https://orcid.org/0000-0003-1459-2046
Figures
Abstract
The decline of overwintering populations of monarch butterflies (Danaus plexippus) in central Mexico has raised concern worldwide. Recent evidence suggests that the fall migration is becoming disrupted by several factors, leading to a decreasing number of monarchs completing their journey to their overwintering sites in central Mexico. One potential factor assumes that there are suitable habitats for species of tropical milkweed as well as for eggs and larvae of monarchs (ELM) laid by adult individuals. We evaluated the current and future habitat suitability of ELM and species of milkweeds in Mexico using ecological niche modeling. We built a three-stage protocol for habitat suitability of ELM, including climatic, biological, and environmental suitability modeling, under two general circulation models (CanESM5 and MPI-ESM1–2-HR), three-time horizons (2030, 2050, and 2070), and the “Middle of the road” socioeconomic pathway (SSP 2 RCP 4.5). We found a consistent southward shift in suitable areas, with climate emerging as the most limiting factor. Climate change projections indicate contrasting potential reductions of 40% and 8% in suitable habitat by 2070, under the CanESM5 and MPI-ESM1–2-HR, respectively. While environmental suitability remained relatively stable—mainly due to the persistence and increase of rainfed agriculture—biological suitability, determined by the distribution of 46 perennial species of milkweed (Asclepias), also shifted southward. This trend suggests that monarchs may increasingly establish reproductive resident populations in northeastern to central Mexico rather than reaching their overwintering sites. Our findings underscore the need to expand similar multifactorial modeling efforts across the U.S. and Canada to have a transboundary collaboration that integrates conservation strategies for monarch migration, accounting for potential shifts in climatic, biological, and environmental suitability under climate change scenarios.
Citation: Sánchez-Cordero V, Castañeda S, Mendoza-Ponce A, Juárez-Jaimes BV, Botello F, Ureta C (2026) Regional risk shifts to monarch butterfly migration due to climate change. PLOS Clim 5(2): e0000802. https://doi.org/10.1371/journal.pclm.0000802
Editor: Jennifer Lee Wilkening, US Fish and Wildlife Service, UNITED STATES OF AMERICA
Received: September 10, 2025; Accepted: December 28, 2025; Published: February 25, 2026
Copyright: © 2026 Sánchez-Cordero et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data is available from Mendeley Data repository https://doi.org/10.17632/dyc26myxzg.1.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The monarch butterfly (Danaus plexippus) migration phenomenon includes millions of individuals travelling thousands of kilometers from their breeding areas in Canada and eastern and central United States to their overwintering sites in central Mexico [1]. In spring, the monarch migration moves northward, involving several generations of monarchs that are established in a wide range of breeding areas in North America. In fall, the monarch migration moves southward, including a single generation of monarchs flying thousands of kilometers to their overwintering sites in central Mexico [2–5]. Most of the overwintering sites of monarchs in central Mexico are dominated by oyamel fir forests at the core zones of the Monarch Butterfly Biosphere Reserve (MBBR) and other nearby sites [5] (Fig A in S1 Text). For over a decade now, a drastic decrease in overwintering populations in the core areas of the MBBR has been observed threatening the conservation of the monarch butterfly migratory phenomenon [5–8]. This decreasing trend in overwintering populations has produced great concern worldwide, and among social organizations, academia, governmental and NGOs agencies in Mexico, USA, and Canada [9]. Overall, it appears to be a consensus that the overwintering population declines are due to a multi-factorial component negatively affecting [10] monarch migration at different geographical scales and life-cycle stages [7,11,12]. Several factors have been examined to explain the decline in overwintering monarch populations, ranging from parasitism affecting their lifecycle, reduction of milkweed populations due to an overuse of pesticides in breeding areas of monarchs, habitat loss in core areas of the MBBR, and land use/cover and climate changes in wide geographical areas throughout the migration route [11,13–16]. These factors act at different scales along the geographical range of monarch migration challenging the establishment of an effective integrated strategy for management and conservation [9,17].
The monarch fall migration along the region of Texas, and the northeast and central Mexico has raised interest as it has been related to the declines in overwintering populations. For example, during the fall migration, monarchs appear to suffer from limited food supply [14,18,19]. The decrease in availability of food resources negatively impacts monarch migrants in providing enough energy to complete their journey to the overwintering site in central Mexico [14,18–20]. This explanation has been related to land use/cover changes (LUCC) in this region disrupting populations of the species of plants that monarchs regularly visit to obtain nectar [3,20]. Nonetheless, a new study found low differences in LUCC along Texas and eastern and central Mexico in the last 30 years [21]. Further, while species of milkweeds (Asclepias) are essential for oviposition and larval feeding, other plant groups also play important roles as nectar sources for adult monarchs during migration, including species from the Compositae (Asteraceae) and Lamiaceae plant families [22,23]. Further, perennial species of milkweed (e.g., not showing seasonal senesce) are also relevant to increasing the likelihood of establishing non-migratory monarch populations [24–26].
A recent study has also demonstrated that roost sites have drastically declined during the fall migration in the last decades suggesting a disruption of monarchs failing to reach the overwintering sites in central Mexico [27]. It has also been observed an increasing proportion of migrant monarchs that appear to establish local breeding areas in northeast and central Mexico year-round. These observations have been partially related to an increase in several small-scale plantations of several perennial native and nonnative tropical species of milkweed as Asclepias curassavica and Calotropis gigantea by local inhabitants as a naive response to provide migrant monarchs with enough hostplants and nectar supply. An additional complication is that there is also increasing evidence that eggs and larvae growing on tropical species of milkweeds can get infected with a protozoan parasite Ophryocystis elektroscirrha that disrupt their life cycle, resulting in lower fecundity of adult monarchs and decreasing their flight efficiency [3,28].
This scenario of small-scale plantations of species of perennial tropical milkweed has led to recommendations from governmental agencies, academia, and NGOs to encourage local inhabitants to stop growing such plant species in their backyards or in their agricultural fields, as this might facilitate migrant monarchs establishing local breeding areas rather than completing their fall migration [28]. The fact that local inhabitants are growing small-scale populations of milkweed assumes that there are suitable habitat conditions for both perennial species of milkweed and eggs and larvae of monarchs to establish. Despite governmental recommendations that can change the attitudes of local inhabitants to stop growing, there is the possibility that perennial species of milkweed habitat suitability under current and climate change scenarios can increase in Mexico favoring the establishment of local year-round monarch populations and potentially disrupting the fall migration process. This is particularly relevant as Mexico holds a high number of perennial species of milkweeds [29].
Previous studies have addressed the geographic projections and interactions of milkweeds and monarchs in different regions using ecological niche models projected as species potential distributions [2,4,13,22]. However, few studies have addressed the potential impact of climate change on the monarch migration phenomenon using geographic projections based on ecological niche modeling of monarchs and species of plants relevant to their life cycle [13]. Here, we addressed the impact of potential geographical shifts in habitat suitability of the eggs and larvae of monarchs (ELM) and in the interactions between perennial species of milkweed—all species of PTM (Asclepias) that occur along the monarch migratory route in Mexico (Fig A in S1 Text)—and ELM under climate change and LUCC scenarios in Mexico. We addressed the apparent propensity of monarchs to establish in areas with species of PTM during their fall migration, interrupting their journey to overwintering sites in central Mexico [27]. Specifically, we build a three-stage protocol to evaluate the climatic, biological and environmental suitability of ELM under climate change scenarios in Mexico, using new climatic (https://pcmdi.llnl.gov/CMIP6/) and LUCC information (adjusted new layers produced for this study; see below), as follows: For the ELM, we projected the (1) climatic suitability, using bioclimatic variables; (2) biological suitability, using a weighted ensemble of 46 species of PTM. We assumed that these species of PTM provide biological suitability for ELM; all species of PTM included in the analyses were perennial, e.g., individual plants can be found year-round [30], and (3) environmental suitability using land use cover changes under climate change scenarios. The transformation of natural landscapes into urban or agricultural areas can further limit available habitats for ELM, compounding the effects of climate change [29]. The projected areas for climatic, biological, and environmental suitability for ELM were overlapped to identify potential distribution shifts of ELM highest habitat suitability under climate change and land use cover changes scenarios. We also discuss the potential impact on the conservation of monarch migration, including management strategies and policies needed for preventing present and future regional disruptions of this phenomenon.
Materials and methods
Distribution models
We compiled georeferenced records of ELM in Mexico from the Red Monarca (https://redmonarca.org/) that has been involved with monarch butterfly conservation for decades, and from Journey North’s dataset from the monarch butterfly and milkweed observations by volunteers across North America (https://journeynorth.org/) [31]. We decided to include ELM records given (1) their close dependence on species of milkweed during these life cycle stages of monarchs, and (2) their highly restricted dispersal capabilities. To model the biological suitability of ELM, a careful review of species and records of North American of Asclepias was conducted by an expert in their taxonomy (V. Juárez, a coauthor) and complemented with GBIF records (www.gbif.org; last accessed February 2025) for Central and South America. All incorrect taxonomic identification records were excluded from the analysis. Records were also excluded when occurring outside known distributions of species of Asclepias, following established protocols [32]. For the ELM distribution models, we used the correlogram to avoid spatial autocorrelation. For species of Asclepias distribution models, we used the method proposed by Assis [33] and kept the variable bio 15 (precipitation seasonality), which is one of the most important climatic variables (based on the expertise of V. Juárez). The method by Assis [33] involves analyzing the correlation of each predictor variable across different geographic distances by identifying the minimum distance at which autocorrelation becomes non-significant. The occurrence records were then spatially thinned using the method by Aiello-Lammens et al. [34]. Records with a minimum nearest neighbor distance less than the average of the minimum non-significant distances found were removed. A total of 46 species of Asclepias with a minimum of 20 occurrence records were included in the analysis. Species of Asclepias with fewer than 20 occurrence records underwent an alternative spatial thinning process. During this procedure, occurrence records with a nearest-neighbor distance ranging from 1 to 20 km were iteratively removed in 1 km increments. The group with the greatest distance that still had at least 20 occurrence records was selected. We combined both groups of species to produce a final occurrence record dataset used for model training (Fig A in S1 Text, Table A in S2 Text).
For ELM and species of Asclepias, we tested 33 different models using three different algorithms, five sets of background-pseudoabsences, and four partitioning strategies (Table 1). Most of our results exhibited a robust performance in terms of validation metrics (TSS, ROC, and BOYCE index). We filtered the best experiment by using the TSS mean value and by applying an expert opinion (with ten randomly chosen milkweed species). The chosen experiment corresponded to the random forest algorithm with 10,000 pseudoabsences selected randomly outside a 2° buffer from presence points, a bootstrap partitioning method with 70% for the training set and 30% for the test set and with 5 replicates (Table 1). Finally, to identify geographic regions where the modeling of each species was likely extrapolated given new climatic variable ranges or combinations, we used the ExDet tool [36]. For each species of Asclepias and for the ELM distribution models, a calibration area was delimited (M) [37] by selecting polygons from the RESOLVE Ecoregions dataset from North, Central and South America [38] that included their occurrence records. We did not only focus on the records from the migratory route given that the ecological niche algorithms perform better when having greater climatic (or environmental information) information to calibrate the model. The 19 reference bioclimatic variables were then spatially constrained by intersecting them with these selected polygons. The variable bio 6 (min temperature of coldest month) was considered relevant for ELM of monarchs based on expert opinion and was retained in all sets. Similarly, we retained bio 15 (precipitation seasonality) for species of Asclepias. To address multicollinearity, we used a two-stage approach [39]. First, variables highly correlated with bio15 (r > 0.7) were identified using the Pearson correlation coefficient and were excluded from the initial group of 19 variables along with bio15. From the reduced group of variables, a second selection was conducted to eliminate variables that still showed a strong linear relationship with each other, assessed through the variance inflation factor (VIF) using the vifcor function from the usdm R package [40]. The remaining variables, along with bio 6 for ELM and bio 15 for species of Asclepias constituted the final predictor set for each calibration area (M) depending on the species being modeled (S3 Text).
Climatic suitability
Once the distribution models for ELM were produced, we obtained the climatic suitability distribution models for the present, and under three future time projections. Baseline and climate change data from the WorldClim 2.1 dataset [41] were obtained at a spatial resolution of 30 arc seconds (~1km). Two different general circulation models (GCM), CanESM5 and MPI-ESM1–2-HR (CMIP 6) were used for our modeling exercises. These two GCM have shown a reliable performance in simulating the climate of the Northern Hemisphere [42], and with higher independence—based on GCMs-codes genealogy models—from the GCM that best simulated climate of the Northern Hemisphere [43,44]. Although the evaluations by the latter studies were conducted with CMIP 5, there is a strong dependency between GCM generations [43]. Kuma et al. [44] worked with CMIP 6 and evaluated GCM. We used these three different references to select the most appropriate GCMs.
Three future time horizons were included: short-term (2030), mid-term (2050) and long term (2070), under the Shared Socioeconomic Pathway (SSP) that represents the average tendency: SSP 2 4.5. We did not use another scenario for comparison as we only had one land use cover change layer (under SSP 2 4.5) that was specifically produced for this study (S3 Text; download in a directory that will be available when potential acceptance of the paper). The land-use cover change layer for the “Middle of the road” scenario allowed the combination with optimistic and pessimistic climate change layers. This approach allowed us to identify the exacerbated effects of climate change on ELM. All Models were projected to Mexico’s geographic extent and converted to binary (presence-absence) maps using the threshold that maximizes the True Skill Statistic (TSS) [45] (S3 Text).
Biological suitability
To obtain the biological suitability distribution models, we built a weighted species-richness maps across all scenarios and time horizons. We grouped species of Asclepias into three classes (high, medium, low) based on their relevance to ELM. To have a first approximation of the importance of the species of Asclepias included to ELM, we conducted a literature review (examining “grey” literature) to count the frequency that a species of Asclepias was associated with monarchs. In addition, the expert knowledge and field observations were also considered (see Table A in S2 Text). We then combined all the presence-absence potential distribution maps of species of Asclepias (using climatic suitability) and generated a weighted sum. Higher values indicated higher biological suitability in regions that (1) contained more species of Asclepias, (2) included species of Asclepias preferred by ELM or (3) included a broad set of species of Asclepias and of these species several are preferred by ELM (see S1). The high-weight areas were considered regions with the highest biological suitability. To identify areas abundant in resources for egg-laying and larval feeding, we categorized the weighted species richness layers into three non-overlapping classes using quantiles. Zones classified into class 3 were considered optimal. These biological optimal areas were afterwards overlaid with ELM climatic optimal maps.
Environmental suitability
A land use and cover change layer was constructed using the method proposed by Mendoza-Ponce et al. [29]. This approach combines socioeconomic and biophysical variables (S3 Text) to generate a probability map of land use and cover changes. The quantity of change is determined based on mean historical transitions from available national maps spanning 1985–2020 [29]. This historical mean transition rate of change was considered as “middle of the road” scenario for land use cover changes. We used the 2020 land use and cover changes map to validate the model, and we used the 2015 land use and cover changes as the latest observed map for modelling iterations. All original land use and cover changes classes (more than 170 classes) were grouped into 14 categories, including eight natural covers: four anthropogenic uses, water bodies, and barren land. We modeled 55 out of 156 possible transitions, excluding changes among natural covers and transitions involving water bodies (S1). The chosen SSP is the average-tendency SSP 2 4.5, and the climate layers used in this study were consistent with those used in ecological niche modeling: CanESM5 and MPI-ESM1–2-HR, along the same time horizons (2030, 2050, 2070 and 2100). To identify geographic areas with environmental suitability for ELM, we had to choose those areas that were environmentally suitable also for species of Asclepias given that milkweeds are their food source and defense mechanism—through cardiac glycosides [42]—, and their main hostplant. These areas included all types of primary and secondary vegetation, along with a 250 m buffer around rainfed agricultural areas and water bodies.
ELM suitability areas
To identify ELM suitability areas, we overlaid the climatic, biological, and environmental suitability maps (Fig 1). The regions where only one category of suitability was present were classified as low suitability areas (e.g., climatic suitability alone, biological suitability alone, or environmental suitability alone). When two categories of suitability overlap, we classified these areas as medium suitability areas (e.g., climatic + biological suitability; climatic + environmental suitability, or biological + environmental suitability). Lastly, the regions with the highest suitability were areas where all three suitability layers (climatic, biological, and environmental) overlapped (Fig 1).
Low suitability: areas where only climatic, biological or environmental suitability were present; Medium suitability: areas where at least two suitability categories overlay; Highest suitability: areas where the three suitability categories overlay. Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
Results
Climatic, biological and environmental suitability for ELM
The areas of high climatic suitability for ELM under the baseline scenario were located on the northeast of Mexico (Fig 2). Overall, there was a decreasing trend in areas under the two future climate change scenarios; the CanESM5 model showed a high decrease (40% between time horizons 2030 and 2070), and the MPI-ESM1–2-HR model showed a low decrease (8% between time horizons 2030 and 2070) in the area of climatic suitability for ELM, respectively (Table 2). The future climatic suitable areas moved southwards along the Mexican plateau to central Mexico (Fig 2; Fig A in S1 Text). Similar projections depicting southern future shifts to central Mexico in climatic suitability for ELM with the GCM: MPI-ESM1–2-HR model were observed (Fig B in S1 Text). There was also a high concordance in trends and shifts in climatic suitability across time horizons (Fig 2 and Fig C in S1 Text).
Baseline: 1970–2000, General Circulation Model CanESM5, time horizons; 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). (*): Climatic suitability only. Dark blue: no climatic suitability, Light blue: light climatic suitability, Dark green: intermediate-light climatic suitability, Green: intermediate-strong climatic suitability, Yellow: strong climatic suitability. Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
The areas of biological suitability for ELM based on the species of Asclepias weighted ensemble occurred mainly in central and southern Mexico under baseline and future climate change scenarios, respectively. Overall, there were discrepancies between the two GCM; the CanESM5 model showed a slight decrease (9% between horizons 2030 and 2070), and the MPI-ESM1–2 model showed an important increase (25% between time horizons 2030 and 2070) in biological suitability for ELM, respectively. Across time horizons, the biological suitability shows a slight southward shift under climate change projections to central and southern Mexico (Fig 3). This trend also remained with the GCM: MPI-ESM1–2-HR model (Fig C in S1 Text).
We assumed that the areas with the highest weighted species richness of perennial tropical milkweed represented the regions of highest biological suitability for ELM of monarchs. These areas were identified based on three criteria: (a) a higher number of species of Asclepias, (b) species of PTM preferred by monarchs, and (c) a higher number of species of Asclepias closely related to monarchs. Baseline: 1970–2000, General Circulation Model CanESM5, time horizons; 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP2-4.5). (*): Biological suitability only. See Methods for more details. Dark blue: non-biological suitability, Light blue: low biological suitability, Dark green: medium biological suitability, Green: medium-to-high biological suitability, Yellow: high biological suitability. Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
Further, areas with environmental suitability included all types of primary and secondary vegetation, along with a 250 m buffer around rainfed agricultural areas and water bodies (Fig 4). The two CGM models showed a decrease of 8% in area of environmental suitability for ELM across time horizon 2030–2070 (Table 2).
The areas that in terms of projecting land use and cover changes hold a suitable environment for ELM: all types of primary and secondary vegetation, along with a 250 m buffer around rainfed agricultural areas and water bodies. See Methods for more details. Baseline: 1970–2000, General Circulation Model CanESM5, time horizons; 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
Lastly, the area of highest climatic, biological, and environmental suitability for ELM is expected in northern and central Mexico. Across the time horizon, there was a slight shift southward under future climate change scenarios (Fig 5). However, a sharp decrease of the highest suitability area was observed for the CanESM5 from the time horizon of 2030–2050 (32%) and from 2030 to 2070 (69%) (Table 2). In contrast, a small increase (7%) in Highest suitability areas was observed under the MPI-ESM1–2-HR from 2030 to 2050 (Table 2). Overall, the area of environmental suitability for ELM was greater than the climatic (93%) and biological (46%) suitability, respectively (Table 2; Figs 2-4). The area of climatic suitability for ELM represents a higher limitation for ELM under future climate change. This limitation is higher under the GCM CanESM5 than the MPI-ESM1–2-HR (Table 2). On the other hand, biological suitability projected a future increase in area compared with climatic suitability, while environmental suitability reflected the largest future suitable areas for ELM (Table 2).
General Circulation Model CanESM5, time horizons; 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). (*) Low suitability depicts areas where only one category of suitability is present (*). Medium suitability depicts areas where at least two categories of suitability overlap (**). High suitability depicts areas where all three suitability categories overlap (***). Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
Discussion
Overall, limited attention has been given to the potential impact of climate change on the monarch migration [but see 13]. Our study provides a multi-stage protocol by integrating climatic, biological, and environmental suitability projections of ELM under baseline and future climate change scenarios in Mexico. The presumed establishment of more resident populations of monarchs in areas along the flyways during the fall migration due to an increase in planting of nonnative species of tropical milkweeds has potentially contributed to a decrease in the number of monarchs completing their journey to the overwintering sites in central Mexico [27]. This explanation assumes areas with habitat suitability for ELM and species of Asclepias in this region, although this has not been fully explored. Thus, it is pertinent to investigate the impact of climate and land use cover changes on areas of habitat suitability for ELM and species of Asclepias. These key conditional scenarios can provide a more comprehensive and nuanced understanding of how potential geographic shifting conditions might impact the monarch fall migration in this region. By combining robust ecological niche modeling techniques with updated climate and land use cover changes, our approach provides insights into the potential future trajectories and impacts on the fall migration of monarchs.
Our results reveal differences in projected losses of climatic suitability for ELM depending on the GCM (General Circulation Model) model: the CanESM5 model forecasts a reduction of almost half in climatically suitable areas for ELM from time horizon 2030–2070, while the MPI-ESM1–2-HR model indicates a decrease of approximately 10% between the same time horizons (Table 2). CanESM5 is a GCM that projects higher temperatures and more precipitation variability [41]. Further, projections of biological suitability showed discrepancies depending on the GCM model: the CanESM5 showed a moderate decrease in biologically suitable areas, but an increase for the species of milkweeds weighted ensemble in the MPI-ESM1–2-HR model (Table 2). The future projection of environmentally suitable areas for ELM was the largest and showed a moderate decrease for both climate change models (Table 2). There was a declining trend in areas with the highest habitat suitability for ELM, where climatic, biological, and environmental suitability overlap, with climate emerging as the most limiting factor across all evaluated time horizons and GCMs. This trend can be relevant as it has been documented that the triggering of reproductive diapause and the onset of monarch southward migration into Mexico are photoperiod, age of milkweed plants, and temperature [13,46–50]. Further, temperature has also been demonstrated to be a key factor determining the flight direction during migration [50,51], and an increase in temperature has been related to a decrease in longevity of monarchs, resulting in a reduction of overlaid eggs [52]. The increase in temperature is the most consistent trend of climate change in the coming decades, although depending on the GCM used is the amount of temperature increase expected [53]. The wide disparity between percentage losses underscores the necessity of employing at least two different contrasting GCMs to adequately capture future climatic uncertainty and variability, to better delineate potential risks to ELM habitat suitability. In this study, we used two different GCMs with adequate performance for North America—including Mexico—[42] and with the greatest distance between them in terms of code-genealogy [43]. The difference between codes helps embracing more climatic variability without having to incorporate multiple GCMs. In terms of precipitation, higher precipitation has been associated with higher population growth. Thus, a reduction in precipitation might result in a decrease in monarch population densities [54].
The climatic and biological suitability for ELM projected a consistent southward shift under future climate scenarios. This displacement is significant as it may alter historical migratory routes by concentrating suitable oviposition sites and host-plant resources—mainly species of Asclepias—in southern regions. Since species of Asclepias are critical as larval host plants, nectar sources for adults, and as cardiac glycoside providers for predator defense [46], changes in their distribution can directly influence monarch reproductive success and migration dynamics [13,55]. While integrating biological and environmental dimensions into modeling provides an ecological framework, it is important to highlight that most existing climate change assessments of species distributions frequently omit these facets [56–58]. Our study was intended not only to evaluate the target species vulnerability to climate change, but also the impact on a biological process, such as the monarch migration phenomenon [13].
Interestingly, environmental suitability—centered on land use cover changes and habitat availability for species of Asclepias—did not appear to constrain ELM distribution under the climate change scenarios examined. This likely reflects a relative stability in land use and cover changes and habitat availability in central and northeastern Mexico as has happened for the last 30 y [21] and suggesting these regions will remain (or increase) in area of habitat suitability for the persistence of ELM and species of milkweeds in the coming decades (Fig 2-5). This implies that despite efforts to reduce small-scale planting of tropical milkweed species to avoid resident populations of monarch butterflies along their flyways in the fall migration, large areas of environmental suitability under current and future climate change scenarios can facilitate the establishment of wild populations of perennial species of milkweeds, providing extensive suitable areas for ELM [27]. Thus, a large-scale spatiotemporal monitoring program is needed to locate and determine the establishment of the resident population in areas along the flyways in the fall migration of monarchs. The monitoring program of monarch migration based on citizen-scientists in this region producing robust datasets on the size of monarch “roosts” for several decades has been a successful example [27; see www.correoreal.mx, 31]. Further monitoring efforts should also include the detection and location of resident populations of monarchs in this region [10].
It is also important to highlight how these future shifts of habitat suitability of ELM and species of Asclepias can impact the spring migration, where overwintering populations move northward. As the northward migration involves multiple generations requiring repeated oviposition and larval development across a range of habitats [1], the sustained environmental suitability for species of Asclepias is essential in supporting the spring migratory cycle northward. Rainfed agriculture is the land cover activity that is expected to increase in area in the next decades in Mexico (Fig D in S1 Text). It is known that rainfed agriculture in Mexico is generally a low-input agriculture that is less likely to affect species of Asclepias and ELM. The species interaction of monarch populations migrating northwards with the presumed increasing resident populations in central and northeastern Mexico poses another challenge in the monarch migration. It has been observed that resident monarchs develop shorter wings, which makes it difficult for monarchs to fly for longer distances [59]. An additional complication is that monarchs that lay eggs on species of tropical milkweed like A. curossavica can get infected with the protozoan parasite O. elektroscirrha that disrupts their life cycle [3,28]. On the other hand, the potential increase of U.S. agricultural land with GMOs and its associated technology represents an additional risk to species of Asclepias, ELM, and adult individuals in their migration [60,61].
The geographical overlap of climatic, biological, and environmental suitability occurred in portions of central and northeastern Mexico, as current and future hotspots of highest ELM suitability (Fig 2-5). Even under present conditions, these key areas occurred (~200 km) south of the U.S.-Mexico border. The predicted southward shift in highest suitability zones under climate change scenarios further accentuates the spatial disconnect from northern monarch populations [13]. These findings raise important conservation concerns: adult monarchs may establish resident and reproductive populations prematurely during the fall migration [27] and potentially disrupt migration to the overwintering sites, as well as reduce population connectivity with northern breeding sites. Consequently, comprehensive conservation strategies must account for the complex interplay among climatic constraints, host-plant distributions, and habitat suitability at different spatiotemporal scales that collectively shape monarch migratory pathways. Future research should use these multifaceted niche modeling approaches to assess habitat suitability in the U.S. and Canada. A holistic transboundary perspective of trinational collaboration that includes climatic, biological, and environmental dimensions can anticipate and mitigate the impacts of climate change on the complex monarch migratory phenomenon [9].
Supporting information
S1 Text. In this supplementary material, it is possible to find four figures representing the climatic, biological, environmental, and multidimensional suitability of the eggs and larvae of monarch butterfly with the General Circulation Model MPI-ESM1–2-HR.
Figure A. Records of the 46 Asclepia species present in the migratory route from Mexico to the north. Black dots: all records from Ascleipa species that were used for the modeling, Orange dots: records from Asclepia species found in the migratory route from Mexico to the north, Blue area: the Monarch Butterfly Biosphere Reserve area. Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/. Figure B. Baseline and future projections under climate change scenarios of climatic suitability for eggs and larvae of monarch butterflies (ELM) along the eastern migratory route in Mexico. Baseline: 1970–2000, General Circulation Model MPI-ESM1–2-HR, time horizons; 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). (*): Climatic suitability only. Dark blue: no climatic suitability, Light blue: light climatic suitability, Dark green: intermediate-light climatic suitability, Green: intermediate-strong climatic suitability, Yellow: strong climatic suitability. Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/. Figure C. Baseline and future projections under climate change scenarios for the areas with biological suitability for ELM. We assumed that the areas with the highest weighted species richness of Asclepias represented the regions of highest biological suitability for ELM of monarchs. These areas were identified based on three criteria: (a) a higher number of species of Asclepias, (b) species of Asclepias preferred by monarchs, and (c) a higher number of species of Asclepias closely related to monarchs. Baseline: 1970–2000, General Circulation Model MPI-ESM1–2-HR, time horizons; 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). (*): Biological suitability only. See Methods for more details. Dark blue: no biological suitability, Light blue: light biological suitability, Dark green: intermediate-light biological suitability, Green: intermediate-strong biological suitability, Yellow: strong biological suitability. Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/. Figure D. Environmental suitability. Areas that, in terms of projected land-use change, will have a suitable environment for ELM: all types of primary and secondary vegetation, along with a 250 m buffer around rainfed agricultural areas and water bodies. See Methods for more details. Baseline: 1970–2000, General Circulation Model MPI-ESM1–2-HR, time horizons. 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/. Figure E. Geographic areas depicting the climatic, biological, and environmental suitability categories for ELM under climate change scenarios. General Circulation Model MPI-ESM1–2-HR, time horizons: 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 (SSP 2 4.5). (*) Low suitability depicts areas where only one category of suitability is present (*). Medium suitability depicts areas where at least two categories of suitability overlap (**). Highest suitability depicts areas where all three suitability categories overlap (***). Figures include a combination of author-generated visualizations and images from open-access sources licensed for reuse under Creative Commons licences. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
https://doi.org/10.1371/journal.pclm.0000802.s001
(DOCX)
S2 Text.
Table A. Evaluated species, number of unique records in the calibration area and in Mexico, and their relative importance for monarch eggs and larvae (ELM). Table B. Changes in climatic suitability for each Asclepias species for each scenario evaluated with and without environmental suitability. Climatic suitability: considering limited-full dispersal of the species, Environmental suitability: optimal land use cover change, 2030: 2021–2040, 2050: 2041–2060, 2070: 2061–2080 under the tendency scenario SSP 245. Table C. Bioclimatic variables that contributed the most to the monarch’s modeling. Political-administrative boundaries were obtained from the Instituto Nacional de Estadística y Geografía (INEGI). Marco Geoestadístico shapefiles. Datos abiertos: https://www.inegi.org.mx/datosabiertos/.
https://doi.org/10.1371/journal.pclm.0000802.s002
(XLSX)
S3 Text. Raster files of future land-use/land-cover maps, as well as ecological niche modeling of Asclepias and monarch butterfly larvae and eggs under current and climate change scenarios.
Link to the drive. https://drive.google.com/drive/folders/1aqanevZZ1Dd_c1u4OjTsEkzFa0ZFlpnp
https://doi.org/10.1371/journal.pclm.0000802.s003
(DOCX)
Acknowledgments
We thank the Instituto de Biología and the Instituto de Ciencias de la Atmósfera y Cambio Climático at UNAM for logistic support.
References
- 1. Reppert SM, de Roode JC. Demystifying Monarch Butterfly Migration. Curr Biol. 2018;28(17):R1009–22. pmid:30205052
- 2. Castañeda S, Botello F, Sánchez-Cordero V, Sarkar S. Spatio-Temporal Distribution of Monarch Butterflies Along Their Migratory Route. Front Ecol Evol. 2019;7.
- 3. Davis AK, de Roode JC. Effects of the parasite, Ophryocystis elektroscirrha, on wing characteristics important for migration in the monarch butterfly. Animal Migration. 2018;5(1):84–93.
- 4. Oberhauser K, Wiederholt R, Diffendorfer JE, Semmens D, Ries L, Thogmartin WE, et al. A trans-national monarch butterfly population model and implications for regional conservation priorities. Ecological Entomology. 2016;42(1):51–60.
- 5. Vidal O, Rendón-Salinas E. Dynamics and trends of overwintering colonies of the monarch butterfly in Mexico. Biological Conservation. 2014;180:165–75.
- 6. Thogmartin WE, Diffendorfer JE, López-Hoffman L, Oberhauser K, Pleasants J, Semmens BX, et al. Density estimates of monarch butterflies overwintering in central Mexico. PeerJ. 2017;5:e3221. pmid:28462031
- 7. Saunders SP, Ries L, Neupane N, Ramírez MI, García-Serrano E, Rendón-Salinas E, et al. Multiscale seasonal factors drive the size of winter monarch colonies. Proc Natl Acad Sci U S A. 2019;116(17):8609–14. pmid:30886097
- 8. López-García J, Navarro-Cerrillo RM, de Lourdes M-DL. Forest land-cover trends in the Monarch Butterfly Biosphere Reserve in Mexico, 1994–2017. Environmental Conservation. 2022;49(4):244–54.
- 9. Diffendorfer JE, Drum RG, Mitchell GW, Rendón-Salinas E, Sánchez-Cordero V, Semmens DJ, et al. The benefits of big-team science for conservation: Lessons learned from trinational monarch butterfly collaborations. Front Environ Sci. 2023;11.
- 10. Federation WW. La población de mariposas monarca del este casi se duplica en 2025. https://www.worldwildlife.org/descubre-wwf/historias/la-poblacion-de-mariposas-monarca-del-este-casi-se-duplica-en-2025. 2025.
- 11. Flockhart DTT, Pichancourt J-B, Norris DR, Martin TG. Unravelling the annual cycle in a migratory animal: breeding-season habitat loss drives population declines of monarch butterflies. J Anim Ecol. 2015;84(1):155–65. pmid:24903085
- 12. Zylstra ER, Ries L, Neupane N, Saunders SP, Ramírez MI, Rendón-Salinas E, et al. Changes in climate drive recent monarch butterfly dynamics. Nat Ecol Evol. 2021;5(10):1441–52. pmid:34282317
- 13. Lemoine NP. Climate change may alter breeding ground distributions of eastern migratory monarchs (Danaus plexippus) via range expansion of Asclepias host plants. PLoS One. 2015;10(2):e0118614. pmid:25705876
- 14. Agrawal AA, Inamine H. Mechanisms behind the monarch’s decline. Science. 2018;360(6395):1294–6. pmid:29930122
- 15. James DG. Monarch Butterflies in Western North America: A Holistic Review of Population Trends, Ecology, Stressors, Resilience and Adaptation. Insects. 2024;15(1):40. pmid:38249046
- 16. Geest EA, Wolfenbarger LL, McCarty JP. Recruitment, survival, and parasitism of monarch butterflies (Danaus plexippus) in milkweed gardens and conservation areas. J Insect Conserv. 2018;23(2):211–24.
- 17. Gustafsson KM, Wolf SA, Agrawal AA. Science-Policy-Practice Interfaces: Emergent knowledge and monarch butterfly conservation. Env Pol Gov. 2017;27(6):521–33.
- 18. Inamine H, Ellner SP, Springer JP, Agrawal AA. Linking the continental migratory cycle of the monarch butterfly to understand its population decline. Oikos. 2016;125(8):1081–91.
- 19. Taylor OR Jr, Pleasants JM, Grundel R, Pecoraro SD, Lovett JP, Ryan A. Evaluating the Migration Mortality Hypothesis Using Monarch Tagging Data. Front Ecol Evol. 2020;8.
- 20. Spaeth KE Jr, Williams CJ, Moranz RA, Taliga C, Rutherford WA, Simpson B. Preferred nectar sources for the monarch butterfly (Danaus plexippus plexippus) along the Great Plains migration pathway. Ecosphere. 2025;16(2).
- 21. Diffendorfer JE, Botello F, Drummond MA, Ancona ZH, Corro LM, Thogmartin WE, et al. Changes in landscape and climate in Mexico and Texas reveal small effects on migratory habitat of monarch butterflies (Danaus plexippus). Sci Rep. 2024;14(1):6703. pmid:38509089
- 22. Kass JM, Anderson RP, Espinosa-Lucas A, Juárez-Jaimes V, Martínez-Salas E, Botello F, et al. Biotic predictors with phenological information improve range estimates for migrating monarch butterflies in Mexico. Ecography. 2019;43(3):341–52.
- 23. Sánchez-Tlacuahuac N, Pimentel-Equihua JL, Espinosa-Hernández V, Vibrans H. What do monarchs feed on in winter? Nectar sources at hibernation sites. J Insect Conserv. 2022;27(1):181–91.
- 24. Clement K, Crawford P. Fall available tropical milkweed (asclepias curassavica l.) may be a population sink for the monarch butterfly. onpr. 2020;20.
- 25. DuBose JG, Hoogshagen M, de Roode JC. The role of a non-native host plant in altering the seasonal dynamics of Danaus plexippus (Lepidoptera: Nymphalidae) development. J Insect Sci. 2025;25(3):13. pmid:40459988
- 26. Juárez-Jaimes V, Rodríguez-Morales LO, Paniagua-Ibáñez M, Hernández-Barón GM, Fishbein M. Riqueza y distribución del género Asclepias (Apocynaceae: Asclepiadoideae) en México. RevMexBiodiv. 2022;93:e933958.
- 27. Davis AK, Croy JR, Snyder WE. Dramatic recent declines in the size of monarch butterfly (Danaus plexippus) roosts during fall migration. Proc Natl Acad Sci U S A. 2024;121(43):e2410410121. pmid:39405357
- 28. Kendzel MJ, Altizer SM, de Roode JC. Interactions between parasitism and migration in monarch butterflies. Curr Opin Insect Sci. 2023;59:101089. pmid:37506879
- 29. Mendoza-Ponce A, Corona-Núñez R, Kraxner F, Leduc S, Patrizio P. Identifying effects of land use cover changes and climate change on terrestrial ecosystems and carbon stocks in Mexico. Global Environmental Change. 2018;53:12–23.
- 30. Woodson RE. The North American species of Asclepias L. Annals of the Missouri Botanical Garden. 1954;41(1):1–211.
- 31.
Sheehan N, Weber-Grullon L. Journey North–monarch butterfly and milkweed observations by volunteer community scientists across Central and North America (1996-2020) ver 1. Environmental Data Initiative. 2021.
- 32. Cao W, Hu L, Gao J, Wang X, Ming Z. A study on the relationship between the rank of input data and the performance of random weight neural network. Neural Comput & Applic. 2020;32(16):12685–96.
- 33. Assis JR. Pipelines to reduce the spatial autocorrelation in species distribution models. theMarineDataScientist. 2020.
- 34. Aiello-Lammens ME, Boria RA, Radosavljevic A, Vilela B, Anderson RP. spThin: an R package for spatial thinning of species occurrence records for use in ecological niche models. Ecography. 2015;38(5):541–5.
- 35. Barbet-Massin M, Jiguet F, Albert CH, Thuiller W. Selecting pseudo-absences for species distribution models: how, where and how many?. Methods in Ecology and Evolution. 2012;3(2):327–38.
- 36. Mesgaran MB, Cousens RD, Webber BL. Here be dragons: a tool for quantifying novelty due to covariate range and correlation change when projecting species distribution models. Diversity and Distributions. 2014;20(10):1147–59.
- 37.
Soberon J, Peterson AT. Interpretation of models of fundamental ecological niches and species’ distributional areas. 2005.
- 38. Dinerstein E, Olson D, Joshi A, Vynne C, Burgess ND, Wikramanayake E, et al. An Ecoregion-Based Approach to Protecting Half the Terrestrial Realm. Bioscience. 2017;67(6):534–45. pmid:28608869
- 39. Dormann CF, Elith J, Bacher S, Buchmann C, Carl G, Carré G, et al. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography. 2012;36(1):27–46.
- 40. Naimi B, Hamm NAS, Groen TA, Skidmore AK, Toxopeus AG. Where is positional uncertainty a problem for species distribution modelling?. Ecography. 2013;37(2):191–203.
- 41. Fick SE, Hijmans RJ. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Intl Journal of Climatology. 2017;37(12):4302–15.
- 42. Altamirano del Carmen MA, Estrada F, Gay-García C. A new method for assessing the performance of general circulation models based on their ability to simulate the response to observed forcing. Journal of Climate. 2021;:1–52.
- 43. Knutti R, Masson D, Gettelman A. Climate model genealogy: Generation CMIP5 and how we got there. Geophysical Research Letters. 2013;40(6):1194–9.
- 44. Kuma P, Bender FA -M., Jönsson AR. Climate Model Code Genealogy and Its Relation to Climate Feedbacks and Sensitivity. J Adv Model Earth Syst. 2023;15(7).
- 45. Allouche O, TsoaR A, Kadmon R. Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS). Journal of Applied Ecology. 2006;43(6):1223–32.
- 46. Rothschild M, Marsh N, Gardiner B. Cardioactive substances in the monarch butterfly and Euploea core reared on leaf-free artificial diet. Nature. 1978;275(5681):649–50. pmid:703827
- 47. Brower LP, Taylor OR, Williams EH, Slayback DA, Zubieta RR, Ramírez MI. Decline of monarch butterflies overwintering in Mexico: is the migratory phenomenon at risk?. Insect Conserv Diversity. 2011;5(2):95–100.
- 48. Flores-Martínez JJ, Martínez-Pacheco A, Rendón-Salinas E, Rickards J, Sarkar S, Sánchez-Cordero V. Recent Forest Cover Loss in the Core Zones of the Monarch Butterfly Biosphere Reserve in Mexico. Front Environ Sci. 2019;7.
- 49. Goehring L, Oberhauser KS. Effects of photoperiod, temperature, and host plant age on induction of reproductive diapause and development time in Danaus plexippus. Ecological Entomology. 2002;27(6):674–85.
- 50. Green DA 2nd, Kronforst MR. Monarch butterflies use an environmentally sensitive, internal timer to control overwintering dynamics. Mol Ecol. 2019;28(16):3642–55. pmid:31338928
- 51. Guerra PA, Reppert SM. Coldness triggers northward flight in remigrant monarch butterflies. Curr Biol. 2013;23(5):419–23. pmid:23434279
- 52. Oberhauser KS. Fecundity, lifespan and egg mass in butterflies: effects of male-derived nutrients and female size. Functional Ecology. 1997;11(2):166–75.
- 53.
Masson-Delmotte V, Zhai P, Pirani A, Connors SL, Péan C, Berger S, et al. Climate Change 2021: The Physical Science Basis. Cambridge University Press. 2021.
- 54. Zipkin EF, Ries L, Reeves R, Regetz J, Oberhauser KS. Tracking climate impacts on the migratory monarch butterfly. Glob Chang Biol. 2012;18(10):3039–49. pmid:28741829
- 55. Spaeth KE Jr, Barbour PJ, Moranz R, Dinsmore SJ, Williams CJ. Asclepiasdynamics on US rangelands: implications for conservation of monarch butterflies and other insects. Ecosphere. 2022;13(1).
- 56. Bellard C, Bertelsmeier C, Leadley P, Thuiller W, Courchamp F. Impacts of climate change on the future of biodiversity. Ecol Lett. 2012;15(4):365–77. pmid:22257223
- 57. Velasco JA, Estrada F, Calderón-Bustamante O, Swingedouw D, Ureta C, Gay C, et al. Synergistic impacts of global warming and thermohaline circulation collapse on amphibians. Commun Biol. 2021;4(1):141. pmid:33514877
- 58. Ureta C, Ramírez-Barahona S, Calderón-Bustamante Ó, Cruz-Santiago P, Gay-García C, Swingedouw D, et al. Evaluation of animal and plant diversity suggests Greenland’s thaw hastens the biodiversity crisis. Commun Biol. 2022;5(1):985. pmid:36115902
- 59. Altizer S, Davis AK. Populations of Monarch butterflies with different migratory behaviors show divergence in wing morphology. Evolution. 2010;64(4):1018–28. pmid:20067519
- 60. Hellmich RL, Siegfried BD, Sears MK, Stanley-Horn DE, Daniels MJ, Mattila HR, et al. Monarch larvae sensitivity to Bacillus thuringiensis- purified proteins and pollen. Proc Natl Acad Sci U S A. 2001;98(21):11925–30. pmid:11559841
- 61. Dively GP, Rose R, Sears MK, Hellmich RL, Stanley-Horn DE, Calvin DD, et al. Effects on Monarch Butterfly Larvae (Lepidoptera: Danaidae) After Continuous Exposure to Cry1Ab-Expressing Corn During Anthesis. Environ Entomol. 2004;33(4):1116–25.