A city-based framework identifies wild Hedychium species suitable for near-nature urban landscaping in China.

Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Sci Rep . 2026 Mar 4;16:11935. doi: 10.1038/s41598-026-37132-7 Search in PMC Search in PubMed View in NLM Catalog Add to search A city-based framework identifies wild Hedychium species suitable for near-nature urban landscaping in China Xiaodong Liu Xiaodong Liu 1 College of Horticulture and Landscape Architecture, Zhongkai University of Agriculture and Engineering, Guangzhou, China Find articles by Xiaodong Liu 1 , Can Lai Can Lai 1 College of Horticulture and Landscape Architecture, Zhongkai University of Agriculture and Engineering, Guangzhou, China Find articles by Can Lai 1 , Yucheng Zhong Yucheng Zhong 1 College of Horticulture and Landscape Architecture, Zhongkai University of Agriculture and Engineering, Guangzhou, China Find articles by Yucheng Zhong 1 , Xiu Hu Xiu Hu 1 College of Horticulture and Landscape Architecture, Zhongkai University of Agriculture and Engineering, Guangzhou, China Find articles by Xiu Hu 1, ✉ Author information Article notes Copyright and License information 1 College of Horticulture and Landscape Architecture, Zhongkai University of Agriculture and Engineering, Guangzhou, China ✉ Corresponding author. Received 2025 May 3; Accepted 2026 Jan 20; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13068998  PMID: 41781411 Abstract Wild plants with high ornamental value, such as Hedychium species, offer untapped potential for enhancing urban biodiversity and sustainability. However, their integration into urban landscapes is hindered by a lack of cultivation protocols and climate uncertainty. This study addresses these challenges by identifying low-maintenance, climate-resilient Hedychium species suitable for near-nature urban landscaping—a strategy critical for reducing irrigation dependency and fostering ecological resilience in cities. In this paper, firstly, we developed an approach to evaluate the near-nature landscaping suitability of four Hedychium species representative of four ecologic types in China based on MaxEnt and GIS, and identified that Kunming is the most suitable metropolises for landscaping the most Hedychium species sustainably in China. Furthermore, potential suitable areas for these species were projected under two distinct climate change scenarios (SSP2-4.5 and SSP5-8.5). The results indicate divergent trends by the 2070s: under the SSP2-4.5 scenario, suitable habitats for H. coronarium, H. villosum , and H. coccineum are projected to expand, whereas H. sinoaureum faces significant habitat contraction. Conversely, under the SSP5-8.5 scenario, a shift occurs where the habitat for H. villosum contracts significantly, while H. sinoaureum achieves a slight expansion. Thirdly, Temperature Seasonality (Bio4) was identified as the main factor affecting the distribution of four species, which gives hints to how to landscaping Hedychium in less favorable areas. This methodology empowers urban planners to identify low-maintenance, climate-resilient species tailored to specific cities, advancing global efforts toward sustainable urban ecosystems. Supplementary Information The online version contains supplementary material available at 10.1038/s41598-026-37132-7. Keywords: Urban sustainability, Low-maintenance landscaping, Climate resilience, Species distribution modeling, Hedychium , Landscaping Subject terms: Environmental sciences, Urban ecology Introduction Urban sustainability initiatives increasingly prioritize green infrastructure that supports biodiversity while minimizing resource inputs. Cities face dual pressures: expanding green spaces to mitigate urban heat islands and conserving water resources. Low-maintenance, climate-resilient plants—particularly those adapted to natural precipitation—are essential for achieving these goals. Wild ornamental species like Hedychiums represent ideal candidates due to their resource efficiency (e.g., reduced irrigation needs) and potential to enhance ecological resilience in urban ecosystems. However, their application in cities remains underexplored, despite their capacity to address challenges like urban heat island mitigation through increased vegetation cover and transpirational cooling. There is a large number of plants with high ornamental value haven’t been introduced into the garden yet. However, there has long been a conflict between the need for a variety of ornamental plants and the limited experience in introducing these plants. It is ideal for a method that could take full advantage of distribution data to predict suitable landscaping areas for wild plants in the circumstance of scarce cultivation practice. Species distribution modeling (SDM) offers a transdisciplinary framework to bridge ecological research and urban horticulture, enabling targeted selection of wild plants for sustainable cityscapes. SDMs are a general suite of models that relate the locations of a species’ known occurrences to sets of underlying environmental and/or climatic variables (e.g., mean annual temperature, annual precipitation, seasonality, slope, etc.) 1 , 2 . The probability of occurrence in relation to the environmental variables can then be interpolated and extrapolated across the broader landscape to produce maps of the species’ predicted potential geographic distributions 3 , 4 . Due to their power and relative ease of use, SDMs have become one of the most widely used tools in conservation biology, biogeography, and ecology 5 , 6 . SDMs have been used to generate predictions of where invasive species may occur under present climatic conditions 7 – 9 as well as shifts in species’ geographic distributions under future climatic change scenarios 10 – 12 . Despite their wide application in ecological and biogeographic studies, SDMs have rarely been applied to cultivated systems or agriculture systems 11 , 13 , 14 . The agricultural community has been hesitant to adopt Species Distribution Models due to the perception that they may have limitations in accurately predicting suitable areas for agricultural crop species. This perception stems from the understanding that environmental conditions on farms can be altered through active management practices like irrigation, thereby enabling crop species to grow in areas that are unsuitable based on ambient climate alone 15 . However, SDMs have been used in those crops with low management requirements, such as forest, banana, coffee, medicines, and zea 16 – 18 . Near-nature landscaping is a cultivated system that fosters plants in the environment closely related to their natural habitat, that is, without changing the soil or fertilizing, no temperature control, no irrigation, just relocates the plants from one area (A) to another (B), where both areas share the same habitat. Under this circumstance, using SDMs to explore suitable near-nature landscaping areas is almost the same as identifying suitable habitats for plant conservation through SDMs. Due to a lack of progress tracking in ecology, hither to now, few horticulture studies are aware of exploring the near-nature landscaping area for plants with ornamental values by SDMs. Climate change is one of the major challenges of our time. According to the IPCC Fourth Assessment Report 19 , global temperatures have risen by approximately 1 °C compared to 200 years ago, and further temperature increases are expected in the future. The temperature in China is going to increase 0.5 ℃ ~ 1.2 °C in future 30 years, and precipitation is also predicted to increase 20 . Plant is expected to play an active role in alleviating climate change in urban areas. There is an increasing need to improve plant diversity in urban areas by cultivating more wild plants. Hedychium J. Koenig (1783: 61), a genus of perennial rhizomatous herbs belonging to the family Zingiberaceae, represents a tropical to subtropical floristic element that typically thrives in warm, humid, and shaded ecological environments. The genus with approximately 50 21 to 80 species 22 , 23 , is widely distributed from India, especially the Himalayas and China through continental SE Asia region to Malesia 23 , 24 . Thirty-three species of this genus have been recorded in China 25 . Hedychium plants are used in perfumery and in ethnomedicine 26 , 27 . Additionally, flowers of some species are edible, and several species have been reported to have antimicrobial and insecticidal properties 26 – 28 . However, Hedychium is probably mostly grown as ornamentals because of its showy, diverse, and aromatic flowers, sweet-scented flowers, and attractive green foliage which are suitable for landscape planting. Although many Hedychium species (including those distributed in China) have been introduced into the gardens in Europe and America, they are mostly been cultivated in greenhouse. In China, most Hedychium species remain in the wild and have not been introduced into gardens or landscape plantings. This paper employs SDM, specifically MaxEnt, to predict suitable landscaping areas for Hedychium in China under climate change, primarily using distribution data supplemented with a few introduction practices. We aimed to 1), explore suitable near-nature landscaping areas for four Hedychium species representing four ecotypes under the current climate, identifying the area with the highest concentration of Hedychium ; 2), analyze the change of suitable areas for these four species under climate change and identify suitable landscaping species of Hedychium for special large cities like Guangzhou under climate change; 3), analysis the primary factors influencing the distribution of Hedychium and provide recommendations for garden strategies in different area of China under climate change. Materials and methods Data source Four species representing four morphological and ecological types of Hedychium distributed in China were selected: Hedychium villosum in Clade I, Hedychium coronarium in Clade II, Hedychium coccineum in Clade III, and the narrow-range endemic species Hedychium sinoaurum representing Clade I 29 . Species in Clade I have thick roots, comparatively small rhizomes, epiphyte, green all the year, flowering mainly from December to March; species in Clade II have comparatively thin roots, large rhizomes, grounded, flowering mainly from June to November; species in Clade III has comparative thin roots, large rhizomes, grounded, flowering mainly in from June to September. Species in Clade IV have thick roots, comparative small rhizomes, epiphyte or grounded, dormant in the winter, flowering mainly from June to September. Our sampling strategy attempted to maximize the geographic coverage of each species. The data collection was based on extensive checks of specimens, references, and tedious field trips during ten years of taxonomic research, which ensured that the great majority of each taxon’s distribution was well represented by the sampling points, and no significant parts of the species’ known ranges were omitted from sampling. For identification, we examined all Hedychium protologues including monographs and taxonomic revisions 30 . The corresponding author Prof. X. Hu undertook the formal identification of the plant material used in this study. Herbarium collections, including type specimens, were consulted at CDBI, HITBC, IBSC, KUN, and PE. Records were scrutinized for synonyms, subspecies, misidentification, missing coordinate data, spatial uncertainty, points in the ocean, and duplicated locality points that we omitted. All the presence data of these study taxa were collected in the geography range of China. Voucher specimens have been deposited in the Herbarium, South China Botanical Garden, Chinese Academy of Sciences (IBSC) under voucher ID number HU200 ( H. coronarium ), HU129 ( H. villosum ), HU015 ( H. coccineum ), and HU099 ( H. sinoaureum ). The use of plants in the present study complied with relevant institutional, national, and international guidelines and legislation. At last, a total of 66, 49, 50, 28 records were used for H. coronarium , H. villosum , H. coccineum , H. sinoaureum to build the model, respectively (Table 1 , Fig. 2 A–D). Although the number of records for H. sinoaureum is relatively small, this species is a narrow-range endemic. Our exhaustive survey confirms that these records cover its full realized niche, making them sufficient for distribution modeling. Table 1. Data source and evaluation of MaxEnt models for 4 species in Hedychium. Scientific name Number of distribution data Number of cultivated data Cultivation performance Mean AUC value by 10 times (AUC train) H. coronarium 66 28 Succeed in all records 0.9668 H. villosum 49 7 Succeed in all records 0.9909 H. coccineum 50 8 Succeed in all records 0.9880 H. sinoaureum 28 4 Succeed in Kunming, and fail in Guangzhou, Mengla and Xiamen 0.9839 Open in a new tab ‘cultivation’ here means cultivated by irrigation and changing the substrate. Fig. 2. Open in a new tab Potential landscaping area of four species under current climate ( A : H. coronarium ; B : H. villosum ; C : H. coccineum ; D : H. sinoaureum ) (Note: warmer colors represent higher suitability scores, while colder colors represent lower suitability scores).This map was created using ArcGIS10.4.1 software. Additionally, Cultivation records by Authors from Guangzhou, Kunming, and Mengl Guangzhou (Guangdong province), Kunming (Yunnan province), and Mengla (Yunnan province) were analyzed to prioritize species requiring minimal irrigation and thriving under natural precipitation—key criteria for urban sustainability 31 . Some other cultivation records were collected from PPBC (Plant Photo Bank of China, http://ppbc.iplant.cn/),which is managed by the Institute of Botany, Chinese Academy of Sciences (IBCAS) and adopts the latest taxonomic systems and CVH(Chinese Virtual Herbarium, https://www.cvh.ac.cn/ ). The introduction performance H. coronarium , H. villosum , H. coccineum , H. sinoaureum were recorded from 28, 7, 8, 4 places respectively (Table 1 , Fig. 2 A–C). Environmental variables In cultivate system, the primary factor influencing plant growth is climate variables 32 . Although distribution prediction made by SDMs means to identify a completely coincide environment for the specific plant, including biotic factors (compete, etc.) and abiotic factors (temperature, precipitation, soil, elevation, slope, etc.). When SDMs were used to predict wild plant distribution in near-nature landscaping, abiotic factors are the main factor in function. Humans can decide where to grow the species (selecting proper temperature and precipitation), to change the soil, to irrigate, etc. At this point, SDMs are more suitable to be applied in cultivation systems than it was designed to be used in ecological aiming. When choosing environmental variables, more variables mean more constraints for the suitable area. Under landscaping circumstances, plants receive more assistance from human beings, allowing for a more optimistic estimation than a conservative one. Thus, we define the near-nature landscaping as cultivating the plant in an area with similar climatic variables to those found in the wild. As such, climate variables (temperature and precipitation) were prioritized to align with urban low-maintenance goals, ensuring selected species thrive without artificial irrigation. In this study. For climatic variables, bioclimatic variables were widely used in the species distribution model, for they are ecologically meaningful variables that explain annual trends, seasonality, and the adaptation of species to extremes of temperature and precipitation 33 . Compared with original climatic variables, bioclimatic variables could avoid the deviation caused by the uneven distribution of weather stations 34 , for they were generated through interpolation of average monthly climate data from weather stations. Also, for this reason, they were suitable for describing the distribution of species across large scales 35 , which fit the needs of this study. In this study, 19 bioclimatic variables were downloaded from the Worldclim dataset (The new Version 2.0, www.worldclim.org ) for the current period. In the Worldclim database, ‘current period’ was defined from 1950 to 2000, and these data have been widely used in creating species distribution models. To analyze the suitable range of 19 bioclimatic variables for four Hedychium , we compared their ranges by both value range and mean value. Jackknife analysis in MaxEnt was conducted to assist in identifying the main factors influencing these four species. To determine the future distribution of the species under different climate scenarios, we used datasets of future climate from the Climate Change, Agriculture and Food Security (CCAFS) website ( www.ccafs-climate.org ). Representative concentration pathways (SSPs) (including SSP126, SSP245, SSP370 and SSP585) are four greenhouse gas concentration trajectories adopted by the Intergovernmental Panel on Climate Change (IPCC) in its Sixth Assessment Report 36 . These pathways are used in climate modeling and research to describe four possible future climates, all of which are considered possible depending on how many greenhouse gases are emitted in the near future. To thoroughly evaluate the impact of climate change and strengthen the robustness of our findings, we employed a multi-scenario approach. This included the contemporary climate (1970–2000) as a baseline, alongside two contrasting future scenarios for the 2070s (2061–2080): SSP2-4.5 (intermediate pathway) and SSP5-8.5 (fossil-fueled development pathway).The selection of these two scenarios was strategically designed for a more comprehensive uncertainty analysis. SSP5-8.5 represents a high-emission baseline with strong climate forcing, serving as a ‘stress test’ to evaluate the robustness of our city-based framework under a severe climate change trajectory. The inclusion of SSP2-4.5, a medium-emission scenario reflecting intermediate mitigation efforts, provides a more plausible and widely referenced future for comparison. The contrast between SSP2-4.5 and SSP5-8.5 allows us to quantify the potential range of impacts arising from different socio-economic developments and climate policy ambitions, thereby significantly enhancing the transparency and practical applicability of our results for urban planners who must prepare for a spectrum of possible futures. For the circulation model, we selected a global circulation model, BCC-CSM2-MR (Medium resolution climate model) Developed by Beijing Climate Center. This model was chosen for its demonstrated performance in simulating the East Asian monsoon climate and its regional calibration using observational data from China, which enhances the applicability of predictions for the study area 37 . Please refer to the article for the specific introduction of BCC-CSM2-MR 38 . All environmental data used in this model had a 2.5-arc minute spatial resolution (also referred to as 5 km spatial resolution). Metropolises are big cities with dense populations, limited land, and a high demand for plant diversity. Wild plants with high ornamental values provide plenty of source for garden plant diversity. For this reason, the identification of suitable landscaping zones and understanding climate impact on the landscaping of wild ornamental plants are important. In this study, we focus on the suitability of Hedychium at metropolis. Since the landscape cultivation in this study was set as the cultivation under near natural conditions, we tended to neglect the urban heat island effect on environmental conditions. Modeling building and evaluation We choose MaxEnt to build the model in this study. Among many SDMs 5 , 39 , MaxEnt is one of the most popular SDMs being used to estimate species ranges concerning environmental predictors and it has consistently performed well in model comparisons with high model specificity and sensitivity 40 , 41 . MaxEnt has been shown to exhibit robustness in predictive power, even when employing a small number of records 40 , 42 . It achieves this by using a regularization method (ℓ1 regularization) which produces models with few non-zero coefficients 43 , 44 . Therefore, MaxEnt was selected to model the distribution of H. sinoaureum specifically to address the challenge of our limited sample size.This encourages parsimony and prevents overfitting better than other commonly used variable-selection methods 41 . Functionally, this approach estimates ‘the multivariate distribution of suitable habitat conditions (associated with species occurrences) in environmental feature-space’ 45 . Moreover, as a presence-only modeling technique, it estimates a taxon’s potential niche, instead of its realized niche 46 , 47 , a characteristic that is essential for the goals of exploring the potential landscaping area for wild plants. In this study, MaxEnt version 3.4.1 (downloaded from http://www.cs.princeton.edu/ ) was applied to build the model 48 . When building the model, we use the default set except select the minimum training presence threshold, which equals, for each taxon, the minimum suitability score associated with a presence record. This threshold, which ensures maximum sensitivity 49 , would give over-optimistic predictions if used for predicting the actual distribution of a taxon (although it has been identified as “conservative” relative to other commonly used thresholds, e.g. 50 , but in our case, the focus was the potentially suitable area (whether or not it belongs to the current distribution of the taxon), making this threshold ideal. Testing was performed to assess the predictive performance by randomly partitioning the sample points into 75% ‘training’ and 25% ‘test’ occurrences, creating a quasi-independent data for model testing 4 , 51 , 52 . The area under the ROC (AUC) was used as the MaxEnt predictive performance metric under the ROC curve. The AUC (area under roc curve) was an effective threshold independent index that can evaluate a model’s ability to discriminate presence from absence (or background) 53 . The receiver operating characteristic (ROC) describes corresponding values for omission error (FPRd horizontal axis) and sensitivity (TPRd vertical axes), with one point for each unique threshold value. AUC is threshold independent and deemed as one of the best evaluation method 54 . In general, AUC values could range between 0.5 and 1.0, which could be divided into five classes 55 . AUC < 0.5 describes models that perform worse than chance and occurs rarely in reality. An AUC of 0.5 represents pure guessing. Model performance is categorized as failing (0.5 ~ 0.6), poor (0.6 ~ 0.7), fair (0.7 ~ 0.8), good (0.8 ~ 0.9), or excellent (0.9 ~ 1) 55 . The closer the AUC was to 1, the better the model performance was 40 , 51 . Each model was run ten times to generate a mean value of AUC. To find out the main environmental variable influencing the near-nature landscaping cultivation mode, the habitat suitability curves of each variable were calculated, and the contributions of each variable to the habitat model were calculated using the software’s built-in jackknife test. The jackknife test (systematically leaving out each variable) was used to measure which were the dominant climatic factors determining the potential distribution of the species 56 . Limiting factor mapping was also applied to explore spatially how the climatic factors most influencing predictions vary across the study area. The percent contribution and permutation importance are important factors that measure the importance of the environmental variables. The permutation importance is up to the final performance of the model rather than the path used in an individual run and therefore is better for evaluating the importance of a particular variable 57 . Mapping and analyzing For predicting near-nature landscaping areas under the current climate, the presence points and 19 variables were run with log output in MaxEnt to define the current suitable locations. For predicting landscaping areas under future climate change, 19 bioclimatices under both the SSP2-4.5 and SSP5-8.5 scenarios for the 2070s were used as environmental variables. The distribution map was obtained based on the average logistic outputs of the runs with the highest AUC values, which are continuous values from 0 to 1, indicating the predicted probability that conditions are suitable for landscaping 46 . The landscaping suitability in the map was divided into ten levels by using the natural internal method. Further, the landscaping suitability in the map was divided into high suitable, suitable, and not suitable according to the cultivation distribution data obtained by the introduction we have done at Guangzhou, Kunming, Jinghong, and photography record from PPBC. A Map of China was downloaded from the China Standard Map Service ( http://bzdt.ch.mnr.gov.cn/ ). Subsequent data processing and visualization were performed with the software ArcGIS10.4.1. For analysis of the area holding the greatest number of Hedychium in China, the predicted map of four species was overlying on GIS by using the Spatial Analyst tool. By comparing the predicted area with current climate, those areas with little change under future climate were deemed as the most suitable area for landscaping that specific species. Further, those areas with suitability changing were warned by this comparison. Results Model evaluation and the factor influencing the distribution of Hedychium As shown in Table s 1 , at last, we sourced 66, 49, 50, and 28 distribution data for H. coronarium , H. villosum , H. coccineum , and H. sinoaureum respectively. The discrimination performance of the MaxEnt model revealed an averaged AUC [training data] value between 0.9668 and 0.9909, which was higher than 0.5 of a random model, this value is recognized as indicating an excellent model 46 and showed that the model was highly reliable and could effectively reflect their distribution under the current and future scenarios. The influential environmental predictor variables (from their Permutation importance and Contribution to 10 model runs) are detailed in Table 2 . Permutation importance is the more effective indicative capable of explaining the environmental variables. Temperature Seasonality (Bio4) emerged as the dominant factor, suggesting that Hedychiums with broader thermal tolerance are better suited for cities experiencing climatic volatility. For H. coronarium , the permutation importance of the other three environmental factors were 5.1% in Bio 6 (Min Temperature of Coldest Month) , 4.5% in Bio 12 (Annual Precipitation), and 3.3% in Bio 3 [Isothermality (Bio 2/ Bio 7) (* 100)] (Table 2 ). For H. villosum , the permutation importance of the other three environmental factors were 19.9% in Bio 12, 15.4% in Bio 3, and 6.0% in Bio 14 (Precipitation of Driest Month). For H. coccineum the permutation importance of the other three environmental factors were 23.8% in Bio 6, 8.1% in Bio 17 (Precipitation of Driest Quarter), and 7.5% in Bio 2 (Mean diurnal range). For H. sinoaureum , the permutation importance of the other three environmental factors were 8.7% in Bio 2; 5.4% in Bio 18 (Precipitation of Warmest Quarter) and 2.1% in Bio 9 (Mean Temperature of Driest Quarter). Table 2. Comparison of bioclimate factors influencing the distribution of four species and their values. Species H. coronarium H. villosum H. coccineum H. sino-aurum Enviromental factor Contribution (%) Permutation importance (%) Contribution (%) Permutation importance (%) Contribution (%) Permutation importance (%) Contribution (%) Permutation importance (%) Bio 2 7.5④ 8.7② Bio 3 3.3④ 10.4② 15.4③ 27.7② 41.1① Bio 4 14③ 76.3① 44.4① 49.9① 49.3① 44.3① 32.7② 80.8① Bio 6 5.1② 1.4④ 23.8② Bio 7 4.3④ 2.8③ Bio 9 2.1④ Bio 12 21② 4.5③ 17.2③ 19.9② Bio 14 6.0 ④ Bio 17 8.1③ Bio 18 2.7④ 5.4③ Bio 19 46.6① 10④ 8.5③ Open in a new tab Number circled indicate the rank in the column; Bio 2: Mean Diurnal Range (max temp—min temp); Bio 3: Isothermality (Bio 2/Bio 7) (* 100) ; Bio 4: Temperature Seasonality (standard deviation * 100) ; Bio 6: Min Temperature of Coldest Month ; Bio 7: Temperature Annual Range (Bio 5- Bio 6) ; Bio 9: Mean Temperature of Driest Quarter ; Bio 12: Annual Precipitation ; Bio 14: Precipitation of Driest Month ; Bio 17: Precipitation of Driest Quarter ; Bio 18: Precipitation of Warmest Quarter ; Bio 19: Precipitation of Coldest Quarter. From the cumulative response curve, the optimal value of Bio 4, H. sinoaureum and H. coccineum have a comparatively large value, which are 456.78 and 337.54 respectively, compared with that of H. coronarium , H. villosum , which are 137.57 ~ 283.67, 44.98 ~ 294.84, respectively (Fig. 1 , Appendix A ). Fig. 1. Open in a new tab The cumulative response curve of two most Permutation important environmental factors of four species. The value range of Bio 3 for H. coronarium is between 53.54 ~ 56.91, H. villosum is between 53.14 ~ 56.54, H. coccineum is between 53.54 ~ 56.98, and H. sinoaureum is 48.13 (Appendix A ). But H. sinoaureum had a smaller optimal value of Bio 3 compared with the other three species. According to the minimum temperature of the coldest month (Bio 6) and the maximum temperature of the hottest month (Bio 5), the optimum values of H. coronarium is 17 ~ 22.7/30.25, H. villosum is 17 ~ 22.5/30.01, H. coccineum 6.67/29.19, H. sinoaureum is 0.04/22.49, indicating that the four kinds of ginger flower have significant differences in their climatic requirements. H. coronarium and H. villosum were distributed in the regions with the highest Bio 6 (17 ~ 22 degrees), H. coccineum at 6.69 degrees, and H. sinoaureum at 0.04 degrees. According to Bio 5 (max temperature of the warmest month), H. sinoaureum was 22.49, which is much lower than the other three species, indicating that it was not adapted to the high temperature in summer, which was consistent with the results of our introduced cultivation experiment in Guangzhou 31 . Further investigation of the annual mean temperature of the four species (Bio 1) showed that the annual mean temperature of the four species was significantly different, and the optimal values were 20.73 for H. coronarium , 25.71–29.95 for H. villosum , 20.07 for H. coccineum , and 15.83 for H. sinoaureum (Appendix A ). According to our introductory experiment in Guangzhou 31 , the growth of H. coronarium and H. villosum was normal, while the growth of H. coccineum was poor, resulting in smaller inflorescence, and H. sinoaureum did not bloom (The mean annual temperature (Bio 1), the minimum temperature in the coldest month (Bio 6), the maximum temperature in the warmest month (Bio 5), the mean difference between the highest temperature and the lowest temperature in the month (Bio 2), and the temperature annual range (Bio 7) in Guangzhou were 17.67 ~ 22.51 °C, 4.98 ~ 10.68°, 28.05 ~ 33.10 °C, 6.26 ~ 8.37 °C, and 21.31 ~ 24.98 °C, respectively.). The possible reasons are the difference between the monthly maximum temperature and the lowest temperature, as well as the small annual temperature difference in Guangzhou. The minimum temperature in the coldest month is 17.32–22.74 for H. coronariums , 17.13–22.56 for H. villosums , 6.69 for H. coccineum , and 0.04 for H. sinoaureum . The results showed that H. coronarium , H. villosum and H. coccineum still had a certain amount of growth in the coldest month, while H. sinoaureum basically stopped growing, which was consistent with the habitat of the first three annual evergreens, and the latter withered the aboveground leaves and stems in winter. In terms of rainfall, the amount of rainfall in the driest month (Bio 14) of H. coronarium and H. coccineum was 15 times that of H. villosum and H. sinoaureum indicating that H. coronarium had a high demand for rainfall in the driest month. The order of 4 species was H. coronarium 202.30–222.20 > H. coccineum 216.70 > H. villosum 12.77 > H. sinoaureum 10.45 (Appendix A ). It is not surprising that H. villosum could continue to grow and flower during low rainfall seasons, as it possesses bold, fresh roots to resistant drought. Suitable landscaping area under current climate Maps of the full models’ predictions for the studied taxa are presented in Fig. 2 . In the maps, areas in warmer colors represent higher suitability scores, while colder colors represent lower suitability scores, and areas in grey color indicate not suitable. The spatial patterns of the four species exhibited by the taxa are varying. They are significantly dissimilar to each other. According to the natural breakpoint method by MaxEnt, the adaptability of H. coronarium was divided into 11 categories from 0 to 1, and the suitability for four species were identified as > 0.027, 0.01, 0.008, 0.01, respectively. According to the cultivation performance, all the areas predicted above this value can be cultivated in near-nature landscaping. There is a large amount of area that can introduce four species out of the natural distribution towards Fig. 2 . Currently, highly suitable landscaping areas (> 0.44) for H. coronarium were evaluated to occur in south Tibet, south Yunnan, central and east Taiwan, central and east Guangdong, south and east Guangxi, east Hainan, east Sichuan, central Chongqing, south and east Fujian. The evaluated areas of moderate suitability (0.22 ~ 0.44) of H. coronarium included south Xizang, south Sichuan, central Yunnan, east Chongqing, Central Guizhou, north Guangdong, Central Guangxi, west Hainan, and west Taiwan. In the highly suitable area, H. coronarium has a longer flower time (e.g., 5-month flower time cultivated at Guangzhou) than in suitable areas (e.g., 3-month flower time cultivated at Kunming). Currently, highly suitable landscaping areas (> 0.4) for H. villosum were evaluated to occur in south Xizang, south and west Yunnan, west Taiwan, west Guangxi, central and south Hainan. The evaluated areas of moderate suitable (0.09 ~ 0.4) of Hedychium villosum included south Xizang, central Sichuan, central Yunnan, south Guizhou, south Guangxi, southwest and east Guangdong, south Fujian, west Hainan, and south Taiwan. Currently, highly suitable landscaping areas (> 0.37) for H. coccineum were evaluated to occur in south Xizang, south and west Yunnan. The evaluated areas of moderate suitability (0.14 ~ 0.37) of H. coccineum included south Xizang, central Yunnan, west Guangxi, and south Guizhou. Currently, highly suitable landscaping areas (> 0.43) for H. sinoaureum were evaluated to occur in south Xizang, almost the whole Yunnan, and south Sichuan. The evaluated areas of moderate suitability (0.16 ~ 0.43) of H. sinoaureum included south Xizang, south and west Yunnan, south Sichuan, west Guizhou, and west Guangxi. Comparison of landscaping suitability under SSP2-4.5 and SSP5-8.5 In the 2070s, the suitable areas of the four Hedychium species are projected to change substantially under both SSP2-4.5 and SSP5-8.5. Although the magnitude of change differs between scenarios, both indicate a consistent shift of suitability towards higher latitudes and/or higher elevations, suggesting that climate-driven redistribution is robust to the choice of emission pathway (Figs. 3 , 4 and 5 ). Fig. 3. Open in a new tab Potential landscaping suitability under the current climate and the SSP2-4.5 scenario (2070s) ( A : H. coronarium ; B : H. villosum ; C : H. coccineum ; D : H. sinoaureum ).This map was created using ArcGIS10.4.1 software. Fig. 4. Open in a new tab Potential landscaping suitability under the current climate and the SSP5-8.5 scenario (2070s) ( A : H. coronarium ; B : H. villosum ; C : H. coccineum ; D : H. sinoaureum ).This map was created using ArcGIS10.4.1 software. Fig. 5. Open in a new tab Spatial delineation and comparison of suitable areas under the SSP2-4.5 and SSP5-8.5 scenarios(2070s) ( A : H. coronarium ; B : H. villosum ; C : H. coccineum ; D : H. sinoaureum ).This map was created using ArcGIS10.4.1 software. Species responses differ. H. coronarium shows relatively high stability: under SSP2-4.5, its suitable area overlaps strongly with the current range (96.52%), and newly suitable areas are mainly located in northwestern Sichuan and eastern Guizhou (Fig. 3 A). Under SSP5-8.5, its suitable area is projected to decrease by ~ 7.62%. The overlap map (Fig. 5 A) further indicates that the stable area remains concentrated in southern China, while scenario-specific suitable areas shift from Central–Eastern China under SSP2-4.5 (red) to southern Tibet and high-elevation areas of western Sichuan under SSP5-8.5 (dark blue), consistent with an upslope shift under stronger warming. H. villosum shows strong scenario dependence. Under SSP2-4.5, its suitable area expands northward into southern Shaanxi and southern Henan (+ 40.56%), with large SSP2-4.5-only suitable areas in western Sichuan, northwestern Hunan, and the Guangdong–Guangxi border region (Fig. 5 B, red). Under SSP5-8.5, suitable habitat contracts markedly in South China, and SSP5-8.5-only suitable areas are much smaller (Fig. 5 B, dark blue), indicating that stronger warming may constrain its expansion. H. coccineum shows the largest net gain in suitable area under future climate conditions. Suitability increases under SSP2-4.5 and expands further under SSP5-8.5 (+ 56.31%). In Fig. 5 C, SSP5-8.5-only suitable areas (dark blue) are more extensive than those under SSP2-4.5 and extend into northern Sichuan, southern Tibet, and southern Henan, supporting its potential for northward and upslope expansion. H. sinoaureum shows clear contraction along the southeastern coast under both scenarios. Notably, Fig. 5 D indicates that SSP5-8.5 yields a larger scenario-specific suitable area than SSP2-4.5 and that additional suitable patches extend inland towards eastern Qinghai. This pattern suggests that under stronger warming, some interior high-elevation and/or higher-latitude regions may become newly suitable despite losses in low-elevation coastal areas. Overall, under SSP2-4.5, H. coronarium, H. villosum, and H. coccineum show net increases in suitable area, with core suitability persisting in southern China and extending into the mid-subtropics, whereas H. sinoaureum contracts in southern Henan and along the southeastern coast. Under SSP5-8.5, H. villosum shows a substantial net reduction, while H. coccineum and H. sinoaureum still show net increases despite local losses within their current ranges. For landscaping, areas highlighted in red in Figs. 3 and 4 should be treated cautiously, whereas green areas represent more promising candidates. Introducing these species into newly suitable areas may be an option to maintain planting performance under future climates, but any implementation should consider local management constraints and potential ecological risks. Urban hotspots for Hedychium species distribution Although Hedychium species show broad potential suitable areas in South and Southwest China, the city-scale analysis indicates that high-suitability habitat is concentrated in a limited number of major cities. We evaluated the suitability of 15 representative cities for four Hedychium species under the current climate and under SSP2-4.5 and SSP5-8.5 scenarios (Fig. 6 ). Fig. 6. Open in a new tab Comparison of landscaping overlaps across all four clades within Hedychium . (Note: 0: unsuitable; 1: H. coronarium ; 20: H. villosum ; 300: H. coccineum ; 4000: H. sinoaureum ; 21: H. coronarium and H. villosum ; 301: H. coronarium and H. coccineum ; 321: H. coronarium , H. villosum and H. coccineum ; 4001: H. coronarium and H. sinoaureum ; 4020: H. villosum and H. coccineum ; 4021: H. coronarium , H. villosum and H. sinoaureum ; 4300: H. coccineum and H. sinoaureum ; 4301: H. coronarium , H. coccineum and H. sinoaureum ; 4320: H. villosum , H. coccineum and H. sinoaureum ; 4321: suitable area owned by four species.).This map was created using ArcGIS10.4.1 software. Kunming and Guiyang consistently show high suitability for all four studied species ( H. coronarium, H. villosum, H. coccineum , and H. sinoaureum ) under the current climate as well as under both future scenarios, indicating strong stability at the city scale. As plateau cities in Southwest China, they have relatively mild and humid climates and may also benefit from a potential climatic buffering effect, which could help maintain suitable conditions for Hedychium diversity. These results suggest that Kunming and Guiyang are priority cities for landscaping applications and germplasm conservation. Suitability in other cities is more heterogeneous and scenario dependent. Under SSP2-4.5, Guangzhou, Nanning,Haikou, and Chengdu are projected to be suitable for H. coronarium, H. villosum, and H. coccineum , whereas Chongqing and Xiamen are mainly suitable for H. coronarium and H.villosum . Under SSP5-8.5, the set of suitable species changes in several cities: in Chengdu, suitability is projected to decrease for H. sinoaureum but increase for H. coccineum , while H. villosum is no longer projected to be suitable in Guangzhou or Chongqing. Lhasa, representing high-altitude cities, is projected to be suitable only for H. coccineum and H. sinoaureum under both scenarios. In addition, Shanghai, Nanjing, Taipei, Changsha, Nanchang, and Wuhan are projected to be suitable only for H. coronarium under both scenarios; however, Shanghai and Nanjing lose a substantial proportion of suitable area under SSP5-8.5, and planting in these cities should therefore be planned cautiously. Overall, under climate change, Kunming and Guiyang emerge as priority cities for Hedychium conservation and landscaping applications, and H. coronarium shows the most consistent suitability across cities and scenarios among the four species. Discussion The effectiveness of SDMs in exploring the landscaping area for wild plants Species distribution models (SDMs) estimate species responses to environmental gradients and are used to make spatial predictions of habitat suitability or probability of species occurrence. By design, SDMs presume a balanced relationship between species and their environment. This means that species are assumed to occur in all the places where the environment is suitable, and they will not be present in areas where the environment is unsuitable 58 . However, this presumption cannot stick to itself because of the interaction between species and is limited by the low immigration ability of plants to occur in the places where it should be. This led to an overoptimistic estimate when SDMs was used to predict the potential distribution of wild plants. Whereas, when SDMs was used to predict the potential landscaping area of wild plants, this presumes can stick to itself. Because when landscaping a wild plant, humans can facilitate their growth in any environment suitable for cultivation, and eliminate the interaction of other species by weeding. Thus, it is feasible to use SDMs to explore landscaping areas for wild plants. For plant not being present in all the places where the environment is suitable means the inefficiency of distribution data, in all the places where the environment is unsuitable this led to the conservative of the model. Although we adopt an optimistic threshold setting and have selected the most extreme climate scenarios as the background, the prediction made by the model will still be smaller than the actual one. In this study, our aim is to predict outcomes before introduction to avoid losses. Therefore, we prefer to provide conservative predictions. All the prediction made by model is decided by the accuracy of data used to build the model. SDMs are not only can deal with distribution data, they can also incorporate with cultivated data. As specific wild plants are introduced in various locations, data collection can expand, allowing for model refinement and the generation of more accurate predictive maps. Sources of uncertainty and error control The uncertainty and error in SDMs are ultimately determined by model objectives. A discrepancy between introduction (cultivation) prediction with distribution prediction when using SDMs is that more environmental factors are under control under cultivation prediction. When SDMs was used to explore landscaping areas of wild plants, it should be avoided that too many environmental factors were chosen to build the model. Because more environmental factors, more limitations for the prediction. Ultimately it will lead to a much smaller prediction range 59 . For plant introduction, the climate is the decisive factor that affects the success 60 , 61 . So, we only choose climate as an environmental factor to maintain transferability for city-based introduction scenarios where many non-climatic factors can be managed or modified. In addition to SDM-related uncertainties, our future projections depend on the climate forcing data derived from Global Circulation Models (GCMs). Although BCC-CSM2-MR has demonstrated good performance in simulating East Asian climate patterns, projections based on a single GCM inevitably carry model-specific structural uncertainties, and inter-model spread is often larger for precipitation-related variables than for temperature 62 . However, the primary objective of this study is to develop a methodological framework for city-based ornamental plant selection rather than to provide precise quantitative predictions of future suitable areas; relative rankings of cities and broad suitability patterns are less sensitive to inter-model variability than absolute area estimates. Meanwhile, our results indicate that temperature seasonality (Bio4) is the dominant predictor for all four species, and temperature projections typically exhibit greater inter-model consistency than precipitation projections, which to some extent reduces the sensitivity and uncertainty associated with GCM selection. Furthermore, our multi-scenario approach incorporating both SSP2-4.5 and SSP5-8.5 captures a substantial portion of the uncertainty arising from different emission pathway Future studies could further enhance robustness by employing multi-GCM ensemble approaches that integrate projections from multiple climate models 63 . Except for the interaction among different species and the limitation of the emigration ability of plants, another reason that gives rise to a smaller suitability range is the incomplete collection of distribution data. In this study, to avoid the error given rise by distribution data collection, we try many sources to ensure the data represent the widest distribution of species. At the same time, wrong data representing the climate type beyond the range of a species would not only give an over-optimistic outcome but mislead the forecasting. We screened every item of the data source to assure the data accuracy and tried to give a reliable prediction. Although the number of occurrence records for some species was limited, previous studies have demonstrated that MaxEnt maintains high predictive accuracy even with small sample sizes 64 .Our results support this robustness. In this study, we achieve a very high AUC value for each four species (0.967, 0.991, 0.988, 0.983) respectively. These results are comparable to other successful applications of MaxEnt; for instance, in predicting the potential distribution of baobab, Sanchez et al. (2010) reported AUC values of 0.879 (global), 0.963 (West Africa), and 0.933 (East Africa) 65 . Furthermore, the consistency between the test AUC and the mean AUC indicates that there was no overfitting. Collectively, these indicators confirm that the models constructed in this study are reliable despite the limited sample size. In most potential distribution predictions, suitability identification was done by a set threshold in the model. However, predictions made by SDMs models are inferences from the analogy of environmental factors (climate). It is inevitable somewhat uncertain. The most reliable method is verifying by experiment 66 . In this study, to give rise to a prediction map closer to reality, we use cultivated data to assist the suitability identification by reclassifying the map according to cultivation performance. This approach enhances the practicality of the result. Species-specific environmental responses and phenological adaptations Distinct Hedychium ecotypes show clear species-specific responses to climatic conditions, likely reflecting differences in genetic background and biological traits. Based on variable contribution and permutation importance (Table 2 ), temperature seasonality (Bio4) emerges as the most consistent key predictor across all four species, indicating that seasonal temperature variability provides a shared climatic context shaping Hedychium distributions.Beyond this common signal, the remaining high-importance environmental variables vary among species, chiefly spanning precipitation-related variables (annual amount and seasonality) and temperature-related variables (winter extremes and isothermality). Under SSP2-4.5 and SSP5-8.5, projected habitat changes were strongly species-specific. H. coronarium showed the highest stability, consistent with a broader climatic tolerance, particularly to seasonal temperature variability. By contrast, H. villosum was highly sensitive to changes in water availability and temperature seasonality; under SSP5-8.5, its suitable area is projected to contract markedly as local hydrothermal conditions shift. H. coccineum may benefit from the relaxation of winter cold limitation (higher Bio6), leading to substantial expansion toward higher latitudes and a net increase in suitable area 67 . Responses of H. sinoaureum were more complex. Under SSP2-4.5, sensitivity to summer heat (lower Bio5 tolerance) may reduce suitability at lower-elevation margins. Under SSP5-8.5, a potential shift in the dominant limiting factor 68 —from winter cold to summer heat—could allow gains at higher latitudes as winter temperatures rise. Combined with its deciduous dormancy strategy, this northward gain may partially offset losses in the south, consistent with a “habitat compensation” effect 69 , resulting in a slight net increase in suitable area under extreme warming. These differences likely reflect long-term adaptation in morphology and phenology. During the dry season, H. coronarium, H. villosum, and H. coccineum remain evergreen and can maintain growth; H. villosum can be particularly active in shoot and leaf production. In winter, H. coronarium and H. coccineum may continue growth or initiate buds when conditions permit, whereas H. villosum can complete flowering and fruiting supported by well-developed belowground organs. In contrast, H. sinoaureum adopts an avoidance strategy and enters dormancy during cold and dry periods. This phenological differentiation helps explain contrasting sensitivities to future climate change and has practical implications for horticulture. In landscape applications, species selection should consider heat tolerance and phenology, together with appropriate microhabitats and substrate management. As representative Hedychium ecotypes in China, our results provide a basis for ex situ conservation, introduction and domestication, and broader horticultural use of the genus. Species select strategies of Hedychium in China under climate change From an urban planning perspective, integrating Hedychiums into cities like Kunming and Guangzhou aligns with sustainable development goals. These species reduce reliance on irrigation, lower maintenance costs, and enhance aesthetic and ecological value. Their resource efficiency—thriving under natural precipitation—aligns with urban water conservation strategies. Furthermore, their adaptability to climatic fluctuations enhances ecological resilience, ensuring urban green spaces remain functional under climate stressors. By increasing vegetation density, Hedychiums also contribute to urban heat island mitigation, as their foliage provides shade and transpirational cooling, thereby reducing surface temperatures in densely built environments. Based on SDM predictions and analyses under climate-change scenarios, we propose a three-tier strategy for urban applications of Hedychium species. (1)Core suitable areas. Under current climatic conditions, most of Yunnan, southern Sichuan, Guizhou, and western Chongqing show high suitability for multiple Hedychium species. Under both SSP2-4.5 and SSP5-8.5, Kunming and Guiyang remain highly suitable for all four study species. We therefore recommend these cities as core centers for conserving Hedychium germplasm resources and as demonstration areas for landscape applications. In these areas, Hedychium can be promoted as key native plants for near-natural public green spaces and roadside plantings, taking advantage of their low-maintenance traits. (2) Stable suitable areas. Hedychium coronarium is the most consistently suitable species for urban greening. In most southern and central cities (e.g., Changsha, Wuhan, and Nanchang), its suitability remains high and it can be widely used in near-natural plantings in gardens and parks. For cities such as Shanghai and Nanjing, where suitability may decline under high-emission scenarios, we suggest prioritizing microclimate management in urban planning to improve tolerance to extreme events and to reduce the risk of large-scale plant mortality. (3) Potential expansion areas. Under both future scenarios, H. coccineum and H. sinoaureum show clear northward and upslope shifts, especially toward the southern Tibetan Plateau and northern Sichuan. Lhasa, as a representative high-elevation city, shows distinct potential and may provide a strategic opportunity for urban greening. For cities outside the current natural ranges but projected to become suitable (e.g., Shanghai for H. coccineum ; Guangzhou for H. villosum and H. coccineum ), we suggest trial introductions as ornamental plants to enrich urban biodiversity. We recommend establishing pilot acclimatization nurseries first to test their performance under local microclimates, followed by gradual upscaling if successful. In addition, for regions where habitat suitability may contract substantially under SSP2-4.5 and SSP5-8.5 (e.g., H. sinoaureum in Guangdong, Fujian, and Guangxi), we suggest a precautionary approach: use these species in flexible or seasonal landscape designs (e.g., seasonal flower borders) rather than as permanent structural plantings, to reduce potential ecological and economic risks associated with future habitat loss. Conclusions In this study, we proved that MaxEnt can be used to explore suitable landscaping areas for wild plants under near-nature cultivation. SDM applications can be expanded to other wild ornamentals, creating a database for city-specific plant selection. We found additional areas where the four species of Hedychium were absent from their natural distribution range. The results of this study are helpful in accelerating the horticulture of Hedychiums , and prioritizing Hedychium species in municipal planting schemes to reduce urban water consumption which has high ornamental value, and for enriching plant diversity in all kinds of gardens in China. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (17.1KB, xlsx) Author contributions Conceptualization, X.H.; data curation, X.L.,C.L.,Y.Z.; formal analysis, Y.Z.; funding acquisition, X.H.; investigation, X.L.,C.L.,Y.Z.; writing, X.L.,C.L.,Y.Z.,X.H.. All authors have read and agreed to the published version of the manuscript. Funding This work has been supported by the National Natural Science Foundation of China (31100510). Data availability The data that support the findings of this study are available from the corresponding author ([[email protected]](mailto:[email protected])) upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Pearson, R. G. & Dawson, T. P. Predicting the impacts of climate change on the distribution of species: Are bioclimate envelope models useful?. Glob. Ecol. Biogeogr. 12 , 361–371 (2003). [ Google Scholar ] 2. Phillips, S. J., Miroslav Dudík & Schapire, R. E. A maximum entropy approach to species distribution modeling. In Proc. of the Twenty-First International Conference on Machine Learning 83 (Association for Computing Machinery, 2004). 3. Anderson, R. P. & Enrique Martı´nez-Meyer. Modeling species’ geographic distributions for preliminary conservation assessments: an implementation with the spiny pocket mice (Heteromys) of Ecuador. Biol. Conserv. 116 , 167–179 (2004). 4. Machovina, B. & Feeley, K. J. Climate change driven shifts in the extent and location of areas suitable for export banana production. Ecol. Econ. 95 , 83–95 (2013). [ Google Scholar ] 5. Franklin, J. Mapping Species Distributions: Spatial Inference and Prediction Biodiversity and Conservation (Cambridge University Press, 2010). [ Google Scholar ] 6. Richardson, D. M. & Whittaker, R. J. Conservation biogeography—Foundations, concepts and challenges. Divers. Distrib. 16 , 313–320 (2010). [ Google Scholar ] 7. Ficetola, G. F., Thuiller, W. & Miaud, C. Prediction and validation of the potential global distribution of a problematic alien invasive species—The American bullfrog. Divers. Distrib. 13 , 476–485 (2007). [ Google Scholar ] 8. Guisan, A. & Thuiller, W. Predicting species distribution: Offering more than simple habitat models. Ecol. Lett. 8 , 993–1009 (2005). [ DOI ] [ PubMed ] [ Google Scholar ] 9. Peterson, A. T. & Vieglais, D. A. Predicting species invasions using ecological niche modeling: New approaches from bioinformatics attack a pressing problem: A new approach to ecological niche modeling, based on new tools drawn from biodiversity informatics, is applied to the challenge of. Bioscience 51 , 363–371 (2001). [ Google Scholar ] 10. Feeley, K. J. & Silman, M. R. Land-use and climate change effects on population size and extinction risk of Andean plants. Glob. Chang. Biol. 16 , 3215–3222 (2010). [ Google Scholar ] 11. Hijmans, R. J. & Graham, C. H. The ability of climate envelope models to predict the effect of climate change on species distributions. Glob. Chang. Biol. 12 , 2272–2281 (2006). [ Google Scholar ] 12. Kearney, M., Simpson, S. J., Raubenheimer, D. & Helmuth, B. Modelling the ecological niche from functional traits. Philos. Trans. R. Soc. B Biol. Sci. 365 , 3469–3483 (2010). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Beck, J. & Sieber, A. Is the spatial distribution of mankind’s most basic economic traits determined by climate and soil alone?. PLoS ONE 5 , 1–10 (2010). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Trnka, M. et al. European Corn Borer life stage model: Regional estimates of pest development and spatial distribution under present and future climate. Ecol. Modell. 207 , 61–84 (2007). [ Google Scholar ] 15. Wittwer, S. H. & Castilla, N. Protected cultivation of horticultural crops worldwide. HortTechnology 5 , 6–23 (1995). [ Google Scholar ] 16. Mertens, A. et al. Conservation status assessment of banana crop wild relatives using species distribution modelling. Divers. Distrib. 27 , 729–746 (2021). [ Google Scholar ] 17. Faraz, M. et al. A Systematic Review of analytical and modelling tools to assess climate change impacts and adaptation on coffee agrosystems. Sustain. 15 , (2023). 18. Evans, J. M., Fletcher, R. J. Jr. & Alavalapati, J. Using species distribution models to identify suitable areas for biofuel feedstock production. GCB Bioenergy 2 , 63–78 (2010). [ Google Scholar ] 19. IPCC. Section “Discussion”: Near-Term Responses in a Changing Climate. Clim. Chang. 2023 Synth. Rep. 42–66 (2023) . 20. Wenjuan, Z., Lin, C., Kejian, D., Xiayu, D. & Yilin, Z. Prediction of potential geographic distribution areas of the maize downy wildew in China by using MAXENT. Plant Prot. 35 , 32–38 (2009). [ Google Scholar ] 21. Wongsuwan, P. & Picheansoonthon, C. Taxonomic Revision of the Genus Hedychium J. Koenig (Zingiberaceae) in Thailand. J. R. Inst. Thail. III , 126–149 (2011). 22. Jain, S. K. & Prakash, V. Zingiberaceae in India: Phytogeography and endemism. Rheedea 5 , 154–169 (1995). [ Google Scholar ] 23. Sirirugsa, P. & Larsen, K. The genus Hedychium (Zingiberaceae) in Thailand. Nord. J. Bot. 15 , 301–304 (1995). [ Google Scholar ] 24. Ashokan, A. et al. Himalayan orogeny and monsoon intensification explain species diversification in an endemic ginger (Hedychium: Zingiberaceae) from the Indo-Malayan Realm. Mol. Phylogenet. Evol. 170 , 107440 (2022). [ DOI ] [ PubMed ] [ Google Scholar ] 25. Mou, F. J., Zhong, Y. C., Hu, X. & Wang, H. S. Hedychium wangmoense (Zingiberaceae), a New Species from Southwest Guizhou China. Ann. Bot. Fenn. 60 , 145–150 (2023). [ Google Scholar ] 26. Gao, L., Liu, N., Huang, B. & Hu, X. Phylogenetic analysis and genetic mapping of Chinese Hedychium using SRAP markers. Sci. Hortic. (Amsterdam) 117 , 369–377 (2008). [ Google Scholar ] 27. Dan, M., Shiburaj, S., Sethuraman, M. G. & George, V. Chemical composition and antibacterial activity of the rhizome oil of Hedychium larsenii. Acta Pharm. 55 , 315–320 (2005). [ PubMed ] [ Google Scholar ] 28. Hartati, R., Suganda, A. G. & Fidrianny, I. Botanical, phytochemical and pharmacological properties of Hedychium (Zingiberaceae)—A review. Procedia Chem. 13 , 150–163 (2014). [ Google Scholar ] 29. Xiu, H., Jian-xun, Y., Nian, L. & Zhi, W. Studies on the wild existence and introduction of ornamental resources of Hedychium in China. Acta Hortic. Sin. 37 , 643–648 (2010). [ Google Scholar ] 30. Bai, L., Hu, X., He, J. & Tian, Z. New records of Hedychium hookeri (Zingiberaceae) from China and Myanmar. Phytotaxa 494 , 237–243 (2021). [ Google Scholar ] 31. Xiu, H., Yaping, K. & Niu, L. Study on propagation, cultivation and maintenance of four landscape species of Hedychium plants. Guangdong Agric. Sci. 38 , 4 (2011). [ Google Scholar ] 32. Amanullah. Agronomy . (IntechOpen, 2020). 33. Booth, T. H., Nix, H. A., Busby, J. R. & Hutchinson, M. F. Bioclim: The first species distribution modelling package, its early applications and relevance to most current MaxEnt studies. Divers. Distrib. 20 , 1–9 (2014). [ Google Scholar ] 34. Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G. & Jarvis, A. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25 , 1965–1978 (2005). [ Google Scholar ] 35. Zhang, M.-G., Slik, J. W. F. & Ma, K.-P. Using species distribution modeling to delineate the botanical richness patterns and phytogeographical regions of China. Sci. Rep. 6 , 22400 (2016). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Fan, X., Duan, Q., Shen, C., Wu, Y. & Xing, C. Evaluation of historical CMIP6 model simulations and future projections of temperature over the Pan-Third Pole region. Environ. Sci. Pollut. Res. 29 , 26214–26229 (2022). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Wu, T. et al. The Beijing climate center climate system model (BCC-CSM): the main progress from CMIP5 to CMIP6. Geosci. Model Dev. 12 , 1573–1600 (2019). [ Google Scholar ] 38. Wu, C., Jiang, P., Ding, C., Feng, F. & Chen, T. Intelligent fault diagnosis of rotating machinery based on one-dimensional convolutional neural network. Comput. Ind. 108 , 53–61 (2019). [ Google Scholar ] 39. Elith, J. & Leathwick, J. R. Species distribution models: Ecological explanation and prediction across space and time. Annu. Rev. Ecol. Evol. Syst. 40 , 677–697 (2009). [ Google Scholar ] 40. Elith*, J. et al. Novel methods improve prediction of species’ distributions from occurrence data. Ecography (Cop.). 29 , 129–151 (2006). 41. Phillips, S. J. & Dudík, M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography (Cop.) 31 , 161–175 (2008). [ Google Scholar ] 42. Yi, Y., Cheng, X., Yang, Z.-F. & Zhang, S.-H. Maxent modeling for predicting the potential distribution of endangered medicinal plant (H. riparia Lour) in Yunnan China. Ecol. Eng. 92 , 260–269 (2016). [ Google Scholar ] 43. Tibshirani, R. Bias, variance and prediction error for classification rules. Monogr. Soc. Res. Child Dev. 79 , 1–14 (1996). [ Google Scholar ] 44. Williams, P. M. Bayesian regularization and pruning using a Laplace prior. Neural Comput. 7 , 117–143 (1995). [ Google Scholar ] 45. Franklin, J. & Miller, J. A. Mapping species distributions: Spatial inference and prediction (Cambridge University Press, 2010). [ Google Scholar ] 46. Phillips, S. J., Anderson, R. P. & Schapire, R. E. Maximum entropy modeling of species geographic distributions. Ecol. Modell. 190 , 231–259 (2006). [ Google Scholar ] 47. Jiménez-Valverde, A. et al. Use of niche models in invasive species risk assessments. Biol. Invasions 13 , 2785–2797 (2011). [ Google Scholar ] 48. Phillips, S. J., Anderson, R. P., Dudík, M., Schapire, R. E. & Blair, M. E. Opening the black box: An open-source release of Maxent. Ecography (Cop.) 40 , 887–893 (2017). [ Google Scholar ] 49. Liu, C., Newell, G. & White, M. On the selection of thresholds for predicting species occurrence with presence-only data. Ecol. Evol. 6 , 337–348 (2016). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 50. Pearson, R. G. et al. Model-based uncertainty in species range prediction. J. Biogeogr. 33 , 1704–1711 (2006). [ Google Scholar ] 51. Fielding, A. H. & Bell, J. F. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ. Conserv. 24 , 38–49 (1997). [ Google Scholar ] 52. Thuiller, W., Brotons, L., Araújo, M. B. & Lavorel, S. Effects of restricting environmental range of data to project current and future species distributions. Ecography (Cop.) 27 , 165–172 (2004). [ Google Scholar ] 53. McPherson, J. M., Jetz, W. & Rogers, D. J. The effects of species’ range sizes on the accuracy of distribution models: Ecological phenomenon or statistical artefact?. J. Appl. Ecol. 41 , 811–823 (2004). [ Google Scholar ] 54. Chen, P., Wiley, E. O. & Mcnyset, K. M. Ecological niche modeling as a predictive tool: Silverand bighead carps in North America. Biol. Invasions 9 , 43–51 (2007). [ Google Scholar ] 55. Swets, J. A. Measuring the accuracy of diagnostic systems. Science 240 , 1285–1293 (1988). [ DOI ] [ PubMed ] [ Google Scholar ] 56. Li, G., Du, S. & Wen, Z. Mapping the climatic suitable habitat of oriental arborvitae ( Platycladus orientalis ) for introduction and cultivation at a global scale. Sci. Rep. 6 , 30009 (2016). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. Songer, M., Delion, M., Biggs, A. & Huang, Q. Modeling impacts of climate change on giant panda habitat. Int. J. Ecol. 2012 , 108752 (2012). [ Google Scholar ] 58. Guisan, A. & Zimmermann, N. E. Predictive habitat distribution models in ecology. Ecol. Modell. 135 , 147–186 (2000). [ Google Scholar ] 59. Hong, Y. Comparison of climate similarity between Australia and China using numerical classification analysis. Sci. Silvae Sin. 42 , 30–36 (2006). [ Google Scholar ] 60. Woodward, F. I. & Williams, B. G. Climate and plant distribution at global and local scales. Vegetatio 69 , 189–197 (1987). [ Google Scholar ] 61. Eldridge, K., John, D., Harwood, C. & Wyk, G. V. Eucalypt Domestication and Breeding (Oxford University Press, 1994). [ Google Scholar ] 62. Knutti, R. & Sedláček, J. Robustness and uncertainties in the new CMIP5 climate model projections. Nat. Clim. Chang. 3 , 369–373 (2013). [ Google Scholar ] 63. Tebaldi, C. & Knutti, R. The use of the multi-model ensemble in probabilistic climate projections. Philos. Trans. R . Soc. A Math. Phys. Eng. Sci. 365 , 2053–2075 (2007). [ DOI ] [ PubMed ] [ Google Scholar ] 64. Pearson, R. G., Raxworthy, C. J., Nakamura, M. & Peterson, A. T. Predicting species distributions from small numbers of occurrence records: A test case using cryptic geckos in Madagascar. J. Biogeogr. 34 , 102–117 (2006). [ Google Scholar ] 65. Sanchez, A. C., Osborne, P. E. & Haq, N. Identifying the global potential for baobab tree cultivation using ecological niche modelling. Agrofor. Syst. 80 , 191–201 (2010). [ Google Scholar ] 66. Larmour, J. S., Whitfeld, S. J., Harwood, C. E. & Owen, J. V. Variation in frost tolerance and seedling morphology of the spotted gums Corymbia maculata, C. variegata, C. henryi and C. citriodora. Aust. J. Bot. 48 , 445–453 (2000). [ Google Scholar ] 67. Chen, C., Hill, J. K., Ohlemüller, R., Roy, D. B. & Thomas, C. D. Rapid range shifts of species associated with high levels of climate warming. Science 33 , 1024 (2011). [ DOI ] [ PubMed ] [ Google Scholar ] 68. Lenoir, J. & Svenning, J. Climate-related range shifts—A global multidimensional synthesis and new research directions. Ecography (2014). [ Google Scholar ] 69. Parmesan, C. & Yohe, G. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421 , 37–42 (2003). [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1 (17.1KB, xlsx) Data Availability Statement The data that support the findings of this study are available from the corresponding author ([[email protected]](mailto:[email protected])) upon reasonable request. Articles from Scientific Reports are provided here courtesy of Nature Publishing Group ACTIONS View on publisher site PDF (4.4 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top