No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Guisan A, Zimmermann NE. Predictive habitat distribution models in ecology. Ecological Modeling
Abstract
With the rise of new powerful statistical techniques and GIS tools, the development of predictive habitat distribution models has rapidly increased in ecology. Such models are static and probabilistic in nature, since they statistically relate the geographical distribution of species or communities to their present environment. A wide array of models has been developed to cover aspects as diverse as biogeography, conservation biology, climate change research, and habitat or species management. In this paper, we present a review of predictive habitat distribution modeling. The variety of statistical techniques used is growing. Ordinary multiple regression and its generalized form (GLM) are very popular and are often used for modeling species distributions. Other methods include neural networks, ordination and classification methods, Bayesian models, locally weighted approaches (e.g. GAM), environmental envelopes or even combinations of these models. The selection of an appropriate method should not depend solely on statistical considerations. Some models are better suited to reflect theoretical findings on the shape and nature of the species’ response (or realized niche). Conceptual considerations include e.g. the trade-off between optimizing accuracy versus optimizing generality. In the field of static distribution modeling, the latter is mostly related to selecting appropriate predictor variables and to designing an appropriate procedure for model selection. New methods, including threshold-independent measures (e.g. receiver operating characteristic (ROC)-plots) and resampling techniques (e.g. bootstrap, cross-validation) have been introduced in ecology for testing the accuracy of predictive models. The choice of an evaluation measure should be driven primarily by the goals of the study. This may possibly lead to the attribution of different weights to the various types of prediction errors (e.g. omission, commission or confusion). Testing the model in a wider range of situations (in space and time) will permit one to define the range of applications for which the model predictions are suitable. In turn, the qualification of the model depends primarily on the goals of the study that define the qualification criteria and on the usability of the model, rather than on statistics alone.
... On the other hand, logistic regression models have been widely used in species distribution modeling but often face challenges with non-linear ecological processes and overprediction in high-suitability areas. Guisan and Zimmermann (2000) emphasized that while logistic regression is robust for presenceabsence data, it may struggle with complex ecological gradients, such as elevation, which aligns with our findings of overpredicted NTFPs presence at higher altitudes. Furthermore, Pearce and Ferrier (2000) identified that logistic regression models tend to overpredict presence in areas with limited absence data, which may explain the lower specificity observed in this study. ...
... The GIS expert system, relying on expert knowledge, outperformed logistic regression in accuracy and efficiency, aligning with findings by Rae et al. (2014), Store et al. (2001), Tabor and Hutchinson (1994) who demonstrated the strength of expert-based models under limited data conditions. Logistic regression, while robust and data-driven (Guisan and Zimmermann 2000;Pearce and Ferrier 2000), requires large, high-quality datasets (minimum 50-250 samples). With field data collection being costly, integrating expert knowledge has become a practical solution (Elith et al. 2009;Phillips et al. 2006), though it introduces subjectivity. ...
This study focuses on the sustainable management of non-timber forest products (NTFPs) in the Narmada, Dang, and Panchmahal districts of Gujarat, India, emphasizing carbon sequestration and carbon credits. NTFPs such as medicinal plants, fruits, nuts, and resins play a crucial role in the local economy and biodiversity conservation. Accurate mapping and assessment of these resources are essential for implementing sustainable management and conservation strategies. Advanced spatial analysis techniques, including geographic information systems (GIS), remote sensing, and logistic regression models, were employed to analyze the spatial distribution of NTFPs. High-resolution Sentinel-2 satellite imagery and field survey data were integrated to create detailed spatial maps while, logistic regression models evaluated environmental factors like soil type, elevation, and climatic conditions affecting NTFPs distribution. The study identified that environmental variables such as litter cover, elevation, and NTFPs type are critical in determining the distribution of NTFPs, with NTFPs type accounting for 68% of the variance in distribution patterns. The logistic regression model and GIS expert system predicted NTFPs distribution with 65.43% and 70.37% accuracy, respectively. The GIS expert system demonstrated a higher specificity rate (47.05%) compared to logistic regression (35.29%), indicating its superior ability to predict NTFP absence. Both models were validated using a sample size of 81, with error matrices generated for comparative analysis. Additionally, the research explored the carbon sequestration potential of NTFPs and their implications for carbon credits. The study recommend machine learning algorithms, particularly the random forest (RF) model, for aboveground carbon stock estimation across different NTFPs regions. The RF model showed superior performance with an R² value exceeding 0.6 across the study areas, compared to multivariate stepwise regression, which had R² values below 0.4. The RF model's accuracy was validated through a comparison with actual field data, achieving a root mean square error of 24.72 t hm⁻² in NTFPs regions. Carbon stocks were observed to range from 50 to 250 t hm⁻², depending on the region's ecological characteristics and topographical variations. The research findings highlight the importance of integrating local ecological knowledge with spatial data and advanced modelling techniques to enhance the sustainable management of NTFPs. The potential for carbon credits offers a financial incentive for conserving NTFPs, promoting sustainable harvesting, and ensuring long-term biodiversity conservation. This study provides a replicable framework for NTFPs management and carbon stock estimation, applicable to similar ecological settings worldwide. The integration of GIS, remote sensing, and machine learning methodologies presents a robust approach for policymakers and stakeholders aiming to optimize NTFPs conservation strategies while advocating carbon credit opportunities for economic development.
... Determining and quantifying the environmental conditions that explain the spatial distribution of a species is a complex task, especially for species with limited distribution and very specific habitat requirements. In such cases, species distribution models (SDMs) can be used to determine and quantify the environmental conditions that explain the spatial distribution of a species (Guisan and Zimmermann, 2000;Peterson et al., 2011), and they have become widely used in the fields of animal ecology and conservation biology (Guisan et al., 2002;Rushton et al., 2004;Estrada et al., 2011;Moreno-Zarate et al., 2020). These explanatory and predictive models can establish statistical relationships between the presence of a species and the variables that could explain its distribution, such as climate, landscape characteristics, and the degree of human influence (Bustamante and Seoane, 2004;Guisan and Thuiller, 2005;Márquez et al., 2011;Rosalino et al., 2019). ...
... Building an SDM involves relating the distribution of the presence and absence of a species in a territory to a series of environmental characteristics in order to identify the areas that are potentially most favourable (Guisan and Zimmermann, 2000). The presence records considered in the SDM in this study were the locations of bearded vulture nests in the Spanish Pyrenees, occupied during the 2021-2022 breeding season. ...
... The first step of the analysis is to construct a realised niche of species, based on its known occurrence in environmental space. Then, the species distribution is estimated in the region of interest, based on the geographical distribution of the environmental space (Guisan and Zimmermann 2000). Since the straight-tusked elephant is an extinct species, our analysis is based on its fossil record and reconstructed palaeoclimate data. ...
... We modelled the environmental niche of the straight-tusked elephant using a binomial generalised linear model (GLM) that predicted the likelihood of occurrence of the straighttusked elephant in a respective raster cell, based on fossil finds and the climate variables of annual mean temperature and annual precipitation (Fig. 1C, D). We implemented explanatory variables as polynomials to allow for a humpshaped relationship between species occurrence and climate (for detailed information on using GLMs in SDM, see Guisan and Zimmermann (2000) and Guisan et al. (2002)). ...
The straight-tusked elephant ( Palaeoloxodon antiquus ) was amongst the largest herbivores once engineering the European landscape on a continental scale. In combination with the glacial-interglacial cycles of the Pleistocene, the species was an integral part of the control regimes that shaped European flora and fauna. With the human-facilitated extinction of the straight-tusked elephant, these landscape-forming processes were lost during the last Glacial-Interglacial cycle. Given today’s climate, could straight-tusked elephants still be part of modern ecosystems in Europe? And if yes, where? Answers to these questions can support nature conservation in preserving species and ecosystems historically adapted to these lost control regimes.
We reconstructed the realised niche of the straight-tusked elephant by allocating a novel compilation of fossil occurrences to either cold or warm stages, based on their assignment to Marine Isotope Stages. Further, we quantified the past potential distribution of the straight-tusked elephant since its extinction and its current potential distribution given the modern climate. Results show that the elephant could have persisted in the Mediterranean Basin until today and that modern climate across Central and Western Europe, excluding the Alps, as well as in the Mediterranean, is highly suitable for its occurrence.
Our results show that, without human-induced extinctions, European fauna would comprise extinct megafauna, acting as ecosystem engineers on a continental scale. Local rewilding initiatives aim at restoring these lost processes, but potentially cannot achieve lasting ecological effects on comparable scales.
The current European climate would still be suitable for the extinct straight-tusked elephant ( Palaeoloxodon antiquus ).
The straight-tusked elephant could have persisted during the last Glacial-Interglacial cycle, considering past climate.
Conserving the ecosystems shaped by the top-down ecosystem functions executed by the straight-tusked elephant in Europe may be especially promising in regions where the elephant could still exist today.
Using reference cold and warm stages, based on Marine Isotope Stages in a Species Distribution Modelling framework, is a promising attempt to overcome dating uncertainties inhibiting more specific niche reconstructions of extinct species.
... Notably, species distribution modeling (SDM) plays a vital role in predicting species ecological requirements together with the possible spatial distribution of species distribution in the fields of regional ecology and biogeography, when limited distribution information is available [12]. A variety of SDM models have been adopted for predicting distribution area, ecological response, and ecological requirements [12]. ...
... Notably, species distribution modeling (SDM) plays a vital role in predicting species ecological requirements together with the possible spatial distribution of species distribution in the fields of regional ecology and biogeography, when limited distribution information is available [12]. A variety of SDM models have been adopted for predicting distribution area, ecological response, and ecological requirements [12]. Among these modeling approaches, bioclimatic envelope models such as CLIMEX, Maximum entropy modeling (Maxent), Domain, and GARP, are extensively used for predicting the potential distributions of species based on their ecological and climatic profiles [13], with the assumption that climate is the primary determinant of the distribution of plant species [14]. ...
Predicting the geographic distribution of a species together with its response to climate change is of great significance for biodiversity conservation and ecosystem sustainable development. Zelkova serrata is an excellent shelterbelt tree species that is used for soil and water conservation due to the fact of its well-developed root system, strong soil fixation, and wind resistance. However, the wild germplasm resources of Z. serrata have been increasingly depleted due to the fact of its weak ability to regenerate naturally and the unprecedented damage humans have caused to the natural habitats. The present work using Maxent aimed to model the current potential distribution of this species as well as in the future, assess how various environmental factors affect species distribution, and identify the shifts in the distribution of this species in various climate change scenarios. Our findings show habitat in provinces in the southern Qinling and Huai river basins have high environmental suitability. Temperature seasonality, annual precipitation, annual mean temperature, and warmest quarter precipitation were the most important factors affecting its distribution. Under a climate change scenario, the appropriate habitat range showed northeastward expansion geographically. The results in the present work can lay the foundation for the cultivation and conservation of Z. serrata.
... We focused on species with at least 60 occurrences in the Swiss forests. This ensured that the number of occurrences was at least 10 times larger than that of the predictors (see Section 2.4.2) and helped to reduce overfitting (Guisan and Zimmermann 2000). Consequently, 162 terricolous macrofungi were left, of which 51 are saprotrophic and 111 are either symbiotic or parasitic fungi. ...
... For GLM, we fitted partial models with species data and environmental predictors using linear and quadratic terms. We ranked the predictors within the three predictor pools using adjusted D 2 as a proxy for predictive power (Guisan and Zimmermann 2000). This was performed with the 'ecospat' R package (Broennimann et al. 2023). ...
Aim
This study aims to (1) test whether mapped soil properties can improve the performance of species distribution models (SDMs) for 162 terricolous macrofungi at a regional level, (2) identify relevant soil predictors for macrofungal regional distribution and (3) quantify the relative importance of soil properties as compared to climate and topography in explaining macrofungal regional distribution.
Location
The forested area (~ 12,000 km ² ) in Switzerland.
Taxon
Terricolous Macrofungi.
Methods
We collected occurrences (presence‐only) for 162 species of terricolous macrofungi, including 111 ectomycorrhizal and 51 saprotrophic species, from the SwissFungi database. We used soil property maps, generated through digital soil mapping at a 25 m resolution, to enhance macrofungal SDMs. For each species, we selected two climate, two topography and two soil predictors by an automated variable selection procedure. We built SDMs with randomised soil properties for performance comparison. We quantified the importance of soil properties based on permutation and variance partitioning. Finally, we projected the SDMs for three representative species at 25 m resolution with and without soil properties to assess the role of soil properties in shaping their biogeographical distributions.
Results
Soil properties significantly improved the median performance of the SDMs across the 162 species. Ectomycorrhizal fungi showed a significantly greater improvement than saprotrophic fungi. On average, our models were able to explain two‐thirds of the variance in macrofungal distribution, of which 11% could be independently explained by soil properties. Air temperature and topographic slope were identified as additional important factors controlling macrofungal distribution. Evident changes in geographical distribution were observed for the three representative species after adding soil properties.
Main Conclusions
High‐resolution digital soil maps significantly improve the predictive accuracy of macrofungal regional distribution. They should therefore be taken into account when modelling the geographical distribution of macrofungi.
... Sistem informasi geografis dikombinasikan dengan penghitungan multivariat digunakan untuk menentukan kesesuaian habitat dan memungkinkan pengelola untuk membuat peta distribusi potensial suatu jenis terancam punah seperti orangutan (Hirzel et al., 2004;Guisan & Zimmermann, 2000;Chefaoui 2005). Long et al. (2008) menggunakan teknologi citra landsat untuk mengukur populasi dan status satwa yang terancam punah di Madagaskar, sedangkan Engler et al. (2004) menggunakan hasil suatu model kesesuaian habitat untuk jenis satwa yang menjadi target konservasi, Wich et al. (2012b) menggunakan model kesesuaian habitat secara global untuk orangutan di Kalimantan. ...
Kabupaten Kapuas Hulu sebagai kabupaten konservasi telah menetapkan daerah koridor satwa yang menghubungkan Taman Nasional Betung Kerihun dan Taman Nasional Danau Sentarum sebagai Kawasan Strategis Kabupaten yang menonjolkan aspek lingkungan. Koridor satwa ini memiliki peranan yang penting bagi pergerakan satwa terutama orangutan dari kedua taman nasional ini. Studi ini dilakukan untuk memahami dampak tata guna lahan terhadap sebaran orangutan, di koridor satwa. Pembukaan jalan, perkebunan skala besar, pembukaan lahan untuk pemukiman, perladangan serta penebangan hutan telah menjadi penyebab terputusnya habitat orangutan. Wilayah yang masih aman sebagai habitat orangutan adalah di dalam kawasan taman nasional. Penelitian ini dilakukan menggunakan teknologi penginderaan jarak jauh untuk memetakan sebaran dan kesesuaian habitat orangutan di kawasan koridor satwa. Tujuh parameter habitat orangutan digunakan dalam analisis spasial kesesuaian habitat. Dari hasil penelitian ini didapatkan bahwa kawasan koridor memiliki tingkat kesesuaian habitat yang tinggi sebesar 49.94%, tingkat kesesuaian sedang sebesar 46.61% dan kesesuaian yang rendah sebesar 3.46%. dan hasil ini ditunjang dengan besaran nilai validasi untuk kelas kesesuaian sedang sebesar 32.29% dan kelas kesesuaian tinggi sebesar 67.71%.
... Species distribution models (SDM) estimate where species are likely to occur based on the relationships between environmental variables and known wild occurrences by producing habitat suitability maps (Elith and Leathwick 2009;Guisan and Zimmermann 2000;Lehmann et al. 2002). Predicting species distribution under changing climate using SDM has been done extensively over the last two decades, and various methods have been proposed Bellard et al. 2012;Elith et al. 2010;Hamann and Aitken 2013;Hijmans and Graham 2006;Pearson and Dawson 2003;Porfirio et al. 2014;Thuiller et al. 2005). ...
Aim
Future species distributions were modelled to assess the distributional changes of 1692 plant species according to several scenarios. Species vulnerability was calculated according to their red list and native statuses using a new index to better understand the most vulnerable groups of species to future conditions and improve conservation actions.
Location
Greater Geneva region between France and Switzerland.
Time Period
From present to 2050.
Major Taxa Studied
A total of 1692 species of plants from all major groups.
Methods
Two different species distribution models were created to combine the effect of climate change at the continental scale and land‐use changes at the regional scale in 2050. Two scenarios of climate change were used (optimistic and pessimistic) and one business‐as‐usual land‐use–land‐cover scenario. Current and future distributions were compared using six spatial indicators combined to create a new vulnerability index to global changes.
Results
More than one‐third of all species assessed showed at least a moderate vulnerability, and more than 10% showed a high vulnerability to global changes, mostly composed of native species. Most exotic species showed moderate benefits to global changes, and one‐third were associated with high benefits. Pessimistic scenarios of climate change exacerbated the trends identified under optimistic scenarios and were associated with higher vulnerability for native species and higher benefits for exotic ones. No clear pattern was found when comparing species vulnerability according to their red list status, questioning its ability to preserve species in the long term.
Main Conclusions
Native species are more vulnerable to global changes, while exotic species benefit from them. Climate is the main driver of future distributional changes in the study area. Current levels of threat fail to inform us of species vulnerability to future conditions, questioning their relevance and supporting the assessments of similar studies to identify the most vulnerable species.
... The spatial distribution of sample points is also an important consideration. If the sample points are clustered within a small area, they may be insufficient to represent the entire study region [27,28]. By introducing a negative distance score into the fitness function, the dispersion of sample points in space is encouraged, thus preventing the over-sampling of a particular environmental condition [29,30]. ...
Soil mapping plays a crucial role in optimizing agricultural production by providing spatially explicit information on soil types and properties, which supports decision-making in precision fertilization, irrigation, and crop selection. Traditional soil mapping methods, which rely on field surveys and laboratory analyses, face challenges related to efficiency and scalability. Although combining legacy soil maps with environmental covariates can reveal soil–environment relationships and improve sampling layouts, low soil spatial variability and significant human activity in plain areas often hinder the effectiveness of existing algorithms, making them sensitive to sample density and environmental variability. This study proposes a genetic algorithm (GA)-based sampling optimization framework tailored to plain areas with low soil spatial variability. By integrating legacy soil maps and environmental covariates, the GA dynamically balances spatial dispersion and environmental representativeness, addressing the limitations of traditional methods in homogeneous landscapes. In a case study conducted in Tongzhou District, Beijing, China, the GA sampling method combined with random forest modeling, applied to soil type mapping, achieved the highest kappa coefficient of 70.25% with 5000 sampling points—an average improvement of 10% over fuzzy C-means clustering and K-nearest neighbor methods. Additionally, field-validated accuracy reached 89.69%, representing a 13% improvement over the other methods. This study demonstrates that the GA-based sampling approach significantly enhances sample representativeness and efficiency, thereby improving the accuracy of digital soil mapping. The proposed method offers an efficient and reliable solution for soil mapping in plain areas, contributing to optimized land use and more informed precision agriculture decisions.
... It uses the combination of species presence data with characteristics such as bioclimatic, edaphic, topographic, and other parameters (Cotrina Sánchez et al., 2021;Guisan & Zimmermann, 2000;Kong et al., 2021). Among various SDM algorithms, the Maximum Entropy (MaxEnt) model (Phillips et al., 2006) combines machine learning and maximum entropy principles to predict the potential distribution areas of species (Elith et al., 2011). ...
Climate change and human activity have significantly contributed to the decline of biodiversity worldwide. Orchids, in particular, are sensitive to disturbances and may respond rapidly to the impacts of climate change than many other plant species. Due to their complex biology and the Original Research Article pressures placed on their habitats, orchids are considered a highly vulnerable group of plants. The lateritic plateaus of northern Western Ghats support a variety of life forms and are quite fragile in nature. The slightest pressure on the habitat affects all the flora and fauna thriving in these microhabitats. Habenaria suaveolens is one of the endemic orchids that can be found growing in these habitats only. In the present study, we examine the effect of climate change on the distribution of threatened H. suaveolens utilizing ecological niche modeling for present and future climatic scenarios (SSP 2-4.5 & SSP 5-8.5) to identify key environmental determinants and population parameters. A total of 25 occurrence records were used to study the SDM model using the MIROC6 global climatic model. The result revealed that only 19.50% (7921 km 2) is highly suitable for the species in the current period. The highly suitable area reduces from 57.99% (SSP245) to 74.12% (SSP585) in 2090. Moreover, the total suitable habitat gets confined to less than 1000 km 2 in future climatic conditions. The major environmental variable that predicts the distribution of the species are Bio17 (Precipitation of driest quarter), Bio1 (Annual Mean Temperature), Bio2 (Mean Diurnal Range), and aspect. Recently, the species is becoming less in population due to man-made activities like quarrying, tourism, grazing, trampling, etc., in and around the lateritic plateaus, leading to disturbances in the habitat. This study has provided a baseline data about the contiguous distribution of the species and its potentially suitable habitats in the future for conservationists to take precautionary steps for the protection of the species. Hence, a strategic plan can be compiled based on this to conserve these highly habitat specific plants.
... Future changes MaxENT (Phillips et al., 2006), GLMs (Guisan and Zimmermann, 2000), Random Forest model (Prasad et al., 2006), etc. ...
The ‘6W principle’ provides a paradigmatic framework for microbial biogeography.
Six key actions such as developing unified theoretical framework were suggested.
The 6W principle should be further refined by the whole community.
A comprehensive understanding of microbial biogeography is essential to elucidate the mechanisms that regulate microbial diversity and facilitate ecosystem functioning. Here, we present a standardised approach for microbial biogeography research, using the ‘6W principles’ of ‘Who’, ‘What’, ‘Where’, ‘When’, ‘Why’, and ‘How’, to provide a paradigmatic framework for its study. The ‘6W principle’ we developed aimed to address the six fundamental questions in microbial biogeographical researches, including the taxonomic and functional identity, abundance and diversity, distribution patterns, movement or evolutionary trajectory, driving factors, and future changes of microbial communities. Some key corresponding actions were suggested to promote the microbial biogeographical research such as constructing high-resolution taxonomic and functional annotation databases, developing absolute-quantitative high-throughput sequencing, increasing sampling coverage, establishing multidimensional time-series monitoring, developing unified theoretical frameworks and advanced biogeographical modelling approaches, and establishing long-term global networking experiments. We call on the community to jointly enrich the connotation and coverage of the 6W principle, in order to promote the further development and exploitation of microbial biogeography in the context of ongoing global change.
... Also, the multi-species umbrella approach has been proposed to better represent the conservation needs of co-occurring species (Kang et al., 2013;Wang et al., 2018). Secondly, our method is based on SDMs, so the results depend on the selected model and the environmental variables inputted (Guisan & Zimmermann, 2000). The availability of more environmental predictive variables in the future could improve the precision of model analyses. ...
Umbrella species are widely used as conservation strategies for the design of reserves. However, empirical data on their potential effectiveness, particularly in the context of climate change, is limited and inconclusive. Here, we evaluated the potential umbrella effectiveness of the giant panda (Ailuropoda melanoleuca) and its reserve network in the Qinling Mountains for the conservation of sympatric golden snub‐nosed monkeys (Rhinopithecus roxellana) under climate change. We modeled their current and future habitat suitability index (HSI) by using the MaxEnt model and analyzed their current and future spatial congruence of HSI. Their suitable areas were also overlaid onto the boundaries and management zones of the reserve network. Subsequently, we designed a series of indices to assess the potential umbrella effectiveness of the reserve network under climate change. Our results indicate that their habitat spatial congruence will remain high in the future. The suitable habitat of giant pandas overlaps substantially with that of golden snub‐nosed monkeys in both present and future. Importantly, the umbrella effectiveness of the reserve network for the golden snub‐nosed monkey will not decrease. However, there are still some protection gaps shared by them outside the reserve network. Overall, our results demonstrate that the giant panda and its reserve network can serve as an effective umbrella for the golden snub‐nosed monkey under climate change, providing theoretical support to the application of umbrella species under climate change.
... El análisis multicriterio (AMC) es un proceso cualitativo que involucra el juicio de expertos y las preferencias de grupos sociales con respecto a la toma de decisiones con criterios múltiples, para seleccionar la opción más adecuada y mejor vinculada con los objetivos establecidos (Adams y Ghaly, 2007;Etxano y Villalba-Eguiluz, 2021). En el caso de los megaproyectos, el proceso de decisión es aún más complejo por las dimensiones económicas y territoriales, que deben vincularse con la sustentabilidad, y la importancia relativa que se asignan es estos factores para las decisiones (Guisan y Zimmermann, 2000). ...
Avances en nuevos modelos del turismo en México: Sustentabilidad, cultura e inclusión como ejes del desarrollo endógeno presenta catorce capítulos de investigación en los cuales se exploran diversas problemáticas relacionadas con el turismo a lo largo del país en destinos clave, tales como Ciudad de México, Jalisco y Quitana Roo. Los autores abordan a través de diferentes estudios de caso y análisis bajo diversas metodologías posibles soluciones. El libro enfatiza la sostenibilidad y la cultura como principales elementos que guían al desarrollo de modelos turísticos responsables en la actualidad en el país, integrando a las comunidades locales como parte de los elementos clave para asegurar que la actividad turística funja como verdadero motor de desarrollo, beneficiando a todos los involucrados.
... Hence, effective species management at both habitat and landscape levels requires a clear understanding of species distributions and the identification of suitable habitats [36]. In this context, species distribution models (SDMs) serve as valuable tools for predicting species occurrence across specific geographic regions, providing critical insights for habitat management and conservation planning [37][38][39][40]. In recent years, the integration of climatic and habitat variables into SDMs has become increasingly important for projecting species distributions, particularly in assessing range shifts in response to climate change [41][42][43][44]. ...
Global warming and anthropogenic threats are significant drivers of biodiversity loss, particularly impacting smaller mammalian species. Hence, this study assessed two overlooked giant flying squirrel species, Petaurista magnificus and Petaurista nobilis, distributed across the transboundary regions of the Eastern Himalayas and Indo-Burma biodiversity hotspots. Utilizing a maximum entropy (MaxEnt) species distribution model, this study delineated suitable habitats within the IUCN-defined extent of both Petaurista species based on two modeling approaches: the habitat–climate model (HCM) and the climate-only model (COM). The models identified suitable habitat coverage of only 3.92% (HCM) and 3.75% (COM) for P. magnificus and 14.17% (HCM) and 10.04% (COM) for P. nobilis. However, as the HCM integrates both environmental and habitat variables, providing a more holistic assessment, it revealed limited biological corridor connectivity within the IUCN-defined extent for both species. Furthermore, the future projections based on the HCM indicate habitat loss of up to 81.90% for P. magnificus and 89.88% for P. nobilis due to climate change, alongside severe fragmentation, leading to the disappearance of viable habitat patches. These remaining suitable patches are expected to shrink and become increasingly isolated in the future due to climate change. Furthermore, centroid shift analyses based on the HCM predict a northwestward shift for P. magnificus and a westward shift for P. nobilis under different climate scenarios. Hence, to address these conservation challenges, the study underscores the necessity for extensive field surveys, genetic assessments, habitat corridor evaluations, and the establishment of transboundary conservation frameworks to formulate an evidence-based species management strategy for both Petaurista species.
... This approach has proven to be particularly valuable in studying organisms that are logistically difficult to sample on a large scale, such as MCEs or deep-sea habitats (Costa et al. 2015;Ross et al. 2015;Nolan et al. 2024). HSMs are a rapid and cost-effective tool that can help unravel two research interests: first, a better understanding of which environmental factors are the most important drivers of a species' distribution and second, predicting the distribution of investigated species over a defined area (Guisan and Zimmermann 2000). To date, modelling approaches are the main methods used to map benthic communities in deep water (Howell et al. 2022). ...
To support conservation efforts, accurate mapping of marine organism community’ distribution has become more critical than ever before. While previous mapping endeavours have primarily focused on easily accessible shallow‐water habitats, there remains limited knowledge about the ecosystems lying beyond SCUBA diving depths, such as mesophotic coral ecosystems (MCEs, ~30–150 m). MCEs are important habitats from an ecological and conservation perspective, yet little is known about the environmental factors that shape these ecosystems and their distribution, particularly in the Indian Ocean region. The goals of this study are to (1) predict the spatial distribution and extent of distinct benthic communities and MCEs in the Chagos Archipelago, central Indian Ocean, (2) test the effectiveness of a range of environmental and topography derived variables to predict the location of MCEs around Egmont Atoll and the Archipelago, and (3) independently validate the models produced. In addition, we compared the MCEs predicted extent in the Archipelago for the models derived from high‐resolution multibeam and low‐resolution GEBCO bathymetry data. Using maximum entropy modelling, all models resulted in excellent (> 0.9) performances, for AUC and threshold‐dependent metrics, predicting extensive and previously undocumented MCEs across the entire Archipelago with, however, differences in the predicted extent between the high‐ and low‐resolution models. Independent validation resulted in fair (> 0.7 AUC) and poor (> 0.6 AUC) performances for the high‐resolution and low‐resolution models, respectively. Photosynthetically active radiation (PAR), temperature, chlorophyll‐a, and topographically derived variables were identified as the most influential predictors. In conclusion, this study provides the first prediction of the distribution of MCEs and their distinct benthic communities in the Archipelago. It highlights their significance in terms of potential extent and response to various environmental factors, supporting decision making for prioritising future survey sites to study MCEs across the Archipelago and targeting ecologically important areas for conservation.
... To minimize the bias related to the data split process, the process was repeated 20 times to construct 20 training datasets and 20 testing datasets. A total of 180 SDMs were developed based on the nine SDM algorithms and 20 training datasets (Guisan and Zimmermann 2000). The Area Under the Curve (AUC) of the Receiver Operating Characteristic and True Skill Statistic (TSS) were calculated with the test dataset to estimate the model predictive accuracy (Allouche et al. 2006;Jiménez-Valverde 2012). ...
Climate and land‐use changes are key factors in the habitat loss and population declines of climate change‐sensitive endangered species. We assessed the climate change effects on the distribution of Parnassius bremeri, a critically endangered wildlife species in the Republic of Korea, in association with food availability (Sedum kamtschaticum and Sedum aizoon), land‐use change, and dispersal limitation. We first predicted the current and future distributions of P. bremeri, S. kamtschaticum, and S. aizoon using the presence/absence data and current (2000) and future climate data (2050, 2100) with BioMod2, an ensemble platform for species distribution model projections. Then, the dispersal capacity of P. bremeri and land‐use change were coupled with SDMs using MigClim. We used future climate and land‐use changes predicted according to the SSP scenarios (SSP1‐2.6, SSP2‐4.5, and SSP3‐7.0) and the dispersal model estimated from previous studies. The current distributional areas of P. bremeri were predicted to be about 10,956 km² without land‐cover coupling and 8.861 km² with coupling, showing land‐cover decreased by about 19% of the suitable habitat. The future predictions under climate change only showed the distribution reduced by 56% and 50% in 2050 and 2100 under SSP1‐2.6, respectively, 55% and 48% under SSP2‐4.5, and 44% and 14% under SSP3‐7.0. Applying land‐use change and dispersal capacity further decreased the future distribution of P. bremeri but trivially (about 0.42% on average). The strict conservation policies and measures for P. bremeri's habitats explain the trivial additional decrease, delaying its habitat loss. However, our results suggest that such efforts cannot halt the climate change‐driven habitat loss trend of P. bremeri. Strong climate mitigation efforts and promoting the species' adaptive capacity are the only ways to reverse the tragic decline of climate‐sensitive species.
... Similarly, (bio12, annual precipitation) and (bio1, annual mean temperature) influenced the distribution probability of P. microcarpa. High altitude places (mountain areas) are typically associated with substantial amounts of precipitation [48] and as a result, the temperature is lower than in the lowlands. The Zagros Mountains undergo significant variations in temperature and precipitation throughout the seasons, characterized by scorching, arid summers and cool, rainy winters. ...
Prunus microcarpa is an endemic species prevalent throughout the highlands of the Kurdistan Region of Iraq. Conservation, introduction, and restoration efforts require an in-depth understanding of the species’ current and future habitat distributions under different climate change scenarios. This study utilized field observations, species distribution modeling, geospatial techniques, and environmental predictors to analyze the distribution and forecast potential habitats for P. microcarpa in the highlands of Iraq. Findings indicate that, according to the global climate models (i.e., BCC-CSM2-MR and MRI-ESM2.0), the reduction in habitat for the species is projected to be more than the potential expansion. Specifically, the area of habitat is expected to reduce by 2351.908 km² (4.6%) and 2216.957 km² (4.3%), while it could increase by 1306.384 km² (2.5%) and 1015.612 km² (2.0%) for the respective climate models. Topographic features such as elevation and slope, climatic conditions, precipitation seasonality, and annual mean temperature relatively shape the distribution of P. microcarpa. The modeling demonstrated good predictive capability (area under the curve (AUC) score = 0.933). The total study area is approximately 51,558.327 km², with around 20.5% (10,602 km²) identified as suitable habitat for P. microcarpa. These findings offer essential baseline information for conservation strategies and provide new insights into where the species currently resides and where it could be found in the future. This underscores how combining distribution modeling with geospatial techniques can be effective, particularly in data-deficient regions like Iraq.
... Environmental data like bathymetry, temperature, and surface primary productivity are obtained in high spatialtemporal resolution with sonars and satellite images, while biodiversity data are sparse and logistically challenging to obtain (Balmford and Gaston, 1999;Heink and Kowarik, 2010), particularly offshore. Accurate inferences of biodiversity based on environmental data are crucial for marine ecosystem monitoring and conservation (Guisan and Zimmermann, 2000;Guisan and Thuiller, 2005;Holon et al., 2018). One challenge of modeling marine biodiversity is that oceanographic processes are dynamic, differ in spatial extent, and interact with each other (Sonnewald et al., 2021). ...
Predicting ecological communities is highly challenging but necessary to establish effective conservation and monitoring programs. This study aims to predict the spatial distribution of nematode associations from 25 m to 2500 m water depth over an area of 350,000 km² and understand the major oceanographic processes influencing them. The study considered data from 245 nematode genera and 44 environmental parameters from 100 stations. Data was analyzed by means of a hybrid machine learning (ML) approach, which combines unsupervised and supervised methods. The unsupervised phase detected that the nematodes were geographically structured in six associations, each with representative genera. In the supervised stage, these associations were modeled as a function of the environmental features by five supervised algorithms (Support Vector Machine, Random Forest, k-Nearest Neighbors, Naive Bayes, and Stochastic Gradient Boosting), using 80% of the samples for training, leaving the remaining for testing. Among them, the random forest was the best model with an accuracy of 86.4% in the test portion. The Random Forest (RF) model recognized 8 environmental features as significant in predicting the associations. Depth, the concentration of dissolved oxygen in the water near the bottom, the quality and quantity of phytodetritus, the proportion of coarse sand and carbonate, the sediment skewness, pH, and redox potential were the most important features structuring them. The inference of each association across the whole study area was based on the modeling results of the 8 significant environmental features. This model still correctly classified 90% of test data. Such findings demonstrated that it is possible to infer the spatial distribution of the nematode associations using only a small set of environmental features. The recommendation is thus to permanently monitor these environmental variables and run the ML models. Implementing ML approaches in monitoring programs of benthic systems will increase our prediction capacity, reduce monitoring costs, and, ultimately, support the conservation of marine systems.
... Nautilids often remain within reef slopes [36], which are a key factor in their distribution. Given the lack of detailed topographic data, distance to land was used as a proxy for topographic features, as it is generally correlated with local topography [37]. Table 1. ...
Nautilus pompilius, a ‘living fossil’ of the oceans, is crucial to the study of biological evolution and paleontology. However, the species’ habitat has been severely impacted by global climate change. Based on this, species distribution models and conservation gap analyses were conducted under current and future climate scenarios. The results revealed that the current habitats for N. pompilius were primarily located in the coastal waters of Australia, Indonesia, and the Philippines. Under the Representative Concentration Pathway (RCP) 4.5 scenario, suitable habitat is projected to decline by 4.8% in the 2050s and 5.3% in 2100s. This loss is expected to intensify under higher emission scenarios, particularly RCP 8.5, where the reduction could reach 15.4% in the 2100s. Conservation gap analysis indicates that while nearly 30% of suitable habitats fall within marine protected areas (MPAs), many vulnerable regions remain unprotected. Future MPA establishment should strategically address these conservation gaps, particularly in coastal waters such as the Gulf of Carpentaria, the Arafura Sea, and the southern edge of the Timor Sea. This study provides critical insights into the distribution patterns and conservation needs of N. pompilius, emphasizing the urgent need for targeted conservation efforts to protect this endangered species.
... By linking species occurrences to environmental variables, ENM predicts species distributions across space and time (Peterson et al. 2011). ENM has broad applications, including forecasting species distributions, informing conservation planning by examining present and future species-environment interactions and managing invasive species (Guisan and Zimmermann 2000;Phillips et al. 2006;Peterson et al. 2011;Cordier et al. 2020). ENM is particularly valuable for IAS management as it provides cost-effective insights for predicting, preventing, and controlling invasions (Guisan and Thuiller 2005;Marcelino and Verbruggen 2015;Cordier et al. 2020). ...
Biological invasions are one of the major threats to biodiversity and ecosystem functioning worldwide. Rose-ringed Parakeets (Psittacula krameri) are well-known invaders and have established populations in over 40 countries outside their native range. Regular records of Rose-ringed Parakeets in Turkey go back to 1975, and the species has established populations in at least ten cities, including in metropolitan areas such as Istanbul, Izmir, and Ankara. To understand the scale of the invasion potential of Rose-ringed Parakeets in Turkey, first, we estimated their current local population size through roost counts carried out in 2020 and 2021 in four cities. Then, we used ecological niche modeling to forecast the present-time and future climatic habitat suitability over three different periods (2041-2060; 2061-2080; 2081-2100). Field surveys and citizen science data revealed the presence of the species in 37 provinces in Turkey. Our counts indicated a rapid population growth from 1355 individuals in 2015 to 6000 individuals in 2021. The ecological niche modeling predicted an overall 50% reduction in spe-cies' potential habitats in Turkey by 2100 due to climate change. Suitable habitats in Central Anatolia are predicted to shrink by the model, while those in Thrace and the coastal areas of Turkey will remain stable. Even though model results indicate a future reduction in suitable habitats for Rose-ringed Parakeets in Turkey, we argue the need for long-term management, particularly in coastal cities where already-established populations are increasing in size.
... For the best-fit nonlinear mixed-effect growth model, the nlraa package (Miguez 2023) was used to estimate marginal R 2 (fixed effects only) and conditional R 2 (fixed and random effects). Performance of non-normal GAMMs was evaluated using the adjusted deviance explained (D 2 adj ), a R 2 adj analog that penalizes for the number of parameters (Guisan and Zimmermann 2000). No reliable method to estimate the predictive performance of a QGAMM is available to our knowledge, such that this yet important metric could not be reported. ...
Although the role of recreational harvest on size structure of declining fish populations is often unclear, bag and size limits are often implemented to prevent overharvest. Long-term monitoring and periodic assessments of stock status then become necessary to evaluate their potential impacts. Based on a long-term gillnet monitoring program in the St. Lawrence River, Québec, Canada, the effects of a 381-545 mm harvest-slot length limit implemented in 2011 were evaluated on walleye (Sander vitreus). Mixed-effects models revealed continued declines in the abundance of large walleyes, size distribution, total annual mortality, and female growth, condition, and size-at-maturity. Expected impacts were mostly not achieved, potentially because of environmental and trophic interaction changes in the St. Lawrence River, in addition to increasing fishing pressure. Our results highlight a need to reassess current walleye fisheries management strategies.
... In this study, the prediction maps were not entirely accurate in delineating the actual presence of YFPs (Fig. 4). This limitation stems from SDMs predicting the Suitability of presence for YFP, reflecting constraints of binary classification in habitat prediction 33 . Suitability distributions are theoretical representations of the potential distribution of species. ...
The Species Distribution Model (SDM) provides a crucial foundation for the conservation of the Yangtze finless porpoise (YFP), a critically endangered freshwater cetacean endemic to China. In this study, we conducted population and habitat surveys, and employed the Random Forest algorithm (RF) to construct SDMs. We found that the habitat preference of YFP shows complex seasonality. Cyanobacteria and total phosphates have been identified as the predominant factors influencing the YFP distributions by affecting prey resources. We emphasize that ascertaining the presence and pseudo-absence points of YFP, in conjunction with the selection of key factors, constitutes the foundational element in the construction of SDMs. We suggest that the incorporation of techniques such as environmental DNA could expand the range of environmental factors, particularly with regard to the distribution of prey resources at the genus or species level. This study provides guidance for the SDMs of YFP and demonstrates the potential of machine learning algorithms in constructing SDMs for the endangered aquatic species.
... In this context, species distribution models (SDMs) have emerged as valuable tools for enabling precise predictions of current habitat conditions and future climatic projections for species [43][44][45][46]. These models utilize existing data on species and their ecological niches across both spatial and temporal scales to provide a deeper understanding of habitat dynamics in respect to climate and habitats [47,48]. This ensemble method utilizes a combination of modeling algorithms to forecast species distributions within different geographic regions and timelines, leveraging the distinct advantages of each model to address various factors that influence distribution patterns [49]. ...
The endemic and critically endangered gharial, Gavialis gangeticus, experienced a severe population decline in its range. However, conservation efforts, notably through the implementation of “Project Crocodile” in India, have led to a significant recovery of its population. The present study employs an ensemble Species Distribution Model (SDM) to delineate suitable habitats for G. gangeticus under current and future climatic scenarios to understand the impact of climate change. The model estimates that 46.85% of the area of occupancy is suitable under the present scenario, with this suitable area projected to increase by 145.16% in future climatic conditions. States such as Madhya Pradesh, Uttar Pradesh, and Assam are projected to experience an increase in habitat suitability, whereas Odisha and Rajasthan are anticipated to face declines. The study recommends conducting ground-truthing ecological assessments using advanced technologies and genetic analyses to validate the viability of newly identified habitats in the Lower Ganges, Mahanadi, and Brahmaputra River systems. These areas should be prioritized within the Protected Area network for potential translocation sites allocation. Collaborative efforts between the IUCN-SSC Crocodile Specialist Group and stakeholders are vital for prioritizing conservation and implementing site-specific interventions to protect the highly threatened gharial population in the wild.
... . Correlative species distribution models, which relate known locations of a species to environmental conditions, provide a method that allows for quantification of current suitable habitat (Franklin 2009). These models can then be projected using forecasts of future climatic conditions to predict where suitable habitat may exist under new environmental conditions (Guisan and Zimmermann 2000). Species distributions can be modeled using a variety of algorithms and combining the outputs of different algorithms using an ensemble modeling approach can be a powerful method to increase the robustness of predictions (Araújo and New 2006;Thuiller et al. 2009). ...
Describing future habitat for sensitive species can be helpful in planning conservation efforts to ensure species persistence under new climatic conditions. The Gila monster (Heloderma suspectum) is an iconic lizard of the southwestern United States. The northernmost range of Gila monsters is the Mojave Desert, an area experiencing rapid human population growth and urban sprawl. To understand current and potential future habitat for Gila monsters in the Mojave Desert, we fit ensemble species distribution models using known locations and current environmental variables known to be important to the species' biology. We then projected future suitable habitat under different climate forecasts based on IPCC emission scenarios. To ensure that Gila monsters would be able to disperse to newly suitable habitat, we fit Brownian Bridge movement models using telemetry data from two locations in Nevada. This model indicated that Gila monsters prefer to move through areas with a moderate slope and higher shrub cover. Modeled current suitable habitat for Gila monsters in Nevada was primarily in rugged bajadas and lower elevations at the bases of mountain ranges. Predictions of potential future habitat suggested that overall habitat suitability through 2082 would remain relatively stable throughout the study area in the lower emissions scenario, but in the high emissions scenario potential habitat is greatly reduced in many lower‐elevation areas. Future habitat areas at higher elevations under the high emissions scenario showed moderate increases in suitability, though occupancy would likely be limited by Gila monster dispersal capabilities. Finally, we determined how well the protected area network of our study area encompassed future Gila monster habitat to highlight potential opportunities to protect this important species.
... To evaluate the relative importance of environmental variables that determine the preference for nesting sites by Barn Swallows, we used generalized additive mixed-effect models (GAMMs) to fit our absence and presence data with a binomial distribution, logit link function and a spatially explicit model structure (Dormann et al. 2007). We used GAMMs because such models allow the addition of latitude and longitude data as smoothing terms to control spatial autocorrelation (Guisan and Zimmermann 2000;Lobo et al. 2010). Prior to building our models, we centered and scaled our study variables to mean 0 and standard deviation 1, which reduces heterogeneity of variance due to different value ranges and measurement units. ...
Context
Urbanization has detrimental effects on biodiversity, yet how species respond to urban planning zoning outcomes and environmental changes at different spatial scales when selecting urban breeding habitats remains understudied. Mitigating such impacts on wildlife is instrumental to create biodiversity-friendly cities while accommodating urban development.
Objectives
We used Barn Swallow nesting site data (2017–2023) collected from a citizen science program to help identify the most influential factors affecting the presence of Barn Swallow nests at site and landscape scales.
Methods
We analyzed the relationship between Barn Swallow nest site selection in urbanized areas of southern China and land use data, including built-up percentages, cropland and waterbodies, as well as environmental factors such as heat-island effect (land surface temperature), noise pollution (road density and road simplicity), artificial light at night (ALAN) and nesting building attributes (year constructed, height and surroundings).
Results
Our findings revealed a positive association between Barn Swallow nest abundance and several anthropogenic factors, including land surface temperature at the site scale, and ALAN and road simplicity at the landscape scale. Our findings indicated the building year also had a negative impact on the Barn Swallow nests.
Conclusions
These results suggest that urban design and revitalization efforts can consider mitigating negative effects by implementing measures to regulate noise pollution and nighttime lighting schemes. Furthermore, urban planning could carefully consider the requirements for biodiversity-friendly architectural elements in new constructions and rezoning process of existing urban districts, such as old residential neighborhoods and urban fringes, to minimize impacts on declining nesting sites in urban areas.
... Comparing logistic regression with other available approaches (e.g., discriminant analysis and neural networks), it exhibits a similar performance for predicting species distributions (Manel et al. 1999), and, unlike other modelling methods, it has the ability to achieve easy transformation of the effects of explanatory variables into odds ratios, which are the probabilities of an outcome occurring given the exposure to some variable (Arkes 2019). Moreover, the logistic regression approach is ideal for species that involve cost and sample size problems for distribution modelling studies (Guisan and Zimmermann 2000;Jeganathan et al. 2004;Rushton et al. 2004;Estes et al. 2008), as is the case of pudu deer due to the type of environment where it lives (remote, thick, moist, etc.), their low densities, and the elusive behaviour (Jiménez 2010). The most likely pudu deer habitat predictors were selected via Akaike's Information Criterion for small sample sizes (AIC c ), which is a relative measure that is useful for comparing a prespecified set of candidate models drawn from a common global model (Burnham and Anderson 2002). ...
The pudu deer Pudu puda is a near-threatened mammal endemic to southern Chile and Argentina. Since this species inhabits mostly human-modified landscapes, addressing the factors that influence habitat use in such environments could aid in their conservation. We evaluated the presence of pudu deer during the spring-summer of 2020–2021 and 2021–2022 at two coastal range sites in the province of Osorno (Huellelhue Community, HC, and Los Riscos, RI) using camera traps. We applied logistic regression (LR) analysis to identify the predictor variables related to vegetation type, human perturbation, and native forest fragmentation that are correlated with the presence or absence of pudu deer as a binary response variable. In HC site, LR analysis identified the most likely top models including the fragmentation variable of the contiguity of patches (AICc = 22.73, Δi = 0.00 and wi = 0.77) with a negative effect (β = -8.573, P < 0.05). In RI locality the top model included the human perturbation variable of the distance to village (AICc = 23.37, Δi = 0.00 and wi = 0.71) with a positive effect (β = 1.805, P < 0.05). Both models accounted for a medium amount of the total variance in the response variable (Nagelkerke R2 = 51–52%). Results suggest that habitat use by pudu deer is affected by human perturbation and native forest fragmentation. Our findings indicate that maintaining a more continuous native forest or areas of native forest located far from human settlements could represent an important strategy for improving the long-term survival of the species.
... In the last decades, several modelling tools have emerged to provide insights into the linkages of the ecological system with the anthropogenic pressures (Guisan and Zimmermann 2000;Fulton 2010), including marine ecosystem models such as Ecopath with Ecosim (EwE onwards). EwE models have been proven to be powerful tools, especially, but not exclusively, linked with fisheries management (Christensen and Walters 2004;Heymans et al. 2016Heymans et al. , 2020. ...
In various parts of the world, the fisheries sector is undergoing a rapid transition due to a combination of ecological and economic stressors. In this context, there is increasing interest in using integrated ecological–economic tools to support ecosystem-based fisheries practices within a value chain framework, aiming to reduce ecological impacts while ensuring socioeconomic benefits. Our study integrates a food web model with a value chain model for an area of the Mediterranean Sea. This case study may provide valuable insights for decision-makers: (1) it allows the calculation of indicators that go beyond the evaluation of species and fisheries incomes, which are relevant for assessing fisheries management; (2) it offers a more comprehensive perspective on what should be considered priority fisheries species and their key characteristics; and (3) it supports the identification of key actors of the fisheries value chain and generates pertinent information to use in vertically integrated decision-making initiatives.
... Truly independent biodiversity data are often difficult to obtain, so researchers typically validate predictive models using subsets of the same dataset. Occurrence data are divided into two categories: (1) training or calibration data for model development, and (2) testing or evaluation data for model validation [109,110]. We used the R package ENMeval to partition occurrence localities into training and testing datasets. ...
Climate change has led to global biodiversity loss, severely impacting all species, including essential pollinators like bees, which are highly sensitive to environmental changes. Like other bee species, A. dorsata is also not immune to climate change. This study evaluated the habitat suitability of A. dorsata under climate change in Pakistan by utilizing two years of occurrence and distribution data to develop a Maximum Entropy (MaxEnt) model for forecasting current and future habitat distribution. Future habitat projections for 2050 and 2070 were based on two shared socioeconomic pathways (SSP245 and SSP585) using the CNRM-CM6-1 and EPI-ESM1-2-HR-1 global circulation models. Eight bioclimatic variables (Bio1, Bio4, Bio5, Bio8, Bio10, Bio12, Bio18, and Bio19) were selected for modeling, and among the selected variables, the mean temperature of the wettest quarter (Bio8) and precipitation of the warmest quarter (Bio18) showed major contributions to the model building and strongest influence on habitat of A. dorsata. The model estimated 23% of our study area as a suitable habitat for A. dorsata under current climatic conditions, comprising 150,975 km² of moderately suitable and 49,792 km² of highly suitable regions. For future climatic scenarios, our model projected significant habitat loss for A. dorsata with a shrinkage and shift towards northern, higher-altitude regions, particularly in Khyber Pakhtunkhwa and the Himalayan foothills. Habitat projections under the extreme climatic scenario (SSP585) are particularly alarming, indicating a substantial loss of the suitable habitat for the A. dorsata of 40% under CNRM-CM6-1 and 79% for EPI-ESM1-2-HR-1 for the 2070 time period. This study emphasizes the critical need for conservation efforts to protect A. dorsata and highlights the species’ role in pollination and supporting the apiculture industry in Pakistan.
... In addition to overcoming the limitations of our approach, future work could build upon these classification maps in various ways. First, rocky land cover classification maps could be used in species distribution modeling [75,76] to identify which rocky habitat patches support species of interest such as the American pika or hoary marmot. By identifying species presence across the region and taking measurements on habitat characteristics, future research could combine field-based measurements with remote sensing data to improve our understanding of the species that rely on rocky habitats. ...
Rocky land cover provides vital habitat for many different species, including endemic, vulnerable, or threatened plants and animals; thus, various land management organizations prioritize the conservation of rocky habitat. Despite its importance, land cover classification maps rarely classify rocky land cover explicitly, and if they do, they are limited in spatial resolution or extent. Consequently, we used random forest models in Google Earth Engine (GEE) to classify rocky land cover at a high spatial resolution across a broad spatial extent in the Cascade Mountains and Columbia River Gorge in Washington, USA. The spectral indices derived from Sentinel-2 satellite data and NAIP aerial imagery, the specialized multi-temporal predictors formulated using time series of normalized burn ratio (NBR) and normalized difference in vegetation index (NDVI), and topographical predictors were especially important to include in the rocky land cover classification models; however, the predictors’ relative variable importance differed regionally. Beyond evaluating random forest models and developing classification maps of rocky land cover, we conducted three case studies to highlight potential avenues for future work and form connections to land management organizations’ needs. Our replicable approach relies on open-source data and software (GEE), aligns with the goals of land management organizations, and has the potential to be elaborated upon by future research investigating rocky habitats or other rare habitat types.
... Цель исследования -создание моделей пространственного распространения Балашёв, Байдашников, 2012;Коцур, 2013;Островский, 2016;Муханов, Лисицын, 2017;Bondareva et al., 2020;Гураль-Сверлова, Гураль, 2020;Земоглядчук, 2020;Egorov, 2021;10.15468 Крамаренко, Сверлова, 2001;Сверлова, Гураль, 2007;Вычалковская, 2008;Рабчук, Земоглядчук, 2011;Kramarenko, Dovgal, 2014;Balashov et al., 2018a;Гураль-Сверлова и др., 2018;Хайленко, 2018;Zhukov et al., Стойко, Булавкина, 2008; Сачкова, 2009; Балашёв, Байдашников, 2010; Балашёв и др., 2013; Коцур, 2013, 2015; Шиков, 2016; Снегин, Артемчук, 2017; Алексанов и др., 2019; Гураль-Сверлова, Гураль, 2020; 10.15468/qxy4mc; 10.15468 Так как данные были получены из разных источников, в некоторых случаях по причине близкого расположения точек находок возможно искажение распределения данных из-за неравномерности их сбора [Guisan, Zimmermann, 2000]. Для того чтобы этого избежать, была проведена процедура пространственного разреживания точек находок каждого вида с использованием пакета spThin в среде R [Aiello-Lammens et al., 2015]. ...
В работе представлены модели распространения девяти чужеродных видов наземных моллюсков (Mollusca, Gastropoda, Stylommatophora) на территории Восточной Европы, а именно на европейской территории России и cопредельных территориях. В качестве предикторов выбраны климатические
переменные, тип землепользования и расширенный вегетационный индекс (EVI). Созданы модели потенциального распространения каждого вида, а также выявлены территории, пригодные для совместного обитания видов-вселенцев. Анализ перекрывания потенциальных ареалов показал, какие виды
могут совместно заселить новые для них территории. Показано, что наиболее благоприятные условия обитания исследуемых видов определены в Причерноморье, Приазовье и на Кавказе, Подольской возвышенности, Среднедунайской и Нижнедунайской низменности. Подтверждена приуроченность наземных моллюсков-вселенцев к антропогенно измененным ландшафтам.
... models. Moreover, selecting environmental variables that directly influence on physiological responses would be ideal but challenging when relating modelled projections to local SRs (Gardner et al. 2019;Guisan and Zimmermann 2000). Therefore, we ran models with different combinations of climatic, soil, and land cover variables derived from multiple global databases (Table S1) to evaluate their relative performance. ...
Species distribution models rely on species' observed geographic distributions, which reflect only subset of the true ecological niche. This inevitably leads to discrepancies between the predictions of habitat suitability (HS) and the actual ecological performance in novel environments beyond the trained range. We examined this limitation by comparing modelled HS with empirical survival rates (SRs) of three Acer species, A. davidii, A. palmatum, and A. pictum, cultivated in the UK botanic gardens. We hypothesise that ex‐situ species with greater niche overlap with native UK/European species will show higher HS, which also correspond to species' SR relative to that of local species. This HS‐SR alignment will then indicate the alignment of species' geographic range and ecological range. We first quantified niche similarity between these East Asian species and UK/Europe native Acer species at both regional and continental scales. MaxEnt models were calibrated using native occurrences with various combinations of environmental variables and model configurations, then projected onto UK regions. Species' SRs were standardised against those of native species using long‐term inventory data. Our results show that niche overlap with native species generally corresponded to predicted HS, while observed SR patterns revealed an inverse relationship. A. davidii, showing high niche overlap and high HS, exhibited the lowest SR. Contrarily, A. pictum, despite showing low niche overlap and predicting most regions unsuitable, demonstrated the highest SR, comparable to native species. This discrepancy was particularly noteworthy as A. pictum shared closer phylogenetic relationships with European species, while A. davidii was more closely related to North American species. The observed phylogenetic signal in SR patterns suggests that intrinsic traits that relate to climate tolerance may be conserved yet masked in the conventional modelling approach. This interdisciplinary approach bridges the gap between macro‐scale predictions and local‐scale individual performance, offering a new perspective on niche conservatism through a phylogenetic framework.
... A total of four separate models were used to examine each combination of variables. The proportion deviance explained by the model was calculated using adjusted R 2 in R package MuMIn (Guisan & Zimmermann 2000;Bartoń 2009). The relationship between breeding success (breeding or non-breeding) and body stores was analysed using a Bayesian generalized mixed-effect linear model with a binomial distribution and logit link function in R package blme (Chung et al. 2013;Bolker et al. 2009;Dobson & Barnett 2018). ...
In many bird species across a variety of taxa, larger individuals with greater body stores are more likely to breed successfully. The migratory Barnacle Goose Branta leucopsis breeds in arctic regions including Greenland, Svalbard, Russia, and the North and Baltic Seas. Larger body size and metabolic body stores (weight relative to body size) have been linked to increased fitness through larger clutch size, social dominance, better overall health and more efficient feeding for the Svalbard and Baltic subpopulations, but this has not yet been shown for the Greenland subpopulation. In Greenland, the geese breed in remote areas inaccessible for study, but observations made on the wintering grounds in northwest Scotland and Ireland can provide important insight into factors affecting trends in abundance. To determine the influence of larger body stores or physical size on breeding success in the Greenland Barnacle Goose subpopulation, a dataset with 60 years of morphometric measurements and field observations at one of the principal wintering grounds in Ireland was analysed. Both males and females with larger body stores were more likely to breed successfully, and this relationship was significant. Males with larger body size were also more likely to breed successfully, although this relationship was not significant. Similar relationships were not seen with pairing success, suggesting that body stores and size may have more direct influence on the ability to raise offspring than on securing a mate. While all of the Barnacle Goose subpopulations showed a dramatic increase in the past 60 years, they are susceptible to pressures on both breeding and wintering grounds (e.g. human–wildlife conflicts and avian influenza), necessitating an understanding of the factors that contribute to their population dynamics.
... In ecology, "habitat suitability" refers to the likelihood that a population of a certain size could persist in a given area, based on this area's biotic and abiotic characteristics (Guisan and Zimmermann, 2000). Identifying sufficient suitable habitat is thus an essential early step in reintroduction planning (Akçakaya et al., 2018;Cianfrani et al., 2013;Cook et al., 2010;Martínez-Meyer et al., 2006;Olsson and Rogers, 2009;Rodríguez-Soto et al., 2011). ...
... Environmental features affecting the number of magpie nests in Zielona Góra and their height were tested by using generalised linear models (GLMs) [29]. A test of variance inflation factor (VIF) was applied to check for potential multi-collinearity among predictor variables, using the "car" package for R [30]. ...
This study investigates the changes in population size, distribution, and habitat preferences of the Eurasian magpie Pica pica in Zielona Góra over 23 years, emphasising the effects of urbanisation and habitat transformation. A comprehensive survey conducted in 2022 identified 953 magpie pairs, with an average density of 8.8 pairs/km² across the current administrative boundaries of Zielona Góra (without forests), and 27.7 pairs/km² in strictly urbanised zones. The highest densities were observed in the old town (36.5 pairs/km²) and residential blocks (34.5 pairs/km²), while peripheral areas, like allotment gardens and industrial zones, showed significantly lower densities. The nests were predominantly located in coniferous trees, especially spruces, marking a shift from the previously favoured poplars. The mean nest height was 11.8 m, varying by habitat type, with the highest nests found in the old town and parks. Environmental factors, such as proximity to trash bins, water sources, and tall trees, were significant predictors of nest density and placement. These findings underscore the magpie’s adaptability to urban environments, influenced by the availability of anthropogenic resources, habitat structure, and surrounding urban features.
The Neotropics realm is a significant biogeographic region in tropical and subtropical areas of the Americas, with high biodiversity and endemism. This region serves as a prime example of how genetics and ecology intersect, especially when considering that genetic diversity in this region plays a crucial role in shaping ecological processes and species interactions. The genetic makeup of species within the Neotropics also influences their ecological roles and functions: genetic variation can affect traits such as reproductive strategies and physiological adaptations, which in turn influence the ecological relationships and species interactions. The intricate relation between genetics and ecology highlights the importance of preserving both the genetic diversity and the ecological integrity of the landscape. Fortunately, with all the data available on the Neotropics, computational approaches such as ecological niche modeling (ENM) can be used to uncover insights into how environmental factors and historical events have influenced the distribution of populations and their genetic diversity, especially under climate change scenarios. ENM provides means for identifying barriers to gene flow, areas of historical isolation, and test biogeographical hypotheses on how environmental factors have affected the genetic diversity of populations. This chapter is aimed at introducing the core concepts of ecological niche models, their intersection with population genetics studies, and applications of ENMs in conservation efforts and biodiversity management studies in the Neotropics to ensure the long-term viability and sustainability of the neotropical ecosystems in the face of human interference and environmental challenges.
Global warming is an undeniable fact occurring in different parts of the world. Climate changes can have irreversible effects on plant communities, particularly on endemic and endangered species. Therefore, it is important to predict the impact of climate change on the distribution of these species to help protect them. This study utilized the MaxEnt model to forecast the impact of climate change on the distributions of two medicinal, edible, and aromatic species, Kelussia odoratissima and Allium stipitatum, in Chaharmahal and Bakhtiari province. The study used the CCSM4 general circulation model along with two climate scenarios, RCP2.6 and RCP8.5, for the 2050s and 2070s to predict the potential impact of climate change on the distribution of the species studied. The research findings indicated that the model performed effectively for prediction (AUC≥0.9). The primary environmental variables influencing species distribution were found to be isothermality (Bio3), soil organic carbon, and pH for A. stipitatum, and soil organic carbon, precipitation seasonality (Bio15), and precipitation of the wettest month (Bio13) for K. odoratissima. The findings suggest that the distribution of the studied species is expected to decline in the 2050s and 2070s due to climate change, under both the RCP2.6 and RCP8.5 climate scenarios. The research indicates that climate change is likely to have a significantly negative effect on the habitats of these species, leading to important ecological and socio-economic impacts. Therefore, our study emphasizes the urgent need for conservation efforts to prevent their extinction and protect their habitats.
The Atlantic oyster drill Urosalpinx cinerea is an introduced predatory gastropod that has negatively impacted Olympia oyster restoration in multiple estuaries on the west coast of the United States. In San Francisco Bay, California, Atlantic oyster drills have a patchy spatial pattern of presence and absence and occur in a range of densities where they are present. This variable population distribution and a limited understanding of their local dispersal history poses a challenge to oyster restoration site selection. To address this dilemma, we evaluated five abiotic habitat factors as potential determinants of drill distribution. In 2017 and 2018, we compared quarterly drill abundance data to substrate composition, elevation, water temperature, salinity, and inundation at eight sites in Richardson Bay, a small embayment in San Francisco Bay. Using generalized linear mixed effects models, we found that amount of coarse substrate and elevation were positively and negatively, respectively, associated with drill population density at the four sites where drills were present. None of the five habitat factors, however, explained the absence of drills from the other four sites. These findings suggest ways for oyster restoration practitioners to select sites that optimize the chances oyster and drill co-existence or minimize the risk of drill invasion and point to the need for extreme caution against accidental introductions of drills to novel areas with suitable habitat. We recommend extensive drill population surveys in regions where Olympia oyster conservation is taking place coupled with additional fine-scale environmental data to better understand Atlantic oyster drill biogeography and to improve the odds of success of future Olympia oyster restoration work.
The global change threats significantly alters the ecological distribution of species across different ecosystems. Species distribution models (SDMs) are considered a widely used tool for assessing the global impact on biodiversity. Recently, remote sensing data have been used in a growing number of studies to predict species distribution and improve SDMs performance. This study evaluates the contribution of spectral indices in species distribution modeling using MaxEnt. We compared models based on spectral indices data (RS-only), environmental variables (EN-only), and their combination (CM) to predict the distribution of three key Mediterranean native species: Thymelaea hirsuta, Ononis vaginalis, and Limoniastrum monopetalum. The combined models (CM) demonstrated superior performance with excellent accuracy measures values compared to other models. Jackknife tests revealed both environmental factors (e.g., distance to coastline, mean temperature of wettest and driest quarters) and spectral indices (e.g., NDWI, LST) contributed substantially to predicting the studied species. The findings emphasize the importance of integrating diverse data sources to improve the accuracy of SDMs, particularly in heterogeneous landscapes like the Mediterranean region. This integrated approach provides a more comprehensive understanding of species spreading patterns and is critical for effective management and conservation strategies.
Understanding the characteristics and key driving factors of plant diversity is of great significance for maintaining biodiversity and the ecosystem. Current studies on plant diversity in the Yellow River Delta are limited to local areas; there is a lack of comprehensive discussion on the spatial heterogeneity of plant diversity and the driving factors at a regional scale. Based on field investigations, this study explored the characteristics of plant composition and diversity under the land–sea gradient, with particular emphasis on the differences of plant diversity under different riverbanks and at a distance from the sea. Using the regression, redundancy, and Mantel test analysis, we analyzed soil properties, environmental factors, and human influence to assess their potential impacts on plant diversity. The results demonstrated that Asteraceae, Poaceae, and Amaranthaceae are the dominant plant families in the Yellow River Delta. As the distance from the sea increases, the community transitions from the monospecies dominance of Suaeda salsa to one dominated by various plants. The species similarity was higher in the adjacent environment and coastal areas. The overall level of plant diversity was not high, and the Margalef, Shannon–Wiener, Simpson, and Pielou index showed a fluctuating downward trend from land to sea. Notably, there was a peak value in the region of 3–17 km and >42 km from the sea. The plant diversity of the main stream bank was higher than that of its tributaries, where the former was more susceptible to human interference and the latter to soil electrical conductivity. In terms of the region, soil electrical conductivity had the greatest influence on plant diversity. This study could provide theoretical support for vegetation restoration and ecological protection in the Yellow River Delta.
Rapid climate change has significantly impacted species distribution patterns, necessitating a comprehensive understanding of dominant tree dynamics for effective forest resource management and utilization. The Camellia subgenus Camellia, a widely distributed taxon in subtropical China, represents an ecologically and economically important group of woody plants valued for both oil production and ornamental purposes. In this study, we employed the BIOMOD2 ensemble modeling framework to investigate the spatial distribution patterns and range dynamics of the subgenus Camellia under projected climate change scenarios. Our analysis incorporated 1455 georeferenced occurrence records from 15 species, following the filtering of duplicate points, along with seven bioclimatic variables selected after highly correlated factors were eliminated. The ensemble model, which integrates six single species distribution models, demonstrated robust predictive performance, with mean true skil l statistic (TSS) and area under curve (AUC) values exceeding 0.8. Our results identified precipitation of the coldest quarter (Bio19) and temperature seasonality (Bio4) as the primary determinants influencing species distribution patterns. The center of species richness for the subgenus Camellia was located in the Nanling Mountains and eastern Guangxi Zhuang Autonomous Region. The projections indicate an overall expansion of suitable habitats for the subgenus under future climate conditions, with notable scenario-dependent variations: distribution hotspots are predicted to increase by 8.86% under the SSP126 scenario but experience a 2.53% reduction under the SSP585 scenario. Furthermore, a westward shift in the distribution centroid is anticipated. To ensure long-term conservation of Camellia genetic resources, we recommend establishing a germplasm conservation center in the Nanling Mountains region, which represents a critical biodiversity hotspot for this taxon.
To address the urgent biodiversity crisis, it is crucial to understand the nature of plant assemblages. The distribution of plant species is not only shaped by their broad environmental requirements, but also by micro-environmental conditions, dispersal limitations, and direct and indirect species interactions. While predicting species composition and habitat identity is essential for conservation and restoration purposes, it thus remains challenging. In this study, we propose a novel approach inspired by advances in large language models to learn the "syntax" of abundance-ordered plant species sequences in communities. Our method, which captures latent associations between species across diverse ecosystems, can be fine-tuned for diverse tasks. In particular, we show that our methodology is able to outperform other approaches to (i) predict species that might occur in an assemblage given the other listed species, despite being originally missing in the species list (+16.53% compared to co-occurrence matrices and +6.56% compared to neural networks) and (ii) classify habitat types from species assemblages (+5.54% compared to expert systems and +1.14% compared to deep learning). The proposed application has a vocabulary that covers over ten thousand plant species from Europe and adjacent countries and provides a powerful methodology for improving biodiversity mapping, restoration, and conservation biology.
Climate change poses significant challenges to the health and functions of forest ecosystems. Ecological niche models have emerged as crucial tools for understanding the impact of climate change on forests at the population, species, and ecosystem levels. These models also play a pivotal role in developing adaptive forest conservation and management strategies. Recent advancements in niche model development have led to enhanced prediction accuracy and broadened applications of niche models, driven using high‐quality climate data, improved model algorithms, and the application of landscape genomic information. In this review, we start by elucidating the concept and rationale behind niche models in the context of forestry adaptation to climate change. We then provide an overview of the advancements in occurrence‐based, trait‐based, and genomics‐based models, contributing to a more comprehensive understanding of species responses to climate change. In addition, we summarize findings from 338 studies to highlight the progress made in niche models for forest tree species, including data sources, model algorithms, future climate scenarios used and diverse applications. To assist researchers and practitioners, we provide an exemplar data set and accompanying source code as a tutorial, demonstrating the integration of population genetics into niche models. This paper aims to provide a concise yet comprehensive overview of the continuous advancements and refinements of niche models, serving as a valuable resource for effectively addressing the challenges posed by a changing climate.
Alpine vegetation found in mountainous areas is reportedly sensitive to the expected heat caused by climate change. Aconitum species, which is an indicator species for monitoring the signal of climate change, also falls under this range. Aconitum heterophyllum and Aconitum balfourii thrive among diverse plant communities and serve as key species marking the onset of the alpine range. The current study uses the maximum entropy (MaxEnt) program to estimate the geographic distribution of Aconitum taxa for the present and the future (RCPs 2.6–8.5 encompassing 2050). The final model that was obtained had an area under the ROC curve AUC of 0.95, demonstrating the model’s resilience. Species distribution modeling (SDM) toolbox was used to evaluate changes in the habitat appropriateness, area expansion, and contraction based on the MaxEnt model. Future forecasts indicate that the Eastern Himalayas would be a more favorable environment for both Aconitum species when compared to the Western Himalayas. According to our best knowledge, this package was used for the first time in the study to evaluate the dynamics of the climatic niches of two critically endangered Aconitum species for the Indian Himalayan ranges. The climatic niche of this alpine species is expected to shift in the future due to alterations in its habitat regions. The Aconitum niche is in danger under scenarios of future climate change, according to the AUC and Jackknife values. The findings of the current study contribute substantially to understanding how climate change affects the Himalayas and have broad implications for the development of scientifically sound adaptation and mitigation measures, as the major factors influencing its distribution are altitude, snow cover, chilling hours for both seeds and tubers, temperature, and rainfall patterns.
Aboveground biomass (AGB) and belowground biomass (BGB) are key components of carbon storage, yet their responses to future climate changes remain poorly understood, particularly in China. Understanding these dynamics is essential for global carbon cycle modeling and ecosystem management. This study integrates field observations, machine learning, and multi-source remote sensing data to reconstruct the distributions of AGB and BGB in China from 2000 to 2020. Then CMIP6 was used to predict the distribution of China under three SSP scenarios (SSP1-1.9, SSP2-4.5, SSP5-8.5) from 2020 to 2100 to fill the existing knowledge gap. The predictive accuracy for AGB (R² = 0.85) was significantly higher than for BGB (R² = 0.48), likely due to the greater complexity of modeling belowground dynamics. NDVI (Normalized Difference Vegetation Index) and soil organic carbon density (SOC) were identified as the primary drivers of AGB and BGB changes. During 2000–2020, AGB in China remained stable at approximately 10.69 Pg C, while BGB was around 5.06 Pg C. Forest ecosystems contributed 88.52% of AGB and 43.83% of BGB. AGB showed a relatively slow annual increase, while BGB demonstrated a significant annual growth rate of approximately 37 Tg C yr⁻¹. Under the low-emission scenario, both AGB and BGB show fluctuations and steady growth, particularly in South China and the northwestern part of Northeast China. Under the moderate-emission scenario, AGB and BGB show significant declines and increases, respectively. In the high-emission scenario, both AGB and BGB decline significantly, particularly in the southwestern and central regions. These results provide valuable insights into ecosystem carbon dynamics under climate change, emphasizing the relatively low responsiveness of AGB and BGB to climatic variability, and offering guidance for sustainable land use and management strategies.
As our understanding of ecology deepens and modeling techniques advance, species distribution models have grown increasingly sophisticated, enhancing both their fitting and predictive capabilities. However, the dependability of predictive accuracy remains a critical issue, as the precision of these predictions largely hinges on the quality of the base data. We developed models using both field survey and remote sensing data from 2016 to 2020 to evaluate the impact of different data sources on the accuracy of predictions for Scomber japonicus distributions. Our research findings indicate that the variability of water temperature and salinity data from field suvery is significantly greater than that from remote sensing data. Within the same season, we found that the relationship between the abundance of S. japonicus and environmental factors varied significantly depending on the data source. Models using field survey data were able to more accurately reflect the complex relationships between resource distribution and environmental factors. Additionally, in terms of model predictive performance, models based on field survey data demonstrated greater accuracy in predicting the abundance of S. japonicus compared to those based on remote sensing data, allowing for more accurate mastery of their spatial distribution characteristics. This study highlights the significant impact of data sources on the accuracy of species distribution models and offers valuable insights for fisheries resources management.
This chapter explores the innovative role of artificial intelligence (AI) in the conservation of terrestrial ecosystems, aligned with the United Nations Sustainable Development Goal 15. Through a detailed review of current AI technologies, key applications such as real-time monitoring, deforestation detection using satellite images, and predictive models for managing environmental threats such as fires and deforestation are discussed. It also examines the challenges inherent in the use of AI, including data accuracy and algorithmic biases, and how these can be mitigated through a multidisciplinary approach that fosters a sustainable coexistence between technology and nature. This study highlights the importance of cross-sector collaboration and ethical regulation in the development of AI technologies that not only respect ecological balances, but also promote the effective and sustainable conservation of terrestrial ecosystems.
Species distribution modelling (SDM) is a method that assists in linking the biotic and abiotic features of the ecosystems. In this study, we developed SDMs for total fish abundance, abundance of Escualosa thoracata (ET), Mugil cephalus (MC), Sardinella longiceps (SL) and Fenneropenaeus merguiensis (FM) based on biotic and abiotic variables using multivariate statistical techniques such as Support Vector Machine (SVM), Regression tree (RT), K-nearest neighbours (kNN), Gaussian process regression (GPR), Random Forest (RF), eXtreme Gradient Boosting (XGB), Cubist, and Quantile random forest (QRF). A combination of abiotic and biotic variables such as salinity, temperature, mud, turbidity, zooplankton, shell debris, sand, and benthos were identified as the significant predictors for the distribution of total fish abundance and species wise fish abundance. The model performance was evaluated based on coefficient of determination (R2), d-index, mean bias error (MBE), Nash–Sutcliffe efficiency (NSE) and root mean squared error (RMSE). A unique average ranking procedure was followed using all the model evaluation indicators to select a best model for a given season or species. The top performing models did not vary much between the seasons and species. We identified the following models performed best for different species respectively during postmonsoon and premonsoon seasons- Cubist and QRF for E. thoracata, RF and GLMNET for M. cephalus, SVM and QRF for S. longiceps, GPR and QRF for F. merguiensis. The major variables influenced the species distribution were benthic substratum (sand, mud, shell debris), chlorophyll-a and depth. Thus, it is evident that though the other variables also influence the distribution and abundance of the species in an estuarine bay, the benthic features, depth and morphometry can have significant effect on the species distribution. Moreover, the advanced multivariate statistical tools can aid in the prediction of the species distribution in the tropical estuaries.
سابقه و هدف: آگاهی از پراکنش مکانی گیاهان به عنوان یکی از مهمترین جنبههای اکولوژی گیاهی، نقش برجستهای در تشریح پایداری اکوسیستم مرتع و طراحی برنامههای مناسب مدیریتی دارد. طیف وسیعی از عوامل محیطی از ویژگیهای جوی گرفته تا خصوصیات سطح زمین، زیربنای دامنه پراکنش گونهها در مقیاسهای محلی تا جهانی طی دو دهه گذشته میباشند. از میان این عوامل، متغیرهای زیست اقلیمی به عنوان یک نوع داده ضروری در مدلسازی پراکنش گونهها در نظر گرفته میشوند. این مطالعه جهت مدلسازی آشیان اکولوژیک بالقوه گونه علف بره (Festuca ovina L.) در مراتع استان اصفهان با استفاده از دو روش مدلسازی الگوریتم ژنتیک (GARP) و قلمرو (DOMAIN) انجام شد.
مواد و روشها: دادههای رخداد گونه مورد مطالعه به روش تصادفی- طبقهبندی شده جمعآوری شدند و 22 لایه محیطی شامل سه متغیر فیزیوگرافی و 19 متغیر اقلیمی مشتق شده از بارندگی و درجه حرارت در فرآیند مدلسازی بهکار رفتند. این متغیرها در محیط نرمافزار ArcMap 10.3 به اندازه پیکسل یک کیلومتر مربعی تبدیل و سپس به صورت نقشههای رقومی وارد فرآیند مدلسازی شدند. مدل الگوریتم ژنتیک در نرمافزار Open Modeler 1.1.0 و مدل قلمرو در نرمافزار DOMAIN 32 اجرا شد. به منظور انتخاب متغیرهای محیطی با اجرای آزمون همبستگی پیرسون، از بین جفت متغیرهای با همبستگی بیش از 0/8، یک متغیر انتخاب و دیگری حذف گردید. در مدل قلمرو، پیشبینی تناسب رویشگاه بر اساس شاخص عدم تشابه یا فاصله اکولوژیک گاور (Gower Metric) انجام گرفت. به منظور اعتبارسنجی و ارزیابی نقشههای رویشگاه بالقوه حاصل از دو مدل، از ماتریس خطا و شاخصهای آماری مشتق شده از آن همچون AUC (سطح زیر منحنی پلات ROC)، ضریب کاپا (Kappa) و صحت کلی (CCR) استفاده شد. در تحلیل حساسیت، آن دسته از متغیرهای محیطی که بیشترین افت را در عملکرد مدل (AUC) داشتند، به عنوان با اهمیتترین متغیرهای محیطی انتخاب شدند.
نتایج: تحلیل حساسیت نشان داد که به ترتیب متغیرهای ارتفاع از سطح دریا، بارندگی گرمترین فصل، میانگین دمای روزانه، ایزوترمالیتی، بارندگی سالانه، دمای متوسط سالانه، شیب، و دامنه تغییرات سالانه دما مؤثرترین متغیرهای محیطی در پراکنش گونه بودند. الگوریتم ژنتیک (0/99=AUC) نسبت به مدل قلمرو (0/67=AUC) از عملکرد مناسبتری برخوردار بود. در بررسی گستره جغرافیایی گونه تحت مدلها پیشبینی شد که رویشگاههای مطلوب (احتمال رخداد بالاتر از 75 درصد) در مدل الگوریتم ژنتیک و مدل قلمرو به ترتیب حدود 7069 و 10187 کیلومترمربع از سطح استان را به خود اختصاص دادهاند. نتایج مدلهای الگوریتم ژنتیک و قلمرو همچنین نشان داد که رویشگاه عمده گونه علف بره در شهرستانهای فریدونشهر و سمیرم میباشد.
نتیجهگیری: مناطقی از شهرستانهای فریدونشهر و سمیرم که مراتع ییلاقی استان را شامل میشوند به همراه ارتفاعات کرکس و موته که جزو مراتع استپی سرد استان میباشند، پتانسیل بالایی جهت حضور گونه F.ovina دارند. تأثیر تغییر اقلیم به همراه چرای بیرویه از این گونه که جزو گونههای خوشخوراک میباشد، میتواند بتدریج باعث کاهش و حذف این گونه و جایگزینی آن با سایر گونهها شود. از یافتههای این مطالعه میتوان جهت تعیین مناطق مستعد پروژههای بذرکاری و کپهکاری با استفاده از این گونه استفاده نمود. پیشنهاد میشود که فعالیتهای زیستی و مدیریتی همچون تغییر کاربری، شدت چرای دام، و رقابت بینگونهای در فرآیند مدلسازی مورد استفاده قرار گیرند تا باعث بهبود پیشبینی زیستگاههای مطلوب گونه گردد.
واژههای کلیدی: رخداد گونه ای، مدل پروفیل، شاخص عدم تشابه Gower، گراسهای فصل سرد
Data science and artificial intelligence are revolutionizing habitat management and biodiversity conservation by providing creative answers for a sustainable future. This study examines the ways in which AI technologies—such as machine learning, predictive modeling, and real-time data processing—are improving our capacity to observe, forecast, and safeguard ecosystems. AI-driven technologies make it possible to thoroughly analyze vast and varied datasets, including species records and climatic data, which improves conservation tactics and facilitates better decision-making. AI aids in the creation of efficient management strategies and early environmental danger detection systems by combining several data sources and using sophisticated analytical techniques. Along with addressing important issues like data quality and ethical implications, the report also outlines potential future paths for using AI in conservation. The ongoing developments in data science and AI have great potential to solve new environmental issues and protect biodiversity for coming generations.
Knowledge of regional-scale patterns of ecological community structure, and of factors that control them, is largely conceptual. Regional- and local-scale factors associated with regional variation in community composition have not been quantified. We analyzed data on woody plant species abundance from 2443 field plots across natural and seminatural forests and woodlands of Oregon to identify and quantify environmental, biotic, and disturbance factors associated with regional gradients of woody species composition; to examine how these factors change with scale (geographic extent) and location; and to characterize and map geographic patterns of species and environmental gradients. Environmental correlates of species gradients, species diversity patterns, and the spatial patterning of woody plant communities varied with geographic extent and location. Total variation explained (TVE) by canonical correspondence analyses (CCAs) was 9-15% at three hierarchical geographic extents: the entire state, two half-states, and five subregions. Our high level of unexplained species variation is typical of vegetation gradient analyses, which has been attributed to landscape effects, stochastic processes, and unpredictable historical events. In addition, we found that TVE in canonical correspondence analysis is confounded by sample size. Large numbers of plots and species, as in our study, are associated with lower TVEs, and we propose a mechanism for this phenomenon. Climate contributed most to TVE (46-60%) at all locations and extents, followed by geology (11-19%), disturbance (6-12%), and topography (4-8%). Seasonal variability and extremes in climate were more important in explaining species gradients than were mean annual climatic conditions. In addition, species gradients were more strongly associated with climatic conditions during the growing season than in winter. The dominant gradient at the state scale was from the lower elevation, moderate, maritime climate along the coast to the higher elevation, drier, continental climate of eastern Oregon. The second canonical axis followed a gradient from the warm, dry, growing seasons of the western interior valleys and eastern Cascade Range to the cooler, wetter mountainous areas. Geologic variables were most strongly correlated with axis 3, and measures of local site and disturbance with axis 4. For most of the state, our findings on the associations of disturbance factors with species gradients were inconclusive due to confounding of land ownership patterns, disturbance histories, and elevation in our sample. Near the coast, where gradients were not confounded, clear-cutting and stand age accounted for only 2 and 1% of TVE, respectively, in partial CCA. Ordinations of our long, regional gradients were influenced more by species presence than by abundance, and few woody species have been totally eliminated from sites by clear-cutting. Within Oregon and for the range of geographic extents we examined, variation in the environmental correlates of species gradients was more strongly associated with geographic location than with geographic extent, although topographic factors explained slightly more variation at smaller geographic extents. The greatest subregional contrast in vegetation character was between eastern and northwestern Oregon, and the Klamath subregion was intermediate. In the drier climate of eastern Oregon, community structure varied at a finer spatial scale, and climatic and topographic moisture were more strongly associated with species gradients than in the moister areas of western Oregon. Topographic effects were weakest and climatic effects strongest near the coast, where climate is moderate. Alpha and gamma diversity were greater in western Oregon, but beta diversity was greater in eastern Oregon and greater for shrubs than for trees. Our findings supported a conceptual model of multiscaled controls on vegetation distribution, and the related notion that local community structure is the result of both regional and local-scale processes. Despite strong ecological contrasts within the region, we were able to synthesize species-environment relations at the regional level. This suggests that apparent conflicts among local vegetation studies can be explained by real ecological differences among places.
This rigorous yet accessible text introduces the key physical and biochemical processes involved in plant interactions with the aerial environment. It is designed to make the more numerical aspects of the subject accessible to plant and environmental science students, and will also provide a valuable reference source to practitioners and researchers in the field. The third edition of this widely recognised text has been completely revised and updated to take account of key developments in the field. Approximately half of the references are new to this edition and relevant online resources are also incorporated for the first time. The recent proliferation of molecular and genetic research on plants is related to whole plant responses, showing how these new approaches can advance our understanding of the biophysical interactions between plants and the atmosphere. Remote sensing technologies and their applications in the study of plant function are also covered in greater detail.
The paper reports the results of the implementation of a range of spatial context scales into spatial models using neural networks. The results are calibrated against those produced on the same dataset using Gaussian maximum likelihood and decision trees. The raster dataset used for modelling is complex, containing hilly terrain and highly disturbed forest. The result suggest that the incorporation of spatial context does improve mapping accuracy. But, the scale of the spatial context must be matched to the theme and the internal complexity of the model. Different spatial context scales do produce better thematic mapping accuracies for each land cover class. This relationship is a function of the spatial homogeneity - and complexity of each land cover class. -Authors
The methodology used to construct tree structured rules is the focus of this monograph. Unlike many other statistical procedures, which moved from pencil and paper to calculators, this text's use of trees was unthinkable before computers. Both the practical and theoretical sides have been developed in the authors' study of tree methods. Classification and Regression Trees reflects these two sides, covering the use of trees as a data analysis method, and in a more mathematical framework, proving some of their fundamental properties.
Multivariable regression models are powerful tools that are used frequently in studies of clinical outcomes. These models can use a mixture of categorical and continuous variables and can handle partially observed (censored) responses. However, uncritical application of modelling techniques can result in models that poorly fit the dataset at hand, or, even more likely, inaccurately predict outcomes on new subjects. One must know how to measure qualities of a model's fit in order to avoid poorly fitted or overfitted models. Measurement of predictive accuracy can be difficult for survival time data in the presence of censoring. We discuss an easily interpretable index of predictive discrimination as well as methods for assessing calibration of predicted survival probabilities. Both types of predictive accuracy should be unbiasedly validated using bootstrapping or cross-validation, before using predictions in a new data series. We discuss some of the hazards of poorly fitted and overfitted regression models and present one modelling strategy that avoids many of the problems discussed. The methods described are applicable to all regression models, but are particularly needed for binary, ordinal, and time-to-event outcomes. Methods are illustrated with a survival analysis in prostate cancer using Cox regression.
The gopher tortoise (Gopherus polyphemus) is a rare species on the Ocala National Forest. It is listed by the state of Florida as a species of special concern and is a federally listed threatened species outside of its Florida range. The Ocala National Forest comprises a significant portion of the remaining undeveloped gopher tortoise habitat. The geographic information system has the potential for predictive modeling of gopher tortoise habitat distribution and gopher tortoise populations. A simple model is proposed to demonstrate the potential of GIS for wildlife management applications. -from Authors
Die Problematik einer vom Menschen mitverursachten Klimaänderung beschäftigt nicht nur die Klimatologie selbst, sondern auch die Natur- und Landschaftsschutzpraxis (Dobson et al., 1989; Halpin, 1994; Peters, 1990; Peters und Darling, 1985; Peters und Lovejoy, 1992). Dabei geht es in erster Linie darum, ob die Artenvielfalt bei möglichen grossräumigen Vegetations-und Landnutzungsveränderungen langfristig erhalten bleibt oder ob mit drastischen Verlusten der Artenzahl gerechnet werden muss. Weiter ist für die Praxis von grossem Interesse, ob die Schutzziele der zwangsläufig ortsgebundenen Natur- und Landschaftsschutzgebiete aufrechterhalten werden können oder ob sie an die veränderten Klimabedingungen angepasst werden müssen. Solche Risikoabschätzungen lassen sich — zumindest was die Vegetation anbelangt — mit Hilfe von computergestützten Modellen durchführen. Risikoabschätzungen sind keine Prognosen, sondern dienen dazu, die ökologischen Auswirkungen von Umweltveränderungen nach bestem Wissen abzuschätzen (Graham et al., 1991).
Discusses data compilation and implementation of an automated Geographic Information System (GIS) for use in identifying and analyzing remaining potential habitat for Gymnogyps californianus within California. Essential condor GIS data layers, including a land use and land cover classification system for use with multispectral satellite data, are identified and grouped hierarchically in a 4 part system. -from Authors
The abstract for this document is available on CSA Illumina.To view the Abstract, click the Abstract button above the document title.
Do quasi-physically based models with more detail perform better than regression or other empirical models? This is a question that was raised many years ago and still remains. In an effort to respond to this question, the author reviews the needs and concerns of users and then divides the large number of models into three classes: (1) screening, (2) research, and (3) planning, monitoring, and assessment. Empirical and causal (physically based) models are contrasted and the advantages and disadvantages of each described. Sources of model uncertainty (properties of data bases, model structure, parameter estimation methods, algorithmic implementation, verification and validation, and future users) that lead to skepticism about models' performance are investigated. Simulation scale and spatial variability are also important considerations. The leading question is then discussed from the perspective of screening, research, and planning, monitoring, and regulatory models.
Within a survey of mountain forest types in the Swiss Alps, we compared three statistical sampling methods, ie stratified random sampling of the investigation area, systematic sampling and unrestricted random sampling. The stratification was based on a digital terrain model and performed using GIS technology. A comparison of the three sampling methods shows that stratified random sampling is an efficient method that guarantees a sample which is proportional to the extension of the environmental types. We investigated the resemblance pattern of releves and found that it differs between purely random and all others. This leads to the conclusion that stratification introduces a bias into the true pattern. A comparison of the efficiency of the three tested sampling strategies shows that for a classification as used at the regional level (10-50km2), stratified random sampling draws an accurate picture of the small scale vegetation pattern at low sampling effort. At lower resolution, unrestricted random sampling should be applied in the present case, as it leads to a more balanced sampling of large scale vegetation patterns. Compared to manual stratification, GIS-based stratification generally leaqds to a considerably higher number of strata. We found that releves from rare attribute combinations neither tend to become outliers in the final data set nor considerably contribute to rare vegetation types. Stratification using GIS technology therefore can be recommended as an efficient technical help even if the number of resulting strata is large and rare strata have to be omitted.
The geographical limits of Nothofagus cunninghamii are highly correlated with climate and appear to be more or less in equilibrium with the climate of the present century in all but one of the areas of its present range. Suitable climates for the species occur in the highlands of NE Victoria and S New South Wales, beyond its present range, and it is possible that it occurred within the predicted area prior to the last ice age. Populations along the NE edge of its present range in the Central Highlands of Victoria may be migrating NE along a narrow corridor of apparently suitable climate to re-occupy the postulated former range. Rate of migration would be extremely slow because of the poor dispersal ability of the species and the adverse impact of recurrent fires.-from Author
Recent research has shown an artificial neutral network (ANN) to be capable of pattern recognition and the classification of image data. This paper examines the potential for the application of neural network computing to satellite image processing. A second objective is to provide a preliminary comparison of training site data inputs and generalized land-cover classification results for conventional supervised classification and ANN classification. An artificial neutral network can be trained to do land-cover classification of satellite imagery using selected sites representative of each class in a manner similar to conventional supervised classification. One of the major problems associated with recognition and classification of patterns from remotely sensed data is the time and cost of developing a set of training sites. This research compares the use of an ANN back propagation classification procedure with a conventional supervised maximum likelihood classification procedure using a minimal training set. When using a minimal training set, the neural network is able to provide a land-cover classification superior to the classification derived from the conventional classification procedure. This research is the foundation for developing application parameters for further prototyping of software and hardware implementations for artificial neural networks in satellite image and geographic information processing.
The objective of this research was to develop a descriptive GIS model to identify potential nesting habitat of greater sandhill cranes in northwestern Minnesota. The modeling approach involved five fundamental steps: generating data layers, describing nest sites, testing for discrepancies between observed and expected distributions of nest sites, generating the model, and assessing the model. Results indicated that some crane pairs nested in sub-optimal and marginal areas despite the apparent availability of optimal habitat. The absence of nesting pairs in optimal habitat may be accounted for by conditions and assumptions inherent in the data and modeling approach, unanswered questions concerning the behavior of nesting cranes, the uncertainty that all nest sites in the study area were known, and the inability to model or detect certain pertinent landscape features and local variables. -Authors
Reproduces and describes pollen diagrams and maps showing the migration of species. Spruce developed over quite large areas about 12 000 BP, no doubt as a result of climatic amelioration. 5000 BP vegetation throughout the south changed from oak forest to forest dominated by pine, and in the north vegetation changes suggested the peak of Holocene climatic warmth had passed. A useful stratigraphic marker is the hemlock decline at 4800 yr BP. -K.Clayton
Rule-based methods are a useful alternative framework for scientific investigation. The incorporation of intelligent functions in a geographic information system provides a means to generate new information. Rules to classify ridges and channels in a triangular irregular network model to estimate slope position demonstrate the usefulness of this framework. A study, using these techniques in Arc/Info, was undertaken to predict forest cover type from topographic position. We can now determine, for example, whether species composition can be predicted from elevation and topographic position alone, or whether added geologic, pedologic, or hydrologic data are necessary in species composition prediction. -Authors
Landsat Thematic Mapper digital data were classified into seven native eucalypt forest type classes using a nonparametric classifier that also calculated the probability of correct classification for each pixel. A digital elevation model, spaced on a 30-m grid, was generated and used to derive terrain features of gradient, aspect, and topographic position, which were geometrically co-registered with the TM thematic images. The thematic maps of forest type, probability of correct classification, and terrain features provided data for the expert system to infer the most likely forest species occurring at any given pixel. The modified thematic map output by the expert system had a higher mapping accuracy than the thematic map produced by the supervised nonparametric, the maximum likelihood, and the Euclidean distance classifier.
Wetland habitats suitable for foraging by the Wood Stork (Mycteria americana) were inventoried and analyzed using remotely sensed imagery, digital image processing, and geographic information system (GIS) techniques. Maps of foraging habitats were created from Landsat Thematic Mapper imagery, one for a 'wet' year and one for a 'dry' year. Change detection, proximity to the Wood Stork Colony, and size of foraging site analyses were performed on the maps using GIS algorithms to obtain quantitative foraging habitat statistics. Results of the analyses indicate a 47 percent reduction in foraging cover during the 'dry' year. The largest concentration of foraging cover was within about one kilometer of the Colony nesting site during both years.
Historic Ursus arctos horribilis sightings in the North Cascades area of Washington were analyzed using the GRASS geographic information system software. A 22-class land-cover database (determined to be 85% accurate) was compared to 91 historic sightings of grizzly bears. The historic sightings were positively associated with the whitebark pine - subalpine larch (Pinus albicaulis - Larix lyallii) and subalpine herb cover types. The sighting locations had similar land-cover richness but land-cover interspersion different from the overall landscape within the study area. Data were used to develop new map layers for relative cover type and diversity selection which were then combined into a map of sighting potential for grizzly bears in the N Cascades. -Authors
The method of trend surface analysis was originally introduced into the earth sciences by Oldham and Sutherland (1955), Miller (1956), Krumbein (1956, 1959), Grant (1957) and Whitten (1959). These authors used the method for the analysis of gravity maps, stratigraphic maps (facies maps), isopach maps and maps representing specific attributes of sedimentary and igneous rocks.
1. Empirical models for spatial distribution of wildlife, given data from a complete census or a random sample of sites, are reviewed briefly. 2. The use of covariates, recorded at different resolutions, for modelling spatial distribution is explored. 3. Presentation of model predictions in map form is discussed. 4. A framework of models for change in spatial distribution, given data from successive surveys, is developed. 5. Methods for quantifying and presenting precision and bias are described. 6. The methods are illustrated for two bird species (green woodpecker Picus viridis and redstart Phoenicurus phoenicurus) and for red deer Cervus elaphus, using data from north-east Scotland.
