2022) that links a simple mechanical model with a stochastic approach, using the estimated maximum compressive stress as a latent variable (Fig. 1). Specifically, we modeled treefall occurrence (yes/no) as the response and used the stress-based latent variable as a mechanistically interpretable predictor. Although this stress is a rough proxy, statistical models can calibrate and incorporate additional covariates, and the stress-based latent variable provides a mechanistically interpretable link between mechanical and stochastic components. Stochastic modeling also requires data on both fallen and non-fallen trees, but storms that cause large-scale damage are rare in any given city, and systematic surveys of treefall incidents are difficult to obtain (Koeser et al. 2020; Koeser et al. 2025). In this study, we developed a simple mechanical model based on tree diameter and height and then fit a regression model using observed treefall data from Typhoons Faxai and Hagibis (2019). The model estimates treefall probability using the maximum 10-min mean wind speed from the nearest meteorological station, with the hypothetical maximum compressive stress treated as a latent variable. Importantly, all predictors (diameter, height, species, and station wind speed) are routinely available in municipal inventories, enabling city-wide application. Using this model, we assessed the treefall risk of every tree under a hypothetical extreme storm corresponding to a 30-year return period. Although treefall data are currently limited, to our knowledge, this study is among the first to predict future treefall probability using empirically observed treefall records and to propose a quantitative approach that practitioners can use to manage treefall risk. 2. Methods 2.1. Study site and trees The research site was an urban university campus of 455531 m 2 with 97 buildings and about 10000 students and staff. The campus is on a coastal terrace in the Tokyo Metropolitan Area (latitude 35.47, longitude 139.59, altitude 50 m). An estuary with saline water is located 1.6 km distant, and the current coastal line after recent land reclamation is 4.4 km. Average temperature is 16.2 °C, average annual precipitation is 1730 mm, and it belongs to the warm-temperate broadleaved forest zone (Miyawaki 1986). The old golf course, with a few remnant forests, became a university campus in the 1970s, and many trees were planted simultaneously to construct the campus. The remnant forest in the study area suggests that the climax forest is dominated by Castanopsis sieboldii (Makino) Hatus. ex T.Yamaz. et Mashiba subsp. sieboldii with subcanopy trees such as Euria japonica Thunb. var. japonica , Neolitsea sericea (Blume) Koidz., and Ilex integra Thunb. All tree surveys on trunk diameter at breast height (DBH) and tree height were made by Mochida Lab from 2010 to 2016 (Supporting Material 1). The maximum tree height in the site was 26 m, and the maximum DBH was 156 cm. The most dominant species was the planted non-native Cinnamomum camphora (L.) J.Presl, and cherry Cerasus spp. followed it. Almost all cherry trees are grafted cultivars, although there are a few native populations. There are native populations of Castanopsis sieboldii in small remnant forests, although this tree was planted in several places. Regionally native evergreen oaks Quercus spp. and Machilus thunbergii Siebold et Zucc. were planted in the 1970s for environmental mitigation and natural forest restoration. We considered non-native Ginkgo biloba L. as a conifer because of its soft wood, together with native Pinus densiflora Siebold et Zucc. and other planted non-native conifers such as Cedrus deodara (Roxb.) G.Don. 2.2. Surveys after typhoons Two large typhoons in 2019 caused severe damage to this region. In Typhoon Faxi, the maximum wind speed was 23.4 m/s (84.2 km/h) on 9 September with the lowest pressure of 969.1 hPa at Yokohama Local Meteorological Observatory (Japan Meteorological Agency 2020a). Immediately after the storm, we surveyed fallen trees before traffic recovery, including trunk-break and uprooting. Few trees were removed to recover traffic before our survey. Because it was difficult to look for old tags fixed in previous tree measurements, height (length of fallen tree), and diameter (DBH) were measured for all 28 fallen trees found (Supporting Material 2). In Typhoon Hagibis, the maximum wind speed was 23.8 m/s (85.7 km/h) on 12 October with the lowest pressure of 966.0 hPa at Yokohama Local Meteorological Observatory (Japan Meteorological Agency 2020b). We surveyed fallen trees immediately after the storm but found no significant treefalls. The maximum wind speeds between the tree measurement and two typhoons were 19.9 m/s in 2012 (Japan Meteorological Agency 2025), and no significant treefalls were recorded. 2.3. Mechanical model We assumed a simple tree structure model. Following the pipe model theory, biomass is assumed to be distributed evenly vertically from ground to the top of the tree (Shinozaki et al. 1964). In this model, a tree is considered a combination of many pipes with a single leaf on top; the lower part of the trunk combines all pipes. The combined pipes branch and form twigs in the upper crown. This assumption has often been used to study tree biomass allocation (Huan et al. 2023). The point of wind drag application was assumed at 75 % height of the tree, approximately the center of the crown. Crown size (m) was estimated by DBH (cm) based on our preliminary survey, as ln crown area = 0.796 × ln basal area + 7.2315 eqn 1 where basal area = π×(( DBH )/2) 2 . The vertical and horizontal profile areas of crowns were assumed to be the same as a sphere. The wind drag (N) was calculated as wind drag = vertical crown profile area × wind drag coefficient × 0.5 × air density × wind speed 2 × gravitational acceleration eqn 2 where the wind drag coefficient was assumed to be 0.75 (Bekkers et al. 2022). We calculated wind drag for each maximum wind speed of two typhoons for all trees. We obtain the bending moment at trunk base from wind drag and tree height as a cantilever beam. We divided the cross-section of the trunk into 20 parts and calculated compressive stress at each region. The maximum value of 20 regions was used as the maximum compressive stress of the tree. Although we can calculate cavities and rotten parts in the trunk using this method, we assumed a simple cylindrical trunk because we do not have detailed information for each tree. We also assumed that Young’s modulus was equal in both tensile and compressive sides so that the neutral plane was the center of the trunk. In general, tensile strength is greater than compressive strength for wood (Senalik and Farber 2021), and we considered only compressive strength in treefall risk evaluation. We did not consider the own weight of the tree because the compressive stress should be less than 255 kPa as the wood bulk density < 1.0 g/cm 3 and the maximum tree height in this site is 26 m, thus smaller than the compressive stress by storm winds. 2.4. Statistical model Logistic regression was applied by assuming the typhoons’ treefall as the objective variable, and maximum compressive stress and qualitative tree types (conifer/broadleaved) as explanatory variables. Because we have a limited number of fallen trees, we distinguished conifer or broadleaved trees instead of species. Because two typhoons gave different maximum compressive stress to the same tree, two data for one tree were used for the regression. Because we could not specify the tags on fallen trees, we could not specify the fallen individuals in the tree measurement dataset. We removed the log-likelihood value assuming “not-fallen” for fallen trees to prevent double-counting the fallen individuals in regression. STAN was used for this adjustment (Stan Development Team, 2025). 2.5. Risk assessment The treefall should be compared to other natural disasters in risk evaluation and decision-making by the general public. In this site, the probability of an earthquake with a seismic intensity of 6 Lower (weaker wooden structures may lean or even collapse) or greater occurring within the next 30 years (approximately one human generation) is about 30% (NIED 2025). Thus, we assumed the return period of the strongest typhoon to be 30 years. The maximum wind speed within 30 years, 33.7 m/s (121 km/h), was estimated by extreme value analysis (Gumbel 1954) based on the historical maximum wind speed at Yokohama Local Meteorological Observatory (Japan Meteorological Agency 2025). Maximum compressive stress by the hypothetical typhoon was calculated for each tree, and treefall probability was estimated by the statistical model obtained from two typhoons. We considered treefall energy in addition to treefall probability. Quite slender trees might have a high treefall probability; however, damage by the fall will be much smaller than that of trees with large biomass. To consider this, we calculated the expected amount of energy by treefall as the potential energy of the tree × treefall probability . Potential energy was calculated as tree weight × gravitational acceleration × tree height / 2. The tree height / 2 represents the average distance that the tree biomass falls. A treefall risk control diagram representing treefall probability and expected amount of energy on a two-dimensional graph plane of trunk diameter and tree height was depicted. We can identify high-risk trees in this diagram. Major plant groups include conifers, the most dominant non-native tree, Cinnamomum camphor , and the most dominant native tree, Castanopsis sieboldii subsp. sieboldii , and native subcanopy trees combining Euria japonica , Neolitsea sericea , and Ilex integra were compared on the treefall risk control diagram. There are various tree height-diameter regression models (Li et al., 2015), and we used a simple model, height = 2 × b ×exp( a × DBH ) /(1 +exp( a × DBH )) – b , to fit species-specific tree growth curves of major plant groups. 3. Results 3.1. Treefall survey In Typhoon Faxai, 22 trees among 13333 broadleaved trees had fallen (0.17%), and six trees among 570 coniferous trees had fallen (1.1%). In Typhoon Hagibis, we did not find fallen trees. The calculated maximum compressive stress in broadleaved trees ranged from 12.6 to 20.3 MPa as the 25th and 75th percentiles (Fig. 2). Those of fallen broadleaved trees ranged from 25.5 to 34.3 MPa. All conifers’ calculated maximum compressive stress ranged from 13.7 to 20.7 MPa, and fallen coniferous trees ranged from 17.8 to 27.2 MPa. The area under the curve (AUC) for predicting treefall during Typhoon Faxai using the calculated maximum compressive stress was 0.94 for broadleaf trees and 0.77 for conifers. 3.2. Statistical model The maximum compressive stress by the mechanical model was a significant factor in predicting treefall (Supporting Material 3). Conifers fell significantly than broadleaved trees. Treefall probability rose above 1% at 24.0 MPa in coniferous trees and 38.4 MPa in broadleaved trees (Fig. 3). 3.3. Risk assessment In the treefall risk control diagram representing treefall probability and expected treefall energy by the strongest typhoon in the next 30 years (Fig. 4), the average trunk diameter (DBH)–Height growth curve closed to 10% treefall probability in Cinnamomum camphora and intersected in conifers at the tripping point changing from height growth to DBH growth (Fig. 4ab). Those of native dominant Castanopsis sieboldii subsp. sieboldii and native subcanopy trees were not close to the 10% treefall contour line (Fig. 4a). Proportion of trees with more than 10% treefall probability was 36.7% in conifer trees and 28.7% in non-native-dominant Cinnamomum camphora (Fig. 4, Supporting Material 4), although less than 10% in other types. If the magnitude of damage was evaluated by considering expected treefall energy (potential energy × treefall probability), the proportion of trees with more than 31.6 kJ was 14.6% in Cinnamomum camphora and 5.4% in conifers (Fig. 4, Supporting Material 5). Cinnamomum camphora and conifers were high-risk taxa, and locally native subcanopy trees were safer taxa. 4. Discussion Trunk diameter (DBH) and height are often measured as basic tree metrics, and treefall risk and the expected magnitude of treefall energy can be estimated by these metrics. Using a treefall risk control diagram based on DBH and height (Fig. 4), we can judge the risk of each tree and take mitigation measures such as removing, cutting down the treetop, or thinning and supporting until enough trunk diameter growth. By considering the maximum compressive stress, we can estimate treefall probability by hypothetical storms of various wind speeds, because maximum compressive stress roughly correlates to the square of wind speed (eqn 2). The same return period to earthquake enables decision makers to take seismic retrofit or measures against treefalls. The model based on the maximum compressive stress has another benefit: we can estimate treefall probability for trunks with cavities using the parameters of our regression model, if trunk cross-section data is available especially for important old trees in parks or gardens. For practitioners, conifers and non-native trees which are taller than locally-native species are better avoided for planting, or tree height monitoring and management are necessary if these species are planted. Canopy and subcanopy species native to the local climax forest were safer taxa for urban green areas. Among tree taxa, the native climax forest in this site is dominated by Castanopsis sieboldii subsp. sieboldii and subcanopy trees, such as Euria japonica , Neolitsea sericea , and Ilex integra, were safer than non-native Cinnamomum camphora and conifers. The maximum height of local climax forest species is smaller than that of some non-native trees and grows parallel to the contour of treefall probability (Fig. 4a). Although tall tree height is the key ecological trait for plants to compete in the community assembly process (Koike 2001), treefall by frequent typhoons may prevent taller species from dominating at this coastal site, and low treefall risk species survive in natural ecosystems. Conifers showed higher treefall risk than broadleaved trees (Fig. 3, Supporting Material 3). Compressive strengths in conifers are generally lower than those of broadleaved trees (Senalik and Farber 2021) because of lower wood density (Salisbury et al. 2023) and might cause conifers’ treefall due to lower stress. Prospects of future research This study aimed to propose a feasible approach for practitioners. However, a modest amount of additional fieldwork could further improve predictive performance. Local environmental conditions, such as dry ridge sites and nutrient-poor soils, may suppress height growth (Muscarella et al. 2020; Siipilehto et al. 2023). As a result, we observed many low-risk individuals even within the high-risk species Cinnamomum camphora (Fig. 4c), as well as several high-risk individuals among otherwise low-risk taxa at our study site. Incorporating site quality and predicting site-specific maximum height for each species may improve risk management in tree-planting planning. The estimated range of maximum compressive stress in failed trees was comparable to published values of compressive strength parallel to the grain (Figs 2 and 3; Senalik and Farber 2021), despite several simplifying assumptions in our analysis. We grouped uprooting together with stem breakage; however, uprooting can involve lower stem stress than stem breakage, even though a strong bending moment at the stem base contributes to both failure modes (Ancelin et al. 2004). In principle, a categorical variable for failure mode (uprooting vs. stem breakage) could be added in the regression model to estimate mode-specific risks. However, because failure occurs via the more likely (weaker) mode, our current model effectively reflects the probability of failure governed by the limiting (weaker) mode for each tree. Wood strength is orthotropic, and actual failure processes are complex (Senalik and Farber 2021). In addition, the wind drag force was necessarily approximated because foliage density, crown shape, and local wind speed vary with topography and surrounding trees and buildings. These sources of variability in both loading and material properties introduce uncertainty in treefall predictions. In this study, such variability was implicitly captured during statistical calibration. Future work could improve accuracy by incorporating city-scale wind simulations (Huang et al. 2024) driven by wind speeds and directions from storms over the past >30 years. It may also be possible to identify additional readily measured tree attributes that improve predictability. Limitations of this study The number of fallen trees in our dataset was limited. Accumulating data on both fallen and non-fallen trees in urban landscapes (Koeser et al. 2025), ideally paired with local wind information, will enable more reliable risk assessment in the future. Declarations Author contributions All authors contributed to the presented study. Conceptualization, and model building, F.K.; tree survey before storms, Y.M.; treefall survey after storms, K.K., formal analysis K.K. and F.K.; writing K.K. and F.K. All authors have read and agreed to the version of the manuscript. Conflicts of interest Not applicable. Open Access CC BY Data Accessibility The data used are available as Supporting Materials. References Ancelin, P., Courbaud, B. & Fourcaud, T. Development of an individual tree-based mechanical model to predict wind damage within forest stands. For. Ecol. Manag. 203 , 101–121. https://doi.org/10.1016/j.foreco.2004.07.067 (2004). ANSI. ANSI A300 Tree Care Standards. (2023). https://treecareindustryassociation.org/business-support/ansi-a300-standards/ Bartens, J., Day, S. D., Harris, J. R., Dove, J. E. & Wynn, T. M. Can urban tree roots improve infiltration through compacted subsoils for stormwater management? J. Environ. Qual. 37 (6), 2048–2057. https://doi.org/10.2134/jeq2008.0117 (2008). Bekkers, C. C. A., Angelou, N. & Dellwik, E. Drag coefficient and frontal area of a solitary mature tree. J. Wind Eng. Ind. Aerodyn. 220 , 104854. https://doi.org/10.1016/j.jweia.2021.104854 (2022). Bowler, D. E., Buyung–Ali, L. M., Knight, T. M. & Pullin, A. S. Urban greening to cool towns and cities: A systematic review of the empirical evidence. Landsc. Urban Plann. 97 , 147–155. https://doi.org/10.1016/j.landurbplan.2010.05.006 (2010). Bratman, G. N., Daily, G. C., Levy, B. J. & Gross, J. J. The benefits of nature experience: Improved affect and cognition. Landsc. Urban Plann. 138 , 41–50. https://doi.org/10.1016/j.landurbplan.2015.02.005 (2015). Grazide, C. et al. Wood heterogeneities and failure load of timber structural elements: a statistical approach. Wood Sci. Technol. 49 , 421–440. https://doi.org/10.1007/s00226-015-0706-z (2015). Gumbel, E. J. Statistical Theory of Extreme Values and Some Practical Applications (National Bureau of Standards, 1954). Huang, C. et al. Pipe Model Can Accurately Estimate Crown Biomass of Larch ( Larix olgensis ) Plantation Forest in Northeast China. Forests 14 , 400. https://doi.org/10.3390/f14020400 (2023). Huang, M., Liao, S., Lou, W., Lin, W. & Kareem, A. Multi-scale simulation of typhoon wind field at building scale utilizing mesoscale model with nested large eddy simulation. J. Wind Eng. Ind. Aerodyn. 249 , 105733. https://doi.org/10.1016/j.jweia.2024.105733 (2024). Jackson, T. et al. Finite element analysis of trees in the wind based on terrestrial laser scanning data. Agric. For. Meteorol. 265 , 137–144. https://doi.org/10.1016/j.agrformet.2018.11.014 (2019). Japan Meteorological Agency. Report on natural phenomena during disasters, No. 2, 2020. (2020a). https://www.jma.go.jp/jma/kishou/books/saigaiji/saigaiji_202002.pdf , Accessed 5 August 2025. Japan Meteorological Agency. Report on natural phenomena during disasters, No. 3, 2020. (2020b). https://www.jma.go.jp/jma/kishou/books/saigaiji/saigaiji_2019/saigaiji_202003.pdf , Accessed 5 August 2025. Japan Meteorological Agency. Historical weather data. (2025). https://www.data.jma.go.jp/stats/etrn/view/annually_s.php?prec_no=47670&year=&month=&day=&view=a3 , Accessed 5 August 2025. James, K. R., Hallam, C. & Spencer, D. Forces on trees during storms and implications for failure. Arboric. Urban Forestry . 40 (4), 219–229. https://doi.org/10.48044/jauf.2014.022 (2014). Jim, C. Y. & Liu, H. H. T. Storm damage on urban trees in Guangzhou, China. Landsc. Urban Plann. 38 (1–2), 45–59. https://doi.org/10.1016/S0169-2046(97)00018-2 (1997). Koeser, A. K. et al. Can professionals gauge likelihood of failure? – Insights from tropical storm Matthew. Urban Forestry Urban Green. 52 , 126701. https://doi.org/10.1016/j.ufug.2020.126701 (2020). Koeser, A. K. et al. Defective or just different? Observed storm failure in four urban tree growth patterns. Forests 14 , 988. https://doi.org/10.3390/f14050988 (2023). Koike, F. Reconstruction of two-dimensional tree and forest canopy profiles using photographs. J. Appl. Ecol. 22 , 921–929. https://doi.org/10.2307/2403240 (1985). Koike, F. Plant traits as predictors of woody species dominance in climax forest communities. J. Veg. Sci. 12 , 327–336. https://doi.org/10.2307/3236846 (2001). Kurose, T. & Koike, F. Detecting behavioral preferences of dog walkers in urban green spaces. Japanese J. Conserv. Ecol. 25 , 109–115. https://doi.org/10.18960/hozen.2002 (2020). Li, Y. Q. et al. Development and Evaluation of Models for the Relationship between Tree Height and Diameter at Breast Height for Chinese-Fir Plantations in Subtropical China. PLOS One . 0125118. https://doi.org/10.1371/journal.pone.0125118 (2015). Luo, Y. & Ai, X. Wind risk assessment of urban street trees based on wind-induced fragility. Disaster Prev. Resil. 1 , 7. http://dx.doi.org/10.20517/dpr.2022.05 (2022). Miyawaki, A. Vegetation of Japan: Kanto (Shibundo), 1986). Muscarella, R., Kolyaie, S., Morton, D. C., Zimmerman, J. K. & Uriarte, M. Effects of topography on tropical forest structure depend on climate context. J. Ecol. 108 , 145–159. https://doi.org/10.1111/1365-2745.13261 (2020). NIED. Japan seismic hazard information station. (2025). https://www.j-shis.bosai.go.jp/en/ , Accessed 5 August 2025. Öztürk, Ş., Yılmaz, K., Dinçer, A. E. & Kalpakcı, V. Effect of urbanization on surface runoff and performance of green roofs and permeable pavement for mitigating urban floods. Nat. Hazards . 120 , 12375–12399. https://doi.org/10.1007/s11069-024-06688-w (2024). Salisbury, A. B. et al. Predictors of tropical cyclone-induced urban tree failure: an international scoping review. Front. Forests Global Change . 6 , 1168495. https://doi.org/10.3389/ffgc.2023.1168495 (2023). Senalik, C. A. & Farber, B. Mechanical Properties of Wood. In Ross R, Wood handbook: Wood as an engineering material. USDA Forest Service. (2021). https://research.fs.usda.gov/treesearch/62200 Shinozaki, K., Yoda, K., Hozumi, K. & Kira, T. A quantitative analysis of plant form-the pipe model theory: I. basic analyses. Japanese J. Ecol. 14 , 97–105. https://doi.org/10.18960/seitai.14.3_97 (1964). Siipilehto, J., Sarkkola, S., Nuutinen, Y. & Mehtätalo, L. Predicting height-diameter relationship in uneven-aged stands in Finland. For. Ecol. Manag. 549 , 121486. https://doi.org/10.1016/j.foreco.2023.121486 (2023). Stan Development Team. Stan: Software for Bayesian Data Analysis. (2025). Available at: https://mc-stan.org/ van Haaften, M. A., Meuwissen, M. P. M., Gardebroek, C. & Kopinga, J. Trends in financial damage related to urban tree failure in the Netherlands. Urban Forestry and Urban Greening, 15 , 15–21. (2016). https://doi.org/10.1016/j.ufug.2015.11.002 Xiao, Q. & McPherson, E. G. Rainfall interception by Santa Monica’s municipal urban forest. Urban Ecosyst. 6 (4), 291–302. https://doi.org/10.1023/B:UECO.0000004828.05143.67 (2002). Zur, S. L., Bahat, S., Amara, R. & Farber, Y. Risk assessment of street trees failure by non-experts: a case study from the North of Israel. Nat. Hazards . 121 , 161–174. https://doi.org/10.1007/s11069-024-06835-3 (2025). Tables Table 1 Trees in the study area before treefall by two Typhoons in 2019. Dominant taxa up to 80% cumulative basal area are shown. Species Status Number of individuals Share in basal area (%) Cinnamomum camphora (L.) J.Presl Planted and naturalized non-native tree 903 21.6 Cerasus spp. Deciduous native populations and many planted trees of cultivars 877 12.4 Castanopsis sieboldii (Makino) Hatus. ex T.Yamaz. et Mashiba subsp. sieboldii Native tree, some planted 1304 12.3 Quercus glauca Thunb. Native tree, mainly planted 2123 7.5 Machilus thunbergii Siebold et Zucc. Native tree, planted with natural reproduction 1365 6.5 Cornus controversa Hemsl. var. controversa Native deciduous tree 635 6.1 Quercus myrsinifolia Blume Native tree, mainly planted with natural reproduction 1174 4.2 Zelkova serrata (Thunb.) Makino Planted native deciduous tree 193 3.6 Ginkgo biloba L. Non-native deciduous gymnosperm included in conifer 125 3.6 Liquidambar styraciflua L. Non-native deciduous tree 59 2.6 Others 5191 19.6 Additional Declarations No competing interests reported. Supplementary Files SupportingMaterial1.csv SupportingMaterials2.csv SupportingMaterials35.pdf Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 13 May, 2026 Reviews received at journal 01 May, 2026 Reviews received at journal 20 Apr, 2026 Reviewers agreed at journal 11 Apr, 2026 Reviewers agreed at journal 30 Mar, 2026 Reviewers invited by journal 28 Mar, 2026 Editor assigned by journal 23 Mar, 2026 Editor invited by journal 23 Mar, 2026 Submission checks completed at journal 20 Mar, 2026 First submitted to journal 20 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9147742","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":633326031,"identity":"a0f7298f-bd29-4b28-866d-260e8c9c5aed","order_by":0,"name":"Kohei Katayama","email":"","orcid":"","institution":"Yokohama National University","correspondingAuthor":false,"prefix":"","firstName":"Kohei","middleName":"","lastName":"Katayama","suffix":""},{"id":633326032,"identity":"0ace8b07-0d5e-402f-b408-233a686bc1d4","order_by":1,"name":"Yukira Mochida","email":"","orcid":"","institution":"Yokohama National University","correspondingAuthor":false,"prefix":"","firstName":"Yukira","middleName":"","lastName":"Mochida","suffix":""},{"id":633326033,"identity":"07ff9a2a-c48a-4756-bed5-9371e061ef44","order_by":2,"name":"Fumito Koike","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBElEQVRIiWNgGAWjYBACCQbGZiAlx8DADBGQg8kwE9BiDFQCUWNMhBawnDFcTWIDIYdJtjc3G3xgMEhc285/8AHjjsPpG273GDD8qGFgN8ehRZrnYHPiDKCWbYeZmQ0Yz6TlbrhzxoCx5xgDsyUO++QkEpsP8/77A9LCJv23zSZ3w40cAwbeBgZmgwN4tPyB2MImwdgmkW4A1ML4F48WaaCWZAaEFpsEkBZmfLZI9hxsNuxhMDAGajE2YGxLM5x5I63gsMwxCZx+kTje/ljiB4OB7LbzBx8+YGw7LM93I3njwzc1Nsm4QgwTKADdA0QSyQZEa5GHuseOeC2jYBSMglEwzAEACdhUG08npxIAAAAASUVORK5CYII=","orcid":"","institution":"Yokohama National University","correspondingAuthor":true,"prefix":"","firstName":"Fumito","middleName":"","lastName":"Koike","suffix":""}],"badges":[],"createdAt":"2026-03-17 10:41:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9147742/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9147742/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108389362,"identity":"fb2bdb50-ed70-4c28-ae47-8b0ce956b3e3","added_by":"auto","created_at":"2026-05-04 06:49:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":306840,"visible":true,"origin":"","legend":"\u003cp\u003eProcedure of treefall risk assessment assuming the strongest hypothetical storm in the return period of 30 years based on the damage by Typhoon Faxai and Hagibis in 2019. Purple texts represent the mechanical model, green represents data from storms in 2019, blue those of return period 30 years, and brown represents obtained statistical model.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/dc4be217cc44a147c3a5dd85.png"},{"id":108493295,"identity":"27b8600c-ae08-4026-b2e7-bcad1df7042f","added_by":"auto","created_at":"2026-05-05 09:59:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":89745,"visible":true,"origin":"","legend":"\u003cp\u003eRelative frequency distribution of the reconstructed maximum compressive stress in Typhoon Faxai. Each histogram is adjusted to 100% in all individuals. Please note that the number of all broadleaved or coniferous trees includes those of fallen trees by Typhoons in 2019.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/51618b19c77bf373bb6c627c.png"},{"id":108389367,"identity":"b24ee1bc-1838-4bfb-b0b4-952f97d44828","added_by":"auto","created_at":"2026-05-04 06:49:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":84954,"visible":true,"origin":"","legend":"\u003cp\u003eEstimated logistic regression model predicting treefall probability from the maximum compressive stress in each tree.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/e31a0e54175918635d525270.png"},{"id":108803997,"identity":"baa9b736-ad37-4d20-89bb-b9d6a2730407","added_by":"auto","created_at":"2026-05-08 15:14:11","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":796539,"visible":true,"origin":"","legend":"\u003cp\u003eTreefall risk control diagrams representing treefall probability (green lines and figures in percentage) and expected amount of treefall energy (purple contour lines and figures ×10\u003csup\u003e3\u003c/sup\u003e kg m G) on the two-dimensional plane of trunk diameter and tree height. The average diameter (DBH)-height curves are shown for the dominant non-native \u003cem\u003eCinnamomum camphor\u003c/em\u003e, the dominant native \u003cem\u003eCastanopsis sieboldii\u003c/em\u003e subsp. \u003cem\u003esieboldii\u003c/em\u003e, and native subcanopy trees combining \u003cem\u003eEuria japonica\u003c/em\u003e, \u003cem\u003eNeolitsea sericea\u003c/em\u003e, and \u003cem\u003eIlex integra\u003c/em\u003e. Vertical fine lines show deviation of individual trees from the average diameter (DBH)-height curve.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/e4ddc6c66e843b1e991a9117.png"},{"id":109204521,"identity":"e7d3fc36-d17b-4e8c-bf8f-c0197d5a2b66","added_by":"auto","created_at":"2026-05-13 15:00:39","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1331629,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/6802ce5a-5bbd-48a0-82ff-0c512e6b6e85.pdf"},{"id":108492759,"identity":"49148ee4-10a9-452a-80b2-7ad060fa4e45","added_by":"auto","created_at":"2026-05-05 09:58:33","extension":"csv","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":291476,"visible":true,"origin":"","legend":"","description":"","filename":"SupportingMaterial1.csv","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/b5f8c50e0edda12d1b7d3a42.csv"},{"id":108389366,"identity":"2def7fa9-7b58-4cc0-a69a-f4f1c2127015","added_by":"auto","created_at":"2026-05-04 06:49:13","extension":"csv","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":909,"visible":true,"origin":"","legend":"","description":"","filename":"SupportingMaterials2.csv","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/eea65f87f777e6d7e29f0bb3.csv"},{"id":108389368,"identity":"bb067c0b-67ef-42ba-8da8-50cc0aca3e30","added_by":"auto","created_at":"2026-05-04 06:49:13","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":182403,"visible":true,"origin":"","legend":"","description":"","filename":"SupportingMaterials35.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9147742/v1/b8098ac1776af9e31ee085b3.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Treefall risk assessment in an urban green area for a hypothetical 30-year return period storm using damage data from Typhoons Faxai and Hagibis in 2019","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eUrban trees play an essential role in environments by supporting human well-being through mental health benefits (Kurose and Koike 2020; Bratman et al., 2015), mitigating high temperatures (Bowler et al., 2010), and reducing flood risk through enhanced rainfall interception and infiltration (Xiao \u0026amp; McPherson, 2002; Bartens et al, 2008; Öztürk et al. 2025). Despite these benefits, fallen trees during storms can cause injuries or fatalities, obstruct traffic, and damage electrical cables and other urban infrastructure (Jim \u0026amp; Liu, 1997; van Haaften et al., 2016).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eReliable and feasible assessment of treefall risk for practitioners is essential for balancing the benefits and risks associated with urban trees (Zur et al. 2025). One typical risk assessment is to survey many aspects such as trunk decay, species, and tree architecture, and estimate the risk by expert judgement (Kane \u0026amp; Ryan, 2004; Smiley et al., 2017; ANSI 2023). Such qualitative risk assessment by specialists worked significantly; however, reliability is limited (Koeser et al., 2020; Koeser et al., 2023).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAmong many causes of tree failures (van Haaften et al. 2021), storm wind is the most important cause in subtropical and warm-temperate regions with Typhoon and Hurricane damage (Koeser et al., 2020; Salisbury et al., 2023). Material and structural mechanics have been applied to assess the likelihood of tree failure by wind (James et al., 2014; Luo and Ai 2022). However, wood strength differs significantly within a tree (Grazide et al. 2015; Luo and Ai 2022), and in urban landscapes, each tree has a unique, complicated form of trunk, branch, and crown (Koike 1985). It is unrealistic to measure the detailed structure of every tree in a city (Jackson et al. 2019) and assess treefall risk using purely mechanistic methods. Because practitioners can readily obtain tree height and diameter, we developed a simple mechanical model to estimate a hypothetical maximum compressive stress for each individual tree as a mechanically derived index of failure susceptibility. Peak wind speed in urban areas varies with local factors such as buildings and other artificial structures (Huang et al. 2024). In most cities, however, the only readily available wind data come from the nearest meteorological station, and site-specific wind conditions at individual trees are difficult to obtain. Given these uncertainties, we adopted a hybrid framework (Luo and Ai 2022) that links a simple mechanical model with a stochastic approach, using the estimated maximum compressive stress as a latent variable (Fig. 1). Specifically, we modeled treefall occurrence (yes/no) as the response and used the stress-based latent variable as a mechanistically interpretable predictor. Although this stress is a rough proxy, statistical models can calibrate and incorporate additional covariates, and the stress-based latent variable provides a mechanistically interpretable link between mechanical and stochastic components. Stochastic modeling also requires data on both fallen and non-fallen trees, but storms that cause large-scale damage are rare in any given city, and systematic surveys of treefall incidents are difficult to obtain (Koeser et al. 2020; Koeser et al. 2025).\u003c/p\u003e\n\u003cp\u003eIn this study, we developed a simple mechanical model based on tree diameter and height and then fit a regression model using observed treefall data from Typhoons Faxai and Hagibis (2019). The model estimates treefall probability using the maximum 10-min mean wind speed from the nearest meteorological station, with the hypothetical maximum compressive stress treated as a latent variable. Importantly, all predictors (diameter, height, species, and station wind speed) are routinely available in municipal inventories, enabling city-wide application. Using this model, we assessed the treefall risk of every tree under a hypothetical extreme storm corresponding to a 30-year return period. Although treefall data are currently limited, to our knowledge, this study is among the first to predict future treefall probability using empirically observed treefall records and to propose a quantitative approach that practitioners can use to manage treefall risk.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cp\u003e\u003cem\u003e2.1. Study site and trees\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe research site was an urban university campus of 455531 m\u003csup\u003e2\u003c/sup\u003e with 97 buildings and about 10000 students and staff. The campus is on a coastal terrace in the Tokyo Metropolitan Area (latitude 35.47, longitude 139.59, altitude 50 m). An estuary with saline water is located 1.6 km distant, and the current coastal line after recent land reclamation is 4.4 km. Average temperature is 16.2\u0026nbsp;°C, average annual precipitation is 1730 mm, and it belongs to the warm-temperate broadleaved forest zone (Miyawaki 1986). The old golf course, with a few remnant forests, became a university campus in the 1970s, and many trees were planted simultaneously to construct the campus. The remnant forest in the study area suggests that the climax forest is dominated by \u003cem\u003eCastanopsis sieboldii\u0026nbsp;\u003c/em\u003e(Makino) Hatus. ex T.Yamaz. et Mashiba subsp.\u003cem\u003e\u0026nbsp;sieboldii\u003c/em\u003e with subcanopy trees such as \u003cem\u003eEuria japonica\u003c/em\u003e Thunb. var.\u003cem\u003e\u0026nbsp;japonica\u003c/em\u003e, \u003cem\u003eNeolitsea sericea\u003c/em\u003e (Blume) Koidz., and \u003cem\u003eIlex integra\u003c/em\u003e Thunb.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll tree surveys on trunk diameter at breast height (DBH) and tree height were made by Mochida Lab from 2010 to 2016 (Supporting Material 1). The maximum tree height in the site was 26 m, and the maximum DBH was 156 cm. The most dominant species was the planted non-native \u003cem\u003eCinnamomum camphora\u0026nbsp;\u003c/em\u003e(L.) J.Presl, and cherry \u003cem\u003eCerasus\u003c/em\u003e spp. followed it. Almost all cherry trees are grafted cultivars, although there are a few native populations. There are native populations of \u003cem\u003eCastanopsis sieboldii\u003c/em\u003e in small remnant forests, although this tree was planted in several places. Regionally native evergreen oaks \u003cem\u003eQuercus\u003c/em\u003e spp. and \u003cem\u003eMachilus\u003c/em\u003e \u003cem\u003ethunbergii\u003c/em\u003e Siebold et Zucc. were planted in the 1970s for environmental mitigation and natural forest restoration. We considered non-native \u003cem\u003eGinkgo biloba\u003c/em\u003e L. as a conifer because of its soft wood, together with native \u003cem\u003ePinus densiflora\u003c/em\u003e Siebold et Zucc. and other planted non-native conifers such as \u003cem\u003eCedrus deodara\u003c/em\u003e (Roxb.) G.Don.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.2. Surveys after typhoons\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTwo large typhoons in 2019 caused severe damage to this region. In Typhoon Faxi, the maximum wind speed was 23.4 m/s (84.2 km/h) on 9 September with the lowest pressure of 969.1 hPa at Yokohama Local Meteorological Observatory (Japan Meteorological Agency 2020a). Immediately after the storm, we surveyed fallen trees before traffic recovery, including trunk-break and uprooting. Few trees were removed to recover traffic before our survey. Because it was difficult to look for old tags fixed in previous tree measurements, height (length of fallen tree), and diameter (DBH) were measured for all 28 fallen trees found (Supporting Material 2). In Typhoon Hagibis, the maximum wind speed was 23.8 m/s (85.7 km/h) on 12 October with the lowest pressure of 966.0 hPa at Yokohama Local Meteorological Observatory (Japan Meteorological Agency 2020b). We surveyed fallen trees immediately after the storm but found no significant treefalls. The maximum wind speeds between the tree measurement and two typhoons were 19.9 m/s in 2012 (Japan Meteorological Agency 2025), and no significant treefalls were recorded.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.3. Mechanical model\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWe assumed a simple tree structure model. Following the pipe model theory, biomass is assumed to be distributed evenly vertically from ground to the top of the tree (Shinozaki et al. 1964). In this model, a tree is considered a combination of many pipes with a single leaf on top; the lower part of the trunk combines all pipes. The combined pipes branch and form twigs in the upper crown. This assumption has often been used to study tree biomass allocation (Huan et al. 2023).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe point of wind drag application was assumed at 75 % height of the tree, approximately the center of the crown. Crown size (m) was estimated by DBH (cm) based on our preliminary survey, as\u003c/p\u003e\n\u003cp\u003eln \u003cem\u003ecrown area\u003c/em\u003e = 0.796 × ln\u0026nbsp;\u003cem\u003ebasal area\u003c/em\u003e + 7.2315\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;eqn 1\u003cbr\u003ewhere \u003cem\u003ebasal area\u003c/em\u003e = π×((\u003cem\u003eDBH\u003c/em\u003e)/2)\u003csup\u003e2\u003c/sup\u003e. The vertical and horizontal profile areas of crowns were assumed to be the same as a sphere. The wind drag (N) was calculated as\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ewind drag\u003c/em\u003e = \u003cem\u003evertical crown profile area\u003c/em\u003e × \u003cem\u003ewind drag coefficient\u003c/em\u003e × 0.5 ×\u003cem\u003eair density\u003c/em\u003e × \u003cem\u003ewind speed\u003c/em\u003e \u003csup\u003e2\u003c/sup\u003e × \u003cem\u003egravitational acceleration\u003c/em\u003e eqn 2\u003c/p\u003e\n\u003cp\u003ewhere the wind drag coefficient was assumed to be 0.75 (Bekkers et al. 2022). We calculated wind drag for each maximum wind speed of two typhoons for all trees.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe obtain the bending moment at trunk base from wind drag and tree height as a cantilever beam. We divided the cross-section of the trunk into 20 parts and calculated compressive stress at each region. The maximum value of 20 regions was used as the maximum compressive stress of the tree. Although we can calculate cavities and rotten parts in the trunk using this method, we assumed a simple cylindrical trunk because we do not have detailed information for each tree. We also assumed that Young’s modulus was equal in both tensile and compressive sides so that the neutral plane was the center of the trunk. In general, tensile strength is greater than compressive strength for wood (Senalik and Farber 2021), and we considered only compressive strength in treefall risk evaluation. We did not consider the own weight of the tree because the compressive stress should be less than 255 kPa as the wood bulk density \u0026lt; 1.0 g/cm\u003csup\u003e3\u003c/sup\u003e and the maximum tree height in this site is 26 m, thus smaller than the compressive stress by storm winds.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.4. Statistical model\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eLogistic regression was applied by assuming the typhoons’ treefall as the objective variable, and maximum compressive stress and qualitative tree types (conifer/broadleaved) as explanatory variables. Because we have a limited number of fallen trees, we distinguished conifer or broadleaved trees instead of species. Because two typhoons gave different maximum compressive stress to the same tree, two data for one tree were used for the regression. Because we could not specify the tags on fallen trees, we could not specify the fallen individuals in the tree measurement dataset. We removed the log-likelihood value assuming “not-fallen” for fallen trees to prevent double-counting the fallen individuals in regression. STAN was used for this adjustment (Stan Development Team, 2025).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.5. Risk assessment\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe treefall should be compared to other natural disasters in risk evaluation and decision-making by the general public. In this site, the probability of an earthquake with a seismic intensity of 6 Lower (weaker wooden structures may lean or even collapse) or greater occurring within the next 30 years (approximately one human generation) is about 30% (NIED 2025). Thus, we assumed the return period of the strongest typhoon to be 30 years. The maximum wind speed within 30 years, 33.7 m/s (121 km/h), was estimated by extreme value analysis (Gumbel 1954) based on the historical maximum wind speed at Yokohama Local Meteorological Observatory (Japan Meteorological Agency 2025). Maximum compressive stress by the hypothetical typhoon was calculated for each tree, and treefall probability was estimated by the statistical model obtained from two typhoons.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe considered treefall energy in addition to treefall probability. Quite slender trees might have a high treefall probability; however, damage by the fall will be much smaller than that of trees with large biomass. To consider this, we calculated the expected amount of energy by treefall as the \u003cem\u003epotential energy of the tree\u003c/em\u003e ×\u003cem\u003etreefall probability\u003c/em\u003e. Potential energy was calculated as\u003cem\u003e\u0026nbsp;tree weight\u003c/em\u003e × \u003cem\u003egravitational acceleration\u003c/em\u003e ×\u0026nbsp;\u003cem\u003etree height\u003c/em\u003e / 2. The \u003cem\u003etree height\u003c/em\u003e / 2 represents the average distance that the tree biomass falls.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA treefall risk control diagram representing treefall probability and expected amount of energy on a two-dimensional graph plane of trunk diameter and tree height was depicted. We can identify high-risk trees in this diagram. Major plant groups include conifers, the most dominant non-native tree, \u003cem\u003eCinnamomum camphor\u003c/em\u003e, and the most dominant native tree, \u003cem\u003eCastanopsis sieboldii\u003c/em\u003e subsp.\u003cem\u003e\u0026nbsp;sieboldii\u003c/em\u003e, and native subcanopy trees combining \u003cem\u003eEuria japonica\u003c/em\u003e, \u003cem\u003eNeolitsea sericea\u003c/em\u003e, and \u003cem\u003eIlex integra\u003c/em\u003e were compared on the treefall risk control diagram. There are various tree height-diameter regression models (Li et al., 2015), and we used a simple model, \u003cem\u003eheight\u003c/em\u003e = 2 ×\u003cem\u003eb\u003c/em\u003e ×exp(\u003cem\u003ea\u003c/em\u003e ×\u003cem\u003eDBH\u003c/em\u003e) /(1 +exp(\u003cem\u003ea\u003c/em\u003e ×\u003cem\u003eDBH\u003c/em\u003e)) – \u003cem\u003eb\u003c/em\u003e, to fit species-specific tree growth curves of major plant groups.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cem\u003e3.1. Treefall survey\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn Typhoon Faxai, 22 trees among 13333 broadleaved trees had fallen (0.17%), and six trees among 570 coniferous trees had fallen (1.1%). In Typhoon Hagibis, we did not find fallen trees. The calculated maximum compressive stress in broadleaved trees ranged from 12.6 to 20.3 MPa as the 25th and 75th percentiles (Fig. 2). Those of fallen broadleaved trees ranged from 25.5 to 34.3 MPa. All conifers’ calculated maximum compressive stress ranged from 13.7 to 20.7 MPa, and fallen coniferous trees ranged from 17.8 to 27.2 MPa. The area under the curve (AUC) for predicting treefall during Typhoon Faxai using the calculated maximum compressive stress was 0.94 for broadleaf trees and 0.77 for conifers.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.2. Statistical model\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe maximum compressive stress by the mechanical model was a significant factor in predicting treefall (Supporting Material 3). Conifers fell significantly than broadleaved trees. Treefall probability rose above 1% at 24.0 MPa in coniferous trees and 38.4 MPa in broadleaved trees (Fig. 3).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.3. Risk assessment\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn the treefall risk control diagram representing treefall probability and expected treefall energy by the strongest typhoon in the next 30 years (Fig. 4), the average trunk diameter (DBH)–Height growth curve closed to 10% treefall probability in \u003cem\u003eCinnamomum camphora\u003c/em\u003e and intersected in conifers at the tripping point changing from height growth to DBH growth (Fig. 4ab). Those of native dominant \u003cem\u003eCastanopsis sieboldii\u003c/em\u003e subsp. \u003cem\u003esieboldii\u003c/em\u003e and native subcanopy trees were not close to the 10% treefall contour line (Fig. 4a). Proportion of trees with more than 10% treefall probability was 36.7% in conifer trees and 28.7% in non-native-dominant \u003cem\u003eCinnamomum camphora\u003c/em\u003e (Fig. 4, Supporting Material 4), although less than 10% in other types. If the magnitude of damage was evaluated by considering expected treefall energy (potential energy × treefall probability), the proportion of trees with more than 31.6 kJ was 14.6% in \u003cem\u003eCinnamomum camphora\u003c/em\u003e and 5.4% in conifers (Fig. 4, Supporting Material 5). \u003cem\u003eCinnamomum camphora\u003c/em\u003e and conifers were high-risk taxa, and locally native subcanopy trees were safer taxa.\u0026nbsp;\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eTrunk diameter (DBH) and height are often measured as basic tree metrics, and treefall risk and the expected magnitude of treefall energy can be estimated by these metrics. Using a treefall risk control diagram based on DBH and height (Fig. 4), we can judge the risk of each tree and take mitigation measures such as removing, cutting down the treetop, or thinning and supporting until enough trunk diameter growth.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBy considering the maximum compressive stress, we can estimate treefall probability by hypothetical storms of various wind speeds, because maximum compressive stress roughly correlates to the square of wind speed (eqn 2). The same return period to earthquake enables decision makers to take seismic retrofit or measures against treefalls. The model based on the maximum compressive stress has another benefit: we can estimate treefall probability for trunks with cavities using the parameters of our regression model, if trunk cross-section data is available especially for important old trees in parks or gardens.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor practitioners, conifers and non-native trees which are taller than locally-native species are better avoided for planting, or tree height monitoring and management are necessary if these species are planted. Canopy and subcanopy species native to the local climax forest were safer taxa for urban green areas. Among tree taxa, the native climax forest in this site is dominated by \u003cem\u003eCastanopsis sieboldii\u003c/em\u003e subsp. \u003cem\u003esieboldii\u003c/em\u003e and subcanopy trees, such as \u003cem\u003eEuria japonica\u003c/em\u003e, \u003cem\u003eNeolitsea sericea\u003c/em\u003e, and \u003cem\u003eIlex integra,\u003c/em\u003e were safer than non-native \u003cem\u003eCinnamomum camphora\u003c/em\u003e and conifers. The maximum height of local climax forest species is smaller than that of some non-native trees and grows parallel to the contour of treefall probability (Fig. 4a). Although tall tree height is the key ecological trait for plants to compete in the community assembly process (Koike 2001), treefall by frequent typhoons may prevent taller species from dominating at this coastal site, and low treefall risk species survive in natural ecosystems. Conifers showed higher treefall risk than broadleaved trees (Fig. 3, Supporting Material 3). Compressive strengths in conifers are generally lower than those of broadleaved trees (Senalik and Farber 2021) because of lower wood density (Salisbury et al. 2023) and might cause conifers’ treefall due to lower stress.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProspects of future research\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study aimed to propose a feasible approach for practitioners. However, a modest amount of additional fieldwork could further improve predictive performance. Local environmental conditions, such as dry ridge sites and nutrient-poor soils, may suppress height growth (Muscarella et al. 2020; Siipilehto et al. 2023). As a result, we observed many low-risk individuals even within the high-risk species \u003cem\u003eCinnamomum camphora\u003c/em\u003e (Fig. 4c), as well as several high-risk individuals among otherwise low-risk taxa at our study site. Incorporating site quality and predicting site-specific maximum height for each species may improve risk management in tree-planting planning.\u003c/p\u003e\n\u003cp\u003eThe estimated range of maximum compressive stress in failed trees was comparable to published values of compressive strength parallel to the grain (Figs 2 and 3; Senalik and Farber 2021), despite several simplifying assumptions in our analysis. We grouped uprooting together with stem breakage; however, uprooting can involve lower stem stress than stem breakage, even though a strong bending moment at the stem base contributes to both failure modes (Ancelin et al. 2004). In principle, a categorical variable for failure mode (uprooting vs. stem breakage) could be added in the regression model to estimate mode-specific risks. However, because failure occurs via the more likely (weaker) mode, our current model effectively reflects the probability of failure governed by the limiting (weaker) mode for each tree.\u003c/p\u003e\n\u003cp\u003eWood strength is orthotropic, and actual failure processes are complex (Senalik and Farber 2021). In addition, the wind drag force was necessarily approximated because foliage density, crown shape, and local wind speed vary with topography and surrounding trees and buildings. These sources of variability in both loading and material properties introduce uncertainty in treefall predictions. In this study, such variability was implicitly captured during statistical calibration. Future work could improve accuracy by incorporating city-scale wind simulations (Huang et al. 2024) driven by wind speeds and directions from storms over the past \u0026gt;30 years. It may also be possible to identify additional readily measured tree attributes that improve predictability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations of this study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe number of fallen trees in our dataset was limited. Accumulating data on both fallen and non-fallen trees in urban landscapes (Koeser et al. 2025), ideally paired with local wind information, will enable more reliable risk assessment in the future.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e All authors contributed to the presented study. Conceptualization, and model building, F.K.; tree survey before storms, Y.M.; treefall survey after storms, K.K., formal analysis K.K. and F.K.; writing K.K. and F.K. All authors have read and agreed to the version of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOpen Access\u003c/strong\u003e CC BY\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Accessibility\u0026nbsp;\u003c/strong\u003eThe data used are available as Supporting Materials.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAncelin, P., Courbaud, B. \u0026amp; Fourcaud, T. Development of an individual tree-based mechanical model to predict wind damage within forest stands. \u003cem\u003eFor. Ecol. Manag.\u003c/em\u003e \u003cstrong\u003e203\u003c/strong\u003e, 101\u0026ndash;121. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2004.07.067\u003c/span\u003e\u003c/span\u003e (2004).\u003c/li\u003e\n\u003cli\u003eANSI. ANSI A300 Tree Care Standards. (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://treecareindustryassociation.org/business-support/ansi-a300-standards/\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003eBartens, J., Day, S. D., Harris, J. R., Dove, J. E. \u0026amp; Wynn, T. M. Can urban tree roots improve infiltration through compacted subsoils for stormwater management? \u003cem\u003eJ. Environ. Qual.\u003c/em\u003e \u003cstrong\u003e37\u003c/strong\u003e (6), 2048\u0026ndash;2057. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2134/jeq2008.0117\u003c/span\u003e\u003c/span\u003e (2008).\u003c/li\u003e\n\u003cli\u003eBekkers, C. C. A., Angelou, N. \u0026amp; Dellwik, E. Drag coefficient and frontal area of a solitary mature tree. \u003cem\u003eJ. Wind Eng. Ind. Aerodyn.\u003c/em\u003e \u003cstrong\u003e220\u003c/strong\u003e, 104854. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jweia.2021.104854\u003c/span\u003e\u003c/span\u003e (2022).\u003c/li\u003e\n\u003cli\u003eBowler, D. E., Buyung\u0026ndash;Ali, L. M., Knight, T. M. \u0026amp; Pullin, A. S. Urban greening to cool towns and cities: A systematic review of the empirical evidence. \u003cem\u003eLandsc. Urban Plann.\u003c/em\u003e \u003cstrong\u003e97\u003c/strong\u003e, 147\u0026ndash;155. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.landurbplan.2010.05.006\u003c/span\u003e\u003c/span\u003e (2010).\u003c/li\u003e\n\u003cli\u003eBratman, G. N., Daily, G. C., Levy, B. J. \u0026amp; Gross, J. J. The benefits of nature experience: Improved affect and cognition. \u003cem\u003eLandsc. Urban Plann.\u003c/em\u003e \u003cstrong\u003e138\u003c/strong\u003e, 41\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.landurbplan.2015.02.005\u003c/span\u003e\u003c/span\u003e (2015).\u003c/li\u003e\n\u003cli\u003eGrazide, C. et al. Wood heterogeneities and failure load of timber structural elements: a statistical approach. \u003cem\u003eWood Sci. Technol.\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 421\u0026ndash;440. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00226-015-0706-z\u003c/span\u003e\u003c/span\u003e (2015).\u003c/li\u003e\n\u003cli\u003eGumbel, E. J. \u003cem\u003eStatistical Theory of Extreme Values and Some Practical Applications\u003c/em\u003e (National Bureau of Standards, 1954).\u003c/li\u003e\n\u003cli\u003eHuang, C. et al. Pipe Model Can Accurately Estimate Crown Biomass of Larch (\u003cem\u003eLarix olgensis\u003c/em\u003e) Plantation Forest in Northeast China. \u003cem\u003eForests\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 400. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/f14020400\u003c/span\u003e\u003c/span\u003e (2023).\u003c/li\u003e\n\u003cli\u003eHuang, M., Liao, S., Lou, W., Lin, W. \u0026amp; Kareem, A. Multi-scale simulation of typhoon wind field at building scale utilizing mesoscale model with nested large eddy simulation. \u003cem\u003eJ. Wind Eng. Ind. Aerodyn.\u003c/em\u003e \u003cstrong\u003e249\u003c/strong\u003e, 105733. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jweia.2024.105733\u003c/span\u003e\u003c/span\u003e (2024).\u003c/li\u003e\n\u003cli\u003eJackson, T. et al. Finite element analysis of trees in the wind based on terrestrial laser scanning data. \u003cem\u003eAgric. For. Meteorol.\u003c/em\u003e \u003cstrong\u003e265\u003c/strong\u003e, 137\u0026ndash;144. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.agrformet.2018.11.014\u003c/span\u003e\u003c/span\u003e (2019).\u003c/li\u003e\n\u003cli\u003eJapan Meteorological Agency. Report on natural phenomena during disasters, No. 2, 2020. (2020a). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.jma.go.jp/jma/kishou/books/saigaiji/saigaiji_202002.pdf\u003c/span\u003e\u003c/span\u003e, Accessed 5 August 2025.\u003c/li\u003e\n\u003cli\u003eJapan Meteorological Agency. Report on natural phenomena during disasters, No. 3, 2020. (2020b). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.jma.go.jp/jma/kishou/books/saigaiji/saigaiji_2019/saigaiji_202003.pdf\u003c/span\u003e\u003c/span\u003e, Accessed 5 August 2025.\u003c/li\u003e\n\u003cli\u003eJapan Meteorological Agency. Historical weather data. (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.data.jma.go.jp/stats/etrn/view/annually_s.php?prec_no=47670\u0026amp;year=\u0026amp;month=\u0026amp;day=\u0026amp;view=a3\u003c/span\u003e\u003c/span\u003e, Accessed 5 August 2025.\u003c/li\u003e\n\u003cli\u003eJames, K. R., Hallam, C. \u0026amp; Spencer, D. Forces on trees during storms and implications for failure. \u003cem\u003eArboric. Urban Forestry\u003c/em\u003e. \u003cstrong\u003e40\u003c/strong\u003e (4), 219\u0026ndash;229. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.48044/jauf.2014.022\u003c/span\u003e\u003c/span\u003e (2014).\u003c/li\u003e\n\u003cli\u003eJim, C. Y. \u0026amp; Liu, H. H. T. Storm damage on urban trees in Guangzhou, China. \u003cem\u003eLandsc. Urban Plann.\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e (1\u0026ndash;2), 45\u0026ndash;59. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0169-2046(97)00018-2\u003c/span\u003e\u003c/span\u003e (1997).\u003c/li\u003e\n\u003cli\u003eKoeser, A. K. et al. Can professionals gauge likelihood of failure? \u0026ndash; Insights from tropical storm Matthew. \u003cem\u003eUrban Forestry Urban Green.\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, 126701. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ufug.2020.126701\u003c/span\u003e\u003c/span\u003e (2020).\u003c/li\u003e\n\u003cli\u003eKoeser, A. K. et al. Defective or just different? Observed storm failure in four urban tree growth patterns. \u003cem\u003eForests\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 988. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/f14050988\u003c/span\u003e\u003c/span\u003e (2023).\u003c/li\u003e\n\u003cli\u003eKoike, F. Reconstruction of two-dimensional tree and forest canopy profiles using photographs. \u003cem\u003eJ. Appl. Ecol.\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 921\u0026ndash;929. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2403240\u003c/span\u003e\u003c/span\u003e (1985).\u003c/li\u003e\n\u003cli\u003eKoike, F. Plant traits as predictors of woody species dominance in climax forest communities. \u003cem\u003eJ. Veg. Sci.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 327\u0026ndash;336. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/3236846\u003c/span\u003e\u003c/span\u003e (2001).\u003c/li\u003e\n\u003cli\u003eKurose, T. \u0026amp; Koike, F. Detecting behavioral preferences of dog walkers in urban green spaces. \u003cem\u003eJapanese J. Conserv. Ecol.\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 109\u0026ndash;115. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18960/hozen.2002\u003c/span\u003e\u003c/span\u003e (2020).\u003c/li\u003e\n\u003cli\u003eLi, Y. Q. et al. Development and Evaluation of Models for the Relationship between Tree Height and Diameter at Breast Height for Chinese-Fir Plantations in Subtropical China. \u003cem\u003ePLOS One\u003c/em\u003e. 0125118. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pone.0125118\u003c/span\u003e\u003c/span\u003e (2015).\u003c/li\u003e\n\u003cli\u003eLuo, Y. \u0026amp; Ai, X. Wind risk assessment of urban street trees based on wind-induced fragility. \u003cem\u003eDisaster Prev. Resil.\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.20517/dpr.2022.05\u003c/span\u003e\u003c/span\u003e (2022).\u003c/li\u003e\n\u003cli\u003eMiyawaki, A. \u003cem\u003eVegetation of Japan: Kanto\u003c/em\u003e (Shibundo), 1986).\u003c/li\u003e\n\u003cli\u003eMuscarella, R., Kolyaie, S., Morton, D. C., Zimmerman, J. K. \u0026amp; Uriarte, M. Effects of topography on tropical forest structure depend on climate context. \u003cem\u003eJ. Ecol.\u003c/em\u003e \u003cstrong\u003e108\u003c/strong\u003e, 145\u0026ndash;159. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/1365-2745.13261\u003c/span\u003e\u003c/span\u003e (2020).\u003c/li\u003e\n\u003cli\u003eNIED. Japan seismic hazard information station. (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.j-shis.bosai.go.jp/en/\u003c/span\u003e\u003c/span\u003e, Accessed 5 August 2025.\u003c/li\u003e\n\u003cli\u003e\u0026Ouml;zt\u0026uuml;rk, Ş., Yılmaz, K., Din\u0026ccedil;er, A. E. \u0026amp; Kalpakcı, V. Effect of urbanization on surface runoff and performance of green roofs and permeable pavement for mitigating urban floods. \u003cem\u003eNat. Hazards\u003c/em\u003e. \u003cstrong\u003e120\u003c/strong\u003e, 12375\u0026ndash;12399. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11069-024-06688-w\u003c/span\u003e\u003c/span\u003e (2024).\u003c/li\u003e\n\u003cli\u003eSalisbury, A. B. et al. Predictors of tropical cyclone-induced urban tree failure: an international scoping review. \u003cem\u003eFront. Forests Global Change\u003c/em\u003e. \u003cstrong\u003e6\u003c/strong\u003e, 1168495. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/ffgc.2023.1168495\u003c/span\u003e\u003c/span\u003e (2023).\u003c/li\u003e\n\u003cli\u003eSenalik, C. A. \u0026amp; Farber, B. Mechanical Properties of Wood. In Ross R, Wood handbook: Wood as an engineering material. USDA Forest Service. (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://research.fs.usda.gov/treesearch/62200\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003eShinozaki, K., Yoda, K., Hozumi, K. \u0026amp; Kira, T. A quantitative analysis of plant form-the pipe model theory: I. basic analyses. \u003cem\u003eJapanese J. Ecol.\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 97\u0026ndash;105. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18960/seitai.14.3_97\u003c/span\u003e\u003c/span\u003e (1964).\u003c/li\u003e\n\u003cli\u003eSiipilehto, J., Sarkkola, S., Nuutinen, Y. \u0026amp; Meht\u0026auml;talo, L. Predicting height-diameter relationship in uneven-aged stands in Finland. \u003cem\u003eFor. Ecol. Manag.\u003c/em\u003e \u003cstrong\u003e549\u003c/strong\u003e, 121486. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2023.121486\u003c/span\u003e\u003c/span\u003e (2023).\u003c/li\u003e\n\u003cli\u003eStan Development Team. Stan: Software for Bayesian Data Analysis. (2025). Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://mc-stan.org/\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003evan Haaften, M. A., Meuwissen, M. P. M., Gardebroek, C. \u0026amp; Kopinga, J. Trends in financial damage related to urban tree failure in the Netherlands. Urban Forestry and Urban Greening, \u003cstrong\u003e15\u003c/strong\u003e, 15\u0026ndash;21. (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ufug.2015.11.002\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003eXiao, Q. \u0026amp; McPherson, E. G. Rainfall interception by Santa Monica\u0026rsquo;s municipal urban forest. \u003cem\u003eUrban Ecosyst.\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e (4), 291\u0026ndash;302. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1023/B:UECO.0000004828.05143.67\u003c/span\u003e\u003c/span\u003e (2002).\u003c/li\u003e\n\u003cli\u003eZur, S. L., Bahat, S., Amara, R. \u0026amp; Farber, Y. Risk assessment of street trees failure by non-experts: a case study from the North of Israel. \u003cem\u003eNat. Hazards\u003c/em\u003e. \u003cstrong\u003e121\u003c/strong\u003e, 161\u0026ndash;174. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11069-024-06835-3\u003c/span\u003e\u003c/span\u003e (2025).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eTrees in the study area before treefall by two Typhoons in 2019. Dominant taxa up to 80% cumulative basal area are shown.\u003c/div\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eSpecies\u003c/div\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eStatus\u003c/div\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNumber of individuals\u003c/div\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eShare in basal area (%)\u003c/div\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eCinnamomum camphora\u003c/span\u003e (L.) J.Presl\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003ePlanted and naturalized non-native tree\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e903\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e21.6\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eCerasus\u003c/span\u003e spp.\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eDeciduous native populations and many planted trees of cultivars\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e877\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e12.4\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eCastanopsis sieboldii\u003c/span\u003e (Makino) Hatus. ex T.Yamaz. et Mashiba subsp. \u003cspan class=\"Italic\"\u003esieboldii\u003c/span\u003e\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNative tree, some planted\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e1304\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e12.3\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eQuercus glauca\u003c/span\u003e Thunb.\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNative tree, mainly planted\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e2123\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e7.5\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eMachilus thunbergii\u003c/span\u003e Siebold et Zucc.\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNative tree, planted with natural reproduction\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e1365\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e6.5\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eCornus controversa\u003c/span\u003e Hemsl. var. \u003cspan class=\"Italic\"\u003econtroversa\u003c/span\u003e\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNative deciduous tree\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e635\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e6.1\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eQuercus myrsinifolia\u003c/span\u003e Blume\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNative tree, mainly planted with natural reproduction\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e1174\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e4.2\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eZelkova serrata\u003c/span\u003e (Thunb.) Makino\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003ePlanted native deciduous tree\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e193\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e3.6\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eGinkgo biloba\u003c/span\u003e L.\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNon-native deciduous gymnosperm included in conifer\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e125\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e3.6\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e\u003cspan class=\"Italic\"\u003eLiquidambar styraciflua\u003c/span\u003e L.\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eNon-native deciduous tree\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e59\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e2.6\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cdiv class=\"SimplePara\"\u003eOthers\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e5191\u003c/div\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cdiv class=\"SimplePara\"\u003e19.6\u003c/div\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"latent variable, maximum compressive stress, mechanical model, return periods, statistical model","lastPublishedDoi":"10.21203/rs.3.rs-9147742/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9147742/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eUrban trees support human well-being, yet their benefits must be balanced against the risk of tree failure. Here, we propose a simple quantitative approach that practitioners can use to assess and manage urban treefall risk. We developed a simple mechanical model based on readily measured trunk diameter and height to calculate a hypothetical maximum compressive stress in the trunk for a given wind speed. This stress-based index was treated as a latent variable linking the mechanical and stochastic components of our framework. We then fitted a statistical model that predicts treefall probability from the latent stress index using empirically observed treefall records from Typhoons Faxai and Hagibis (2019). Unmodeled sources of variability and other unaccounted-for factors were implicitly incorporated through statistical calibration. The fitted model showed a statistically supported association between the latent stress index and treefall probability (broadleaved tree AUC\u0026thinsp;=\u0026thinsp;0.94, conifer AUC\u0026thinsp;=\u0026thinsp;0.77). Using extreme value analysis, we estimated the 30-year return level of maximum wind speed at the study site and assessed treefall risk for individual trees under this scenario. The model indicated higher risk for tall non-native trees (relative to local native species) and for conifers. Although available treefall data remain limited, to our knowledge, this study is among the first to predict future treefall probability using empirically observed treefall records and to provide a quantitative framework for practitioners to manage treefall risk.\u003c/p\u003e","manuscriptTitle":"Treefall risk assessment in an urban green area for a hypothetical 30-year return period storm using damage data from Typhoons Faxai and Hagibis in 2019","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 06:49:08","doi":"10.21203/rs.3.rs-9147742/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-13T05:57:07+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-02T02:31:52+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-20T14:33:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"329743350069074206182703429340763143828","date":"2026-04-11T21:09:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"274194024170048051327440655158764382927","date":"2026-03-30T14:45:16+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-28T17:20:45+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-23T14:55:02+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-23T12:48:30+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-20T10:01:43+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-03-20T09:37:18+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"de300a53-fbcb-4964-b5b3-4e28c496277e","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-13T05:57:07+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-02T02:31:52+00:00","index":47,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":67413386,"name":"Biological sciences/Ecology"},{"id":67413387,"name":"Earth and environmental sciences/Ecology"},{"id":67413388,"name":"Earth and environmental sciences/Natural hazards"}],"tags":[],"updatedAt":"2026-05-19T03:53:55+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 06:49:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9147742","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9147742","identity":"rs-9147742","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}